Language/OS - Multiplatform Resource Library/LANGUAGE OS.iso/dsp/dspgroup/cis_dsp.arc/CIS_DSP5.TXT

Text File | 1988-08-25 | 81.7 KB | 2,225 lines

#: 74527 S17/TAPR NNC/DSP
    28-Apr-88  01:03:02
Sb: #74458-#HF Modem Stuff
Fm: Franklin Antonio, N6NKF 76337,1365
To: Barry McLarnon VE3JF 71470,3651 (X)

Post the address and $$$ to get copy of the IEE HF conference proceedings.

*** There is a reply:   74695

*** More ***

#: 74695 S17/TAPR NNC/DSP
    02-May-88  20:15:16
Sb: #74527-HF Modem Stuff
Fm: Bob McGwier N4HY 74615,1366
To: Franklin Antonio, N6NKF 76337,1365 (X)

Institute of Electrical Engineers Savoy Place London WC2, UK I don't see "hide
nor hair" of a price anywhere. Fourth International Conference on HF Radio
Systems and Techniques.

Try to imagine an IEEE journal or proceedings with Jets, Ships, Helicopters
bombing a radar installation and featuring the bridge of a warship with a
multitude of HF antennas coming out here.  It was a decent conference. Bob



*** More ***

#: 74694 S17/TAPR NNC/DSP
    02-May-88  20:15:06
Sb: #74458-HF Modem Stuff
Fm: Bob McGwier N4HY 74615,1366
To: Barry McLarnon VE3JF 71470,3651 (X)

I also attended this conference and have a copy of the proceedings.  I don't
volunteer to make a copy of the whole manual but I will make a copy of the
table of contents available to those who have been talking about the HF stuff.
I enjoyed the conference and as usual there was one or two good papers and the
rest were a re-hash with a new wrinkle or just a rehash altogether.  The
exception to this were the papers people gave on hardware implementations of HF
adaptive terminals.   DSP was a topic of many papers. Some official of the RSGB
and I had a race to jump up and shout at a person from France.  He was talking
about HF radar and about his monster transmitter and antenna.  He boldly
presented his frequency as 14.110 Mhz. We pointed out to him that this was an
exclusive amateur allocation in France and he just shrugged his shoulders.  It
is not only the Soviets doing woodpecker apparently ;-( Bob

#: 74700 S17/TAPR NNC/DSP
    02-May-88  21:38:55
Sb: #Schematics
Fm: Lyle Johnson, WA7GXD 76246,565
To: DSPers

Now that Dayton is over, let's get those comments coming!  I would like to put
the schematics to bed and get Eric cranking on laying out the prototypes before
I leave next week for Canada.

I have the pulse swallower from Mike Brock and will get it massaged into the
diagram of choice...

Lyle

#: 74721 S17/TAPR NNC/DSP
    03-May-88  08:58:34
Sb: #74700-Schematics
Fm: Bob McGwier N4HY 74615,1366
To: Lyle Johnson, WA7GXD 76246,565

Lyle, the first thing I noticed was that no privision was made for changing the
anti-aliasing and conditioning filters.  What is adequate for an SSB rig of
course as you know, won't pass 9600 BPS data or CD output to be interpolated
and resampled for DAT ;-) ;-).  We need that put on a header, or configurable
in some other manner.  I think Dan is working on this isn't he?  Tell Dan also,
that I will call him this week and sorry that I have been gone for two weeks.
Eric gave me the message but my arrival for my twenty minute talk and then
return to DC didn't go as planned.  I will have other comments later today. Bob




#: 74701 S17/TAPR NNC/DSP
    02-May-88  21:58:43
Sb: Schematics
Fm: Lyle Johnson, WA7GXD 76246,565
To: DSPers

CHucks comments have been uploaded to DL17 as CHUCK.DSP

Lyle

#: 74732 S17/TAPR NNC/DSP
    03-May-88  13:59:56
Sb: #74721-Schematics
Fm: Franklin Antonio, N6NKF 76337,1365
To: Bob McGwier N4HY 74615,1366

I think this has been discussed before, but here goes again... If we claim this
thing will "do everything", and "it's only software", etc. (I, of course, have
never made such claims, but YOU have.) Then... the thing should be built so
that you can change from one application to another without removing screws,
opening box, moving jumpers, remembering that you've done so so that you can
remember to change 'em back before running that other application that requires
'em back the other way, etc. So how come you're suggesting antialiasing filter
configured by a HEADER?

#: 74736 S17/TAPR NNC/DSP
    03-May-88  16:42:57
Sb: #74527-HF Modem Stuff
Fm: Barry McLarnon VE3JF 71470,3651
To: Franklin Antonio, N6NKF 76337,1365

I can't find any info on the price, but I think they will accept orders that
are not prepaid.  Orders go to:

Institution of Electrical Engineers
P.O. Box 26, Hitchin, Herts.  SG5 1SA
United Kingdom

If you want to try phoning, the number is 0462 53331.  Tell 'em you want
Conference Publication No. 284.

Barry

PS: Nice meeting you at Dayton!  Maybe next time we can get organized a
little better.  BTW, the numbers I gave you re HF packet lengths were a
crock... way too long.  I was thinking in terms of bytes rather than bits
(Malibu racetracks aren't the best place to hold a technical discussion :-)

[76246,565]
CHUCK.DSP                 02-May-88 6536               Accesses: 2

    Keywords: DSP TAPR AMSAT DSP1

    chucks comments on dsp schematics mailed on 20 april for dsp1 project


Comments on schematics and comments dated 20 April, 1988.

OK, I give up!!!  I'll change my position.  There seems to be a continued
appetite for options and jumpers.  Instead of being against all options and
jumpers, I'll only oppose one class of options and jumpers and support all
others.  I OPPOSE ALL OPTIONS AND JUMPERS WHICH, WHEN IMPLEMENTED, CAUSE
APPLICATIONS WHICH DO NOT USE THEM TO NO LONGER FUNCTION.  This will allow
most of the suggested options and jumpers.  The problem I am trying to avoid
is the one we created in TNC1 and 2 where both VHF and HF were supported, but
could only be switched between by opening up the box.  This kind of thing would
be an even larger problem in a "multi-mode" device.  If we "must" have jumpers
which do not satisfy the above criterion, they should be software "jumpers."
Another way to state this is that all options and jumpers must be "upward
compatible."  I'll ignore the marketing complexity this philosophy will cause.

The first violation of the above philosophy is in the area of sample clocking
the AD7569's.  Although this jumper block does provide the ultimate in
flexibility, I don't see what functional advantage it offers.  If there is any,
it should be software selectable.  I suspect that if the number of combinations
offering advantages is greater than one, it is still a small number and would
not cost much to make software selectable.

It looks like an optional second AD7569 or the two-channel MUX would be "upward
compatible."  Question:  Does this imply an optional radio port with optional
watch-dog timer, optional filters, etc?

Optional 16-bit I/O ports and/or memory swapping seem to pass the "upward
compatibility" test.

We would do ourselves a big favor if both DSP boards were identical!  It might
require a jumper (Ya, I know, I'm the one against jumpers) or two so they can
tell each other apart.

The 8-bit I/O port "for serial data and miscellaneous V40 I/O, or lighting
LEDs, reading (DIP) switches, etc." seems inadequate.  I'm sure we will want
another eight, even if we keep the number of LEDs to a minimum.

I don't see the 8-bit port(s) "for communication between the processors" on the
V40 board.  I know I don't see so good, but I did look hard.  Would the second
port for DSP1 be optional?

Add the BB real-time clock chip to the list of options.  The battery and control
circuitry are already on the board.

DMAing both ports of the SCC would only be needed when two HIGH speed ports are
needed.  This only leaves the V40 on-chip UART which can't be used when both
SCC ports are DMAed.  This presents an application restriction.  If two DMAed
ports are required, there can be no serial port for communication with a PC,
etc.  The Centronics port could be considered as an alternative.  It could be
that if the application is such a HIGH speed data pusher, the Centronics port
would be BETTER than a serial port.

I like the V40 memory configuration.  This should avoid what we put the users
through to upgrade their TNC2's to larger memory chips.

Lyle, if you didn't draw the Centronics port, what is that in the lower-right
corner of the V40 schematic?  I think we should include it.  Not only is it
useful for the reasons you mentioned, I think those who want to do the spectrum
analyzer will like the high speed access to a PC.  Also, note my comments
regarding DMAing both ports of the SCC.

SWITCHES:  I agree that we must have a way for this thing to come up running a
pre-defined application at power-up time.  I also believe the thumbwheel-style
attached to the V40 is the right way to go.  It is the V40 that will load the
DSP application code so it MUST know the switch setting.  I see the switch
having 16 positions (4 bits of data).  Settings 1-15 would select applications
which normally do not require an attached terminal.  This does not preclude
assigning an application requiring a terminal to one of these numbers.  Setting
0 would be used to query an attached terminal for the application to be run.
This menu would include ALL applications, even those specifiable by switch
settings 1-15.  This would make the upper limit to the number of applications
installed dependent only on the amount of memory the V40 has to hold them.
   Power-off-on seems adequate to do a reset.  Pads for a reset button should
be provided in a convenient place and front panel space provided.  The IBM PC
didn't have this and it drove those who like to twiddle the bits crazy.  I'm
sure our box will come up much faster than a PC, but...  If we identify a
switch (and even stock it) which is pc board mounted, this could be another
infamous option.

The radio ports should be on the DSP board.  This seems like the easiest way to
keep the analogue signals clean on their way to/from the AD/DA chip (is this our
number one priority?).  I think the up-down lines also come from the DSP (?).
The PTT line may be a bit of a problem.  I think it has its origin from the
V40.  An interface line from the V40 board to the DSP board(s) will have to be
provided.  Since the number of DSP boards is variable, the watch-dog timer
should be on the DSP board to avoid another option.
     How many radio ports are on each DSP board?  It looks like the number two
is shaping up.  That makes four PTT lines, total.

The PTT lines should double for the CW keying.  Including the watchdog timer in
the CW circuit seems like a good idea.  One CW keyer on the DSP board (keyed by
the V40) seems adequate.  With two DSP boards, you get two.  Can anyone thing
of a use for more than one?

Which serial port to use for a user interface?  This is a tough question.
Using jumpers to select between them violates the criteria I stated at the
beginning.  It is also a cop-out (if a decision is not fun to make, don't make
it - just add jumpers and let things go every which-way).  Maybe I don't
understand, but it looks to me like one of the hardware handshaking lines is
provided by the V40 async port (SRDY!).  Either way, I vote for using the V40
async port.

