The picture of the PCB in my book does have '72-2174-0-1' shown as silkscreen text and FWIW, my GFA-555 has that PCB installed.
G²
Looks like we do have the same manual, then. But my GFA-555 uses the older board and like I said, there are some parts differences. I've contacted Adcom in hopes of finding the right info.
FWIW, when this amp is running again I'm going to rewrite/summarize the thread so it is more useful. I'm learning a lot and I think I can pass that on. Tech writing is really all I have to offer the forum (and a donation, of course)!
cheers.
Last edited:
Shopping for a variac on the ebay. Wish me luck... A 'scope is just not in the budget right now, sadly.
Moving forward (hoping I still can), I'm wondering what my next step is. Did I introduce a new problem with the replacement components? Or is this how the initial fault is manifesting now that some components are replaced?
Moving forward (hoping I still can), I'm wondering what my next step is. Did I introduce a new problem with the replacement components? Or is this how the initial fault is manifesting now that some components are replaced?
Well... it would certainly help.
Someone like Anatech, who doubtlessly worked on a ton of these amps, might recognize the issue and tell you a fix. Otherwise, it's back to more first priciples of tracking down the cause.
Short of a variac, the old light bulb(s) in series with the power cord trick would probably reduce B+ enough to run the amp for further diagnosis without the oscillations causing damage... or you could try reconfiguring the transformer to run at 240V, which should drop the 62V rails down to about 40V...
How about something like this for a scope:?
Protek P-3502C 20Mhz Oscilloscope HC Protek 2-Channel
Cheers,
Paul
Someone like Anatech, who doubtlessly worked on a ton of these amps, might recognize the issue and tell you a fix. Otherwise, it's back to more first priciples of tracking down the cause.
Short of a variac, the old light bulb(s) in series with the power cord trick would probably reduce B+ enough to run the amp for further diagnosis without the oscillations causing damage... or you could try reconfiguring the transformer to run at 240V, which should drop the 62V rails down to about 40V...
How about something like this for a scope:?
Protek P-3502C 20Mhz Oscilloscope HC Protek 2-Channel
Cheers,
Paul
I'll send Anatech a PM and see if he's willing to read this thread.
I'm keeping my eyes peeled for a used scope, but there's just no budget for it at the moment.
Tonight I'll check the transistors I replaced to make sure they are working (not damaged in installation). Logically, it seems like the over-heating of R22 must be caused by something I did, rather than being a symptom of the original fault. Otherwise, wouldn't I have seen the overheating, or at least evidence of it, before?
To that end, can i use the transistor referred to in post #82 to replace the 970's?
I'm keeping my eyes peeled for a used scope, but there's just no budget for it at the moment.
Tonight I'll check the transistors I replaced to make sure they are working (not damaged in installation). Logically, it seems like the over-heating of R22 must be caused by something I did, rather than being a symptom of the original fault. Otherwise, wouldn't I have seen the overheating, or at least evidence of it, before?
To that end, can i use the transistor referred to in post #82 to replace the 970's?
Last edited:
I would say, no. The Fairchilds are rated 50mA, while the 2SA970's are 100mA devices.
I would get a very bright light and a magnifying glass - bigger the better - and go over that PCB very carefully looking for bad solder. There is something strange going on that sound more like bad connections than bad parts.
As for your KSA992 transistor, the 50mA isn't that big of a deal. Q8 is normally not conducting and Q9 runs at 5mA (or less.
G²
Hi Fred, all,
I have to say that things seem to be getting confused and a bit out of hand here. Stop doing work now and let's collect ourselves.
Do not start replacing parts, especially the signal transistors. That goes for everyone here. The very first thing you have to do is to know what you are up against. That means that you must get into testing some components properly.
Testing transistors, power or otherwise, with the diode function will help you find outright short circuits or open junctions. That's about it. The resistance function is not too helpful beyond that. Beta measurements will be very temperature dependent, so they are a rough guide only. The digital beta testers in some meters are also a very rough guide. Remember that beta is also dependent on collector current as well. What you need now is a proper transistor tester, something like a Heathkit IT-18 is very effective. It's the one meter of my 6 or 7 that I always use first. The parameters that you need to pay attention to are C-E and C-B leakage. Beta tests must be in the normal range, to high or too low for the average part of that number and the transistor should be rejected. I have posted a jig to use for matching signal transistors in a few places here. Do not start replacing anything else until you have the amplifier repaired properly. You may be creating other faults.
Buying a "variac" is essential, and it will save transistors and PC boards. You should be using a really good meter like a Fluke or one of the new Agilent meters. Also, an oscilloscope is a required piece of equipment for this work. That is non-negotiable, see if you can find someone who can assist you on that. Your soldering iron must actually be a soldering station that holds a constant temperature. Getting good solder joints using an iron that plugs into the wall is not really possible.
