home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!olivea!decwrl!csus.edu!netcom.com!netcom!mjohnson
- From: mjohnson@netcom.Netcom.COM (Mark Johnson)
- Newsgroups: comp.lsi.testing
- Subject: Re: Ring oscillators for characterization?
- Message-ID: <MJOHNSON.92Sep3080501@netcom.Netcom.COM>
- Date: 3 Sep 92 16:05:01 GMT
- References: <2374@devnull.mpd.tandem.com>
- Organization: Netcom Online Communications Service
- Lines: 42
- In-Reply-To: black@mpd.tandem.com's message of 3 Sep 92 11:28:32 GMT
-
- In my experience, ring oscillators are among the BEST
- vehicles for process monitoring. Generally the
- most useful ones are a series of _loaded_ oscillators:
-
- (i) Oscillator with long thin fingers of N+ diffusion
- loading down every node. Monitors effect of
- N+ junction perimeter capacitace
-
- (ii) Oscillator with fat rectangular boxes of N+
- diffusion loading down every node. Monitors
- effect of N+ junction perimeter capacitace
-
- (iii) Oscillator with long meander track of minwidth
- metal-1 loading down every node. Monitors
- effect of metal-1 capacitance.
-
- (iv) Oscillator with a 3-input mux in series between
- each inverting stage. Monitors effect of
- mux delay (mosfet "R" & "C", junction C).
-
- (v) Oscillators made of NAND gates, inverters,
- NOR gates. Shows effect of PMOS stacks,
- NMOS stacks.
-
- etc.
-
- Admittedly the measurements you get are indirect
- indicators of process deviations. However they are
- DIRECT indicatiors of product performance.
-
- The tough part is getting the automatic tester to
- do the measurements. About the best arrangement
- I've seen for this was at Vitesse (GaAs vendor).
- They do it simply: the oscillators are always
- located in the same location in the scribe lane,
- so they have dedicated probecard pins for the
- oscs. These go to a GPIB-interfaced frequency
- counter. The frequency counter does the
- measurement and presents its result to the tester
- over GPIB. Costs a couple hundred bux.
-
- MJ
-