LED's:  So far, this seems like the only thing that might cause the DSP
board(s) to require front panel space.  I would therefore suggest that all LEDs
be on the V40 board and let the DSP board(s) request the V40 to turn them on
and off.  If you specify any as PTT, does that mean we need to dedicate four
LEDs to this (and maybe four others as DCD)?  I would suggest we pre-define as
few as possible.

#: 74745 S17/TAPR NNC/DSP
    03-May-88  21:32:45
Sb: #74732-#Schematics
Fm: Bob McGwier N4HY 74615,1366
To: Franklin Antonio, N6NKF 76337,1365 (X)

Good question.  I don't care if we change it from the front panel, it just
seemed that the original suggestion we made (months ago) that all this be put
on a "game cartridge" got dropped by the wayside.  The "game cartridge"
approach was when the DSP card had no host and we wanted to get software to it
without a kludgy loader.  Then we thought better of not allowing folks to
implement a better mousetrap by not having a loader.  This lead almost
inevitably in my opinion to the general purpose microprocessor in place of
making it either IBM PC specific or limiting it to what could be done by
plugging it into a tnc-2.  When we got the GP host the only need for the "game
cartridge" was the filters and it just slid under the table. It is clear to me
that one set of filters is not sufficient.  I would be happy if we had filters
selectable with cutoffs that were okay for SSB radios, and then selectable
upwards.  It is not as if this is the first time we thought about it.  A header
is cheaper than a switch but your right about the lack of class in a header.
Bob



*** There are replies:  74750, 74771

*** More ***

#: 74750 S17/TAPR NNC/DSP
    03-May-88  23:01:25
Sb: #74745-Schematics
Fm: Franklin Antonio, N6NKF 76337,1365
To: Bob McGwier N4HY 74615,1366

One thought:  programmable lowpass filter ala Gould S3528.  Lyle will probably
disapprove of the price.  Last time i looked they were around $25., but i'm not
real sure of that number anymore.  (These are the chips in the SuperSCAF in few
months ago QST article.)

*** More ***

#: 74771 S17/TAPR NNC/DSP
    04-May-88  10:08:32
Sb: #74745-Schematics
Fm: Lyle Johnson, WA7GXD 76246,565
To: Bob McGwier N4HY 74615,1366

I am meeting with Dan and Eric tomorrow at lunchtime to get Dan's progress on
the filter issue.  I agree that we don't want something that has to be changed
by opening the cabinet.  The filters shown are admittedly sleazy; this is not
because I want to use sleazy filters, only that filters aren't my bag!

The problem with a "game port" is that we have to become mechanical engineers
and pay very close attention to "tempest shielding" which means more money.
Remember all the grief the computer manufacturers went through to get part 15
certification back in the early '80s to get the noise out. The Atari 400 and
800 are amazing in this regard, and the C-64 has 4800 bps diskettes as a
result!  We are talking about a big hole through which two high-speed
processors are trying to push digital noise!

As a minimum I would think we want filters useable in a voice bandwidth
application with the option to switch them out and use some sort of external
filter.  As a minimum.

Anyway, I'll report tomorrow on Dan's progress.  Meanwhile, what about the
shared DSP memory circuits, resolution of the DSP I/O address space vs. TBLW
usage, 2-channel analog MUX vs. dual analog I/O ports vs. single analog I/O
port, etc?  :-)

Lyle

#: 74814 S17/TAPR NNC/DSP
    05-May-88  01:51:25
Sb: Dan and Lyle Mtg
Fm: Lyle Johnson, WA7GXD 76246,565
To: DSPers

Dan came over tonight and the results of our discussions were:

1) We will look into using a SCF like an MF5 for anti-aliasing on the way in
and another on the way out of the AD7569.  Use the 8254 output to drive a fixed
divider as well as the MF5 (or section of MF10).  USe output of divider to
drive the AD7569.  This puts 3 dB point of filter at a fixed ratio to the
sampling frequency.  Make provision to bypass the whole thing in case someone
wants to.

2) Drop the memory swap scheme and look into a 16-bit wide FIFOed output port
from a cost perspective.

3) Give the TMS320 handshake latches for communications with the V40 like the
V40 has.

4) Use RIO5 (currently unused) to read a status register which would include
the handshake signals, etc.

5) Buffer the Centronics port or otherwise protect it from static, etc.

Lyle


#: 74816 S17/TAPR NNC/DSP
    05-May-88  08:28:10
Sb: #74750-Schematics
Fm: Barry McLarnon VE3JF 71470,3651
To: Franklin Antonio, N6NKF 76337,1365

Time flies!  The SuperSCAF article was actually about 2 years ago.  I know
because that's the length of time the one I breadboarded has been sitting under
the desk in my shack, waiting for me to package it up.

The idea of using the S3528 crossed my mind too, but I dismissed it since I
knew Lyle would have a fit about the cost.  Maybe it's not TOO bad though... I
just had a look at the catalog from a local retail clip shop, and the unit
price of the S3528 is listed at $11.59 CAN$, which is about $9.30 US$ or so. I
suspect that they could be had for $5 or so with a bit of finagling.  Lyle will
still have a fit, but it's worth looking into.

Also, it is a 7th-order elliptic, so the group delay could be a little much for
certain applications.  And it is certainly overkill for the D/A reconstruction
filter.  On the other hand, I HATE the thought of changing headers when
changing applications!

#: 74834 S17/TAPR NNC/DSP
    05-May-88  13:39:28
Sb: #74816-Schematics
Fm: Franklin Antonio, N6NKF 76337,1365
To: Barry McLarnon VE3JF 71470,3651 (X)

The IF crystal filters in most HF rigs are 8th order chebyshev already. Many HF
rigs have two of them in series, and use them both in the xmit chain and recv
chain.  So that makes 4 of 8th order chebyshev filters (all about the same
bandwidth) in series between xmit and rcv modems already.

I had no idea the S3528's were down to such a low price.  I've never seen them
listed in anybody's catalog.  I was planning to buy some from a local
distrubutor, who would, no doubt, try to charge me full price.  Re the catalog
you referred to... is this a place that's used to selling to amateurs in low
quantities?  If so, how about sharing nam,addr,ph#.

*** More ***

#: 74869 S17/TAPR NNC/DSP
    05-May-88  21:21:48
Sb: #74823-Schematics
Fm: Bob McGwier N4HY 74615,1366
To: Barry McLarnon VE3JF 71470,3651 (X)

Ditto on the switches. Bob



*** More ***

#: 74868 S17/TAPR NNC/DSP
    05-May-88  21:21:29
Sb: #74771-Schematics
Fm: Bob McGwier N4HY 74615,1366
To: Lyle Johnson, WA7GXD 76246,565 (X)

RIGHT.  I uploaded on all that earlier.  If we can just switch them out of
line, I will be happy.  Tell Dan I am spending 12 hours a day at work and will
be out of town all next week.  I will be glad when summer and the special
project starts out in California so that I can get away from the beck and call
of the folks jjjuuussttt far enough away that it is a pain in the butt to go
see them but too far for it to not waste 4-5 hours everytime I have to do it.

Bob




#: 74835 S17/TAPR NNC/DSP
    05-May-88  13:52:34
Sb: HF Modem Design
Fm: Franklin Antonio, N6NKF 76337,1365
To: ALL

For those of you interested in HF modem design, here's a 10 year old article
that helps you visualize the kinds of fading that can occur on HF channels. The
text of the article isn't very useful.. it's mostly about their measurement
technique, but the 12 pictures they call "Fadeograms" are great.  These are
similar to sonograms used in speech analysis.  These guys xmitted a multitone
signal at one location, then rcvd at another location, putting the audio output
of an SSB rig into a sonogram machine.  Result is a picture where horizontal
axis is time, vertical axis is freqency, and gray scale is amplitude.

The Fadeogram, A Sonogram-Like Display of the Time-Varying Frequency Response
of HF-SSB Radio Channels, by J.H.J Filter, et al.  June 1978, IEEE Trans. on
Communications, pp 913-917.

#: 74865 S17/TAPR NNC/DSP
    05-May-88  20:56:32
Sb: DSP comments
Fm: Bob McGwier N4HY 74615,1366
To: ALL

I have uploaded my comments along with those of Paul Newland, AD7I, hardware
father of the TNC-2.  He is still in their pushing for AMTOR as you will be
able to read but his comments are a different prespective and need to be
considered.  I have read Chucks comments and I like most of what I have read
with the exception of the LED question.  I would rather not have to figure out
a "status"protocol between the v40 and the TMS320C1X that says "go light up the
gobbly gook LED" when the entire operation of the LED is determined by what the
DSP chip sees.  This will enable us to have modem disconnects to leave the DSP
and still have LED readout on lock, DCD, level, etc.




#: 74870 S17/TAPR NNC/DSP
    05-May-88  21:22:02
Sb: #74764-#HF Modem Stuff
Fm: Bob McGwier N4HY 74615,1366
To: Franklin Antonio, N6NKF 76337,1365 (X)

Didn't I tell you that I had the manuals for the assembler and the emulator and
would send them?  If not I boo boo'd.  I remember running around the last
minute trying to find you so that we could hand the board off and even though I
meant to tell you I might have forgotten.

Bob



*** There is a reply:   74890

*** More ***

#: 74890 S17/TAPR NNC/DSP
    06-May-88  00:12:00
Sb: #74870-#HF Modem Stuff
Fm: Franklin Antonio, N6NKF 76337,1365
To: Bob McGwier N4HY 74615,1366 (X)

You probably told me, and the importance of the matter didn't sink in. Now that
i've installed the board, ground down my BNC connectors, etc, it becomes real
obvious.

*** There is a reply:   74947

*** More ***

#: 74947 S17/TAPR NNC/DSP
    07-May-88  00:47:14
Sb: #74890-#HF Modem Stuff
Fm: Bob McGwier N4HY 74615,1366
To: Franklin Antonio, N6NKF 76337,1365 (X)

Those BNC's are a real joke are they not?  He could have (and we all should)
replace them with the ones with the long neck but naaaaaa, that would be the
right way and it's more fun to curse and nash our teeth as you do an FFT and
see crap from the computer all over it. ;-)

Bob



*** There is a reply:   75019

*** More ***

#: 75019 S17/TAPR NNC/DSP
    08-May-88  14:44:53
Sb: #74947-HF Modem Stuff
Fm: Franklin Antonio, N6NKF 76337,1365
To: Bob McGwier N4HY 74615,1366

The real problem is NOT the BNC connectors, but the fact that the board is 1/4"
shorter than it's supposed to be.  If the board were the right length, the
BNC's would stick out the back correctly, and there'd be no problem.  I suspect
that the reason the board is 1/4" short is so that the BNC's don't stick out
the back the expected amount so that it is POSSIBLE to get the board into a PC.
The combination of the BNC's sticking out and the board being full length at
the other end (ie end of board at front of computer is max length from the
connector) make the board almost impossible to insert.  I almost replaced the
BNCs with RCA jacks.  This would make it possible to insert, and would allow
the plate to be moved 1/4" farther away from the board, so ya wouldn't have 't
bend it.  Ah well.. water under the damn.