The next steps are critical. On a pad of paper, write down in order each thing you have checked and what the result was. Do not write down any guesses or assumptions, that goes somewhere else. So, go mind dead and measure all your junctions and observations (facts only). This will keep you on track. You will not be making any tests under power, so discharge all the filter caps in the power supply.
If you have had one power transistor in a bank fail, you must replace the set with a matched set. It's a good time to put current devices in. I'd recommend the new On Semi MJ21195 and MJ21196. As a bonus, they should match fairly closely. Buy something like two extra per polarity to allow for matching. This is less than you would need with other part numbers. For TO-220 driver transistors, use MJE15032 and MJ15033. I have attached the data sheet for you.
Okay, I'm really tired now and it's way past my bed time. You have some work to do Fred, and anyone else following along.
One more bit of advice. I seldom disagree with djk, he knows what he is doing. I will say that I can't support the installation of much larger filter capacitors than are original though. On top of that, you can test them easily in operation (you need an oscilloscope for this). I find that most main filter capacitors are actually working just fine, and some are in better shape than new ones. Again, don't go changing things before you get the amp settled down. Also, your DC offset it too high. Matching transistors will bring that down.
Fred, you can get a good used 'scope for less than $200. If you plan to repair this amp or anything else, you really have to get your hands on one. It will save you money quite easily. You also need some good X10 probes for it, glue the switchable kind in the X10 position.
-Chris
I have to say that things seem to be getting confused and a bit out of hand here. Stop doing work now and let's collect ourselves.
Do not start replacing parts, especially the signal transistors. That goes for everyone here. The very first thing you have to do is to know what you are up against. That means that you must get into testing some components properly.
Testing transistors, power or otherwise, with the diode function will help you find outright short circuits or open junctions. That's about it. The resistance function is not too helpful beyond that. Beta measurements will be very temperature dependent, so they are a rough guide only. The digital beta testers in some meters are also a very rough guide. Remember that beta is also dependent on collector current as well. What you need now is a proper transistor tester, something like a Heathkit IT-18 is very effective. It's the one meter of my 6 or 7 that I always use first. The parameters that you need to pay attention to are C-E and C-B leakage. Beta tests must be in the normal range, to high or too low for the average part of that number and the transistor should be rejected. I have posted a jig to use for matching signal transistors in a few places here. Do not start replacing anything else until you have the amplifier repaired properly. You may be creating other faults.
Buying a "variac" is essential, and it will save transistors and PC boards. You should be using a really good meter like a Fluke or one of the new Agilent meters. Also, an oscilloscope is a required piece of equipment for this work. That is non-negotiable, see if you can find someone who can assist you on that. Your soldering iron must actually be a soldering station that holds a constant temperature. Getting good solder joints using an iron that plugs into the wall is not really possible.
The next steps are critical. On a pad of paper, write down in order each thing you have checked and what the result was. Do not write down any guesses or assumptions, that goes somewhere else. So, go mind dead and measure all your junctions and observations (facts only). This will keep you on track. You will not be making any tests under power, so discharge all the filter caps in the power supply.
If you have had one power transistor in a bank fail, you must replace the set with a matched set. It's a good time to put current devices in. I'd recommend the new On Semi MJ21195 and MJ21196. As a bonus, they should match fairly closely. Buy something like two extra per polarity to allow for matching. This is less than you would need with other part numbers. For TO-220 driver transistors, use MJE15032 and MJ15033. I have attached the data sheet for you.
Okay, I'm really tired now and it's way past my bed time. You have some work to do Fred, and anyone else following along.
One more bit of advice. I seldom disagree with djk, he knows what he is doing. I will say that I can't support the installation of much larger filter capacitors than are original though. On top of that, you can test them easily in operation (you need an oscilloscope for this). I find that most main filter capacitors are actually working just fine, and some are in better shape than new ones. Again, don't go changing things before you get the amp settled down. Also, your DC offset it too high. Matching transistors will bring that down.
Fred, you can get a good used 'scope for less than $200. If you plan to repair this amp or anything else, you really have to get your hands on one. It will save you money quite easily. You also need some good X10 probes for it, glue the switchable kind in the X10 position.
-Chris
Chris,
Many thanks for your reply. It's true. I'm a terrible project manager. So I agree, it's time for a code-freeze. I'll document what's been done so far, write down all my measurements and do my job as tech writer.
I'm shopping for a variac and scope. I can see I won't learn more if I don't. I'll tell the wife it's an investment. And I'll not comment further on the woeful state of my personal economy.
