Swamping in Output Circuits?
This has been a fun thread, but I'm out of here.
The output circuits of most power amps are not common collector. That would require the speaker to be connected to one of the high or low voltage rails. It isn't, it's grounded. And the approximate value of the drive at the output is also at ground level.
The output circuit is an emitter follower, which is a common emitter circuit with the load on the emitter side, instead of the collector side.
Also, I've never seen an output circuit that was "swamped", unless you have a different definition than the one I was taught in school. In order for the output devices to be "swamped", the emitter resistors would have to have values of 2 to 4 ohms. That's not my experience. Most of the time, the emitter resistors are fractional values between 0.12 ohms and 0.47 ohms. Essentially, they are a dead short. That's because they are only intended to be safety devices and current limiters.
There is a reason why the output cannot be "swamped" at the emitters in an emitter follower output. The reason is damping factor. Most solid state amplifiers brag about that. Damping factor is the ratio of the speaker impedance to the impedance of the amplifier output. For example, an amplifier with a damping factor of 100 into an 8 ohm speaker has an effective output impedance of 8/100ths of an ohm.
There's no way to achieve that if the emitters are "swamped" with a 2 to 4 ohm resistor. No, not even with some major negative feedback.
Now, about this matching business. I was pretty young in 1964, when I started doing stereo. The first couple of times that I replaced output devices, I got lucky and everything worked OK. It was beginners luck and fate favoring the foolish.
It took me a while to figure out what those little dabs of paint on the factory original output devices were all about. The brand new transistors/Darlingtons/MOSFETs that I was buying didn't come with those dabs from the OEM supplier. So how do they get there?
Well, on the manufacturing line, there is a tech who marks the output devices for their matching tolerance. For example, let's say that a certain manufacturer expects the output devices to cover a range of hFEs from 1200 to 2200. The tech will pre-test the devices and mark them with a dab of paint to indicate the hFE code: 1200 to 1300 is black, 1300 to 1400 is brown, 1400 to 1500 is red, etc.
Then the installer on the line simply picks two devices with the same coloured dab of paint and installs them, knowing that the hFEs will match within 100. Pretty slick, huh?
You know, come to think of it, there actually was a solid state amplifier output circuit that was "swamped". Bob Carver built it to win a bet. The bet was that he had to produce a solid state amplifier with the same transfer function as a Futterman direct coupled tube amplifier. He won the bet by lowering the damping factor of the transistor amplifier down to the same damping factor as the tube amplifier. And he did that by increasing the value of the emitter resistors until the output was "swamped". I think the designer was Bob Carver. Is that right?
This has been a fun thread, but I'm out of here.
The output circuits of most power amps are not common collector. That would require the speaker to be connected to one of the high or low voltage rails. It isn't, it's grounded. And the approximate value of the drive at the output is also at ground level.
The output circuit is an emitter follower, which is a common emitter circuit with the load on the emitter side, instead of the collector side.
Also, I've never seen an output circuit that was "swamped", unless you have a different definition than the one I was taught in school. In order for the output devices to be "swamped", the emitter resistors would have to have values of 2 to 4 ohms. That's not my experience. Most of the time, the emitter resistors are fractional values between 0.12 ohms and 0.47 ohms. Essentially, they are a dead short. That's because they are only intended to be safety devices and current limiters.
There is a reason why the output cannot be "swamped" at the emitters in an emitter follower output. The reason is damping factor. Most solid state amplifiers brag about that. Damping factor is the ratio of the speaker impedance to the impedance of the amplifier output. For example, an amplifier with a damping factor of 100 into an 8 ohm speaker has an effective output impedance of 8/100ths of an ohm.
There's no way to achieve that if the emitters are "swamped" with a 2 to 4 ohm resistor. No, not even with some major negative feedback.
Now, about this matching business. I was pretty young in 1964, when I started doing stereo. The first couple of times that I replaced output devices, I got lucky and everything worked OK. It was beginners luck and fate favoring the foolish.