#: 74871 S17/TAPR NNC/DSP
    05-May-88  21:22:13
Sb: #74814-Dan and Lyle Mtg
Fm: Bob McGwier N4HY 74615,1366
To: Lyle Johnson, WA7GXD 76246,565 (X)

I REALLY like the thought of the centronics as a high speed (cheap) port to the
PC, but . . .  Most PC parallel are "one way" ports (see Phil's article on the
parallel interface for driving the ICOM radios by computer in the AMSAT
proceedings from Dallas for example).  I doubt those that are bidirectional
come standard so this would have to be added as an optional port. Bob




#: 74872 S17/TAPR NNC/DSP
    05-May-88  21:26:08
Sb: DSP comments
Fm: Bob McGwier N4HY 74615,1366
To: Lyle

Forgot to comment on the second set of notes. I like the memory map. 128K
shared with the DSP and V40 seems adequate. Glad you thought about INTACK,
let's not have another PC-100 Agree on making the V40 generic Re:
interrupts/hardware, that is to say, don't assume TMS320C1X for the V40 board
design and the necessary extras on the DSP board. Thats all folks. Bob

#: 74892 S17/TAPR NNC/DSP
    06-May-88  00:17:22
Sb: #Filters an FIFOs
Fm: Lyle Johnson, WA7GXD 76246,565
To: DSPers

I am running a price check on the S3528P filter as well as getting a couple
samples.  Meanwhile, the MF10 can be set up to be a tracking filter along with
a divider chip for about $3.50 for a fourth order filter function.

Thinking in terms of "options", would it be reasonable to assume the second
analog port to be an option, and include provision for a second 8254 timer chip
and phase shifter to drive it?  Of course, parts wouldn't be included as a
"base" unit.

What if we drop the shareable memory and use a FIFO for buffering an output
stream, and make the FIFO an option as well.  It would only be used for TMS320
to TMS320 communication in most instances anyway (I think).  Anyhow, it would
cost about $20 to have 512 words of FIFO, so I think it would have to be an
option.

Comments?

Lyle

*** There is a reply:   74949

*** More ***

#: 74949 S17/TAPR NNC/DSP
    07-May-88  00:56:16
Sb: #74892-#Filters an FIFOs
Fm: Bob McGwier N4HY 74615,1366
To: Lyle Johnson, WA7GXD 76246,565 (X)

I can't think of an application where I could use the FIFO to share between
processors in the DSP-1.  So long as it is an option, and the extra traces
don't cost $$$, go for it.  I don't think we need shared memory between
processors for the DSP-1 but we will for next generation.  O the modems we can
build with (say) two 56001 or two TMS320C25's.  WOW!

Bob



*** There is a reply:   74983

*** More ***

#: 74983 S17/TAPR NNC/DSP
    07-May-88  19:27:20
Sb: #74949-#Filters an FIFOs
Fm: Lyle Johnson, WA7GXD 76246,565
To: Bob McGwier N4HY 74615,1366 (X)

Dan was thinking of an application where, say, DSP #1 did an FFT and got rid of
the data in a hurry to DSP #2 which did some complcated filtering or ??
algorithm on the FFT data, finally passing the results out to the V40.

Anyway, a couple square inches and a few traces should cost in the region of
dimes.

Lyle

*** There is a reply:   74995

*** More ***

#: 74995 S17/TAPR NNC/DSP
    07-May-88  20:17:00
Sb: #74983-Filters an FIFOs
Fm: Bob McGwier N4HY 74615,1366
To: Lyle Johnson, WA7GXD 76246,565 (X)

I don't object to having the availability of this as an option.  If we don't
hurry up and get to camera ready copy I am afraid the cost of this thing is
gonna be $500.  Don't want that. Bob

#: 74950 S17/TAPR NNC/DSP
    07-May-88  01:30:55
Sb: #A new toy?????
Fm: John Conner WD0FHG 72165,743
To: ALL

Ok ... Since I have not seen any thing new since DSP4 I will just toss this out
and see if any one wants it.  I have a simple 2 channel scope program that
interfaces with the Dalanco board. It can record and view data files ( same as
format as FROM320.C ). Send the data to the 32010 for processing and display
both the input and output signals.  Run the debugger, the assembler, and your
small editor what ever it is.  Works on CGA and (better) EGA. And is full of
great examples of how not to write in C......

The ARC file is about 60K. If anyone is interested say so and I'll try to put
it where you can download it. It should be on winfree tomorrow.  Maybe Bdale
will have better luck trying to get this version to tomcat than the last time
we tried two weeks ago.

Winfree (Bdales UNIX sys) now has a 2400 baud slip link at [phone]. The
Gateway is 44.32.0.8, winfree is 44.32.0.1.

Next weekend is the Rocky Mountain Packetfest.  Andy wants me to say a few
words about DSP (and they will be few).  Just how much of what is going on here
is public?

john

*** There is a reply:   74984

*** More ***

#: 74984 S17/TAPR NNC/DSP
    07-May-88  19:27:27
Sb: #74950-#A new toy?????
Fm: Lyle Johnson, WA7GXD 76246,565
To: John Conner WD0FHG 72165,743 (X)

John,

I tend to treat this stuff as discussions in the back room, not for public
dissemination, etc.  However, may primary concern in this regard is that we
don't tell folks specs (since the design is still being thrashed, much less
tested) dates (boy! has TPR gotten burnt on THAT!) and prices (we od course
will go over budget on this...).

Lyle

*** There are replies:  75007, 75022

*** More ***

#: 75007 S17/TAPR NNC/DSP
    08-May-88  00:09:26
Sb: #74984-A new toy?????
Fm: John Conner WD0FHG 72165,743
To: Lyle Johnson, WA7GXD 76246,565 (X)

That is what I figured. Just wanted confirmation.  As for Crest factors ??? It
is almost understandable to me.  I'm sure I could find two others that might
also (Bdale and Bill Flyn or Jan King...he is talking on Pacsat).

John

*** More ***

#: 75022 S17/TAPR NNC/DSP
    08-May-88  14:57:08
Sb: #74984-#A new toy?????
Fm: Franklin Antonio, N6NKF 76337,1365
To: Lyle Johnson, WA7GXD 76246,565 (X)

Gee... Tom and Bob did give out the specs at Dayton...  They said it would do
EVERYTHING.

*** There is a reply:   75043

*** More ***

#: 75043 S17/TAPR NNC/DSP
    09-May-88  00:15:41
Sb: #75022-#A new toy?????
Fm: Lyle Johnson, WA7GXD 76246,565
To: Franklin Antonio, N6NKF 76337,1365 (X)

Hmmm.

There we go misrepresenting ourselves again.  This one will do ALMOST
everything.  The next generation unit is the one that will do EVERYTHING!  :-)

Lyle

*** There is a reply:   75044

*** More ***

#: 75044 S17/TAPR NNC/DSP
    09-May-88  00:58:17
Sb: #75043-A new toy?????
Fm: Franklin Antonio, N6NKF 76337,1365
To: Lyle Johnson, WA7GXD 76246,565

Thanks for clarification.  By the way, thanks for the hotel room.  Had you been
there, you would have enjoyed it.

[74615,1366]
DSP                       05-May-88 7963               Accesses: 2

    Keywords: DSP SCHEMATIC COMMENTS

    N4HY comments on the DSP-1.  I also included at the end comments from AD7I
    who saw a talk I gave at Trenton Computer festival.


I am going to write my comments on Lyle's latest mailing and this will be
followed by the comments of AD7I, Paul Newland.

On your document dated 20 April and title DSP-1 schematic Notes.

Bullet 2.  It is no great pain to not be able to TBLW the bottom 256 words
in any of the code that I have written.  Just put the ORG for the table at
the end of memory or the things you will be modifying at the end and forget
them.

Bullet 3.  Please try and work up some way, whether is from NKF/HHV diagrams
or from your own design, a way to adjust (IN EITHER DIRECTION) the sample
clock phase.  Thus a swallower isn't adequate, I need a "maker" as well.
It is really the difference in having to significantly oversample, so as to
maintain at least one sample near the maximal opening of the eye and noise
tolerance on clock recovery at its best.

I would prefer pads/traces/NO PARTS and a full second port to the MUX.

Bullets 6&9  I think the high bandwidth port is much preferable to the shared
memory if we have to choose.  Please make sure that we include the traces and
enough glue etc so that (if I wanted to) I could make a card that plugs into
my PC and talk that way to my computer.  I do not envision myself writing
applications where I will need the capability of shared memory on the
TMS320C15 board but I DO see such a need for the DSP-2 where can get real
shared memory and wait state the processors that aren't up on the shared
memory if someone else is accessing them rather than turning the bloody
processor off (320C1X)

Bullet 10 655 uSec is 80% of a bit time at 1200 BPS.  I do not envision this
ever being useful in the stuff we could really use the shared memory with the
V40 for, leave it out altogether.

Bullet 14.  RE: Header connector.  I can see Tom firing up his megawatt
transmitter on 20 meters & smoke and brimstone coming from the DSP-1 from
RF or either from his head as the DSP croaks in this environment just for the
lack of shielded plugs.  You talk about not being able to scrimp later for
similar considerations, don't do it here.

Bullet 15: V40 or bust.

Bullet 17: Hook both 8530 channels so that DMA can be used.  Where we need
to use the asynchronous port on the V40, we will go back to interrupt driven
handling of the second port.

Bullet 18, 1/4 meg is enough and I am glad you are leaving it flexible.

Bullet 20:  I am in favor of the centronics parallel port.

Bullet 21:  I would like to indicate lock (DCD, etc.) directly from the DSP
and also level from it as well.  So make these LEDS come from the DSP board.

B22: I think front panel selectable audio filters are needed as per other
messages

B25: Like your thinking on the interfaces to the radios.  Tell your thoughts
RE: channel occupancy to ARRL HQ who send long transmissions on one frequency
;-)  They would benefit greatly by our HF work so I don't think I would want
to time that out.  Can't this be made flexible?

B27:  YEAH! Switcher! NO GROUND LOOPS THRU P.S.

B28:  The up/down clicker is needed.

B29: I like your thoughts on decoupling!

