I do not use simulators, except the one between my ears.

So the old slide rule is still up and running?

You can also add a totem-pole output if you need more drive, although it's not rail-rail.

 

Nope but i still have a few slide rules kicking around, somewhere. My desktop is dying a slow death from imminent HD failure, but this old lappie refuses to die, for which I am very grateful. I finally have the means to save a little money to eventually upgrade my hardware. All I can do for now is dream up these circuits, which is what I do pretty much. It kinda sorta keeps me busy.

like I said earlier, one of my design goals is driving MOSFETs, I am not a fan of totem pole outputs.This is not TTL.

 

This is not TTL.

No.
But it's a simple way to provide an open-collector output comparator with a push-pull output for more drive.

 

OK, i really want to get away from using two comparators as comparators Nothing is ever exactly alike, even on the same die. It is very similar to the first scheme I drew and ultimately rejected. I am using the prorogation delay to prevent the shoot through issue. Your approach brings the issue back (I think). I do not use simulators, except the one between my ears.

Have you guys noticed we are up to 20 posts over this silly little circuit concept?

Hi,

"ultimately rejected"
I for one did not say that the circuit was rejected, i agreed that it was a good idea conceptually. That is like an initial design thought that has some specific goal in mind.

What i was doing was looking at the pitfalls of taking the initial idea to full fruition and some ideas how to overcome those pitfalls. So your idea still carries though, and not rejected. I dont know of many circuits that are first thought of that work out in practice, until at least some practical aspects are considered, weighed out, and modified, but the initial idea is what starts the ball rolling and that's your idea. I think it is good otherwise i would not spend so much time on it. I think it is well worth 20 posts or 100 posts. It's both interesting and very useful.

I dont see how you are using the propagation delay to mitigate the shoot through or possible shoot through, maybe you could explain that a little more.

Yes, the input anti parallel connection has it's problems too, and has to be matched to the application. If we have different input offsets then the two outputs could be at the same state for hours. That would be rare, but we have to consider everything if we want to develop a drop in replacement for a rail to rail device.

The way i handled it in the TTL version was to simply place a current limiting resistor in series with the collector. The resistor is made low enough to still get near rail to rail output yet still limit current during the shoot through period. For this circuit though we can probably come up with a better idea.

Also keep in mind that for full H bridge circuits,shoot through is never allowed, even for 10ns. The problem is not only about the higher current it's also about the dynamics of the power supply rail itself which gets pulled down (spike on the power line) when the shoot through occurs. That means it could create a glitch in any associated logic system.

The main point though i think is that the original idea is great, we just have to work on it a little that's all in order to perfect it.

 

The TI LMx39x, LM2901xx Quad Differential Comparators data sheet states:

View attachment 155064

So it would appear that significant shoot-through currents should be avoided for those devices, at least at higher operating frequencies.

I agree that connecting the comparators with inputs in anti-parallel rather than series, will result in a shorter shoot-through time.

Below is the LTspice simulation of the two circuits with an LM339/LM393.
As can be seen, the shoot-through current is much higher for the series circuit.

Note that I used a high-speed saturated-switch transistor.
Using a general-purpose BJT such as a 2N3906 or 2N2905 will result in much higher and longer feed-through current pulses, due to their long saturated recovery time, as shown in the second simulation.

View attachment 155067

View attachment 155068

Hello,

Good point about the other edge shoot through. The storage time can be as much as 10us. That's very significant. What might help 'a little' is to add a base emitter resistor that helps to turn the transistor off.

 

U2a is in full control of Q1, just as it controls U2b, Q1 will turn on/off from U2a, There is a race condition in play here. If U2b is even a bit behind Q1 switching time then No shoot through can occur, I have been working on this concept for a very long time. I fully accept I can be wrong, but in this case I do not think I am.

 

U2a is in full control of Q1, just as it controls U2b, Q1 will turn on/off from U2a, There is a race condition in play here. If U2b is even a bit behind Q1 switching time then No shoot through can occur, I have been working on this concept for a very long time. I fully accept I can be wrong, but in this case I do not think I am.

Hi again,

I am not following how U2a controls Q1, are we looking at the original circuit with U2a on top?

I took your original circuit and inserted a series resistor Rs in series with the collector of Q1 so we can keep track of the 'output' of Q1 as well as the two comparators. With no load, if there is ANY current in Rs that means we have a shoot through condition. The condition exists for the time difference between device delays TdQ and TdA. Note that TdB has no ill effect except when we consider slew rate and then it does have an effect also, but i ignore that for now.

The equivalent logic circuit shows the devices drawn as inverters and one buffer which can also be an inverter.
With no load, the only way Rs can see current flow is if the delays TdQ and TdA are different during a switch transition. Since the transistor is faster TdQ is smaller than TdA, and therefore when there is a switch transition Rs conducts briefly. The time it conducts here would be the time t=TdA-TdQ. This can occur on either edge so both edges have to be checked.

Note Rs can be for example 1 ohm, 0.1 ohm or even 0.001 ohm for convenience.

 

Sorry I am still fighting brain fog along with massive depression, which makes me fairly absent minded. Kinda like Hitch Hikers Guide to the universe, My mind is frequently absent. You are right, of course. I figure Q1 is going to be first to switch no matter what.

 

Sorry I am still fighting brain fog along with massive depression, which makes me fairly absent minded. Kinda like Hitch Hikers Guide to the universe, My mind is frequently absent. You are right, of course. I figure Q1 is going to be first to switch no matter what.

Hello again,

Oh sorry to hear about that. I hope you can find some good meds, maybe Welbutrin? Maybe you tried that already maybe not, just thought i would mention that one.

I still think this circuit is so cool though so i hope you dont give up on it. I cant tell you the number of times i wished i had a comparator with a strong pullup rather than the usually 'soft' resistor pullup. I think it is also a very interesting topic so i am happy you brought it up.

 

Actually there will be several spin off thread coming from it. I have am appointmnt with a neurologist and a psychiatrist scheduled. You have probably noticed I am not a shy person in the slightest. I am not concerned about privacy, I have nothing to hide.

 

Actually there will be several spin off thread coming from it. I have am appointmnt with a neurologist and a psychiatrist scheduled. You have probably noticed I am not a shy person in the slightest. I am not concerned about privacy, I have nothing to hide.

Hi,

Oh that's great. I wish you all the luck with your appointments and successful treatment.