home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!spool.mu.edu!sdd.hp.com!zaphod.mps.ohio-state.edu!cis.ohio-state.edu!ucbvax!terranet.cts.com!taob
- From: taob@terranet.cts.com (Brian Tao)
- Newsgroups: comp.sys.apple2
- Subject: 14-MHz TWGS/ZipGS upgrade info
- Message-ID: <9209142056.aa07361@generic.UUCP>
- Date: 15 Sep 92 00:52:40 GMT
- Sender: usenet@ucbvax.BERKELEY.EDU
- Organization: The Internet
- Lines: 895
-
- [The following is a fairly long compilation of a message thread on the
- GNO mailing list regarding upgrading TransWarp GS and ZipGS accelerators
- with the Sanyo 14-MHz 65816 chip. I'm not sure how it was started, but
- it is being reposted here for the benefit of comp.sys.apple2 readers who
- aren't subscribed to gno-l. All but the first occurence of a person's
- signature was deleted to save space. Articles are separated by '===' and
- followed by the date of the posting and the address of the postter.]
-
- - Brian
-
-
- ===
- Date: Sat, 29 Aug 92 13:01:46 -0400
- From: cdeschu@calvin.sfc.lehigh.edu (Chris Deschu)
-
- Well, all the components came in, and here's the results...
-
- It's running nicely at 13.75mhz, and I had it running briefly at 15mhz by
- cranking the power supply up to 5.73vdc. I couldn't keep it there because
- (1) the cache ram test fails (2) the rest of the computer doesn't like the
- high voltage levels. So I'm back at 13.75, but I'd bet it would still work
- at 14 or 14.5mhz, anyone got a 58mhz crystal? :) This is what dhrystone sez:
-
- Dhrystone(1.1) time for 50000 passes = 43
- This machine benchmarks at 1162 dhrystones/second
-
- I think I can live with that for now. I wonder if just faster ram would
- fix it, that's relatively cheap, right? Til next time,
-
- Chris Deschu (cdeschu@calvin.sfc.lehigh.edu)
-
- ===
- From: Jawaid Bazyar <bazyar@csn.org>
- Date: Sat, 29 Aug 92 11:14:20 MDT
-
- > Well, all the components came in, and here's the results...
- >
- > Dhrystone(1.1) time for 50000 passes = 43
- > This machine benchmarks at 1162 dhrystones/second
-
- You bum. And here I just got my Zip cranking 590 dhrystones. My graphics
- are just as fast as yours, though. :)
-
- --
- Jawaid Bazyar | Ask me about the GNO Multitasking Environment
- Procyon, Inc. | for the Apple IIgs!
- bazyar@cs.uiuc.edu | P.O Box 620334
- --Apple II Forever!-- | Littleton, CO 80162-0334 (303) 933-4649
-
- ===
- Date: Sat, 29 Aug 92 13:57:55 -0400
- From: cdeschu@calvin.sfc.lehigh.edu (Chris Deschu)
-
- Nobody ever really explained why the ZipGS is faster at graphics, to me at
- least... Jawaid, what speed are you getting 590 at, I was getting 750 or so
- at 8.5 mhz, and I've got a few extra crystals lying around.
- A note to the adventuresome: don't leave your voltage at 5.7 volts,
- it makes the computer lock up randomly (even in proterm). I've been running
- at 5.5 for months now fine, I think I'll leave it at 5.6 for awhile and see
- what happens... bye...
-
- ===
- Date: Sat, 29 Aug 1992 14:11:42 -0500
- From: Derek Taubert <derek@slab.slip.uiuc.edu>
-
- This should be in a FAQ somewhere, but I've been detached for a while...
-
- > It's running nicely at 13.75mhz, and I had it running briefly at 15mhz by
- > cranking the power supply up to 5.73vdc. I couldn't keep it there because
- > ...
-
- Where did you order the 65816 from and how much did you pay for it?
-
- ===
- Date: Sat, 29 Aug 92 16:57:59 -0400
- From: cdeschu@calvin.sfc.lehigh.edu (Chris Deschu)
-
- Well, I'm not sure if that last mail went thru or not, something was acting
- very wierd, so here we go again.
-
- To all that may be interested:
-
- High speed 65c816 processors are available from:
-
- The Western Design Center Inc.
- 2166 East Brown Road
- Mesa, AZ. 85213
- (602) 962-4545 (I personally think they should have gotten -6502 :)
-
- Talk to Deborah Lamoree when you call...
- (She told me quite a few people, including me, have tried to bribe her to
- select a really nice and fast one by hand, and she won't do it, so don't
- bother
- with it unless you are going to offer a lot more than the rest of us, or you
- know something we don't :)
-
- Crystals are available from: (crystal info courtesy of Mike Horwath)
-
- DigiKey, number is: 800-344-4539
-
- Part numbers are:
-
- Mhz Part # Price
- ------ ------- -----
- 32 SE1101 3.60
- 33.33 SE1102 3.60
- 36 SE1103 3.60
- 40 SE1104 3.60
- 42 SE1105 3.60
- 46 SE1106 3.60
- 48 SE1107 3.60
- 50 SE1108 3.60
- 55 SE1109 3.60
- 60 SE1110 3.75
-
- Order one that is four (4) times the speed you want your twgs to go.