I am not qualified to comment on your schematics and hope others will do that.
I am limiting my comments that I as a programmer would like to have for the
applications we are talking about being able to support.

THANKS for your efforts.

Bob


>From thumper.bellcore.com!karn  Wed Apr 27 00:31:14 1988 remote from bellcore
Received: from thumper.bellcore.com
        by bellcore.bellcore.com (3.2/smail2.5/07-29-87)
        id AA07479; Wed, 27 Apr 88 00:31:12 EDT
Received: by thumper.bellcore.com (5.54/4.7)
        id AA20926; Wed, 27 Apr 88 00:27:49 EDT
Date: Wed, 27 Apr 88 00:27:49 EDT
From: [e-mail] (Phil R. Karn)
Message-Id: <[e-mail]>
To: idacrd!mac
Subject: FYI - from AD7I

>From bcruucp Tue Apr 26 23:22:26 1988
Received: by thumper.bellcore.com (5.54/4.7)
        id AA20123; Tue, 26 Apr 88 23:22:24 EDT
Received: by bellcore.bellcore.com (3.2/smail2.5/07-29-87)
        id AA06404; Tue, 26 Apr 88 23:25:40 EDT
Date: Tue, 26 Apr 88 23:25:38 EDT
From: homxb![e-mail]
Message-Id: <[e-mail]>
Received: by att-ih.ATT.COM (smail2.5 - att-ih)
        id AA20093; 26 Apr 88 21:48:34 CDT (Tue)
To: bellcore![e-mail]
Status: R

Phil --

I was really impressed by Bob McGwier's presentation on DSP
demodulation techniques at Trenton.  I was also interested
in the new TAPR device that's an upgrade of the PK-232.  I
would like to offer some suggestions for features and
capabilities of the box; most will have already been
included but there may be some others that haven't been
considered.  Let me briefly run through them.

1.   Lots of status LEDs.  One of the things I really enjoy
     on the PK-232 is that there are enough status LEDs to
     tell the user quite a bit about what is going on
     regardless of what mode is in use.  I hope that this
     feature will be retained on the new controller.

2.   To ensure that the box is hardware-capable with AMTOR
     and coherent CW operation, there should be some
     interrupt source that's accurate to +/- 30 ppm (i.e.,
     crystal controlled with a netting trimmer) that can
     interrupt the system at at least one of the following
     rates:  800, 900, 1000, 1100 or 1200 times per second.
     Best choice would be 1000 or 1200 times per second.

3.   Include a printer port and also consider an additional
     (or shared) parallel port for an optional front panel
     1x16 character LCD display.  Perhaps the data port can
     be shared between the printer and the LCD display.
     Such a display might be useful if the box is used for
     SELCALL functions.  I could see the box used on HF SSB
     nets to notify stations that others are looking for
     them.  The dispaly would be useful to provide more
     information than simply that someone has "rung their
     bell", in those cases when a terminal or printer can't
     easily be dedicated full time to the box.

4.   As part of the selective calling encoder/decoder
     function described above, it might also be valuable to
     provide a single bit port to drive an optional buzzer
     or sounder for such SELCALL functions.

5.   You might want to consider providing an 8 pin socket to
     house one of the National 8 pin 5 volt EEPROM devices.
     The processor communicates with the EEPROM via a serial
     interface and most of the leads can be shared with LEDs
     or the printer port.  I did this on a controller that
     I'm working on and it's a real handy device.  The
     advantage of such a part is more for stand alone
     applications where you want to send someone
     configurations already set up (i.e., the user doesn't
     have a terminal).  The EEPROM can be programmed in one
     box, removed from the box, suffed into conductive foam,
     mailed to another person and then installed in his or
     her box.  The second party doesn't need to have a
     terminal to program the device (again, this is useful
     where the box is used only as a SELCALL device).  The
     EEPROM device also seems to be less likely to "forget"
     its contents than some of controllers that opted not to
     use the TNC2 battery circuitry (no, I don't mean to
     imply that using this device would be better than
     battery backed ram -- BBRAM is just too wonderful to
     give up).  Although not of interest to us, the EEPROM
     has great advantages in commerical products, especially
     for distribution of cryto keys to ignorant users.

I might have more thoughts in the future but this is all
that's come to mind in the past few days.  If more comes
along, I will forward it on to you.  Please say hello to the
TAPR crowd at Dayton for me.  Sorry I won't be there to see
everyone.

Paul Newland, ad7i

#: 75185 S17/TAPR NNC/DSP
    11-May-88  15:27:02
Sb: #74834-Schematics
Fm: Barry McLarnon VE3JF 71470,3651
To: Franklin Antonio, N6NKF 76337,1365

Sorry for the delay in replying... been outta circulation for a few days. The
local place with the S3528's is Active Components, an over-the-counter
operation which is part of the Future Electronics distribution empire.  Their
prices are usually nothing to write home about, so I would be very surprised if
you could not find a considerably better price in your area.

Regarding your comments on all the high-order filters present in HF radios: I
didn't have HF applications in mind when I made the comment about the S3528's
Chebyshev response... I see no problem whatever in using them for HF stuff. It
would only be in higher-speed applications where the group delay could be a
problem.  It could be equalized in the DSP chip though.

#: 75295 S17/TAPR NNC/DSP
    13-May-88  05:05:18
Sb: MAC320 ???
Fm: Franklin Antonio, N6NKF 76337,1365
To: Bob McGwier 74615,1366

Still waiting for the documentation.  Figured what the heck, and tried to run
it sans documentation tonite.  I typed the command "MAC320 xxxx.asm", and to my
utter surprise, it tries to read from my empty floppy drive A:. Is this what's
supposed to happen?  What am i missing?

Further Del-Spry question:  Interrupts.  Board can use INT2 or INT7.  These
are, of course, both used on my AT.  How do the rest of you get around this? Do
none of the extant programs use interrupts?


#: 75300 S17/TAPR NNC/DSP
    13-May-88  10:18:04
Sb: #74871-Dan and Lyle Mtg
Fm: Barry McLarnon VE3JF 71470,3651
To: Bob McGwier N4HY 74615,1366

The standard PC Centronics interface IS bidirectional, even though not all of
the individual ports involved are.  For example, the first printer port uses
the I/O addresses 378, 379, and 37A.  The port at 378 is an 8-bit output, which
can't be changed without hacking the board (the output bits can be read
though).  The port at 379 is a 5-bit input, likewise unidirectional.  The port
at 37A is nominally a 4-bit output, but it can also be used as an input since
it can also be read, and the outputs are open-collector drivers.  To use it as
an input, you just have to turn the drivers on first.  There are a number of
software packages around which do bidirectional file transfers through the PC's
parallel ports.  I have one called File Shuttle, and it is remarkably fast.  In
benchmarks done by some of the PC mags a few months back, it was significantly
faster than the transfers done via the serial port at the max rate of 115 kbs.
So a Centronics interface can come in very handy if you do the appropriate
cabling and programming.

#: 75337 S17/TAPR NNC/DSP
    13-May-88  22:05:05
Sb: #75295-#MAC320 ???
Fm: Bob McGwier N4HY 74615,1366
To: Franklin Antonio, N6NKF 76337,1365 (X)

Sorry.  I didn't get it off before I went off to DC for a week.  I am so far
behind on correspondence that it will take me a week to catch up.  Too much
traveling for the company.

If you have the source on drive C

mac320 source.cce

This looks for source.asm on the c drive and puts the hex on the c drive and
doesn't dump the entire listing.

I will mail it tommorrow.

Bob



*** There is a reply:   75362

*** More ***

#: 75362 S17/TAPR NNC/DSP
    14-May-88  01:27:32
Sb: #75337-MAC320 ???
Fm: Franklin Antonio, N6NKF 76337,1365
To: Bob McGwier N4HY 74615,1366 (X)

what's the magic incantation to get a listing?

*** More ***

#: 75357 S17/TAPR NNC/DSP
    14-May-88  01:21:50
Sb: #75295-MAC320 ???
Fm: John Conner WD0FHG 72165,743
To: Franklin Antonio, N6NKF 76337,1365 (X)

I'm not using any interrupts here to the PC ..... However my scope program uses
the interrupt from the PC to the 320.  This is jumper J4.  Have not seen any
one else using interrupts for anything yet.

John

#: 75348 S17/TAPR NNC/DSP
    13-May-88  23:03:43
Sb: #75300-Dan and Lyle Mtg
Fm: Bob McGwier N4HY 74615,1366
To: Barry McLarnon VE3JF 71470,3651 (X)

Now that I read this more carefully, I see what you are talking about.  I am
encouraging the addition of the parallel port as an option.

Bob

#: 75818 S17/TAPR NNC/DSP
    22-May-88  15:39:36
Sb: #75300-Dan and Lyle Mtg
Fm: Phil Karn, KA9Q 73210,1526
To: Barry McLarnon VE3JF 71470,3651

Barry,

The typical PC parallel printer port can be easily modified to support
bidirectional operation.  I did that here to drive my Icom 271/471/751
computer control busses.  I described the modification in the AMSAT
Dallas proceedings over a year ago.

It's easy because the output buffer is a 74LS374, a tri-state buffer
that has its output enable/ pin grounded.  All you do is lift this pin
and connect it to a conveniently spare output bit in the 74LS174 control
latch, which becomes the read/write direction control bit.  (You may
also have to wire the input of this latch to the appropriate data bus
line, it depends on the card).  Then you can read input data with the
data readback feature you mentioned.  Because of the polarity of the
output enable signal, the interface defaults to output mode after a
reset, so it should still work fine as a normal printer port.

The clone printer boards seem to use much the same circuit as IBM.

Phil

#: 76572 S17/TAPR NNC/DSP
    07-Jun-88  09:41:18
Sb: #DSP on AO10
Fm: Alberto E. Zagni I2KBD 71360,3467
To: Bob N4HY 74615,1366 (X)

Dear Bob, a few notes on my DSP work. I have modified your SPECT1K program by
addind: 1) Running average smoothing on every 16 bins 2) Circular window
detector out of 8 FFTs 3) Statistical decision and display for scattering out
of 4 bins. I tried the whole on Oscar 10 and I am able to detect my echos with
less than 100 mW uplink, with signal completely disappearing from my audio at
2-3 W uplink level, while being clearly out of the noise with 50+ W. I have
heard of a DSP4 disk, is it a dream or true ;-) Last question: do you plan to
work on PSK decoder for Phase 3C? (real soon now!) Best regards, Alberto I2KBD

*** There is a reply:   76575

*** More ***

#: 76575 S17/TAPR NNC/DSP
    07-Jun-88  10:28:33
Sb: #76572-#DSP on AO10
Fm: Bob McGwier N4HY 74615,1366
To: Alberto E. Zagni I2KBD 71360,3467 (X)

The decoder for Phase III C 400 bps Biphase PSK is done.  We now have over' 8
diskettes of stuff.  Tom and I just do not have time to duplicate 360 diskettes
for all the members of the team.  We are working out a way for the TAPR office
to send these diskettes out.  Don't feel left out, the other folks over here
are still awaiting these last four diskettes also; other than those who have
actually seen us at some place.  I hope this gets worked out soon.  Send a note
to Tom and remind him also.

