I used some kind of foamy bathroom cleaner and a toothbrush. You could use Simple Green or anything like that. I think there were some physically leaking electrolytics or some other contamination on the boards so that's what lead me to seriously suspect the PCB. The thing is that whatever is wrong with these amps it is exceptionally Odd, confusing all usual troubleshooting approaches. PCB material is rated for volume resistivity and FR4 is something around pico Ohms a centimeter which means if you can measure it with a megohm meter there is something wrong. Conceivably there could have been a problem with the resin used in those boards but ultimately the kind of problem this could cause would depend on the impedance of the input/VAC nodes which if I recall weren't high enough anywhere for it to figure into play. At some point you have to start suspecting some pretty far fetched things to find a problem like this, but a defect anywhere in the output stage seems even futher out there because anything causing a big leakage current or gain assymetry would get an offset in drive from the voltage stage and you probably wouldn't notice it at idle unless you were probing around inside the amp. I'd seriously recommend that you try to get Adcom service on the phone to save yourself some headache and possibly get yourself a running amp. It seems to be such a widespread problem I doubt they will not understand what you're talking about.
Last edited:
You can buy matched sets of transistor in any quantity from Bryston. Check some of their older designs to see if they used a transistor suitable for replacement in the your Adcom.
On some occasions Bryston sent me small quantities of output transistors for free. Of course I owned and operated a repair shop at the time. Bryston even paid the shipping.
On semi will send out free samples if you tell them your from an electronics sound company and they believe you.
I've only repaired a few of the GFA series but that was many year ago.
In general most power amplifiers blow up because of bias failure or thermal runaway. If this is the case with you amp then usually, but not always, the damage can back up as far as the VAS stage and sometimes further. I would begin by finding shorted components and either remove them or isolating them in some fashion. This will allow you to identify any other damaged components. Check all the resistors and transistors from the VAS forward. Be sure your bias generator is working properly or you might have to do a second repair on the amplifier.
Do you have or have access to a variac? If not then you can put a low power incandescent light bulb in series with the primary side of the power transformer to limit current to the amplifier until you're sure everything is working and you can adjust the bias. This will do no harm and can save you a set of output transistor in case you miss something.
David.
On some occasions Bryston sent me small quantities of output transistors for free. Of course I owned and operated a repair shop at the time. Bryston even paid the shipping.
On semi will send out free samples if you tell them your from an electronics sound company and they believe you.
I've only repaired a few of the GFA series but that was many year ago.
In general most power amplifiers blow up because of bias failure or thermal runaway. If this is the case with you amp then usually, but not always, the damage can back up as far as the VAS stage and sometimes further. I would begin by finding shorted components and either remove them or isolating them in some fashion. This will allow you to identify any other damaged components. Check all the resistors and transistors from the VAS forward. Be sure your bias generator is working properly or you might have to do a second repair on the amplifier.
Do you have or have access to a variac? If not then you can put a low power incandescent light bulb in series with the primary side of the power transformer to limit current to the amplifier until you're sure everything is working and you can adjust the bias. This will do no harm and can save you a set of output transistor in case you miss something.
David.
Go read the entire thread and you'll find he's been all over the amp and has bought a scope, variac, audio generator and more. He's making a very serious effort at sorting this out but it has been one of those 'tough dogs'. However it ends up, he's agreed to send me the old semis because I want to see what they do on the curve tracer - a gadget you don't need often but when you do, it clears up a lot of confusion. We've loaded the DC supplies with a 150 Watt lamp to look for power supply ripple and it turns out MINE has some minor problems with the power supply worse than Fred's (which is excellent). If he wanted to send me the entire amp I'd do the labor for free just to see what in blazes is going on.
G²
There is too much thread here for me to read it all.
Has anyone addressed the DC offset servo built around IC151/IC101? If I had a DC offset problem with one of these amps and all else seemed to pass then I would be focused on the servo.
Also molex connectors are notorious for failing. You could try removing them and soldering the wires directly to the board. Once removed you can get a good look at the pads, landing and traces to check for cracks. As was pointed out in another post.
At this point there is no harm in disconnecting the working channel and swapping parts over to the faulty channel. If the output boards are identical and the connectors are wired the same, you can swap an entire output board from one input/driver board to the other. This will answer up some questions.
The risk in doing this is you may end up with two blown channels instead of one.