- like a 55mhz crystal will give you 13.75mhz (probably the fastest you will
- want to order) You should also get the 32k cache upgrade if you haven't yet.
- Good Luck!
-
- ===
- From: Tim Meekins <meekins@cis.ohio-state.edu>
- Date: Sat, 29 Aug 92 22:26:58 EDT
-
- Chris didn't mention prices in his nice post, but when I upgraded to
- 13.75MHz I paid $103 for the entire upgrade. As far as I know Chris
- and I are the only people to have this work with no problem, and the
- only thing we have that noone else does is an AE power supply. But I really
- don't know if thisis the reason or not.
-
- Don't worry about getting the "best" CPU from WDC, I got mine from a batch
- that was bought, and mine was the last chip after everyone picked through the
- good ones, so the ones with the bad statistics still work great.
-
- TIM MEEKINS Ask me about GNO/ME! - MultiTasking on a IIGS!
- meekins@cis.ohio-state.edu Procyon, Inc. 1005 N Kingshighway, Suite 309
- meekins@bluemoon.uucp Cape Girardeau, MO 63701 (314) 334-7078
-
- "Computers aren't intelligent, they just think they are."
-
- ===
- From: toddpw@cco.caltech.edu (Todd P. Whitesel)
- Date: Sat, 29 Aug 92 21:32:53 PDT
-
- Ok, now that I am on gno-l again, I can look forward to gobs of mail about
- whatever's going on... Might as well respond to everything in one letter.
-
- 1. Running your 65816 higher than 5.5v is asking for trouble -- the buffer
- chips in the TWGS and the IIgs are TTL and the design tolerances on
- those don't guarantee they will survive anything above 5.5 volts.
- There's going to be some drop, so feeding the 65816 5.6 would probably
- work since the rest of the machine will still see 5.5, but I wouldn't
- go above 5.6 under any circumstances.
-
- 2. Why the Zip's graphics performance grinds the TWGS into the dirt -- this is
- because of the "latch-on-write" bus interface of the Zip. Put simply,
- the Zip hardware is capable of coprocessing a single byte write to
- the IIgs without slowing down the CPU, provided that the next few
- cycles run from cache. Of course, a two-byte write will pause the CPU
- until the first byte is written, but the second byte gets written in
- the background -- this lets standard graphics code run much faster and
- makes MVP/MVN nearly as fast as PEA/PEI (disgusting, eh?).
-
- > High speed 65c816 processors are available from:
- > The Western Design Center Inc.
- > (She told me quite a few people, including me, have tried to bribe her to
-
- This bit about bribing doesn't make sense -- the report I heard was that WDC
- simply bought a huge shipment of 14mhz Sanyo chips and has been distributing
- those...
-
- > Crystals are available from: (crystal info courtesy of Mike Horwath)
- > DigiKey, number is: 800-344-4539
-
- DigiKey has a $25 minimum order.
-
- Todd Whitesel
- toddpw @ cco.caltech.edu
-
- ===
- From: mfrankow@hamline.edu (Michael Frankowski)
- Date: Sun, 30 Aug 92 0:05:17 CDT
-
- > > Crystals are available from: (crystal info courtesy of Mike Horwath)
- > > DigiKey, number is: 800-344-4539
- >
- > DigiKey has a $25 minimum order.
-
- That is a falsehood. DigiKey doesn't care what your order price is. When I
- had to order some things I couldn't get localy, I ordered through there, and
- my bill was only $11.50, so, I think you are wrong there.
-
- > Todd Whitesel
- > toddpw @ cco.caltech.edu
-
- Mike
-
- ===
- Date: Sun, 30 Aug 92 14:34:28 CDT
- From: Michael Crane <cranem@edsi.plexus.com>
-
- > Nobody ever really explained why the ZipGS is faster at graphics, to me at
- > least... Jawaid, what speed are you getting 590 at, I was getting 750 or so
- > at 8.5 mhz, and I've got a few extra crystals lying around.
-
- I have to wonder also. I'm running my 7MHz TWGS with a 32MHz crystal
- for an effective speed of 8MHz, and *I* get a Dhrystone rating of 704.
- That must be an awfully slow ZipGS you've got there, eh? <smile>
-
- --
- Mike Crane Bang: uunet.uu.net!edsi.plexus.COM!cranem@uunet.UU.NET
- Internet: cranem@edsi.plexus.com | Michael_Crane@ntc.com | mike566@aol.com
- GEnie: M.CRANE AOL: MIKE566 "Who needs drugs...we've got UNIX!" - Me
- (Internet addresses are listed in order of preference of use.)