Bob



*** There is a reply:   76603

*** More ***

#: 76603 S17/TAPR NNC/DSP
    07-Jun-88  23:30:27
Sb: #76575-DSP on AO10
Fm: John Conner WD0FHG 72165,743
To: Bob McGwier N4HY 74615,1366 (X)

Bob

If you can get the disks to Colorado about the middle of next week, I'll get
the things copied ( or get Andy to ;-> ) and send them out.  Andy has my
CopyIIPC board and was going to use it to do the next NET release. He leaves in
the AM for a week and it would take that long to get the 360 disks anyway.....

John

#: 76909 S17/TAPR NNC/DSP
    17-Jun-88  01:54:49
Sb: DSP HArdware Status
Fm: Lyle Johnson, WA7GXD 76246,565
To: All

The 320C15 section schematic appears to be done (!) and will go to Eric in the
next couple days.  It just has to be plotted.  Dan and Eric and I will get
together to figure out the best way to get the prototype board layout done in
such a way as to get it into the hands of software developers ASAP.  Some sort
of easy way to hang it on a PC bus will be looked at, minimum.

The A/D and D/A filters are based on the S3528 filters used in the superSCAF
filter from QST a couple years back.  The filters turned out to be CHEAP ($3.30
each) and save a lot of grief.  Yes, they will be programmable by the DSP chip.
In fact, now that everything (filters, sample clock, etc.) is under control of
the DSP chip, the old idea of making it hang on a TNC without a V40 processor
as a "better modem" becomes feasible.  This simply means, to me, that the V40
part can be made better since it will be "optional" for certain applications.
This also helps drive down the price of the basic unit if a user elects to
forego the V40 part.

I will post a list of features and etc. in a day or so.  Just wanted you all to
know that things are in fact alive and progressing here.  I am hoping that Eric
will be in layout before the end of next week.

The results of the AMSAT project will probably modify the V40 unit a bit, so
stay tuned...

Lyle

#: 77059 S17/TAPR NNC/DSP
    21-Jun-88  00:46:45
Sb: NKFSA 0.97 release
Fm: Franklin Antonio, N6NKF 76337,1365
To: ALL

I have just uploaded to LIB 17 the latest version (0.97) of my spectrum
analyzer program NKFSA.  This program computes real-time spectra of an input
signal, doing all computation in the PC.  Features are...

1.  Runs on any IBM PC/XT/AT/PS2 or clone with CGA/EGA/VGA & DOS or OS/2 2.
Requires NO coprocessors or DSP chips. 3.  Contains interface for Dalanco-Spry
Mdl 10 board. 4.  Hooks and instructions for user-written interface to any A/D
board. 5.  Supports color on EGA/VGA displays.  Colors user-settable. 6.
Supports mouse, if you have one. 7.  Online Help. 8.  Selectable windowing
(rectangular, Hamming, Blackman) 9.  Special input mode for short data sets (ie
FIR filter evaluation) 10. It even comes with documentation!

#: 77237 S17/TAPR NNC/DSP
    24-Jun-88  23:27:35
Sb: #77156-#NKFSA 0.97 release
Fm: Franklin Antonio, N6NKF 76337,1365
To: Bob McGwier N4HY 74615,1366 (X)

Does Tom read mail on tomcat.gsfc.nasa.gov?  I uploaded latest version (0.97)
of NKFSA to tomcat, and sent him mail asking him to delete prior version (0.90
i think).  Instead, latest version appears to have been deleted.  Never got
reply mail, so i don't know if he goofed, or if the new file i ftp'd got
garbled (as has happened several times before), or if..

*** There is a reply:   77328

*** More ***

#: 77328 S17/TAPR NNC/DSP
    27-Jun-88  03:31:04
Sb: #77237-#NKFSA 0.97 release
Fm: Bob McGwier N4HY 74615,1366
To: Franklin Antonio, N6NKF 76337,1365 (X)

He knows.  He hasn't been there much.  He and I had a weekend working meeting
on the west coast.  He has been taking time off to finish the PACSAT receiver
and the module command/control board (personal days).  This means he hasn't
been at work where TOMCAT is located.  He does know that your latest ftp didn't
make it or something like that.

Bob

#: 77332 S17/TAPR NNC/DSP
    27-Jun-88  05:25:43
Sb: #NKFSA 0.97 SW
Fm: Alberto E. Zagni I2KBD 71360,3467
To: Franklin Antonio N6NKF 76337,1365 (X)

Dear Franklin, I downloaded your NKFSA 0.97 program that is great! On my PS/2
model 70 with color display is very nice. I also tried on my DSP system based
on 12 MHz AT with coprocessor, Dalanco/Spry board and EGA display to show Oscar
13 PSK beacon:super ;-) Best 73 and many thanks from Italy Alberto I2KBD

*** There is a reply:   77342

*** More ***

#: 77342 S17/TAPR NNC/DSP
    27-Jun-88  15:10:34
Sb: #77332-NKFSA 0.97 SW
Fm: Franklin Antonio, N6NKF 76337,1365
To: Alberto E. Zagni I2KBD 71360,3467

Thank you for the kind words.  We're all looking at the OSCAR 13 beacon lately!


#: 77333 S17/TAPR NNC/DSP
    27-Jun-88  05:36:42
Sb: EME DSP!
Fm: Alberto E. Zagni I2KBD 71360,3467
To: Bob McGwier N4HY 74615,1366

Dear Bob, do you survive on holiday? (i'm in my office now...) Any news of Tom,
I left him a message more than a week ago, but he didn't read it. The disks
also seems to have been eaten by some black hole... FINALLY I GOT MY ECHOS FROM
THE MOON WITH SPECT_KBD SOFTWARE!!!! The last version of modified SPECT1K now
keys PTT thru parallel port for 2 seconds, than FFT for another 2 seconds for
any echos, adding the resulting spectra and then FFT another 2 seconds to
subtract any anwanted signal around. The peak from the moon, with auto doppler
adapting is clear. Conditions were 120W on 433.010 and 2 yagis. I will send you
the final version ASAP. Best 73, Alberto I2KBD

#: 77362 S17/TAPR NNC/DSP
    28-Jun-88  00:05:28
Sb: #77329-#NKFSA 0.97 release
Fm: Bob McGwier N4HY 74615,1366
To: Franklin Antonio, N6NKF 76337,1365 (X)

My office mate (Lloyd Welch) told me that Viterbi was working at Qualcomm.
True?

Bob

*** There is a reply:   77395

#: 77395 S17/TAPR NNC/DSP
    28-Jun-88  14:37:33
Sb: #77362-#NKFSA 0.97 release
Fm: Franklin Antonio, N6NKF 76337,1365
To: Bob McGwier N4HY 74615,1366 (X)

Yes.  Andy Viterbi was one of the founders of LINKABIT in 1969.  I joined
LINKABIT in 1972.  LINKABIT was sold to M/A-COM in 1979.  Became M/A-COM
LINKABIT.  Name then changed to M/A-COM Telecommunications. Mid-1985, seven of
us left to form QUALCOMM.  About that time, what was LINKABIT was split into
three pieces.  The VideoCipher division was sold to General Instruments.  The
remaining commercial stuff was sold to Hughes, and became Hughes Network
Systems.  The remaining military stuff is now M/A-COM Government Systems.
QUALCOMM began with seven people in July '85.  Irwin Jacobs, Andy Viterbi, Andy
Cohen, Harvey White, Franklin Antonio, Klein Gilhousen, & Dee Coffman.  Now
have approx 150 employees.

*** More ***

#: 77363 S17/TAPR NNC/DSP
    28-Jun-88  00:05:43
Sb: #77333-EME DSP!
Fm: Bob McGwier N4HY 74615,1366
To: Alberto E. Zagni I2KBD 71360,3467 (X)

Alberto:

I am sorry, I wish this were a holiday.  This is grinding work out here in this
last bastion of slave labor, sunny southern California ;-).  I will be here on
this project all summer.  Tom swears that the diskettes will be done real soon
now.   I wish I could help you.  I just don't even have all the latest stuff.
I don't keep a copy of all that stuff on my hard disk and I didn't bring any of
my floppy collection with me.

GREAT on the spectrum analyzer.  It will soon be time for us to hang a neat set
of detection statistics on the bins and have a real QSO.

Bob

#: 78499 S17/TAPR NNC/DSP
    20-Jul-88  03:36:35
Sb: DalancoSpry RFI
Fm: Franklin Antonio, N6NKF 76337,1365
To: Bob McGwier 74615,1366

Bob, you mentioned that you initially had some RFI problems with the
Dalanco/Spry board.  You didn't mention what you did to fix 'em. I have
recently noticed that i can hear the D/S board at certain places on 2m and 10m,
but only when it runs certain software.  I don't know what s going on yet.  Any
info on your RFI experiences with the D/S board appreciated.

#: 78531 S17/TAPR NNC/DSP
    21-Jul-88  01:30:39
Sb: #78499-#DalancoSpry RFI
Fm: Bob McGwier N4HY 74615,1366
To: Franklin Antonio, N6NKF 76337,1365 (X)

Sure, I have NEVER completely removed them all.  First I did not allow the
grounds of the connectors to touch the computer since those grounds float. Do
you know Courtney Duncan?  He has reverse engineered schematics for us on
these.  The only other thing I did was to put toroidal cores everywhere and to
finally tempest . .  errrrr I mean EMI proof the CPU with conductive paint in
the inside and scrape the paint off the edges where the sliding top touches the
base, etc.  The usual tricks.  The worst problem for me is the harmonic of the
320 CPU clock that occurs in the middle of the WX satellite band.  If I
remember anything else I have done, I will pass it along.  WE ARE NOT GOING TO
REPEAT THIS with the DSP-1.  Quietness is a must.