This is a nice amplifier and it would be a shame to use it for a sub.
Has anyone addressed the DC offset servo built around IC151/IC101? If I had a DC offset problem with one of these amps and all else seemed to pass then I would be focused on the servo.
Also molex connectors are notorious for failing. You could try removing them and soldering the wires directly to the board. Once removed you can get a good look at the pads, landing and traces to check for cracks. As was pointed out in another post.
At this point there is no harm in disconnecting the working channel and swapping parts over to the faulty channel. If the output boards are identical and the connectors are wired the same, you can swap an entire output board from one input/driver board to the other. This will answer up some questions.
The risk in doing this is you may end up with two blown channels instead of one.
This is a nice amplifier and it would be a shame to use it for a sub.
Hi Fred,
Looks like we are going 'round and 'round again.
Okay, when transistors are replaced with types that are not 100% original numbers, you can get instability in the circuit. It's not the fault of an Adcom engineer, but rather just a part of servicing. Years ago there were different setups depending on whether an amp used Motorola parts, or Fairchild (faster) parts. Look back in older service manuals to see. The Crown (Amcron in Canada) DC300 manual provides an excellent example of this, and I'm more than sure that Nelson could provide first hand horror stories on this subject.
Without the diff pair installed, the amplifier may do just about anything once you provide current to turn other stuff on. In this instance, the test you are trying to perform will not tell you much of anything. Put the amp back together and allow it to run as designed in order to do testing.
There are only two reasons why an amplifier will draw excessive current assuming you don;t have leaky or shorted transistors. The bias circuit has failed or is out of range for your set of drivers and outputs. Testing to see if there is action on the bias control, and varying conduction through the bias transistor will answer that one. The other possibility is HF oscillation (you would hear it if the frequency was lower). The zobel network tends to act as a canary in a coal mine in this situation. If the zobel has been overheated, you are pretty much guarantied to have oscillation issues somewhere in the system. The oscilloscope will let you know when connected to the "hot" output terminal.
To clean the PCB, remove all the parts in the area around the electrolytic capacitors. Clean the area with Simple Green or similar, PCB cleaning sprays will not affect this nasty stuff. Scrub, rinse, scrub, rinse ... until your test indicates the stuff is gone. I use an ultrasonic cleaner with Simple Green. This usually works after a couple cycles. Clean these parts and test them before you repopulate the board with them. New capacitors of course!
Fred, you need to stay focused on what you need to do. Do write down what you observe only, and what steps you have taken and parts that have been replaced. Japanese transistors have a different pin out than US or Pro Electron (European) transistors do. Always confirm everything, don't trust your memory or someone else's word. That includes what I am telling you.
Any time you seem to be thrashing about, check yourself. This is the time when you've lost track. You're new at this, so some uncertainty is expected. Relax, confirm your facts and wait for the parts. Don't try anything else until you have finished the test or procedure you are currently working on. You will only become confused.
Once you have the parts installed again, check for oscillation and DC offset. If these things are okay, increase your supply voltage while watching the current draw. These amplifiers should not oscillate when at reasonable lower supply voltages, really low voltages can in themselves cause nifty problems.
Lastly, Nelson knows a fair amount about the basic design of these amplifiers. Listen to the man, please.
-Chris 🙂
Looks like we are going 'round and 'round again.
Okay, when transistors are replaced with types that are not 100% original numbers, you can get instability in the circuit. It's not the fault of an Adcom engineer, but rather just a part of servicing. Years ago there were different setups depending on whether an amp used Motorola parts, or Fairchild (faster) parts. Look back in older service manuals to see. The Crown (Amcron in Canada) DC300 manual provides an excellent example of this, and I'm more than sure that Nelson could provide first hand horror stories on this subject.
Without the diff pair installed, the amplifier may do just about anything once you provide current to turn other stuff on. In this instance, the test you are trying to perform will not tell you much of anything. Put the amp back together and allow it to run as designed in order to do testing.
There are only two reasons why an amplifier will draw excessive current assuming you don;t have leaky or shorted transistors. The bias circuit has failed or is out of range for your set of drivers and outputs. Testing to see if there is action on the bias control, and varying conduction through the bias transistor will answer that one. The other possibility is HF oscillation (you would hear it if the frequency was lower). The zobel network tends to act as a canary in a coal mine in this situation. If the zobel has been overheated, you are pretty much guarantied to have oscillation issues somewhere in the system. The oscilloscope will let you know when connected to the "hot" output terminal.