-
- ===
- From: mfrankow@hamline.edu (Michael Frankowski)
- Date: Sun, 30 Aug 92 20:43:33 CDT
-
- > > > DigiKey has a $25 minimum order.
- > >
- >
- > > That is a falsehood. DigiKey doesn't care what your order price is. When
- > > I had to order some things I couldn't get localy, I ordered through there,
- > > and my bill was only $11.50, so, I think you are wrong there.
- >
- > No, it is a truehood. :-) Catalog #924, July-August 1992, page 128:
- >
- > o Handling Charge
- > A $5.00 handling charge applies to all orders less than $25.00.
-
- That is a handling charge, that is NOT what Todd had said.
-
- (can't quote literally, but)
-
- "DigiKey won't ship orders less than $25"
-
- (or something to that effect)
-
- ===
- Date: Sun, 30 Aug 92 22:37:42 -0400
- From: cdeschu@spiff.sfc.lehigh.edu (Chris Deschu)
-
- Well Tim,
- Looks like you are king of the mountain again, I'v been getting some
- wierd problems in proterm, like strange characters popping up in scrollback,
- sometimes it says 'restart $01' I try to run it... dropping back to 12.5 or
- whatever seemed to make those go away, but it's still acting strange in
- general. I'm gonna order faster ram and that chipset from AE (ick!) and see
- what happens... Does anyone know a good place to order really fast TWGS
- RAM for the 32k cache? My chips currently are these:
-
- MOSEL
- MSG2256A-35NC
- 9124 U 5S
-
- So so far, I have 35ns ram... Do they make it any faster in that
- configuration?
- Please let me know, I want to order it asap, of course... I checked with JDR
- Microdevices and DigiKey's catalogs and don't seem to find them. I have
- another
- catalog from Mouser Electronics, I think they might have it. At least, they
- have a whole page full of SRAMs... There are three of these things on the
- 32k cache board. How do you get 32k from 3 256k chips? Does anyone know what
- the config of those chips are as in ( 32k x 8 ) or something else?
-
- Thanks everyone, I'll keep you posted...
-
- ===
- From: Tim Meekins <meekins@cis.ohio-state.edu>
- Date: Sun, 30 Aug 92 23:41:43 EDT
-
- interesting...I was talking to Steve Chiang last weekend, and he said ProTerm
- is a big culprit of not working at 14 MHz....I guess you've backed up his
- statement. In all honesty, I've never tried ProTerm yet, so maybe mine won't
- work either.
-
- Also, I really don't know anything about RAM chips..ask around....
-
- ===
- Date: Mon, 31 Aug 1992 01:25:40 -0400 (EDT)
- From: Gregory Ross Thompson <gt0t+@andrew.cmu.edu>
-
- cdeschu@spiff.sfc.lehigh.edu (Chris Deschu) writes:
- > ... Does anyone know a good place to order really fast TWGS
- > RAM for the 32k cache? My chips currently are these:
- >
- > MOSEL
- > MSG2256A-35NC
- > 9124 U 5S
-
- I think that should be MS6, not MSG. Anyway, Mosel makes a 25ns
- chip, but not in the low power configuration like the one you've got.
- The part number is:
- MS62256-25
- Or you could go for a Sony chip:
- CXK58258B-25LL
- Here's an interesting thing. The LL means Low-Low power. They also
- have a chip with only one L, for Low power. There's no regular power
- chip (no L). Why didn't they make one with no L, and one with one L?
- "Well, ours go to eleven"...
-
- Or there's the Paradigm chip: PDM41256L-25
- Or the VLSI Tech chips: VT62832{H,HL,UHL}-25
- VT62832UHL-20
-
- If you wanted to go totally whacko, you could get Sony's
- CXK58258A-15 Now that's a fast SRAM...
-
- -Greg T.
- InterNet: gregt+@cmu.edu
- or: gt0t+@andrew.cmu.edu
- BITNet : R746GT0T@CMCCVB
- AOL : GRThompson
-
- ===
- Date: Mon, 31 Aug 92 01:45:54 -0400
- From: cdeschu@calvin.sfc.lehigh.edu (Chris Deschu)
-
- Greg T.,
- That's quite a collection of chips you have there. Are all of the ones
- you mentioned of the correct number of pins/width, etc. to fit on a TWGS?
- I looked in 3 of my electronic catalogs and couldn't find any of the ones you
- mentioned. Where would you suggest ordering them from? Maybe Microprocessors
- Unlimited? Thanks,
-
- ===
- From: toddpw@cco.caltech.edu (Todd P. Whitesel)
- Date: Sun, 30 Aug 92 23:06:26 PDT
-
- > That is a handling charge, that is NOT what Todd had said.
- > (can't quote literally, but)
- > "DigiKey won't ship orders less than $25"
-
- Nope. I said they have a $25 minimum order. Taken literally, that implies the
- quote above, but with mail order places "minimum order" becomes a term meaning
- 'smallest order with no restrictions' -- some mail order places won't ship
- under their minimum order, like JameCo, but DigiKey apparently just docks you
- extra.
