I'm trying a design with parallel CFP output stage and I have a dilemma.
Since the CFP stage l has two outputs in parallel, I need emitter resistors to take care of bias differences in the transistors. The trouble is the connection of the driver emitters to the output collectors. The output collectors have resistors connecting them to the output (the output collectors become the "emitter" of the CFP stage.)
How can I properly connect the emitters of the drivers to the collectors of the outputs so that they will be stable and still have enough feedback to be unity gain? Could I use low value resistors that are higher than the emitter resistor values and just connect one to each output collector and both of their other ends to the driver emitters?
If anyone needs more clarity in my question here, just ask, and I'll post a diagram of what I have and what I'm trying to achieve.
Since the CFP stage l has two outputs in parallel, I need emitter resistors to take care of bias differences in the transistors. The trouble is the connection of the driver emitters to the output collectors. The output collectors have resistors connecting them to the output (the output collectors become the "emitter" of the CFP stage.)
How can I properly connect the emitters of the drivers to the collectors of the outputs so that they will be stable and still have enough feedback to be unity gain? Could I use low value resistors that are higher than the emitter resistor values and just connect one to each output collector and both of their other ends to the driver emitters?
If anyone needs more clarity in my question here, just ask, and I'll post a diagram of what I have and what I'm trying to achieve.
You must definately keep the resistors between the collectors and the output . There are two schools of thought about the resistors between the emitters and the rails. One is you need them to avoid "current hogging"; typically o.1R. The other (per experiments by Douglas Self) is that you don't because the PCB traces themselves have enough resistance.
I've done it both ways and think Self is probably right. At any rate I used a thermocouple attachment to my DMM to convince me there was no significant temperature difference between devices (each was on it's own heatsink), which should indicate an absense to hogging.
However, just because I got away with it once doesn't mean it will work next time. I'm waiting now for a couple of PCB I laid out to come back. I went to some trouble to be sure that the power trace to each output device was the same length. It will be interesting to see how it works out.
I've done it both ways and think Self is probably right. At any rate I used a thermocouple attachment to my DMM to convince me there was no significant temperature difference between devices (each was on it's own heatsink), which should indicate an absense to hogging.
However, just because I got away with it once doesn't mean it will work next time. I'm waiting now for a couple of PCB I laid out to come back. I went to some trouble to be sure that the power trace to each output device was the same length. It will be interesting to see how it works out.
I was already quite certain and determined to keep the emitter resistors on there, since this monster has the capability of really frying if something goes wrong.
Right now, I have the amplifier working (it sounds very good).
And as it is, there is a wire from the emitter of the driver connected to the collector of only one of the output transistors.
This is just to provide 100% feedback (hence compound feedback pair). And the amp tests fine and works properly.
I'm just wondering if it's important to have feedback from both output collectors or is the one from one of them enough? I mean, the amplifier works very well, is stable into low loads, sounds very good, and is clean. I don't think I need to make anything better but I just want to get an idea of the practices other people have been using.
EDIT: Another reason I should keep the resistors is that, even though Self mentions PCB traces, I'm using heavy point to point wiring on the output transistors and the wires will have less than .01ohms resistance ...
Right now, I have the amplifier working (it sounds very good).
And as it is, there is a wire from the emitter of the driver connected to the collector of only one of the output transistors.
This is just to provide 100% feedback (hence compound feedback pair). And the amp tests fine and works properly.
I'm just wondering if it's important to have feedback from both output collectors or is the one from one of them enough? I mean, the amplifier works very well, is stable into low loads, sounds very good, and is clean. I don't think I need to make anything better but I just want to get an idea of the practices other people have been using.
EDIT: Another reason I should keep the resistors is that, even though Self mentions PCB traces, I'm using heavy point to point wiring on the output transistors and the wires will have less than .01ohms resistance ...
Ok. I think I see what you mean. Never thought of doing it that way. I think it would not be a problem so long as each of the parralleled output devices are behaving virtually identically. But if that is not the case, then . . .
It would seem to me more reassuring to have both output devices in the feedback network, but if it is working why fix it. However, I don't think you can assume this will work with a second amp since it depends on the output devices being quite similar, more similar than just being within manufacturer's spec.
It would seem to me more reassuring to have both output devices in the feedback network, but if it is working why fix it. However, I don't think you can assume this will work with a second amp since it depends on the output devices being quite similar, more similar than just being within manufacturer's spec.
Well, I doubt these transistors are perfectly matched or anything...
I just took them out of the box and put them in, not matched.
The amp works well and has a stable output.
I think that, being the feedback loop, as long as the specs are close, the circuit works fine.
Does anyone have more ideas on this one???
I just took them out of the box and put them in, not matched.
The amp works well and has a stable output.
I think that, being the feedback loop, as long as the specs are close, the circuit works fine.
Does anyone have more ideas on this one???
I asked this very same question only a couple of months ago!
The consensus was that you use emitter resistors for each device, then you have 2 options for the collectors.
1. Join them all together then connect this node to a single collector resistor.
2. Use individual collector resistors and use a feedback resistor for the driver stage for each o/p device. The common node here is the emitter of the driver transistor.
edit: check out 'Resistors in the output stage' thread, on page 2 is my question with the schematic, and all the answers.
The consensus was that you use emitter resistors for each device, then you have 2 options for the collectors.
1. Join them all together then connect this node to a single collector resistor.
2. Use individual collector resistors and use a feedback resistor for the driver stage for each o/p device. The common node here is the emitter of the driver transistor.
edit: check out 'Resistors in the output stage' thread, on page 2 is my question with the schematic, and all the answers.
Well, that's what I've been thinking.
The amplifier is highly stable, it sounds excellent, and the
numbers all add up.
I sat in my math class today and trying to actually "use"
the time, I did some serious work with my design and
I found that by theory, it should actually be stable and
proper. As long as the output transistors are the same,
the result seems to be fine.
Either way, I'm certainly happy with the amplifier, I couldn't
have asked for a better result being that this my first design
that I did right out of my head.
The sucker is nice too, when it's finished, it'll be four channels
at 500W per channel. It is very juicy, and smooth.
My ears are satisfied, and really, that's what matters.
The amplifier is highly stable, it sounds excellent, and the
numbers all add up.
I sat in my math class today and trying to actually "use"
the time, I did some serious work with my design and
I found that by theory, it should actually be stable and
proper. As long as the output transistors are the same,
the result seems to be fine.
Either way, I'm certainly happy with the amplifier, I couldn't
have asked for a better result being that this my first design
that I did right out of my head.
The sucker is nice too, when it's finished, it'll be four channels
at 500W per channel. It is very juicy, and smooth.
My ears are satisfied, and really, that's what matters.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.