Bipolar power intergration with CMOS Logic Chips

[liquids]

I have a bipolar supply.

My CMOS logic chips run one sided though: one side is V+, the other is ground.

I use bipolar powered op amps into an actual comparator (LM311) - a go between the op amps and CMOS logic.

If I run the LM311 comparator off the bipolar power, it would be easier for me...rather than using Vref as I've been doing...


With an LM311, if Pin 4 goes to V- and Pin 1 goes to ground, will the comparator's output still swing from 0V to V+?


I guess what confuses me is that while I think that is true from what I read, but for example, the EHX DOM (http://www.freeinfosociety.com/media/images/4113.jpg) uses a LM311 to drive a 4013, but the LM311 is powered as single sided. Maybe there is more reason to that particular situation than I understand....but I don't want to blow any more chips up than I already do.   

[soffa]

Liquids...

Have you looked at the EHX rackmount guitar synth schematic? I think it will do what you want.

[liquids]

You sound like StephenGiles!      

I've poured over that schem dozens of times over...but I have a hard time following it...my brain is generally not linear (I am ADD if you will) so it makes it difficult to glean much from it historically...but I'll keep trying, at least once more in the near future now here at your suggestion.    

[Cliff Schecht]

I deal with problems similar to this in the lab all of the time. People for whatever reason don't seem to understand that single supply chips don't like to receive negative voltage on their inputs. It makes many IC's do weird things if they don't go poof immediately.

The solution is very simple, you can use either a single Zener diode or a Schottky from ground to input (cathode to ground, anode to input) and another from input to supply (cathode to input, anode to supply). The Schottky diodes clamp the signal to the supplies that the logic input sees so that you are clearly passing the logic thresholds. For the Zener it's just choose one that is rated at the logic IC's supply voltage and slap it in from input to ground (cathode to ground, anode to input).

[liquids]

Yeah.  I've done that at points, but I just squirm at that work around.  I'd rather go long hand than use diodes for close enough...I don't totally understand why.

I'll keep trucking.

If I could get my depressed butt to the workbench and scope the output of a comparator arranged in various configurations, I'd know if I simply NEED a vref at the inputs to make it swing ground to V+ at it's output (op amp or regular based) or can get the comparator to do so without one and without 'gimmickery,'  pardon the lack of a word with less negative connotations.

Likewise, the LM311 and it's dual/quad kin are the only true comparator I have.  If not this one, I imagine there is a comparator chip out there that takes bipolar power but swings single sided for the sake of CMOS logic.  Got to be.

[Cliff Schecht]

I don't consider Schottky diode clamps a workaround. They are a safety precaution to protect your logic chips really. Many IC's (especially microprocessors and such) have Schottky diodes built into the I/O pins to protect them from overvoltage situations. They are current limited to a point but these monolithic diodes can't handle a lot of power (they have to be small both to keep capacitance down and to keep size down in layout) and are pretty easy to blow up. A set of Schottky clamps cost $0.02 and can handle over 500mA of pulsed current. They make it simple to protect your chips.

If you run your comparator and logic off of the same positive supply with the comparator also getting negative voltage then you do need to clamp the output, at least with your standard LM311 type comparator. An open collector style comparator comparator is what you are looking for (LM3x9 series aka LM339). Take a look at the datasheet and notice the output stage is a BJT with the collector floating and emitter grounded (giving you a logic level output with a bipolar supply). You connect a resistor from supply to the collector and take signal off of the collector right into your logic gates. Requires one resistor per comparator output. You might even be able to find a chip that has this output stage with the resistor built in so you don't have to muck with that extra resistor in layout (sorry I don't know a part # for anything like this). If a solution like this doesn't do it for you, then I can't help ya!

[Earthscum]

I was looking at the datasheet on Friday for the 311 and noticed that it would directly interface with the 7474 TTL Flip Flop, possibly without even using a resistor? Never really looked at the LM311 much until this thread. Looks to be quite a useful chip.

So, sorry for the small hijack...Q: if I run this on 12V, and I'm interfacing to a 5V circuit, the output of this would be: resistor to 5V, ground to "ground" of the next stage, and out to the next stage, and that will keep the swing within that 5V rail?

[liquids]

I don't consider Schottky diode clamps a workaround. They are a safety precaution to protect your logic chips really. Many IC's (especially microprocessors and such) have Schottky diodes built into the I/O pins to protect them from overvoltage situations. They are current limited to a point but these monolithic diodes can't handle a lot of power (they have to be small both to keep capacitance down and to keep size down in layout) and are pretty easy to blow up. A set of Schottky clamps cost $0.02 and can handle over 500mA of pulsed current. They make it simple to protect your chips.

If you run your comparator and logic off of the same positive supply with the comparator also getting negative voltage then you do need to clamp the output, at least with your standard LM311 type comparator. An open collector style comparator comparator is what you are looking for (LM3x9 series aka LM339). Take a look at the datasheet and notice the output stage is a BJT with the collector floating and emitter grounded (giving you a logic level output with a bipolar supply). You connect a resistor from supply to the collector and take signal off of the collector right into your logic gates. Requires one resistor per comparator output. You might even be able to find a chip that has this output stage with the resistor built in so you don't have to muck with that extra resistor in layout (sorry I don't know a part # for anything like this). If a solution like this doesn't do it for you, then I can't help ya!

Hmm.  Well...elsewhere PRR rightly pointed me to the TI datasheet.  I only had page 1 of the fairchild datasheet printed out previously...and the TI datasheet is a bit more thorough.
Anyhow...looks like the outputs are even labeled differently...for the LM311 in the TI, it shows and labels pins 1 and 7 as the output emitter and collector.  I mean...convincingly, if the emitter is grounded and the collector has a pull up resistor, and the collector output is tapped...output can't swing below ground, right?   And if I want a pulse that swings the whole supply rail, I can connect emitter to the negative rail and get a rail to rail pulse...
Cliff, you're far more advanced than I am.  I said this all so as not to fry any CMOS.  But I don't see how I'd be endangered, with that info....?

[Cliff Schecht]

Yes the emitter gets grounded which means the output voltage can't go below a Vce drop above ground (about 0.2V above ground which is *perfect* for your logic chip). Also you can size the resistor from collector to positive supply to drop the voltage down to standard logic levels. Something like a 10k resistor is enough to limit the swing below 5V from a 9V supply.

[liquids]

Cool.

The Logic chips are running the same 12 V+ as the comparator, so all should be well even with a small resistor value, AFAIK.  The fewer voltages to run around the pcb the better...