-
- Face it, we were both partially wrong, and the main disagreement was over
- terminology. We've got better things to do, so let's drop it.
-
- ===
- From: toddpw@cco.caltech.edu (Todd P. Whitesel)
- Date: Sun, 30 Aug 92 23:08:50 PDT
-
- > Looks like you are king of the mountain again, I'v been getting some
- > wierd problems in proterm, like strange characters popping up in scrollback,
-
- ProTerm has known problems when you are running very fast. I have a 10.5 mhz
- Zip and if I run ProTerm with it cranked all the way up I get lockups when I
- do a 14400 connect. The solution in my case is to crank the Zip down to 50%
- of max speed to run ProTerm.
-
- ===
- Date: Mon, 31 Aug 92 04:55:12 -0400
- From: cdeschu@spiff.sfc.lehigh.edu (Chris Deschu)
-
- Todd,
- That's rather interesting. I was using the TWGS@8.5mhz with ProTERM
- for quite awhile, and it worked fine. Even at 14mhz, it wasn't locking up,
- it just got wierd garbled characters in the scrollback. I mean I could be
- sitting there reading, and all of a sudden a '@' or other character would
- suddenly appear in the middle of the screen, and become part of scrollback.
- The other wierd thing it did was the clock would start appearing a few lines
- down from where it usually is in the top right corner, and again, screw up
- the scrollback... It's like the memory map for the text page is moved over
- a few bytes somehow. Also, the mouse pointer would sometimes (only on some
- parts of the screen) go off the edge of the screen, and reappear on the
- other side... Very wierd. later,
-
- ===
- From: toddpw@cco.caltech.edu (Todd P. Whitesel)
- Date: Mon, 31 Aug 92 3:32:34 PDT
-
- > That's rather interesting. I was using the TWGS@8.5mhz with ProTERM
- > for quite awhile, and it worked fine. Even at 14mhz, it wasn't locking up,
- > it just got wierd garbled characters in the scrollback. I mean I could be
-
- Remember, I'm using a Zip at 10.5mhz, so I'm bound to get different failures.
- We know from comparing notes with people that ProTerm has problems running at
- really high speeds (probably has to do with software timing loops and mucking
- about directly with the SCC), so you have every reason to expect strange
- things to happen. Just slow down the accelerator while you're in ProTerm and
- you won't have any problems. That's what _I_ do and it works fine. 5 mhz is
- plenty fast for ProTerm and it seems quite happy at that speed.
-
- ===
- Date: Mon, 31 Aug 1992 11:46:08 -0400 (EDT)
- From: Gregory Ross Thompson <gt0t+@andrew.cmu.edu>
-
- cdeschu@calvin.sfc.lehigh.edu (Chris Deschu) writes:
- > Greg T.,
- > That's quite a collection of chips you have there.
-
- Why, thank you... I try.
-
- > Are all of the ones
- > you mentioned of the correct number of pins/width, etc. to fit on a TWGS?
- > I looked in 3 of my electronic catalogs and couldn't find any of the ones
- you
- > mentioned. Where would you suggest ordering them from? Maybe Microprocessors
- > Unlimited? Thanks,
-
- They're all available in 28 pin DIPS, with tri-state outputs. Some
- may only be available in .3in (300 mil) or .6in (600 mil) packages. I
- don't know which size the TWGS takes. I know that my Zip takes 600
- milers, but someone told me that the newer Zips take the 300's. For
- those of you who don't know how big/small a 600/300mil chip is, the
- 300's are the size of the usual simple TTL gate chips (quad AND gates,
- etc), and the 600's are twice as wide.
-
- I got all of those part numbers out of IC Master (the ECE Tome).
- Just look up 32768x8 general purpose SRAMs.
-
- I'd order 'em straight from the manufacturer's distributors, just
- 'cause that's the only place I know of...
-
- ===
- From: mfrankow@hamline.edu (Michael Frankowski)
- Date: Mon, 31 Aug 92 11:21:18 CDT
-
- > Remember, I'm using a Zip at 10.5mhz, so I'm bound to get different
- failures.
- > We know from comparing notes with people that ProTerm has problems running
- at
- > really high speeds (probably has to do with software timing loops and
- mucking
- > about directly with the SCC), so you have every reason to expect strange
- > things to happen. Just slow down the accelerator while you're in ProTerm and
- > you won't have any problems. That's what _I_ do and it works fine. 5 mhz is
- > plenty fast for ProTerm and it seems quite happy at that speed.
-
- With a TWGS, we can't change our speed like that. We either run at full clock
- speed or we run at 2.8mhz. (or we can change the crystal, which is what I
- do most of the time).
