PNP or NPN Transistors

[Nigel Goodwin]

Which is consistent with what I'm saying.
As far as I can tell, Nigel believes that a sink is an electron sink, and a source is an electron source. The industry standard is just the opposite.

No, I believe no such thing, it makes no difference which it is - a sink 'sinks' current down to ground, and a 'source' sources current from a rail either above, or below ground.

As far as I'm aware there is no industry standard?, and the link I provided supported my point of view - sinking or sourcing isn't dependent on your point of view regarding current, which you seem to think it is?.

I notice you said you generally think of 'conventional current', this is obvious throughout this thread - I rarely think of either, it has no bearing on almost all ciruits, I simply think of current flowing from a higher point to a lower point, I've no interest in if the higher point is positive higher, or negative higher.

 

[Nigel Goodwin]

Perhaps this might help?, a diagram shorn of all distractions:

 

[3v0]

I went back to the post which started this debate. We were talking about NPN vrs PNP transistors.

How about this:

You can use a NPN to sink a device.

You can only use a PNP to source a device if the base voltage is >= the collector voltage. If it is not you need to use a totem pole setup (2 transistors) which is more expensive.

Had I said You can only use a PNP to source a device I would have been wrong. What I said was You can only use a PNP to source a device if the base voltage is >= the collector voltage. The sentance does not say that only PNP transistors can be used to source. Instead it tells you when you can use a PNP to do so.

A lot get posted but we need to take some care to read what is written.

Regarding current: The current flow debate is about as old as the study of electricty. When I was in school EE had it one way and physics the other. I have real doubts we are going to solve it here.

EDIT: Nigel, I do like the last diagram you posted.

 

[Nigel Goodwin]

Regarding current: The current flow debate is about as old as the study of electricty. When I was in school EE had it one way and physics the other. I have real doubts we are going to solve it here.

That's my point, I'm not trying to 'solve' it, just make the point that it's of no concern.

EDIT: Nigel, I do like the last diagram you posted.

Took a bit of editing!

BTW, anyone wondered why there were never 'NPN' and 'PNP' valves?, not that those terms apply of course. Would a 'PNP' valve even be possible?, it is for FET's as we all know.

 

[Pommie]

Would I be right in thinking that you can take most circuits and reverse them. That is, swap NPN for PNP, reverse diodes & electrolytics and reverse the supply connections.

Mike.

 

[dknguyen]

Would I be right in thinking that you can take most circuits and reverse them. That is, swap NPN for PNP, reverse diodes & electrolytics and reverse the supply connections.

Mike.

I seem to remember us doing that in our first analog circuits course with op-amp circuits. Something about flipping supplies and changing all the MOSFETs or something. I forget what it did exactly or why we did it. I think it came up with a functionally identical circuit, except if that's the case I dont' know why we bothered doing it at all.

 

[Nigel Goodwin]

Would I be right in thinking that you can take most circuits and reverse them. That is, swap NPN for PNP, reverse diodes & electrolytics and reverse the supply connections.

Yes, makes no difference at all - where it obviously fails though is IC's, where you can't get an inverted IC but you could probably arrange the circuit suitably anyway with most IC's.

 

[jpanhalt]

Would a 'PNP' valve even be possible?, it is for FET's as we all know.

It's hard to envision positive holes in a vacuum; however, if you consider positive carriers as meeting the definition of "PNP", then a plasma switch might come close to a "valve" in which the carriers are positive. I don't know of any that use protons for switching; although, proton beams are well known. Here is one site that discusses plasma switches:
**broken link removed**

It is obviously not for the portable radio
John

 

[Pommie]

Yes, makes no difference at all - where it obviously fails though is IC's, where you can't get an inverted IC but you could probably arrange the circuit suitably anyway with most IC's.

But, if manufacturers swapped there internal circuits could you get a "pnp" 741? Are there any negative rail op amps?

The other place it fails is with MOSFETS. Why do P channel MOSFETS need a gate voltage above the supply rail whereas PNP transistors don't? Which bit of "positive" don't they understand.

Mike.