I touched on this earlier in the thread. The electrolyte that can leak out of some capacitors used in these amps is colourless. You can smell it if you allow your soldering iron to heat the suspected area up. Just touch the PCB and move the tip around a bit. If electrolyte is there, you will smell it. Cleaning PC boards must be done with proper cleaners and after some parts have been pulled from the board first. Contact cleaner, WD-40 and similar fluids will destroy capacitors and some transistors over time. If you sprayed cleaner on the PCB with capacitors mounted - replace those parts before you go any further.
To clean the PCB, remove all the parts in the area around the electrolytic capacitors. Clean the area with Simple Green or similar, PCB cleaning sprays will not affect this nasty stuff. Scrub, rinse, scrub, rinse ... until your test indicates the stuff is gone. I use an ultrasonic cleaner with Simple Green. This usually works after a couple cycles. Clean these parts and test them before you repopulate the board with them. New capacitors of course!
Fred, you need to stay focused on what you need to do. Do write down what you observe only, and what steps you have taken and parts that have been replaced. Japanese transistors have a different pin out than US or Pro Electron (European) transistors do. Always confirm everything, don't trust your memory or someone else's word. That includes what I am telling you.
Any time you seem to be thrashing about, check yourself. This is the time when you've lost track. You're new at this, so some uncertainty is expected. Relax, confirm your facts and wait for the parts. Don't try anything else until you have finished the test or procedure you are currently working on. You will only become confused.
Once you have the parts installed again, check for oscillation and DC offset. If these things are okay, increase your supply voltage while watching the current draw. These amplifiers should not oscillate when at reasonable lower supply voltages, really low voltages can in themselves cause nifty problems.
Lastly, Nelson knows a fair amount about the basic design of these amplifiers. Listen to the man, please.
-Chris 🙂
Hi davada,
Take a breath there. Fred can't do anything until he gets it back together.
Troubleshooting is not that hard when the fault is catastrophic in nature.
-Chris
Take a breath there. Fred can't do anything until he gets it back together.
Rarely a point of failure for one. Besides, a failure here will cause a DC fault only. My diagram of the GFA-555 does not show a DC servo. Anyone, where is the post with the diagram fr this amp again please? I did look, but missed it - twice.
Molex has two lines. The ones you are referring to, and the better ones like is used in this amp. I have never had a problem with these connectors in an Adcom yet. Also, please don't solder the wires right to the PCB. It's hard enough to work on these without doing that!
Never do this, for any reason! If you are chasing a small issue and both channels still work, then okay. A situation like this demands you keep good away from bad. Never try parts from the good side to the bad side either!
Troubleshooting is not that hard when the fault is catastrophic in nature.
You have to make this point more strongly, like as in .... No! Never do this!
-Chris
Hi davada,
Your desire to help is appreciated, don't get me wrong.
By now you should know about open fuses and anything burnt. It's time to apply power carefully. Depending on the unit, age and fault (sometimes the customer also), you may just plug it in, or you may soft start it in some way. Switching power supplies can be started gradually, and if not you can use an external power supply for the purpose. I can go up to +/- 70 VDC with a variac and old Marantz power supply subassembly. The amp was a cosmetic write-off, plus the damage. I would never strip a good amp that could be returned to service.
When doing your first power-up procedure, keep a sharp eye and a keen nose. Listen carefully as well, most caps may hiss just before they blow. Now for the important point. Each time you make a change in something you are working on, you should follow the "first power-up" procedure. This can save a great deal of damage due to a short, crushed wire or solder splash. Even a connection you forgot to reconnect could cause damage.
So a multiple fault repair is only more time consuming, and perhaps more difficult to decode, but the way you go about it remains the same.