-
- ===
- From: unknown@ucscb.UCSC.EDU (tuu)
- Date: Mon, 31 Aug 1992 09:51:51 PDT
-
- Does ProTERM work with a 14 MHz Zip if I slow the Zip down before
- I run it? I could live with that.. Proterm's definitely fast enough at
- 2.8... it's only desktop stuff (and compiling) that I want a Zip for..
- so I'd live with the annoyance of having to slow down the Zip every time I
- ran ProTERM..
-
- If ProTERM won't even run with a slowed down 14 MHz zip, then I'd consider not
- getting a Zip at all (or a slower one maybe)
-
- I sure wish Greg Schaeffer started logging on again so maybe we could get
- an answer/small patch out of him why it doesn't work.
-
- --
- /unknown@ucscb.ucsc.edu used ADB trackball WANTED Apple II
- Forever!\
- \Mail me:Get INFOCOM GAMES RERELEASED for //s, ULT VI GS done,for UNIX GS
- info/
- What the world needs now is another folk singer like I need a hole in my head.
-
- ===
- From: alb@cognos.com (Al Belyea)
- Date: Mon, 31 Aug 1992 14:29:26 -0400
-
- > DigiKey, number is: 800-344-4539
- >
- >
-
- Anybody have a non-800 number for DigiKey? Just in case the 800 number
- doesn't work from Canada (and they usually don't).
-
- BTW, are the crystals that you listed the mini size crystals that fit
- right into the TWGS or are they the full size ones that need to be
- modified?
-
- I currently have the 32K cache, with 35 ns SRAM. The Schmoo plot that
- I got from WDC seems to indicate that my new 65c816 will run reliably
- (loaded) at about 11.5 MHZ - is the 35 ns SRAM fast enough to work at
- 11.5 MHZ?
-
- Also, I tried calling AE about getting the High-Performance GAL chips
- that are mentioned in the TWGS upgrade article. They said that they've
- handed the marketing of those chips over to LRO Computers. I called
- them and they said that they don't have them yet - call back in a
- couple of weeks.
-
-
- Allan Belyea UUNet: alb@cognos.UUCP 3755 Riverside Dr.
- Cognos Incorporated BITNET: alb@cognos.COM P.O. Box 9707
- (613) 738-1338 x3822 ^ preferred Ottawa, Ontario
- "This is a sword. You hold _THIS_ end." Sun Wolf CANADA K1G 3Z4
-
- ===
- Date: Mon, 31 Aug 92 15:24:53 -0400
- From: cdeschu@spiff.sfc.lehigh.edu (Chris Deschu)
-
- An easy way to use the long rectangular crystals with a TWGS is to
- take an 16 pin dip socket, cut 8 pins off of one end, and solder a
- jumper wire between the 8th pin (cut off) and the 4th pin (not cut
- off) on each side. then voila, you have an adapter. I've been using it
- that way for a long time, and it works great, just make sure you plug
- it in right, or you'll fry the crystal...
-
- Um, I haven't called them recently, but the last time I checked,
- the GAL set part number was 10MHZPALKIT from AE, and they did have it
- then...
-
- ===
- From: Jawaid Bazyar <bazyar@csn.org>
- Date: Mon, 31 Aug 92 14:37:05 MDT
-
- >Also, I tried calling AE about getting the High-Performance GAL chips
- >that are mentioned in the TWGS upgrade article. They said that they've
- >handed the marketing of those chips over to LRO Computers. I called
- >them and they said that they don't have them yet - call back in a
- >couple of weeks.
-
- Oh wonderful! So the LRO Horror Story (tm) continues! Please keep
- me informed as to the exact status of this transaction. It's extremely
- important.
-
- ===
- From: davewh@microsoft.com
- Date: Mon, 31 Aug 92 14:05:00 PDT
-
- | > That's what _I_ do and it works fine. 5 mhz is
- |
- | Of course, TelCom is happy at 14MHz or any speed for that matter. Hmm.
-
- So's Z-Link (well, actually, my Zip only goes to 8MHz). Hmmm. Ain't nothing
- like going "by the book." Tends to keep one in line with the hardware...
-
- ===
- Date: Mon, 31 Aug 92 17:10:30 -0400
- From: cdeschu@spiff.sfc.lehigh.edu (Chris Deschu)
-
- I tried calling AE & LRO myself, and AE's like duh, we don't know anything
- about them anymore, call LRO. Couldn't even tell me who makes them...
- LRO on the other hand said duh, we don't have them yet, we don't know when
- we will have them either, call back in a few weeks. Sound familiar? I must
- say that they did get my 32k cache out promptly, but other than that, I'm
- not impressed with them. They also have absolutely no idea at all when prizm
- was going to be out, nor could they tell me who was demoing it at KFest.
- Somebody PLEASE tell me where I can order those ram chips from for
- the TransWarp... I don't want to live with a half baked gs for a few weeks.
-
- ===
- From: Jawaid Bazyar <bazyar@csn.org>
- Date: Mon, 31 Aug 92 15:33:48 MDT
-
- > So's Z-Link (well, actually, my Zip only goes to 8MHz). Hmmm. Ain't nothing
- > like going "by the book." Tends to keep one in line with the hardware...