 

[Nigel Goodwin]

But, if manufacturers swapped there internal circuits could you get a "pnp" 741? Are there any negative rail op amps?

I see no reason why the entire circuit inside a 741 couldn't be inverted, but there's probably not much point, as they are designed to be used on split supplies anyway.

The other place it fails is with MOSFETS. Why do P channel MOSFETS need a gate voltage above the supply rail whereas PNP transistors don't? Which bit of "positive" don't they understand.

Think of the example of a valve - a triode is turned fully ON by default, in order to turn it OFF you need to apply a negative voltage (with respect to the cathode) to the grid. Now think of a 'P channel' valve, cathode to positive supply, you would need to take the grid positive of the positive supply to turn it OFF.

 

[dknguyen]

How do things like the MAX912 work? It's a comparator, not an op-amp. But it's dual and single supply but only has a +V and -V pins with no ground pin. But somehow it knows to always output either 0V or 3V irrespective of whether duals supplies or single supplies are being used. How does it know that an output LO is equal to halfway between the power rails (when dual supplies are used) and to output a voltage equal to the negative rail (single supplies are used and the negative power pin is connected to ground).

I can't see how an single/dual supply op-amp with just two power pins knows the difference either. Like if the inputs are equal to each other, how does it know whether to output a voltage equal to the negative power (using single supply where the negative rail is ground) rail or half-way between the two power rails (using single dual supplies).

EDIT: Nvm about the op-amp bit. THe op-amp just does everything relative to it's power rails. The description is when an op-amp is being used open-loop as a comparator and it acts funny when the voltage on the inputs is equal

 

[Roff]

So, apparently you guys agree with the labels in the schematic below?

 

[dknguyen]

The rule is flip the voltage supplies. Not reverse their polarity.

 

[Roff]

The rule is flip the voltage supplies. Not reverse their polarity.

They are not reversed. All I did was change which rail was called "chassis". Both circuits deliver exactly the same amount of current to the load.

 

[Nigel Goodwin]

So, apparently you guys agree with the labels in the schematic below?

Yes, sink or source depends on the supply rail polarity employed.

 

[Roff]

Yes, sink or source depends on the supply rail polarity employed.

That's remarkable. That seems to mean that I can make identical ramps (except for offset voltage) with either a current source or a current sink, depending on how I define ground. Agreed?

 

[Roff]

The description is when an op-amp is being used open-loop as a comparator and it acts funny when the voltage on the inputs is equal

Funny in what way?

 

[Nigel Goodwin]

That's remarkable. That seems to mean that I can make identical ramps (except for offset voltage) with either a current source or a current sink, depending on how I define ground. Agreed?

I might be falling in a trap here?, but one way you're rapidly charging the capacitor, then slowly discharging it, the other way you're slowly charging it, then rapidly discharging it - works either way.

However, I can't help but notice your diagrams have expanded from sink and source to constant current sinks and sources.

 

[Roff]

I might be falling in a trap here?, but one way you're rapidly charging the capacitor, then slowly discharging it, the other way you're slowly charging it, then rapidly discharging it - works either way.

However, I can't help but notice your diagrams have expanded from sink and source to constant current sinks and sources.

Well, if sinking and sourcing are defined for logic gates and switching transistors, shouldn't the same definitions apply for current sources and sinks?
As I think you pointed out, I believe sink and source refer to current direction, while you believe it has to do with which rail is ground. In many systems, ground is arbitrary, and may not even be defined. Does that mean we have to label a rail before we can decide whether a switch (e.g., transistor) is sourcing or sinking? Let's say I have an ordinary TTL open collector gate. I think we agree that, if I connect it to +5V and GND, the output is sinking. Now, let's connect it to GND and -5V (emitter to -5V). If I understand you correctly, according to your way of thinking the output will now be sourcing (we could draw it upside down if you like GND to be on the bottom).
In my world, the output is still sinking. I'm getting that sinking feeling. Are you getting that sourcing feeling?

 

[Nigel Goodwin]

In my world, the output is still sinking. I'm getting that sinking feeling. Are you getting that sourcing feeling?

I'm getting that 'going down to open a bottle of cider feeling'