It took me a while to figure out what those little dabs of paint on the factory original output devices were all about. The brand new transistors/Darlingtons/MOSFETs that I was buying didn't come with those dabs from the OEM supplier. So how do they get there?
Well, on the manufacturing line, there is a tech who marks the output devices for their matching tolerance. For example, let's say that a certain manufacturer expects the output devices to cover a range of hFEs from 1200 to 2200. The tech will pre-test the devices and mark them with a dab of paint to indicate the hFE code: 1200 to 1300 is black, 1300 to 1400 is brown, 1400 to 1500 is red, etc.
Then the installer on the line simply picks two devices with the same coloured dab of paint and installs them, knowing that the hFEs will match within 100. Pretty slick, huh?
You know, come to think of it, there actually was a solid state amplifier output circuit that was "swamped". Bob Carver built it to win a bet. The bet was that he had to produce a solid state amplifier with the same transfer function as a Futterman direct coupled tube amplifier. He won the bet by lowering the damping factor of the transistor amplifier down to the same damping factor as the tube amplifier. And he did that by increasing the value of the emitter resistors until the output was "swamped". I think the designer was Bob Carver. Is that right?
Re: Swamping in Output Circuits?
An amplifier can be built using either common collector or common emitter circuits as outputs, but common collector is by far the most common.
Now, any amp that's designed such that the minimum gain of the output transistors is even close to what the drivers can reliably drive is going to have substantial problems with a variant load, such as a loudspeaker.
However, I'm sure amplifiers from way back sometimes had "creative" ciruits that could go bust if mismatched devices were used, due to cost constraints.
Rune
Excuse me, but that's not what I've been taught. In my world an emitter follower is a common collector circuit.chewrock said:
An amplifier can be built using either common collector or common emitter circuits as outputs, but common collector is by far the most common.
Now, any amp that's designed such that the minimum gain of the output transistors is even close to what the drivers can reliably drive is going to have substantial problems with a variant load, such as a loudspeaker.
However, I'm sure amplifiers from way back sometimes had "creative" ciruits that could go bust if mismatched devices were used, due to cost constraints.
Rune
Hi Chewrock,
With regard to emitter resistors.
They are not swamped, that is something to do with water.
They do not need to be 2r to 4r.
They are not a dead short.
They are not intended to be safety devices and current limiters.
Their main purpose is to maintain stability in the output device current and secondary (I think) to define the current gain more accurately.
They can be used as tapping points for a current or VI protection detector system, but this is optional and the correct circuit operation does not depend on the protection circuit.
Something was lacking in the way electronics knowledge was imparted by your school teacher!
With regard to emitter resistors.
They are not swamped, that is something to do with water.
They do not need to be 2r to 4r.
They are not a dead short.
They are not intended to be safety devices and current limiters.
Their main purpose is to maintain stability in the output device current and secondary (I think) to define the current gain more accurately.
They can be used as tapping points for a current or VI protection detector system, but this is optional and the correct circuit operation does not depend on the protection circuit.
Hi chewrock,
-Chris
That is very basically correct. Those amplifiers did a couple other things, but there was a resistor in series with the output outside the feedback network. Notice there were no non-linear fuses, just a "linear" resistor (I think it was wire wound). The TFM series adopted these concepts.
-Chris
Hi again,
Thanks very much for all of the replies, they've been very helpful.
I finally got my new batch of sanken sap15s and 100ohm trimmer pot replacements through today. I put the trimmers in the halfway 50ohm position, everything was installed correctly with mylar insulators, good thermal paste and all the old damaged components were replaced.
Powered it up and it works, sound and everything!
With no manufacturer circuit specs I'm a bit lost (I tried forums and ringing the company with no luck).
The board has just one trimmer per channel. I've been examining the sap15 datasheet as that's my only source for now. Like the datasheet, the amp circuit has the trimmer between the diode outputs of the pair of transistors, the reason for this is to adjust so that the current between these points measures 2.5mA. This will also have an affect on the idling current which should measure 40mA where the emittor points meet.
Still with me?