Early machines like MCI or Scully may have used the bad, cheap old Molex connectors. To be honest with you, I have a great dislike for that brand even now which I am trying to get over. The other side of the problem was that early machines ran a lot of current through connections that were pretty obviously not up to the job. The milky plastic housing would melt pretty well, then allow more terminals to short. Nice. Card edge connectors were the other popular connection. The other issue was the cigarette / cigar smoke that permeates everything from "back in the day". Everyone smoked or toked, maybe both. I'm pretty sure they all inhaled. This leaves a coating over everything. Contacts, connectors, controls and switches - all of it went to heck. Edge connectors are easy to clean (do not use an eraser!!!), but how do you get into a housing full of pins and circular receptacles. Spraying doesn't work, just in case you were wondering. Neumann gold diaphragms didn't stand a chance in that environment. Now, couple all this with the fact that stuff always breaks at night and you have one grumpy studio tech.
Those early Molex connectors that most people used were garbage, but they also had a more expensive line out a bit later. They (the manufacturers) kept buying the cheap crap. The Japanese machines used Japanese connectors which were far better than the North American types that used Molex. Even the Europeans had better connectors (even though many looked ghastly, like made in a garage) than what we were using in the 'States. I didn't help that the contacts tended to move about in the Molex connectors, they could be quite difficult to mate together. Often, the connection wasn't pushed fully home and they separated. Nice. So a technician would blame the connector, rather than the fact that he was rushing and did not confirm the plugs were properly mated.
Many reasons to dislike the early, cheap Molex connectors. But the newer stuff is fine, just avoid the bad old types. You have to recognize change and revise your opinions when you can no longer support a point of view.
-Chris
I wouldn't say that. It's always nice to have another set of eyes on a problem. Of course, the hard part is getting the lay of the land since this is an on-going situation. Continuity seems to be the most difficult issue in any thread. Remote diagnostics doesn't make it any easier either.
Your desire to help is appreciated, don't get me wrong.
Well, that explains it. It just a easily could have been me, so don't feel bad about it.
Well, not really. It's true that there are often more than one thing wrong with an amplifier that has a few years under it's belt. You get updates from the factory that either were not done, or not done properly. Then you possibly get "struck by technician" where someone in a panic keeps on going without thinking. I get a lot of those, always have it seems. Lastly (I hope) you get all the age related issues and possible damage to PCBs due to dropping. Anything can happen. Still, that doesn't really change how you troubleshoot. The same steps are followed, but they are more important now. Give the thing a good look over, after blowing out the dust bunnies first. I like to see if there are any cap cans or burnt debris before doing anything. Smell the unit, both before and after you remove the lid. Simply note anything you can see on paper. It's critical to write down observations only as you go along. There is plenty of time to draw conclusions later, keep an open mind for now.
By now you should know about open fuses and anything burnt. It's time to apply power carefully. Depending on the unit, age and fault (sometimes the customer also), you may just plug it in, or you may soft start it in some way. Switching power supplies can be started gradually, and if not you can use an external power supply for the purpose. I can go up to +/- 70 VDC with a variac and old Marantz power supply subassembly. The amp was a cosmetic write-off, plus the damage. I would never strip a good amp that could be returned to service.
When doing your first power-up procedure, keep a sharp eye and a keen nose. Listen carefully as well, most caps may hiss just before they blow. Now for the important point. Each time you make a change in something you are working on, you should follow the "first power-up" procedure. This can save a great deal of damage due to a short, crushed wire or solder splash. Even a connection you forgot to reconnect could cause damage.
So a multiple fault repair is only more time consuming, and perhaps more difficult to decode, but the way you go about it remains the same.
Well, I R one of those. 🙂
Early machines like MCI or Scully may have used the bad, cheap old Molex connectors. To be honest with you, I have a great dislike for that brand even now which I am trying to get over. The other side of the problem was that early machines ran a lot of current through connections that were pretty obviously not up to the job. The milky plastic housing would melt pretty well, then allow more terminals to short. Nice. Card edge connectors were the other popular connection. The other issue was the cigarette / cigar smoke that permeates everything from "back in the day". Everyone smoked or toked, maybe both. I'm pretty sure they all inhaled. This leaves a coating over everything. Contacts, connectors, controls and switches - all of it went to heck. Edge connectors are easy to clean (do not use an eraser!!!), but how do you get into a housing full of pins and circular receptacles. Spraying doesn't work, just in case you were wondering. Neumann gold diaphragms didn't stand a chance in that environment. Now, couple all this with the fact that stuff always breaks at night and you have one grumpy studio tech.