-
- Oh yeah... for everyone who doesn't know, Dave Whitney here is
- at least partially responsible for TelCom (he provided the serial drivers
- from Z-Link). Take a bow, Dave :)
-
- ===
- From: davewh@microsoft.com
- Date: Mon, 31 Aug 92 15:22:00 PDT
-
-
- | Oh yeah... for everyone who doesn't know, Dave Whitney here is
- | at least partially responsible for TelCom (he provided the serial drivers
- | from Z-Link). Take a bow, Dave :)
-
- TAA DAA! Thank you thank you thank you. And for my next trick...
-
- ===
- From: mfrankow@hamline.edu (Michael Frankowski)
- Date: Mon, 31 Aug 92 18:58:57 CDT
-
- > Anybody have a non-800 number for DigiKey? Just in case the 800 number
- doesn't
- > work from Canada (and they usually don't).
-
- There is no normal number listed, but they are in e 218 area code (very north
- Minnesota). Address is:
-
- Digi-Key
- 701 Brooks Ave S
- POBOX 677
- Theif River Falls, MN 56701-0677
-
- Use 1-218-555-1212 for information to get the normal number if need be.
-
- > BTW, are the crystals that you listed the mini size crystals that fit right
- > into the TWGS or are they the full size ones that need to be modified?
-
- Those are the ones that fit into a TWGS with no modifications.
-
- > I currently have the 32K cache, with 35 ns SRAM. The Schmoo plot that I got
- > from WDC seems to indicate that my new 65c816 will run reliably (loaded) at
- > about 11.5 MHZ - is the 35 ns SRAM fast enough to work at 11.5 MHZ?
-
- Yes, 35ns should be fast enough even for 14. Chris Deschu is trying faster
- chips to see if it makes a difference.
-
- ===
- Date: Mon, 31 Aug 1992 20:37:27 -0500
- From: Derek Taubert <derek@slab.slip.uiuc.edu>
-
- Just out of curiosity, what numbers are on the stickers on everybody's
- TWGS chips? The version number on my EPROM on the 32K card is 1.8S,
- the rams are Toshiba TC55328P-35 (for now), and all of the GAL chips
- are labeled TWGSxy-N where x is 1-8 and y is A for all x except 2
- (B) and 3 (C). How about somebody with the faster GAL set?
-
- ===
- From: toddpw@cco.caltech.edu (Todd P. Whitesel)
- Date: Mon, 31 Aug 92 18:52:53 PDT
-
- > Does ProTERM work with a 14 MHz Zip if I slow the Zip down before
- > I run it? I could live with that.. Proterm's definitely fast enough at
- > 2.8... it's only desktop stuff (and compiling) that I want a Zip for..
- > so I'd live with the annoyance of having to slow down the Zip every time I
- > ran ProTERM..
-
- Yep. This is what I do. 5.25 mhz (50% of 10.5) seems to work fine for me.
-
- > I sure wish Greg Schaeffer started logging on again so maybe we could get
- > an answer/small patch out of him why it doesn't work.
-
- Oh well...
-
- ===
- From: toddpw@cco.caltech.edu (Todd P. Whitesel)
- Date: Mon, 31 Aug 92 18:54:56 PDT
-
- >Also, I tried calling AE about getting the High-Performance GAL chips
- >that are mentioned in the TWGS upgrade article. They said that they've
- >handed the marketing of those chips over to LRO Computers. I called
- >them and they said that they don't have them yet - call back in a
- >couple of weeks.
-
- Uh oh. I've heard some pretty venomous things about LRO lately, mainly with
- regard to certain nonexistent Visionary cards (i.e. LRO has none, yet they are
- "selling" them). I'd be careful about LRO were I you.
-
- ===
- Date: Tue, 1 Sep 92 6:52:00 CDT
- From: Michael Crane <cranem@edsi.plexus.com>
-
- > Anybody have a non-800 number for DigiKey? Just in case the 800 number
- > doesn't work from Canada (and they usually don't).
-
- (218) 681-6674
-
- > BTW, are the crystals that you listed the mini size crystals that fit right
- > into the TWGS or are they the full size ones that need to be modified?
-
- I don't happen to recall what was listed, but you just need to make sure
- that you get "1/2 TTL/CMOS" oscillators. They come in metal and plastic
- bodies...I believe the plastic ones work just fine and are cheaper.
-
- > Also, I tried calling AE about getting the High-Performance GAL chips
- > that are mentioned in the TWGS upgrade article. They said that they've
- > handed the marketing of those chips over to LRO Computers. I called
- > them and they said that they don't have them yet - call back in a
- > couple of weeks.
-
- Thanks for the info. I may be needing some of those myself very soon.
-
- ===
- From: alb@cognos.com (Al Belyea)
- Date: Tue, 1 Sep 1992 10:52:40 -0400
-
- >>> DigiKey, number is: 800-344-4539
-
- Good news! The 800 number for DigiKey works from Canada!
