ECL: Why'd they use negative voltages?
in [Design]
|
From: krw on 7 Jun 2010 21:23 On Mon, 07 Jun 2010 17:54:10 -0700, Joerg <invalid(a)invalid.invalid> wrote: >Joel Koltner wrote: >> Thanks John, that's quite informative. >> >> "John Larkin" <jjlarkin(a)highNOTlandTHIStechnologyPART.com> wrote in >> message news:di3r0613v9lkk2p8dq06snde2hgopt87fr(a)4ax.com... >>> Newer stuff, like EclipsLite, works at 3.3 volts, and some at 2.5. I >>> do mixed-mode PECL and cmos/FPGAs off a +3.3 volt supply. >> >> That's kinda what prompted the question -- Joerg has been pointing me at >> some of the high-speed logic from the likes of Micrel, and I was reading >> up some in the old MECL System Design Handbook so that I hopefully won't >> embarrass myself too badly when I go to use some of it. :-) >> > >If you have the old MECL databooks treasure them, don't let them end up >in the recycling. Both Motorola and Fairchild did a most excellent job >explaining stuff in there and I assume many of the authors are no longer >with us on earth. > >If you do a good job bypassing the supply and keeping leads short there >isn't much that can go wrong. Except for one thing: Some series have >only half the logic swing and that almost bit me once. They're very forgiving about bypassing too. That's one of the big advantages of ECL. Current is higher, but switching current is far lower than TTL, for instance.
From: John Larkin on 7 Jun 2010 21:24 On Mon, 07 Jun 2010 17:54:10 -0700, Joerg <invalid(a)invalid.invalid> wrote: >Joel Koltner wrote: >> Thanks John, that's quite informative. >> >> "John Larkin" <jjlarkin(a)highNOTlandTHIStechnologyPART.com> wrote in >> message news:di3r0613v9lkk2p8dq06snde2hgopt87fr(a)4ax.com... >>> Newer stuff, like EclipsLite, works at 3.3 volts, and some at 2.5. I >>> do mixed-mode PECL and cmos/FPGAs off a +3.3 volt supply. >> >> That's kinda what prompted the question -- Joerg has been pointing me at >> some of the high-speed logic from the likes of Micrel, and I was reading >> up some in the old MECL System Design Handbook so that I hopefully won't >> embarrass myself too badly when I go to use some of it. :-) >> > >If you have the old MECL databooks treasure them, don't let them end up >in the recycling. Both Motorola and Fairchild did a most excellent job >explaining stuff in there and I assume many of the authors are no longer >with us on earth. > >If you do a good job bypassing the supply and keeping leads short there >isn't much that can go wrong. Except for one thing: Some series have >only half the logic swing and that almost bit me once. Another gotcha: some parts really don't like having their inputs pulled all the way up to Vcc. So you have to make a logic "1", a junction drop down maybe, and pipe that all over the place. John
From: krw on 7 Jun 2010 21:26 On Mon, 07 Jun 2010 18:12:37 -0700, D Yuniskis <not.going.to.be(a)seen.com> wrote: >krw(a)att.bizzzzzzzzzzzz wrote: >> On Mon, 07 Jun 2010 17:47:28 -0700, D Yuniskis <not.going.to.be(a)seen.com> >> wrote: >> >>> Hi Joel, >>> >>> Joel Koltner wrote: >>>> I realize it was the early '60s and all, but why does ECL generally use >>>> 0V for VCC and -5.2V for VEE, rather than, oh, say... 5V for VCC and 0V >>> You also had a Vbb of around -1.2V (?). Lines were typically terminated >>> to this. >> >> -1.4, IIRC (but maybe not, since it's been almost 25 years since I used MECK >> 10K). > >Dunno. This was 35 years ago. At the time, I didn't realize it >wasn't "normal practice" :> > >>>> for VEE? Something related to how things were done when toobs ruled? (I >>>> realize that you can almost always run ECL off of 5V/0V -- and >>>> apparently this was popular practice at one time?) >>>> >>>> And why 5.2V anyway? (Granted, 5.2V is no stranger than 6.3V filament >>>> transformers, I suppose...) >>> Dunno. But, it was fast and ate gobs of power. In the >>> mid 70's I worked on a processor (i.e., what nowadays >>> would