Bob

*** There is a reply:   78544

*** More ***

#: 78544 S17/TAPR NNC/DSP
    21-Jul-88  12:49:11
Sb: #78531-DalancoSpry RFI
Fm: Franklin Antonio, N6NKF 76337,1365
To: Bob McGwier N4HY 74615,1366 (X)

I have torroids on absolutely everything here.  The case of my PC has
reasonable looking RF gasketing, (it's a real IBM-AT).  I've been thinking of
putting the Dalanco/Spry on an extender & probing around to find the source of
the noise.  When i run NKFSA, it is audible at certain places on 2m & 10m, and
looking at the audio output of the radio on a scope, it looks like a little
spike once per NKFSA cycle.  I wonder if this has anything to do with the fact
that i halt the TMS320 once per cycle to slurp out the sample buffer?  I wonder
also whether i have the energy to go after the problem!

#: 78785 S17/TAPR NNC/DSP
    26-Jul-88  00:40:32
Sb: New DSP Disks
Fm: John Conner WD0FHG 72165,743
To: ALL

Andy, N0CCZ, and I are preparing to ship a new set of DSP disks late this week.
From a quick look they appear to contain all of the files on TOMCAT. If you
have any thing that you would like to include in this shipment please get the
files to me by midnight Wed.  You can put them on CIS, or somewhere accessable
via UUCP or ARPA, just tell me so I can get Bdale to go get them. (TOMCAT is
not a good place at the moment the connection from COS seems to have diodes in
it. )

John

#: 78892 S17/TAPR NNC/DSP
    29-Jul-88  00:21:40
Sb: DSP Disks
Fm: John Conner WD0FHG 72165,743
To: Alberto E. Zagni I2KBD 71360,3467 (X)

Alberto:

I was told that you may see Tom, W3IWI, this weekend.  If so please tell him
that the DSP Disks will go in to the mail late Friday or first thing Sat. There
will be 13 disks.  The 12 he sent to Andy and one that I Created. Since he
didn't send a corrected mailing list we are sending the disks to those listed
as #1.xx in the Jan/Dec Team Players list with some corrections received by
Courtney, N5BF.  ( Yes, you get a set even if he brings you one.)  Anyone else
that should get one will have to wait until he gets back and can give us the
addresses. Looking forward to seeing the changes you have made for the weak
signal work.  I will be sending out updated disks about every 90 days or so.
The next one will be about Oct. 1.  I hope we can shake some more stuff out of
people by then.

John, WD0FHG

#: 78973 S17/TAPR NNC/DSP
    31-Jul-88  12:31:00
Sb: #DSP Software
Fm: John Conner WD0FHG 72165,743
To: ALL

Friday, July 29, Andy and I mailed 22 sets of disks to memberes of the DSP
Group who have the Dalanco board.  There are 13 disks in the set and they
contain all the DSP software (I think) on TOMCAT.

There are 8 sets still to be shipped that needed addresses.  These will go out
as soon as we get some more disks ( Mon or Tue ). The following people will be
getting their disks late; W5SXD, KT7D, N4QQ, KN6W, NN2Z, KA0YRE, and KB5MU.

I will post a complete mailing list later this week. The list is only the
people who got disks this time.  If corrections are needed let me know.

LOOK FOR YOUR DISK IN THE MAIL EARLY THIS WEEK!!!!!

John

*** There are replies:  78999, 79079

*** More ***

#: 78999 S17/TAPR NNC/DSP
    31-Jul-88  17:09:11
Sb: #78973-#DSP Software
Fm: Franklin Antonio, N6NKF 76337,1365
To: John Conner WD0FHG 72165,743 (X)

Oh no!  If you ship KB5MU's disks late, he'll want copies of mine! Here's
KB5MU's address:  Paul Williamson, 6583 Edmonton Ave, San Diego, CA, 92122.

*** There is a reply:   79042

*** More ***

#: 79042 S17/TAPR NNC/DSP
    01-Aug-88  10:44:26
Sb: #78999-#DSP Software
Fm: John Conner WD0FHG 72165,743
To: Franklin Antonio, N6NKF 76337,1365 (X)

Franklin:

DON'T LET HIM HAVE COPIES!!!!|->

His will go out in todays mail.  Let me know when yours arrive. We sent them on
a first class snail.

John

*** There is a reply:   79058

*** More ***

#: 79058 S17/TAPR NNC/DSP
    01-Aug-88  16:05:22
Sb: #79042-DSP Software
Fm: Franklin Antonio, N6NKF 76337,1365
To: John Conner WD0FHG 72165,743 (X)

Mine arrived about 15 minutes ago.  Nice box.

*** More ***

#: 79079 S17/TAPR NNC/DSP
    02-Aug-88  00:10:13
Sb: #78973-DSP Software
Fm: Lyle Johnson, WA7GXD 76246,565
To: John Conner WD0FHG 72165,743 (X)

John,

I spoke with Mike Parker, KT7D, on the phone today.  Please mail his diskettes
to:

Mike Parker, KT7D 8525 East Cloud Road Tucson AZ  85715

Cheer,

Lyle

#: 79025 S17/TAPR NNC/DSP
    01-Aug-88  02:55:13
Sb: #DSP Hardware
Fm: Lyle Johnson, WA7GXD 76246,565
To: DSPers

Seeing the software message prompts me to let you know the status of the
hardware!

A lot of energy has been expended lately in the PACSAT CPU project, but TAPR
has acquired an AT clone for PC layout and it is now at Chuck Green's (N0ADI)
house.  Chuck will be doing board layout with guidance from Eric, N7CL.

While Chuck and Eric are busy with the 320C15 board, I will try my hand at
doing the V40 board in parallel so maybe we can have prototype hardware in
about 6 weeks!  Maybe sooner!

Meanwhile, has anyone tried masking the A/D and D/A on the DS-10 boards to
simulate an 8 bit converter?  What are the results in modem applications like
we'll see in packet?  Surely someone has tried this!

Please let me know before the 8-bitter gets cast in concrete (the mud is
starting to set!).

Lyle

*** There is a reply:   79041

*** More ***

#: 79041 S17/TAPR NNC/DSP
    01-Aug-88  10:44:18
Sb: #79025-#DSP Hardware
Fm: John Conner WD0FHG 72165,743
To: Lyle Johnson, WA7GXD 76246,565 (X)

Lyle:

Is Mike Parker's address X9 Ranch, Vail, AZ 85614?  And does anyone else down
there need a set of disks?

John



*** There is a reply:   79080

*** More ***

#: 79080 S17/TAPR NNC/DSP
    02-Aug-88  00:14:06
Sb: #79041-#DSP Hardware
Fm: Lyle Johnson, WA7GXD 76246,565
To: John Conner WD0FHG 72165,743 (X)

John,

I could use a set!  My address is Lyle Johnson, WA7GXD, 9991 East Morrill Way,
Tucson AZ  85749.

I am interested so I can get an idea of program sizes and hardware resources
used, that sort of thing.

Spoke with CHuck tonight and next week I think we'll be cookin' with gas on the
PC board layout!

Cheers,

Lyle

*** There is a reply:   79097

*** More ***

#: 79097 S17/TAPR NNC/DSP
    02-Aug-88  11:29:15
Sb: #79080-DSP Hardware
Fm: John Conner WD0FHG 72165,743
To: Lyle Johnson, WA7GXD 76246,565 (X)

Lyle:

Will add you to the list.  Have about 5 sets still to go out. Andy is waiting
for a disk shipment so we can finish up.

Glad to here the pc board is starting to move.  One that I'm doing for MONEY is
stuck at the moment. I have developed a bug in EEIII that keeps me from
plotting schematics.  Guess I will reload everything and try again.

John

#: 79111 S17/TAPR NNC/DSP
    02-Aug-88  13:10:28
Sb: DSP Disks - notes and ma
Fm: John Conner WD0FHG 72165,743
To: ALL

I have uploaded ARC files of both the mailing list used for the DSP Disks and
the notes file from disk DSP-0.  If you were sent disks and they do not show up
(most in the US should have them by now), let me know.

John

#: 79230 S17/TAPR NNC/DSP
    04-Aug-88  01:43:16
Sb: DSP disks
Fm: Bob McGwier N4HY 74615,1366
To: ALL

Thanks to John Conner and Andy, everyone should have DSP disks by now in the US
and soon outside (vagaries of the mail).  I am really pretty proud of these
diskettes that the Colorado folks have taken the trouble to put together!!  The
overseas folks who don't get to see Tom or I three or four times a year (or
should I say suffer through seeing us ;-).  Lyle is really cracking the whip
down there in Tuscon and making me feel a little behind so I think it is time
to crank up here again also and tell you what I have been doing and thinking
about for our software needs.  I am thoroughly disgusted with the FFT routines
we sent out with round one and with those that came from Delanco Spry.  They
are radix 2 single butterfly routines, and we are not using any tricks to
compactify real signals into half size complex FFT's in order to get savings.
I have written a radix 8, two butterfly and a radix 4 three butterfly (i.e.
partially unrolled in the best places to do so) and they are significantly
faster than the routines originally shipped. I got disgusted when Franklin's
routines on a 2.5 MIPS processor (my PC-AT clone) held its own (just a little
slower) than the 6.25 MIP optimized processor and the FFT routines we had done
there.  The reasons are now quite clear and thanks to Franklin for the kick in
the ass.  In addition, Alberto's recent success on satellite station EME with
those lousy routines we have now, forced me to think more about that and other
weak signal routines.  I have the great honor of having some of the most
incredible people working for me. Lloyd Welch, Dick Goldstein, Ed Posner, Vince
Poor, and Maureen Quirk are all here working on a project that I am the nominal
leader of.  These are some of the best information theorists and signal
detection and estimation people in the world.  They have given me great aid in
figuring out what to do with the crap that gets sent back over EME paths and
weak signal paths on VHF.  In addition, Tom wrote me a treatise before he left
for Italy. Goldstein is the first person to image Venus by radar using
interferometry


#: 79231 S17/TAPR NNC/DSP
    04-Aug-88  01:43:38
Sb: #DSP disks II
Fm: Bob McGwier N4HY 74615,1366
To: All

series edited by John Thomas of Princeton) will become a standard in that
field.  He and Lloyd are my office mates.  How can you go wrong with that kind
of talent around?  At any rate, I am thinking hard now about how the
statistical decision algorithms should work for the detection and then
estimation on the SSLLOOWWW FSK.  After I put the final clean up on the new FFT
routines and document them as I have the piii.asm and the psk modem code on the
diskettes, then I will send them in for diskette 13. A huge development effort
is going to have to take a back seat to PACSAT for which I have BBS and
housekeeping responsibilities (Harold is kernal and Skip is I/O) and the first
available day for launch is Jan. 1.  I thought I would stimulate some
conversation here now that John has sent stuff out.

CHEERS!