Those early Molex connectors that most people used were garbage, but they also had a more expensive line out a bit later. They (the manufacturers) kept buying the cheap crap. The Japanese machines used Japanese connectors which were far better than the North American types that used Molex. Even the Europeans had better connectors (even though many looked ghastly, like made in a garage) than what we were using in the 'States. I didn't help that the contacts tended to move about in the Molex connectors, they could be quite difficult to mate together. Often, the connection wasn't pushed fully home and they separated. Nice. So a technician would blame the connector, rather than the fact that he was rushing and did not confirm the plugs were properly mated.
Many reasons to dislike the early, cheap Molex connectors. But the newer stuff is fine, just avoid the bad old types. You have to recognize change and revise your opinions when you can no longer support a point of view.
-Chris
Have you measured all the node voltages on both channels and then written them down on the circuit diagram? Then post it up here (do the bad channel in RED and the good channel in BLACK).
You are very lucky - you have a working channel - use that to help guide your efforts. There's enough expereinced guys around here who can help you - but without some measurements it cna get a bitt difficult.
You are very lucky - you have a working channel - use that to help guide your efforts. There's enough expereinced guys around here who can help you - but without some measurements it cna get a bitt difficult.
Fred,
I believe that we were finally making some progress. There is absolutely no reason why the output transistors/stage should not pass the test that I suggested here unless they are damaged - I don't know what all the panic is about.
One more tip, anytime I remove a part I try to test as many things as possible that are isolated by the removal. Any other components that are left with only one lead connected are essentially out of circuit.
The full voltage leakage test that I am suggesting will find gross problems and let me also suggest that while you have most of the outputs isolated test them with your IT-18 as a more sensitive but low voltage test.
Let's rethink some of this. You built up the entire amp with mostly new driver board parts, no substitutions and it had high DC offset and would not swing to the rail. This suggests something is wrong in the output stage - let's finish testing it.
The leaking electrolytic problem was more with the 555-II and you think that the output was shorted to cause the problem, so it is not likely a board problem. Certainly, clean it if you want, and make sure you de-flux the board when you replace parts.
I don't know what all the fuss is about testing it with a load, certainly we want to do that also but I'll mention a few possible reasons since I believe that you want to learn as we go. Most receivers with a speaker switch run with no load when in the off position and it is reasonable to expect an amp to work fine without an output load. There are two reasons that might cause a problem. When it is loaded the unregulated PS voltages will "sag" as more current is drawn and this helps to avoid exposing any of the devices to twice the supply voltage. When it idles the output stage and VAS see 1/2 the supply voltage and when a signal is supplied that swings the output to the rails they see 2X one rail - loaded this will be a lower voltage. This might be an issue if the devices do not have enough breakdown voltage, although it is a time bomb to build an amp this way, but you are replacing all the marginal devices (mainly the drivers) with devices having plenty of margin. Keep the max voltage low at +/- 70 until you get the new drivers installed.
The other reason for only testing the amp loaded is that the loop stability is altered without a load. The amp uses negative feedback however each stage in the amp has limited bandwidth and as they roll-off at HF they add more phase shift. As the total additional phase shift reaches 180 degrees what was negative feedback becomes positive. The well known condition for oscillation is that the phase shift be 360 deg, so that the signal is back in phase, and the gain = one or more so that oscillation is sustained rather than dying down. Running the amp unloaded also alters the internal feedback loop gain, however it should be designed to work under this condition. When we look at an amp's stability we look for phase margin this is how much less phase shift than 360 deg we have when the gain around the loop = one, and the gain margin which is how many dB less than one the gain is when the phase shift is 360 deg. If we open the feedback loop as we did it should not be possible for the amp to oscillate on a main loop level. It is also possible to have localized oscillations in the output stage for example. This is usually at very high frequency were the wiring inductance combined with the internal device capacitances can be enough to form a high freq oscillator - usually in the MHz range.
You ran one channel unloaded and as I understand it, it worked fine as it should.
I have been suspicious of the bias circuit and you can see that each device has the driver transistor providing feedback around it with the high output transistor's capacitance providing phase shift. The bias circuit oscillated in simulation when I removed C5. No capacitor is perfect and they actually self resonate due to winding and lead inductance. It is possible that the amp is sensitive to the type of cap used there. 100 pF from collector to base on the bias transistors killed this oscillation even without C5 - it is analogous to the Cdom cap (C2) on the VAS and you should add them if we see any HF oscillations of the output stage alone. I have made up a rough LTSpice simulation of the amp.