-
- >
- > Those are the ones that fit into a TWGS with no modifications.
- >
- > Yes, 35ns should be fast enough even for 14. Chris Deschu is trying faster
- > chips to see if it makes a difference.
- >
-
- Thats's great news - thanks for the info.
-
- Now with the crystals on their way, I'll be able to get my TWGS up to
- 9 or 9.5 MHZ almost immediately. However, according to the upgrade
- article, I need to replace the 4 GAL chips with the high-performance
- ones if I want to run at 10 MHZ or more.
-
- BTW, when I called AE and asked about the High Performance GAL chips,
- they said that LRO had been handling them for about a year now. And
- when I first asked LRO about them, they said huh? I had to explain
- what I wanted very carefully and slowly.
-
- Allan.
-
- ===
- From: alb@cognos.com (Al Belyea)
- Date: Tue, 1 Sep 1992 11:01:52 -0400
-
- Re: High Performance GAL chips for TWGS and LRO:
-
- > Uh oh. I've heard some pretty venomous things about LRO lately, mainly with
- > regard to certain nonexistent Visionary cards (i.e. LRO has none, yet they
- are
- > "selling" them). I'd be careful about LRO were I you.
- >
-
- Well, if someone knows where to get these High Performance GAL chips for the
- TWGS from somewhere other than AE or LRO, I'll be the first person in line.
-
- BTW, I didn't lay any money or credit card with LRO yet. I just asked them if
- they had them in stock or not.
-
- ===
- From: alb@cognos.com (Al Belyea)
- Date: Tue, 1 Sep 1992 10:57:37 -0400
-
- >Just out of curiosity, what numbers are on the stickers on everybody's
- >TWGS chips? The version number on my EPROM on the 32K card is 1.8S,
- >the rams are Toshiba TC55328P-35 (for now), and all of the GAL chips
- >are labeled TWGSxy-N where x is 1-8 and y is A for all x except 2 (B)
- >and 3 (C).How about somebody with the faster GAL set?
-
- Going from memory, my 32K EPROM and SRAM are the same as yours, but I have an
- older set of GAL chips. My GAL #3 is rev B I think.
-
- The upgrade article says you need GAL #3 Rev E to run at 10 MHZ or higher.
-
- ===
- Date: Tue, 1 Sep 1992 01:59:33 -0400 (EDT)
- From: Gregory Ross Thompson <gt0t+@andrew.cmu.edu>
-
- Someone asked how to get Sony's mega-fast chip. All I've got is a
- phone number for 'em. Try 714-220-9100. They'll probably tell you to
- call some semi-local place.
-
- Once again, the part number is: CXK58258A-15
-
- Also, VLSI Tech makes one: VT62832UHL-15 [call 408-434-3100 for them]
-
- Zip's charging $399 to upgrade to 14MHz. Seems to me it'd cost
- $99 for the '816 (I think), a few bucks for crystals, and some amount
- for the SRAM. Wouldn't 35ns be enough?
-
- ===
- Date: Tue, 1 Sep 92 13:38:39 -0400
- From: cdeschu@spiff.sfc.lehigh.edu (Chris Deschu)
-
- Allan,
- You mean you actually got LRO to admit they had them, and they
- didn't tell you to call back in a few weeks? Who did you talk to? I need the
- fast GALs for my Transwarp too... Thanks,
-
- ===
- From: alb@cognos.com (Al Belyea)
- Date: Tue, 1 Sep 1992 16:23:02 -0400
-
- > Allan,
- > You mean you actually got LRO to admit they had them, and they
- > didn't tell you to call back in a few weeks? Who did you talk to? I need the
- > fast GALs for my Transwarp too... Thanks,
-
- No, I said that I called them, that it took them a while to understand what I
- was talking about, and then they said Oh we don't have them yet. Call back in
- a couple of weeks.
-
- Has anyone ever gotten these Hi-Performance TWGS GAL chips from LRO within the
- past year?? (which according to AE is when they were turned over to LRO).
-
- ===
- Date: Fri, 4 Sep 92 20:13:52 -0400
- From: cdeschu@spiff.sfc.lehigh.edu (Chris Deschu)
-
- Ok, heres the stats on my TWGS:
-
- ROM: 1.8s
- Cache: 32k
- GALs: 1A1, 2B1, 3D1, 4B1, 5A1, 6A1, 7A1, 8B1
-
- I don't know of anyone who actually has the fast GAL set.