Well, i tried adjusting a trimmer on a channel, this seemed to have absolutely no effect on either the DC offset OR the forward current. The voltage across the trimmer increases or decreases as it's changed so that the current across the trimmer is always about 3.6mA (not 2.5mA).
The DC offsets by the way are 14mV on the left channel and 23mV on the right.
So.... any ideas? As you can see i'm not really that clued up
Cheers,
Andy
Thanks very much for all of the replies, they've been very helpful.
I finally got my new batch of sanken sap15s and 100ohm trimmer pot replacements through today. I put the trimmers in the halfway 50ohm position, everything was installed correctly with mylar insulators, good thermal paste and all the old damaged components were replaced.
Powered it up and it works, sound and everything!
With no manufacturer circuit specs I'm a bit lost (I tried forums and ringing the company with no luck).
The board has just one trimmer per channel. I've been examining the sap15 datasheet as that's my only source for now. Like the datasheet, the amp circuit has the trimmer between the diode outputs of the pair of transistors, the reason for this is to adjust so that the current between these points measures 2.5mA. This will also have an affect on the idling current which should measure 40mA where the emittor points meet.
Still with me?
Well, i tried adjusting a trimmer on a channel, this seemed to have absolutely no effect on either the DC offset OR the forward current. The voltage across the trimmer increases or decreases as it's changed so that the current across the trimmer is always about 3.6mA (not 2.5mA).
The DC offsets by the way are 14mV on the left channel and 23mV on the right.
So.... any ideas? As you can see i'm not really that clued up
Cheers,
Andy
Sorry, me again.
The emittor resistors are included inside the transistor package, and are 0.22ohms, how would you measure the voltage across this to get the quiescent current? Because both emittors in the pair meet (creating the amplified output), this means that the resistors are in parallel (2*0.22ohms), correct? How would this effect the quiescent current?
Cheers again,
Andy
The emittor resistors are included inside the transistor package, and are 0.22ohms, how would you measure the voltage across this to get the quiescent current? Because both emittors in the pair meet (creating the amplified output), this means that the resistors are in parallel (2*0.22ohms), correct? How would this effect the quiescent current?
Cheers again,
Andy
Hi Andy,
You live a charmed life. You could have instantly lost the new parts.
It sounds as though there are some other defective parts as well. Don't turn the amp on again until you have checked everything. Try to use some way of limiting the current in case of a fault.
The bias transistor (if there is one) may be partially shorted, resistor open (check the low value parts). You may have a predriver blown and leaky, that took a while for me to find once.
So test all the other transistors in that channel. Take a picture first so you get everything back properly aligned.
A schematic would really come in handy right now.
-Chris
You live a charmed life. You could have instantly lost the new parts.
It sounds as though there are some other defective parts as well. Don't turn the amp on again until you have checked everything. Try to use some way of limiting the current in case of a fault.
The bias transistor (if there is one) may be partially shorted, resistor open (check the low value parts). You may have a predriver blown and leaky, that took a while for me to find once.
So test all the other transistors in that channel. Take a picture first so you get everything back properly aligned.
A schematic would really come in handy right now.
-Chris
Hi anatech
"You live a charmed life. You could have instantly lost the new parts.", hehe, well i'd run out of options
I spent many painfull hours last week checking EVERY single transistor on the board, of course all damaged resistors and caps were replaced, there isn't much else on the board, so there was a fairly low chance of something major going wrong
Basically i was pretty certain there were no other problems on the board.
Thanks for the reply, I measured and adjusted and as the datasheet recommends, set each emitter current to 40ma. Without any service manuals I can't think what else to do, it's not a hugely expensive amplifier so it's not worth getting it professionally repaired. The DC offset doesn't seem too bad, I'll have to find some cheap speakers to test out the sound.
Thanks for the help,
Andy
"You live a charmed life. You could have instantly lost the new parts.", hehe, well i'd run out of options
I spent many painfull hours last week checking EVERY single transistor on the board, of course all damaged resistors and caps were replaced, there isn't much else on the board, so there was a fairly low chance of something major going wrong
Basically i was pretty certain there were no other problems on the board.