Bob



*** There is a reply:   79317

*** More ***

#: 79317 S17/TAPR NNC/DSP
    05-Aug-88  19:08:47
Sb: #79231-#DSP disks II
Fm: David Toth VE3GYQ 72255,152
To: Bob McGwier N4HY 74615,1366 (X)

Bob, did those guys write a book re radar imaging of Venus. I seem to have
misplaced such a text someone in my ultra-neat den. Just wondered if they
authored such a beastie ...
73, Dave

*** There is a reply:   79326

*** More ***

#: 79326 S17/TAPR NNC/DSP
    05-Aug-88  19:47:25
Sb: #79317-DSP disks II
Fm: Bob McGwier N4HY 74615,1366
To: David Toth VE3GYQ 72255,152 (X)

Not sure.  Dick Goldstein was the first to do so along with Howard Rumsey as
the resident mathematician and programmer (he is also in our little group this
summer).

Bob

#: 79473 S17/TAPR NNC/DSP
    08-Aug-88  21:23:09
Sb: #79461-Tuesday Directions
Fm: Bob McGwier N4HY 74615,1366
To: Franklin Antonio, N6NKF 76337,1365 (X)

I will be leaving here at 12 Noon driving to Jack's house where I will pick up
Jack and Tom KA6SOX.  I need to give the G3RUH modems to Tom for immediate
testing and need to talk to him about that and thought we could do that on the
trip down.  I am hoping that between 12 and 2 the freeway is passable ;-) .  I
don't plan to stay long at EJR's QTH on the way down since we are coming there.
BTW, I FINALLY got the XREF, XLINK, etc. and that will make my life much
easier.  To bad xlink can't seem to use MS libraries or at least I haven't
figured out how to make it do that yet.  This will make code development for
the PS-186 much easier as well (I have it with me).  I have some very nice
prints of our CPU activities here in LA and will be showing them off.  I simply
have no way of judging the ETA but will make my way at best speed from El
Segundo on the 405 to Jack's(he lives just off I-5).

See you tommorrow.

Bob

#: 79553 S17/TAPR NNC/DSP
    10-Aug-88  16:38:30
Sb: TMS320E15 burning
Fm: Barry McLarnon VE3JF 71470,3651
To: John Conner WD0FHG 72165,743

John, in the DSP group notes you mailed out recently, you mentioned the need
for an el cheapo E15 EPROM burning adaptor.  I have such a beastie sitting here
in front of me... we put it together at work when we got some of the early
samples of the E15 and couldn't get an adapter from TI.  It consists of a small
chunk of perfboard ("universal" IC type) with a 40-pin ZIF socket mounted on
top, and two 14-pin rows of headers underneath which can drop into the 28-pin
socket of the PROM burner.  Wiring diagram for the adapter is available from
the TI literature... we used wirewrap wire (soldered) for our "one of" adapter,
but a PC layout should be duck soup for someone with the facilities.

Hope this is useful to you.

73, Barry

#: 79554 S17/TAPR NNC/DSP
    10-Aug-88  16:42:22
Sb: S17 Mini-directory
Fm: Barry McLarnon VE3JF 71470,3651
To: All

For your convenience and edification, here is a mini-directory of the folks
who have S17 access, reproduced from the message headers (if you aren't
listed here, it's probably because you've never left any messages in S17).
Please bring any errors or omissions to my attention.  I will re-post the
list as changes occur.

Tom Clark W3IWI 71260,3640
Jon Bloom, KE3Z 71350,1100
Bob McGwier N4HY 74615,1366
Courtney Duncan, N5BF 72617,2623
Paul Williamson, KB5MU/6 75265,367
Mike Brock WB6HHV 76246,546
Harold Price, NK6K 71635,1174
Franklin Antonio, N6NKF 76337,1365
skip wb6ymh 72345,1725
Lyle Johnson, WA7GXD 76246,565
Phil Karn, KA9Q 73210,1526
ANDY FREEBORN  N0CCZ 73177,1317
John Conner WD0FHG 72165,743
David Toth VE3GYQ 72255,152
Barry McLarnon VE3JF 71470,3651
Alberto E. Zagni I2KBD 71360,3467

#: 79558 S17/TAPR NNC/DSP
    10-Aug-88  20:08:23
Sb: #DSP Disks
Fm: John Conner WD0FHG 72165,743
To: Bob McGwier N4HY 74615,1366 (X)

Bob:

All of the disks have been sent.  DB2GM's went out Monday. P.O. said it should
arrive by next Monday!  If you can get a message to him it might be worth
while.

Thanks for returning the call.  Sorry I was not at home.  By monday you had
answered the questions anyway.

John

*** There are replies:  79565, 79584

*** More ***

#: 79565 S17/TAPR NNC/DSP
    10-Aug-88  22:18:21
Sb: #79558-DSP Disks
Fm: Bob McGwier N4HY 74615,1366
To: John Conner WD0FHG 72165,743 (X)

GREAT!

Thanks a WHOLE bunch from Tom and I and I hope the rest of the folks find time
to say thanks as well.

Bob



*** More ***

#: 79584 S17/TAPR NNC/DSP
    11-Aug-88  00:17:32
Sb: #79558-DSP Disks
Fm: Lyle Johnson, WA7GXD 76246,565
To: John Conner WD0FHG 72165,743 (X)

John,

The disks came on Monday.  Whew!  I appreciate your efforts in sending them
out.

Eric and Chuck and getting together Friday PM to get things started on the
hardware end of things -- layout.  I'll be cooking on the schematics this
weekend so we should be into board layout next week.

Lyle


#: 79559 S17/TAPR NNC/DSP
    10-Aug-88  20:08:32
Sb: #FIR Code
Fm: John Conner WD0FHG 72165,743
To: Franklin Antonio, N6NKF 76337,1365 (X)

Franklin:

How much of the code from the Parks and Burris book do you have in electronic
form?  Do you think it would be worth while to try to get permission to
distribute it from the authors or the Publisher?

Any Ideas for Courtney's DSP Test Bench. ( NO I have not volunteered ... but I
might ).

Also, is the hard copy part of NKFSA working? It doesn't seem to here. But, I
may be doing something wrong.

John



*** There are replies:  79589, 79605

*** More ***

#: 79589 S17/TAPR NNC/DSP
    11-Aug-88  02:29:15
Sb: #79559-FIR Code
Fm: Franklin Antonio, N6NKF 76337,1365
To: John Conner WD0FHG 72165,743 (X)

The only thing i've typed in from Parks & Burris' book is the one program i
mentioned, (which they call by a different name btw).

The hard copy stuff in NKFSA works here & at KB5MU's place.  Describe what
you're doing, and the configuration there.

*** More ***

#: 79605 S17/TAPR NNC/DSP
    11-Aug-88  13:56:37
Sb: #79559-#FIR Code
Fm: Franklin Antonio, N6NKF 76337,1365
To: John Conner WD0FHG 72165,743 (X)

My previous reply was very terse.  Was talking on phone & radio while
composing. Yes, i think it would be worth while to get permission to distribute
the version of EQFIR (that's what they call the FIR program i typed in.  i
called it MACPARK which seemed more mnemonic) from the Parks & Burrus book.
This version of EQFIR is easier to use than the old IEEE DSP book version which
N4HY distributed.  If you would take on obtaining the permission required, that
would be great.  I would then upload source & executable immediately. There are
many cases where EQFIR produces a bogus result, or fails to converge, and the
stuff that prints out at the end of the EQFIR run doesn't really tell you
whether it succeeded really.  That's why it's nice to pop the result into NKFSA
and observe the frequency response.

Parks is prof at Cornell & Burrus is Chairman of EE & CS at Rice U.

I didn't understand your reference to Courtney's test bench.  Did i miss
something?

*** There is a reply:   79632

*** More ***

#: 79632 S17/TAPR NNC/DSP
    11-Aug-88  23:25:33
Sb: #79605-FIR Code
Fm: John Conner WD0FHG 72165,743
To: Franklin Antonio, N6NKF 76337,1365 (X)

Franklin:

Will see what I can do on getting the necessary go ahead.  Will try a simple
landline approach first. May have to go to the publisher.  We shall see....

The Test Bench is an Idea Courtney tossed out in a letter he sent with the
Schematics of the Dalanco board.  If you didn't get a copy, and your name was
not on the list he sent me when I started this, you or I or someone should ping
him to get you a copy.  His idea is a set of programs and notes to make general
use of of the Dalanco board.  Something along the lines of my DSPSCOPE program
with nice front end and anything else useful ( like NKFSA and MACPARK.......).
Also, info about test probes, external filters and amps.  Basicly, a general
purpose scope/sa/filter..... unit for the shack.

I still use the DSPSCOPE program but it sure needs a lot of work.  Scaling and
zero shifting of each channel, swaping input channels to the board for chaining
processing, mouse controls, data window selection (it takes a while to process
64k) and on and on.......

Will let you know about the programs as soon as I can find something out.

John

samples).

#: 79595 S17/TAPR NNC/DSP
    11-Aug-88  08:43:00
Sb: #79553-TMS320E15 burning
Fm: John Conner WD0FHG 72165,743
To: Barry McLarnon VE3JF 71470,3651 (X)

Barry:

If I can get the plotting routines to run again on EEIII I'll take a stab at a
layout.  I have the TI docs. Did you need to make any changes??

I have been using the TI module and a JDR Prom Burner with out any problem's on
a "commerical project" (maybe some money someday).  So far it will decode and
transmitt the test signal I included with DSPSCOPE.  I still need to do some
work on the bit detection parts but .....

Maybe I can get back to DSPSCOPE here in a week or so.  I actually use it a
fair bit and need to make a bunch of changes. Still it does let me play with
recorded data.

John

#: 79621 S17/TAPR NNC/DSP
    11-Aug-88  21:36:25
Sb: #Manchester FSK on PACSAT
Fm: Bob McGwier N4HY 74615,1366
To: All

Franklin, Mike, and others:

The Japanese had it well thought out to begin with.  They wanted people to be
able to send the modulated data with these %&*^(*&^&$%^*)(&(*&^^%$ PM
transmitters we have as well as true FM.  It we dumpl4pe manchester encoding,
we will get next to NADA out and we get few users.  I think none of us really
wants that.   Lyle has described his tests and without exception he answered
all my questions in a way that leads me to believe that we have missed
something about sending both Manchester and NRZI to a chip that expects
Manchester and NRZ in that if you are half a bit time off you don't get hosed
or either they don't really understand how their chip works.  I haven't sat
down and run through a bunch of state transitions but I will.

The real bottom line is that we can't live without the manchestering and this
forces us to take it into account in the fire code detector for the spacecraft.
I know that wasn't the answer you wanted to hear but I am giving JUST THE FACTS
MAAM.