I believe that we were finally making some progress. There is absolutely no reason why the output transistors/stage should not pass the test that I suggested here unless they are damaged - I don't know what all the panic is about.
One more tip, anytime I remove a part I try to test as many things as possible that are isolated by the removal. Any other components that are left with only one lead connected are essentially out of circuit.
The full voltage leakage test that I am suggesting will find gross problems and let me also suggest that while you have most of the outputs isolated test them with your IT-18 as a more sensitive but low voltage test.
Let's rethink some of this. You built up the entire amp with mostly new driver board parts, no substitutions and it had high DC offset and would not swing to the rail. This suggests something is wrong in the output stage - let's finish testing it.
The leaking electrolytic problem was more with the 555-II and you think that the output was shorted to cause the problem, so it is not likely a board problem. Certainly, clean it if you want, and make sure you de-flux the board when you replace parts.
I don't know what all the fuss is about testing it with a load, certainly we want to do that also but I'll mention a few possible reasons since I believe that you want to learn as we go. Most receivers with a speaker switch run with no load when in the off position and it is reasonable to expect an amp to work fine without an output load. There are two reasons that might cause a problem. When it is loaded the unregulated PS voltages will "sag" as more current is drawn and this helps to avoid exposing any of the devices to twice the supply voltage. When it idles the output stage and VAS see 1/2 the supply voltage and when a signal is supplied that swings the output to the rails they see 2X one rail - loaded this will be a lower voltage. This might be an issue if the devices do not have enough breakdown voltage, although it is a time bomb to build an amp this way, but you are replacing all the marginal devices (mainly the drivers) with devices having plenty of margin. Keep the max voltage low at +/- 70 until you get the new drivers installed.
The other reason for only testing the amp loaded is that the loop stability is altered without a load. The amp uses negative feedback however each stage in the amp has limited bandwidth and as they roll-off at HF they add more phase shift. As the total additional phase shift reaches 180 degrees what was negative feedback becomes positive. The well known condition for oscillation is that the phase shift be 360 deg, so that the signal is back in phase, and the gain = one or more so that oscillation is sustained rather than dying down. Running the amp unloaded also alters the internal feedback loop gain, however it should be designed to work under this condition. When we look at an amp's stability we look for phase margin this is how much less phase shift than 360 deg we have when the gain around the loop = one, and the gain margin which is how many dB less than one the gain is when the phase shift is 360 deg. If we open the feedback loop as we did it should not be possible for the amp to oscillate on a main loop level. It is also possible to have localized oscillations in the output stage for example. This is usually at very high frequency were the wiring inductance combined with the internal device capacitances can be enough to form a high freq oscillator - usually in the MHz range.
You ran one channel unloaded and as I understand it, it worked fine as it should.
I have been suspicious of the bias circuit and you can see that each device has the driver transistor providing feedback around it with the high output transistor's capacitance providing phase shift. The bias circuit oscillated in simulation when I removed C5. No capacitor is perfect and they actually self resonate due to winding and lead inductance. It is possible that the amp is sensitive to the type of cap used there. 100 pF from collector to base on the bias transistors killed this oscillation even without C5 - it is analogous to the Cdom cap (C2) on the VAS and you should add them if we see any HF oscillations of the output stage alone. I have made up a rough LTSpice simulation of the amp.
Last edited:
PB2 I don't think there is a stability issue with the output stage being loaded or not.
It just a matter to see if one half is broken.
The one concern I have about stabilty is when the diff amp and VAS stage are re-built with the highest beta devices. I would be interested in playing around with your LTspice model for this.
It just a matter to see if one half is broken.
The one concern I have about stabilty is when the diff amp and VAS stage are re-built with the highest beta devices. I would be interested in playing around with your LTspice model for this.
I agree, as I said I believe that it should work fine no load and I was just explaining the far out possibilities to give some insight. I don't expect even the newer parts to cause an issue, well ... perhaps the Cdom cap will need adjustment if the device capacitance is much lower. Device models are an issue and I've just used some models that I know are good. I could use a 2SC2240 SPICE model if anyone has it handy or knows where to find it. Actually I need all the driver board models and outputs if anyone has them. I used andy's output models. I'm running out now and will post the SPICE file later tonight. I'll probably post it to a new thread to make it easier to find. Do you use LTspice?