Thanks for the reply, I measured and adjusted and as the datasheet recommends, set each emitter current to 40ma. Without any service manuals I can't think what else to do, it's not a hugely expensive amplifier so it's not worth getting it professionally repaired. The DC offset doesn't seem too bad, I'll have to find some cheap speakers to test out the sound.
Thanks for the help,
Andy
Re: Swamping in Output Circuits?
Out of here or not, I suspect curiosity will bring you back, even if there is no response..
Correct me if you like, but please relax and try have fun here!! Else you miss the whole point...
Out of here or not, I suspect curiosity will bring you back, even if there is no response..
Just for grins, I looked this up (even though I know this). Common collector = collectors connected to the rails, i.e, emitter follower.chewrock said:
I used the term 'swamped', which is probably incorrect. Part of the reason for the resistors is to reduce the effects of -re'- (dynamic emitter resistance) on the operation of the circuit, and to provide a means to measure bias current.
Correct me if you like, but please relax and try have fun here!! Else you miss the whole point...
Hi,
there are many readers here who are learning about electronics, myself included.
If a poster spouts balderdash and no-one corrects the error then pity us poor souls who believe the diatribe just posted.
If you are not sure, then say so. Do not publish guesswork and purport it to be factual.
Your views are always welcome, even if only to get a discussion started.
there are many readers here who are learning about electronics, myself included.
If a poster spouts balderdash and no-one corrects the error then pity us poor souls who believe the diatribe just posted.
If you are not sure, then say so. Do not publish guesswork and purport it to be factual.
Your views are always welcome, even if only to get a discussion started.
Hi again
Right, last post i promise. Finally managed to get straight through to cambridge audio, and they sent me through to richer sounds engineering department who apprently fix cambridge audio products.
Now, he said that in this design the (N channel, NPN transistor) should be adjusted so that the quiescent current is 7mA.
I will adjust it to that, but it seems very low from the basic knowledge i have, don't normal amplifiers have a factory setting quite a bit higher than this? more like 50mA.
Thanks,
Andy
Right, last post i promise. Finally managed to get straight through to cambridge audio, and they sent me through to richer sounds engineering department who apprently fix cambridge audio products.
Now, he said that in this design the (N channel, NPN transistor) should be adjusted so that the quiescent current is 7mA.
I will adjust it to that, but it seems very low from the basic knowledge i have, don't normal amplifiers have a factory setting quite a bit higher than this? more like 50mA.
Thanks,
Andy
You may not need a high idle current. Adjusting for minimum THD at high frequencies may yield values of bias current in the 2~5 mA range. Try it.
This doesn't work well with some designs. The "spec" in the manual will give you a value that is guarantied to allow the unit to meet spec for different combinations of components and ambient temperatures. A few designs do better with higher values.
If unknown, I will generally set between 10 mA and 50 mA. You need to watch the heatsink temperature and whether the bias has a temperature coefficient or not. Just did an amplifier that I set at 30 mA and found it increased to 70 mA with light running. I had to back it off to 10 mA. That gets it up to about 35 mA. The bias compensation circuit could use some work I think.
-Chris
This doesn't work well with some designs. The "spec" in the manual will give you a value that is guarantied to allow the unit to meet spec for different combinations of components and ambient temperatures. A few designs do better with higher values.
If unknown, I will generally set between 10 mA and 50 mA. You need to watch the heatsink temperature and whether the bias has a temperature coefficient or not. Just did an amplifier that I set at 30 mA and found it increased to 70 mA with light running. I had to back it off to 10 mA. That gets it up to about 35 mA. The bias compensation circuit could use some work I think.
-Chris
Hi Andy,
I have been reading this thread with interest and have a suspicion that the A5i (is there an A5 without the i, btw?) and the earlier A300 V2 (and A500RC – which is basically = A300 V2 with uprated SAPs and a remote control circuit) are very similar amplifier designs to the A5i. Therefore, I hope that the schematic of the A300 V2 (and A500RC) output stage in this thread is familiar to you.