-
- I just upgraded the cache RAM to 20ns chips, and now it seems that it
- will run ok at 15mhz, although I am still experiencing wierd problems,
- it seems to just all of a sudden crash & total lockup after 20 mins,
- even when the TWGS is set to 2.8mhz. I haven't tried it without the
- TWGS yet, I guess that is next, but it passes the TWGS CDA test for
- over 6 hours with no failures, the claris memory test with no
- failures, the internal diagnostic says system good, and apple's dealer
- diagnostics run 50 iterations with no failures. So I am sort of at a
- loss at this point. Any ideas? Maybe I shouldn't have turned the
- voltage up that high :(... But check out these dhrystone results:
-
- Dhrystone(1.1) time for 50000 passes = 40
- This machine benchmarks at 1250 dhrystones/second
-
- Gee, maybe if we all call LRO at the same time, and ask for the GALs,
- they'll actually do something about it... later,
-
- ===
- Date: Sat, 5 Sep 92 15:26:34 -0400
- From: cdeschu@spiff.sfc.lehigh.edu (Chris Deschu)
-
- Well, I got the benchmark program from Mike's BBS and here's the results:
-
- [15.00Mhz]
- [Result of the 1 runs Averaged (base speed = 100%)]
-
- Base Speed time Accelerated time Expected Time Improvement Actual MHz
- --------------- ---------------- ------------- ----------- ----------
- 10 iterations of the Sieve of Erastothenes.
- 410.00 99.00 69.51 175.9% 10.53
- 500 strings selection sort.
- 1151.00 262.00 194.88 177.2% 11.16
- 100 iterations of floating point test.
- 472.00 93.00 85.69 180.3% 13.82
- 15000 Gamm units (FPE instruction mix) in
- 1535.00 315.00 274.17 179.5% 13.06
- 10 iterations of Fibinacci calls in
- 2006.00 509.00 339.59 174.6% 10.01
- 5000 itereation Integer math in
- 1553.00 308.00 263.04 180.2% 12.81
- -------------------------------------------------------------------------
- Average 117.9% 11.90
-
- Where it says Base Speed time, that's at 2.8mhz.
- Here's some other info that was included with the test:
-
- To obtained optimal performance number, run the benchmark program with
- no init/DAs. System disk version will also affect the tests. System 6
- will give the best time.
-
- @Stock GS *Transwarp +Transwarp @Zip GSX @Zip GSX @Zip GSX
- 2.8 mhz 7.5mhz 8 mhz 9 mhz 8 mhz 8 mhz
- 32k Cache 32k Cache 32k Cache 16k Cache 32k Cache
- ======= ========= ========= ========= ========= =========
- Sieve 413.40 166.20 153.40 152.50 170.00 170.33
- String 1161.20 450.80 422.60 406.50 466.67 444.33
- FPE Test 1 479.50 176.00 168.20 188.70 214.67 203.67
- FPE Test 2 1561.90 579.00 560.40 612.80 699.00 657.00
- Fabinacci 2023.90 772.60 733.40 777.80 853.67 853.33
- Integer 1567.00 549.00 517.20 616.00 683.67 680.00
-
- Effective Avg x2.57 x6.83 x7.20 x6.75 x6.00 x6.17
-
- *Stock GS *Stock GS
- 1.0 mhz 2.8 mhz
- ========= ==========
- 1042.60 411.00
- 2923.20 1153.60
- 1285.40 473.00
- 4112.50 1538.80
- 5093.80 2010.80
- 3945.60 1557.00
-
- Effective Avg x1.00 x2.59
-
- (All numbers should be within +/-1%)
- ( * - denotes numbers obtained by me [Hman] under System 6 w/ no init/DAs)
- ( + - are obtained under System 5.04)
- ( @ - compiled by Goose w/ System 5.04)
-
- Since numbers for the Zip are obtained under Sys5.04 this table is not a fair
- comparison. Update is needed.
-
- [Goose] :
- Both machines look good. It does seem the Transwarp handles the math much
- better though
-
- Either way you go you are ripping. The Zip has price and less heat on it's
- side. While the TW seems to have a edge on the speed. Wish someone had a 10/64
- Zip to test. I would like to see the figures. I would also like to see one of
- the 10 mhz TW's they had the text files out about six months to a year back.
-
- Ok, NEXT.... Soenke, where can we obtain Pepo-Bench? :) Thanks,
-
- ===
- Date: Fri, 11 Sep 92 19:18:40 -0400
- From: cdeschu@spiff.sfc.lehigh.edu (Chris Deschu)
-
- Good News!
- I just called LRO again, cause I felt like buggin them, and apparently
- the GAL chipsets are now available. She told me they don't have them in stock,
- it will take a week and a half to order them from AE. At least they
- acknowledge
- that they exist now. Anyway, I ordered one, it's $79.00 + shipping, and we'll
- see what happens. Hopefully, I'll recieve it within the next few months...
-
- Later,
- Chris Deschu
-
- ===
-
- (Whew! Sorry about the wrapped lines interspersed here and there...)
- --
- Brian T. Tao =*8-) | ::::::::> tcomtaob@chasm.scar.utoronto.ca <::::::::
- Dept. of Exobiology | :::::> generic!terranet!taob@zoo.toronto.edu <:::::
- University of Toronto | "Though this be madness, yet there is method in't."
-