Cool , I might need some guidance on opening the loop.
I like starting there to see where the dominate pole is and were others may lay and sensitivity to bias etc.
I like starting there to see where the dominate pole is and were others may lay and sensitivity to bias etc.
Last edited:
.model Q2sc2240 NPN(Is=99.13f Xti=3 Eg=1.11 Vaf=422.2 Bf=352.8 Ise=1.179p
+ Ne=1.782 Ikf=.4704 Nk=.9631 Xtb=1.5 Var=100 Br=1.663 Isc=555.1p
+ Nc=1.796 Ikr=5.85 Rc=.2032 Cjc=7.561p Mjc=.2472 Vjc=.3905 Fc=.5
+ Cje=5p Mje=.3333 Vje=.75 Tr=10n Tf=1.295n Itf=1 Xtf=0 Vtf=10)
+ Ne=1.782 Ikf=.4704 Nk=.9631 Xtb=1.5 Var=100 Br=1.663 Isc=555.1p
+ Nc=1.796 Ikr=5.85 Rc=.2032 Cjc=7.561p Mjc=.2472 Vjc=.3905 Fc=.5
+ Cje=5p Mje=.3333 Vje=.75 Tr=10n Tf=1.295n Itf=1 Xtf=0 Vtf=10)
.model Q2sa970 PNP(Is=465.4f Xti=3 Eg=1.11 Vaf=57 Bf=407.6 Ise=4.683p Ne=2.051
+ Ikf=.3998 Nk=1.192 Xtb=1.5 Var=100 Br=1 Isc=465.4f Nc=1.048
+ Ikr=6.032 Rc=2.343 Cjc=11.59p Mjc=.4014 Vjc=1.155 Fc=.5 Cje=5p
+ Mje=.3333 Vje=.75 Tr=10n Tf=1.252n Itf=1 Xtf=0 Vtf=10)
+ Ikf=.3998 Nk=1.192 Xtb=1.5 Var=100 Br=1 Isc=465.4f Nc=1.048
+ Ikr=6.032 Rc=2.343 Cjc=11.59p Mjc=.4014 Vjc=1.155 Fc=.5 Cje=5p
+ Mje=.3333 Vje=.75 Tr=10n Tf=1.252n Itf=1 Xtf=0 Vtf=10)
Hi Pete,
However ....
So, if you are concerned about this transistor, add some bypass caps from base to emitter and collector to emitter. Shouldn't cause any trouble after that.
Should be interesting to see what the spice model shows. You could play with the value for C5 and see what it does. Andy may have the models you want as well.
-Chris
Yes, but it's easy for someone to lose track of what they are doing when they are new to troubleshooting.
The circuit has become unstable and liable to give a false result. It's better to ensure that amplification can not occur (killing the oscillation). Right now you have active drivers, bias and possibly Vas. This can really confuse things.
There's no panic. If anything, I was suggesting that Fred cool it until he has the parts. Thinking while going over his notes can be more valuable than continually trying other things. I'd like Fred to keep a clear head - even though you understand were you want to go.
Fred mentioned that he had cleaned the PCB. It was a good time to review the proper procedure and to alert him to possible component damage that might just be giving him grief right now. Tying up loose ends.
Nelson had a valid query. Loading the output may have allowed the oscillations to stop. Don't forget that the output line is a high AC impedance point when the diff pair is missing. This in itself may cause stray signal pickup and oscillation. The other common problem is caused by heat sinks that are floating, not grounded anymore.
Exactly!
However ....
Does not apply right now. The amplifier is not running in a normal mode and there isn't an intentional negative feedback loop. Nothing about this situation will give you valid information.
Sure, that may occur. Mostly C5 reduces the switching distortion when the amp is running, and it also reduces the AC impedance around that transistor. However, C5 can be reduced to 0.1 uF or even lower and it will still kill oscillation. It may even work as well as the original value for distortion reduction.
So, if you are concerned about this transistor, add some bypass caps from base to emitter and collector to emitter. Shouldn't cause any trouble after that.
Should be interesting to see what the spice model shows. You could play with the value for C5 and see what it does. Andy may have the models you want as well.
-Chris
I suggested it because there is a finite probability that you will
discover that 1/2 the output stage doesn't drive the load at all.
That information would alter the course of the investigation.
😎