As I have a working A300 V2 I could probably measure the current through the pots for you and post it in this thread for your guidance. (I'm not prepared to unsolder them yet so I'd have to measure the p.d. across and resistance of the pots in-situ and calculate the current).
If you need them for comparison, I have the full schematics for the A300 V2, which I could email to you.
By the way, how did you measure the DC offset? Is it correct to just stick a DMM on mV DC range across the speaker terminals with the volume set to min?
Good luck and regards,
Rob.
I have been reading this thread with interest and have a suspicion that the A5i (is there an A5 without the i, btw?) and the earlier A300 V2 (and A500RC – which is basically = A300 V2 with uprated SAPs and a remote control circuit) are very similar amplifier designs to the A5i. Therefore, I hope that the schematic of the A300 V2 (and A500RC) output stage in this thread is familiar to you.
As I have a working A300 V2 I could probably measure the current through the pots for you and post it in this thread for your guidance. (I'm not prepared to unsolder them yet so I'd have to measure the p.d. across and resistance of the pots in-situ and calculate the current).
If you need them for comparison, I have the full schematics for the A300 V2, which I could email to you.
By the way, how did you measure the DC offset? Is it correct to just stick a DMM on mV DC range across the speaker terminals with the volume set to min?
Good luck and regards,
Rob.
Hi,
the DC output offset is usually measured with both the speakers and the input disconnected.
You need to add a shorting plug to the input to ensure that zero volts (both DC and AC) is fed into the amp.
You can do an approximate check on DC by turning down the source to zero input signal and leaving the speakers connected, but only as a quick check to see if re-setting is necessary.
the DC output offset is usually measured with both the speakers and the input disconnected.
You need to add a shorting plug to the input to ensure that zero volts (both DC and AC) is fed into the amp.
You can do an approximate check on DC by turning down the source to zero input signal and leaving the speakers connected, but only as a quick check to see if re-setting is necessary.
Hi all,
I know you're all on the edges of your seat waiting for the next thrilling installment of "elaar tries to fix an amplifier with no knowledge of amplifiers", so here it goes
Well, either the cambridge audio engineer told me porkies or was thinking of something else, the quiescent current can't go as low as 7ma, with a 100ohm trimmer it can be adjusted between 20mA and around 80mA.
I set one channel to 20mA and the other to 40mA to see if I could hear the difference in the speakers. Unfortunately the only speakers I had to hand are poor and I can't admit to noticing any difference. I then set it to 20mA vs 60mA, after about an hour of playing music at half volume, the 20mA channel heatsink was hot but I could keep my hand on that heatsink indefinitely, whilst on the 60mA channel i could keep my hand on the heatsink base for about 5secs before it got too hot. This seemed too hot for my liking, so I think about 40mA seems about right.
I also have an A5 amplifier that is broken (as well as this A5i), the trim pots on there measure between 45-55ohms, which is in the correct range for it to be about 40mA, so it seems this is probably the right spec.
I know you're all on the edges of your seat waiting for the next thrilling installment of "elaar tries to fix an amplifier with no knowledge of amplifiers", so here it goes
Well, either the cambridge audio engineer told me porkies or was thinking of something else, the quiescent current can't go as low as 7ma, with a 100ohm trimmer it can be adjusted between 20mA and around 80mA.
I set one channel to 20mA and the other to 40mA to see if I could hear the difference in the speakers. Unfortunately the only speakers I had to hand are poor and I can't admit to noticing any difference. I then set it to 20mA vs 60mA, after about an hour of playing music at half volume, the 20mA channel heatsink was hot but I could keep my hand on that heatsink indefinitely, whilst on the 60mA channel i could keep my hand on the heatsink base for about 5secs before it got too hot. This seemed too hot for my liking, so I think about 40mA seems about right.
I also have an A5 amplifier that is broken (as well as this A5i), the trim pots on there measure between 45-55ohms, which is in the correct range for it to be about 40mA, so it seems this is probably the right spec.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.