I don't think the sections are connected other than the substrate and supplies. But if 0V is your negative supply you'd technically be outside abs max on this chip.
But seriously, I would not do this. The LM319 can start doing a tarantella dance inside. I have seen EMI cases where the root cause was similar, mostly some unintended "ring oscillator" in an FPGA.
-- Regards, Joerg http://www.analogconsultants.com/
Yes.
You think wrong. Most varieties of LM339 and LM324 share a common PMOS current mirror distribution system.
That's down-right liverwurst ;-)
Found a nice Jewish diner... apple sausages for breakfast :-)
If you can not draw an LM339 schematic from memory (I can), you should not comment ;-) ...Jim Thompson
[On the Road, in New York]
--
| James E.Thompson, CTO | mens |
| Analog Innovations, Inc. | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| Phoenix, Arizona 85048 Skype: Contacts Only | |
| Voice:(480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |
I love to cook with wine. Sometimes I even put it in the food.
One of my more brilliant colleagues had the nick-name "Puker" - "barfer"
- because of his effect on design review panels (which happened to be the rest of the guys he was working with). He was named as inventor on some 25 patents, so the effect was clearly transitory.
The extended opinion includes phrases like - "unless you know what you are doing" and "unless you can live with a fairly slow and cranky comparator, or an op amp that needs a great deal of compensation".
Prudent.
-- Bill Sloman, Nijmegen
I brought the 3.3V regulator onto this board, so I had someplace to park one of the inputs, and grounded the other. Grounding one will be fine--the current source won't saturate, so the bias won't collapse.
The schematic (page 7 of the NS datasheet ) shows the bias current source going "to the other section", which was what got me going in the first place. I find it a bit hard to believe that a comparator whose input stage is turned off and whose internal bias has collapsed is likely to oscillate, but I suppose weirder things have happened.
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 845-480-2058 hobbs at electrooptical dot net http://electrooptical.net
It's pretty amusing, watching people's reactions when you do something weird like that. A dozen years ago, when I first presented the Footprints concept (*) to the August Technical Arbiters at IBM Watson Lab, there were two sorts of reactions, which were like photographic negatives of each other. Some folks thought that using LEDs as ultralow leakage switches and biasing them with photocurrent was cool, and some of them thought it was horrible. Nobody could say it didn't work, though. ;)
Cheers
Phil Hobbs
(*)
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 845-480-2058 hobbs at electrooptical dot net http://electrooptical.net
It's worth squinting hard at the datasheets, though; a bit of hysteresis is good for LVDS but can be a very bad thing in a comparator application. MC1489 wasn't the same as MC1489A, as I recall, because sometimes you need that hysteresis. And sometimes you abhor it (but it WAS documented).
In the old days you could also analyze the component-level schematic for an IC. It's annoying that 'modern' design documents omit the schematic, as though it was a redundancy. It isn't. It's a map of the territory, and VERY important for anyone who will ever stray from the well-beaten trail.
I haven't seen it with the LM319 but with other comparators of similar architecture. Are you sure the bias collapses? The 18k resistor supplying the bias goes straight to V+.
-- Regards, Joerg http://www.analogconsultants.com/
It sure looks like it to me, just from saturation of Q6 and Q7.
When Q1 and Q2 turn off, the other Q2 (probably supposed to be Q22) turns into a 600 ohm resistor to (V-) + a Vbe drop, which will drag down the base of Q21. That might be enough to pull Q6 and Q7 out of saturation, but I sort of doubt it.
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 845-480-2058 hobbs at electrooptical dot net http://electrooptical.net
That'll cause a voltage divider of 3k on the upper side and 2k plus
250ohms (well, in really twice that because the current splits up) on the lower side. Plus Vcesat but that's low. So the emitters of both Q8 and Q9 would hover slightly above half V+, assuming you have V- grounded.
But anyhow, I don't quite see how that would starve the circuitry around Q18, Q19 and Q20. As far as I can see the emitter of Q20 is the only node piped over to the other half.
-- Regards, Joerg http://www.analogconsultants.com/
Hmm, well, the emitters of Q3 and Q4 will put the bases of Q6 and Q7 a Vbe below V+, so they only need 200 uA to saturate.
Cheers
Phil Hobbs
-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics 160 North State Road #203 Briarcliff Manor NY 10510 845-480-2058 hobbs at electrooptical dot net http://electrooptical.net
Yes, you are right, that ought to saturate Q8 and Q9 pretty well. The output stage could still sing the blues though. But I don't know if it would in this compator.
And still, Q18 through Q20 should remain unaffected.
-- Regards, Joerg http://www.analogconsultants.com/
I take a lot of business away from people who are prudent.
John
[...]
You, too? The chip job I came back from yesterday is one where a major semiconductor company chickened out on us.
8-D-- Regards, Joerg http://www.analogconsultants.com/
Isn't that fun, when you do things that prudent people decided couldn't be done? People treat you like a magician.
The trick is carefully calibrated risk. I have recently bailed on a couple of projects that were, potentially, quicksand nightmares... one insane arbitrary waveform generator, one fast water-cooled multi-kilovolt pockels cell driver. Older and wiser, I guess.
John
:-)
The Pockels driver would have piqued my interest. An engineer in a German NG had a similar issue. I suggested to him to look at the new IXYS 4kV FET (Mouser finally has some) and TV flyback tubes. To my utter amazement they could (maybe still can) buy the skinny 50's era PL81 tube over there, NOS, around $2. It was also used in very early "portable" TVs. IIRC that can handle 7kV but you have to drive it at over 100V on the gate and probably also drive the screen grid hard.
-- Regards, Joerg http://www.analogconsultants.com/
:-)
Interestingly, the company that's providing the driver is German, starts with "B". The problem with toobs is that the peak current requirement is around 70 amps or something crazy like that. All I have to do now is provide the logic drive to the beast.
IXYS acquired DEI and their amazing fast low-inductance mosfets. They have some nice parts in the 1KV range, and they's have to be stacked into radical h-bridges. The project I'm tangled in has so much to do, I bailed on that one.
If the German one blows up or something, we could look into it again, after we get the small stuff out of the way. Fast power stuff is fun, but it's labor intensive.
John
:-)
Don't remember one that starts with "B", the project in the NG is at a university.
Not sure how many peak amps they can do but I've got tubes here that can stomach well north of 5kV and are the size of a fat glass of pickles. The plate connector is a mean looking 10mm diameter steel rod coming out of the top.
Check out this little dude, Digikey now has stock as well. But I'd pot the pins at that voltage level:
I have something similar coming up soon but it has to do with a (very) rapid snuffing of a magnetic field. That may require kilovolts as well.
-- Regards, Joerg http://www.analogconsultants.com/
. To my utter
"Gate" ?
Jim
That's what they called the control grid back in the 50's era.
It's one of those dual-gate toobs.
John
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.