Bob



*** There are replies:  79639, 79699

*** More ***

#: 79639 S17/TAPR NNC/DSP
    12-Aug-88  05:02:35
Sb: #79621-#Manchester FSK on PACSAT
Fm: Franklin Antonio, N6NKF 76337,1365
To: Bob McGwier N4HY 74615,1366 (X)

This became clear to me after i spoke with you.  The manchester encoding does
two things of benefit.  First, it makes the output of the modem able to go thru
the audio input stages of radios, which are not DC coupled.  Second, it because
every bit is DC balanced, it makes a simple AC coupled slicer on the output of
the discriminator work nicely.  OK.  I can deal with this.

Now, can u explain why they put manchester encoding on the RUDAK downlink?

*** There are replies:  79693, 79708

*** More ***

#: 79693 S17/TAPR NNC/DSP
    13-Aug-88  01:15:44
Sb: #79639-Manchester FSK on PACSAT
Fm: Tom Clark W3IWI 71260,3640
To: Franklin Antonio, N6NKF 76337,1365 (X)

RUDAK is Manchester only when it runs the 400 bps "Deutschland uber alles"
format. This is precisely the AMSAT-DL PSK command/telemetry standard.
According to Hans Peter Kuhlen (who built it) RUDAK also has a mode where it
runs non-Manchester NRZI 1200 bps PSK with the same standard used on JAS-1, the
same standard our PACSATs will use on the downlink, the same standard supported
by both the TAPR and G3RUH PSK modems, and the same standard supported by
N4HY's DSP demod. I hope that they won't use the 400 bps standard, but we have
little control over what they do.

*** More ***

#: 79708 S17/TAPR NNC/DSP
    13-Aug-88  10:56:22
Sb: #79639-Manchester FSK on PACSAT
Fm: Bob McGwier N4HY 74615,1366
To: Franklin Antonio, N6NKF 76337,1365 (X)

NO I can't other than Karl decided that it would be easier to recover clock. In
fact RUDAK can generate 1200 bps PSK and I hope that it goes to this soon!

Bob



*** More ***

#: 79699 S17/TAPR NNC/DSP
    13-Aug-88  03:11:18
Sb: #79621-#Manchester FSK on PACSAT
Fm: Lyle Johnson, WA7GXD 76246,565
To: Bob McGwier N4HY 74615,1366 (X)

Bob, Mike, Franklin,

see my upload on telemail tonite on harris 6409. bottom line is that it will do
the job jus' peachy keen...

Lyle

*** There is a reply:   79731

*** More ***

#: 79731 S17/TAPR NNC/DSP
    13-Aug-88  15:15:13
Sb: #79699-Manchester FSK on PACSAT
Fm: Franklin Antonio, N6NKF 76337,1365
To: Lyle Johnson, WA7GXD 76246,565 (X)

Franklin isn't on telemail.  No problem there, 'cause Mike's the guy who was
looking into Harris 6409.  However, i don't think Mike's on telemail either.

#: 79690 S17/TAPR NNC/DSP
    13-Aug-88  01:15:24
Sb: #79554-S17 Mini-directory
Fm: Tom Clark W3IWI 71260,3640
To: Barry McLarnon VE3JF 71470,3651 (X)

You forgot Rick Hambly, WB2TNL/3

#: 79924 S17/TAPR NNC/DSP
    16-Aug-88  10:13:33
Sb: #79595-TMS320E15 burning
Fm: Barry McLarnon VE3JF 71470,3651
To: John Conner WD0FHG 72165,743

Nope, we didn't require any changes from the wiring diagram given in the TI
docs.  Support for the ROM protect feature is not shown in the docs, but we
didn't need it anyway.

Barry

PS: Let me add my thanks for the super job you and Andy are doing... FB!


#: 79925 S17/TAPR NNC/DSP
    16-Aug-88  10:13:52
Sb: #79690-S17 Mini-directory
Fm: Barry McLarnon VE3JF 71470,3651
To: Tom Clark W3IWI 71260,3640

Do you have Rick's UID handy?  I knew of his existence, but have never seen a
message from him.

#: 79958 S17/TAPR NNC/DSP
    17-Aug-88  01:06:10
Sb: #79925-#S17 Mini-directory
Fm: Tom Clark W3IWI 71260,3640
To: Barry McLarnon VE3JF 71470,3651 (X)

Sorry -- I seem to have trashed the scrap of paper that had Rick's ID. I'll
give him a "tickle" to log in. I did send him a copy of TAPCIS. He has good
access to TOMCAT thru the SLIP link.

One more thing: I seem to have lost the "DSP Annals" disk you gave me at
Dayton. I was going to merge it into the stuff I sent to John/Andy but
alas no luck. I think maybe WA4ONG got off with it when we were juggling
stuff between our cars coming back from Dayton but who knows.

Any progress on HF modem software. You've been very quiet lately on the
topic! What with the PACSAT launch in January, I just haven't had much time
to bug people to keep the DSP work coming. Consider yourself bugged!

FYI -- mailed you a copy of MBL 5.12 Monday AM. Given the US/Canadian postal
interface, you should see it before autumn officially begins.


*** There is a reply:   79979

*** More ***

#: 79979 S17/TAPR NNC/DSP
    17-Aug-88  10:28:49
Sb: #79958-S17 Mini-directory
Fm: John Conner WD0FHG 72165,743
To: Tom Clark W3IWI 71260,3640

Tom -- Barry

I am considering putting TAPCIS on the OCT version of DSP-0. Think it would be
worth the space????  I will put the directory of UIDs on it in any case.

Tom are there any others who should get copies of the DSP disks?

John




#: 79959 S17/TAPR NNC/DSP
    17-Aug-88  01:06:17
Sb: #79924-#TMS320E15 burning
Fm: Tom Clark W3IWI 71260,3640
To: Barry McLarnon VE3JF 71470,3651 (X)

FYI -- a 320-15 (non eprom type) worked just fine in the D-S board with no
changes needed.



*** There is a reply:   79980

*** More ***

#: 79980 S17/TAPR NNC/DSP
    17-Aug-88  10:29:00
Sb: #79959-TMS320E15 burning
Fm: John Conner WD0FHG 72165,743
To: Tom Clark W3IWI 71260,3640

Barry:

I'm starting to do a layout of the adapter today.  Will probably have artwork
by the weekend.  Will try to get some (4-10) boards made up in the next week or
two. Will send you one to try when we get them.

John

#: 79981 S17/TAPR NNC/DSP
    17-Aug-88  10:29:07
Sb: #DS Schematics
Fm: John Conner WD0FHG 72165,743
To: Franklin Antonio, N6NKF 76337,1365 (X)

Franklin:

Did Courtney ever get you a copy of the Dalanco schematics?

John



*** There is a reply:   79989

*** More ***

#: 79989 S17/TAPR NNC/DSP
    17-Aug-88  12:26:55
Sb: #79981-DS Schematics
Fm: Franklin Antonio, N6NKF 76337,1365
To: John Conner WD0FHG 72165,743

Nope.  I have a copy of their preliminary version which i got from KB5MU some
time ago, but i've never rcvd a mailing from Courtney.  It may be that Courtney
hasn't logged in to CI$ since he asked for my address a couple of weeks ago.

#: 79990 S17/TAPR NNC/DSP
    17-Aug-88  12:31:05
Sb: Direct Digital Synth
Fm: Franklin Antonio, N6NKF 76337,1365
To: Tom Clark 71260,3640

Way back at the TAPR meeting in February, you asked me for a copy of the
QUALCOMM DDS chip data sheet.  I told you they weren't final yet & i'd get you
a copy.  When Bob McGwier was in San Diego last week, i gave him two copies.
One was for you.  Chips now available.  New product announcements have already
appeared in a couple of mags. Advertisements to appear soon.

#: 80121 S17/TAPR NNC/DSP
    19-Aug-88  10:04:36
Sb: #79958-S17 Mini-directory
Fm: Barry McLarnon VE3JF 71470,3651
To: Tom Clark W3IWI 71260,3640

OK, I'll ship you another copy of the CIS DSP disk, with a post-Dayton update.
My long-awaited direct access to ARPAnet still hasn't come through, so I'll
have to resort to the postal system one more time.  I'm looking forward to
trying out MBL 5.12.  Some folks have pressed me to try RLI, but a guy's gotta
have some brand loyalty ("I like the flavor of this new MBL Lite...").

Excuses, excuses... I've been quiet on the topic of HF modems for the plain and
simple reason that I haven't accomplished much lately.  DSP has taken a back
seat to a flock of non-ham projects (geez, I gotta get my priorities
straightened around :-), avoiding marital discord, and getting those 56 kbs
modems on the air.  I also lost the host system I was using (reclaimed by my
xyl so she could do some work) and had to stuff the D-S board into my BBS PC,
making it very awkward to use (I never did get EEMS to work in that machine). I
do have most of the code for an m-ary FSK modem written, and hope to have to
have something to start testing on-air before the snow flies. [Note I did not
specify where the reference location for snow flying might be :-)].

73, Barry

#: 80271 S17/TAPR NNC/DSP
    21-Aug-88  16:08:17
Sb: #WT6G Sat Netwrkng Ideas
Fm: Franklin Antonio, N6NKF 76337,1365
To: Bob McGwier 74615,1366 (X)

After hearing your talk @ SANDPAC, WT6G got enthused, and wrote up some
comments re satellite networking.  I've uploaded them to S17. Put them in S17
rather than the usual satellite section just so wouldn't get babble from the
masses.  (would hate to divert them from the THENET/NETROM feud; they're having
too much fun.)

If anyone has comments in response to WT6G's ideas, you can send 'em to me, and
i'll be glad to electronically forward them.  WT6G doesn't compuserve (smart
guy!).

*** There is a reply:   80296

*** More ***

#: 80296 S17/TAPR NNC/DSP
    21-Aug-88  22:55:00
Sb: #80271-WT6G Sat Netwrkng Ideas
Fm: Bob McGwier N4HY 74615,1366
To: Franklin Antonio, N6NKF 76337,1365 (X)

Thanks!  I will download them.

Bob

#: 80382 S17/TAPR NNC/DSP
    23-Aug-88  21:21:53
Sb: Rick's UID
Fm: Rick Hambly WB2TNL 72117,3530
To: Barry McLarnon VE3JF 71470,3651 (X)

I see what appears to be interest in my UIC, so here it is.  I do check in
regularly, but I seldom send any messages.  If I can be of help send messages
here or to WB2TNL @W3ZH or to ...!mimsy!arinc!rick

Rick Hambly