be a CPU "chip") that drew 100A (MECL III and 10K). >>> "Bus bars" for power were 3/4" square copper shafts. >>> Instruction cycle time was 8ns. By comparison, I think a >>> 7404 (inverter) takes *7* ns just to change the state of >>> its output. >>> >>> If you "slipped" when pulling/installing a chip, the legs >>> would vaporize before the power supply would even hiccup. >>> (needless to say, you removed all jewelry -- belt buckles, >>> eye glasses, rings, etc. -- when working on it) >> >> I worked on a test system that had a 100A -4V *linear* HP supply in it. The > >The Vbb supply was a *shunt* regulator. A bunch of Lambda power >supplies driving a pair of *big* diodes "selected at test". >A colossal waste of power. > >> thing was the size of a small refrigerator. The other engineer did the 1600 >> pin pin-driver logic in proprietary ECL (-4V). I did the clock drivers in >> MECL 10K. I only needed about 50A for the 64 clocks. ;-) > >This was the brains of a "600-pin tester". Programmable power >supplies to the UUT. Programmable thresholds for the input >comparators. 1ns timing resolution. It was just "insane". Yep, all that and 1600 pins. ;-) OTOH, it was a DC logic tester (clocks ran at system speed - 25ns).
From: John Larkin on 7 Jun 2010 21:26 On Mon, 07 Jun 2010 19:51:13 -0500, "krw(a)att.bizzzzzzzzzzzz" <krw(a)att.bizzzzzzzzzzzz> wrote: >On Mon, 7 Jun 2010 17:44:55 -0700, "Joel Koltner" ><zapwireDASHgroups(a)yahoo.com> wrote: > >>Thanks John, that's quite informative. >> >>"John Larkin" <jjlarkin(a)highNOTlandTHIStechnologyPART.com> wrote in message >>news:di3r0613v9lkk2p8dq06snde2hgopt87fr(a)4ax.com... >>> Newer stuff, like EclipsLite, works at 3.3 volts, and some at 2.5. I >>> do mixed-mode PECL and cmos/FPGAs off a +3.3 volt supply. >> >>That's kinda what prompted the question -- Joerg has been pointing me at some >>of the high-speed logic from the likes of Micrel, and I was reading up some in >>the old MECL System Design Handbook so that I hopefully won't embarrass myself >>too badly when I go to use some of it. :-) > >Cool stuff. ...and easy to use. Just be careful when doing wired-OR. It seems like a natural in ECL, until you start to think about the speed of light. John
From: Joerg on 7 Jun 2010 21:33
Joel Koltner wrote: > "Joerg" <invalid(a)invalid.invalid> wrote in message > news:875jfpFtf6U1(a)mid.individual.net... >> Joel Koltner wrote: >>> Seems like a belt is an awfully inexpensive addition for the >>> flexibility it provide in terms of being able to locate and size your >>> generator independently of the engine itself, to a large extent! >> But it can snap, and it usually does so on a Saturday night in the >> boonies. About 30 miles past the sign that read "No services next 60 >> miles". Sez Murphy :-) > > Oh, alright -- I suppose the likelihood of enough teeth breaking off the > generator's drive gear so as to completely disable it is pretty much nil... > It didn't have any gears, it sat right on the drift shaft, behind the the fan (also on that shaft). The points were also right there and those did require the occasional TLC. The outside of the fan had the receptacle for the crank which I frequently used because the price of a 6V battery was kind of high for a student. > Do you currently carry about one of those "emergency" replacement > belts... just in case? (E.g., > http://www.amazon.com/V-320-EMERGENCY-FAN-BELT/dp/B000KKN8UY ) > Cool. Not sure it they'd work on my car though, it has a pretty wide belt. > I've never had a belt fail in my entire life, although I have had a > radiator hose spring a leak while in the road... and then on another car > the (plastic) radiator itself crack... > I had belts snap on me. And the occasional alternator fall out of the car. Chrysler screwed up the mount by using an aluminum strut. After making my own from steel that never happened again. -- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM. |