I have the usual Weller iron station (WES51) that works well. I can go fast and accurate with it and stuff doesn't get hot. I've gone over the work I did on this board with a magnifying work-lamp and it looks shiny and clean, with no bridges or cold joints. I'll post some macro pix to make sure I'm not delusional about my ability to at least solder.
I actually recently got an IT-18 tester, but I can't find a manual for it, just the calibration instructions. If anyone has a bead on a manual, I'd be happy to pay and/or scan.
cheers all,
Fred
Many thanks for your reply. It's true. I'm a terrible project manager. So I agree, it's time for a code-freeze. I'll document what's been done so far, write down all my measurements and do my job as tech writer.
I'm shopping for a variac and scope. I can see I won't learn more if I don't. I'll tell the wife it's an investment. And I'll not comment further on the woeful state of my personal economy.
I have the usual Weller iron station (WES51) that works well. I can go fast and accurate with it and stuff doesn't get hot. I've gone over the work I did on this board with a magnifying work-lamp and it looks shiny and clean, with no bridges or cold joints. I'll post some macro pix to make sure I'm not delusional about my ability to at least solder.
I actually recently got an IT-18 tester, but I can't find a manual for it, just the calibration instructions. If anyone has a bead on a manual, I'd be happy to pay and/or scan.
cheers all,
Fred
Re-boot, Step 1
Okay, step 1 of the re-boot. Below I've documented everything that's been done thus far, and known measurements. Just so everyone is on the same page.
Repair of Adcom GFA-555 (1986).
Model Info: This is the 1986 version of the GFA-555, using Front End PCB 72-2174.
Summary:
Initial fault was diagnosed as B-E shorted Q7 due to long-term heat failure. The following components were subsequently replaced: Q4, Q7, Q8, R7, R8.
After replacement, R22 immediately over-heated when amp was powered on, without a load.
Initial Fault Conditions:
Blown fuse (F1) on left channel Output PCB II. Fuse only blows under load.
82 vdc stable measured at speaker output jacks on left channel (functioning right channel measures 41 mvdc)
Known Initial Measurements
Measurements made on amp in initial state (i.e no parts removed or replaced, nothing disconnected):
Voltage across R15, R16 and R14
R16:
good channel: 598 mv
bad channel: 96 mv
R14:
good channel: 620 mv
bad channel: 549 mv
R15:
good channel: .1mv
bad channel: 350mv
Q4 (2SC2912) and Q7 (2SA1210) In circuit diode test (using BK 2707A):
Q7:
Bad Channel:
B+/C- O/L
B-/C+ .580
B+/E- .000
B-/E+ .000
E+/C- O/L
E-/C+ .580
Good Channel:
B+/C- O/L
B-/C+ .607
B+/E- .530
B-/E+ O/L
E+/C- O/L
E-/C+ .625
Q4:
Bad Channel:
B+/C- .662
B-/C+ O/L
B+/E- .666
B-/E+ O/L
E+/C- .703
E-/C+ O/L
Good Channel:
B+/C- .656
B-/C+ O/L
B+/E- .661
B-/E+ O/L
E+/C- .733
E-/C+ O/L
Q8 out of circuit testing (tested with Q7 removed, R7 and R8 disconnected):
Bad Channel:
B+/C- O/L
B-/C+ .626
B+/E- O/L
B-/E+ .641
E+/C- 1.968
E-/C+ O/L
Work Done to date
Flushed Front End PCB with Techspray iso alcohol and Deoxit Contact Cleaner Wash
Replaced: Q4, Q7, Q8, R7, R8
Removed and reinstalled: Q11
Reflowed all above joints.
No work on the good (R) channel was done.
Available Test Equipment
Heathkit IT-18 Transistor Checker
BK 2707A DVM
Elenco XP-720 Power Supply
Velleman K7000 Signal Tracer/Injector
Velleman K8065 Audio signal generator
Tektronix 422 oscilloscope (soon)
Variac (soon)
Okay, step 1 of the re-boot. Below I've documented everything that's been done thus far, and known measurements. Just so everyone is on the same page.
Repair of Adcom GFA-555 (1986).
Model Info: This is the 1986 version of the GFA-555, using Front End PCB 72-2174.
Summary:
Initial fault was diagnosed as B-E shorted Q7 due to long-term heat failure. The following components were subsequently replaced: Q4, Q7, Q8, R7, R8.
After replacement, R22 immediately over-heated when amp was powered on, without a load.
Initial Fault Conditions:
Blown fuse (F1) on left channel Output PCB II. Fuse only blows under load.
82 vdc stable measured at speaker output jacks on left channel (functioning right channel measures 41 mvdc)
Known Initial Measurements
Measurements made on amp in initial state (i.e no parts removed or replaced, nothing disconnected):
Voltage across R15, R16 and R14
R16:
good channel: 598 mv
bad channel: 96 mv
R14:
good channel: 620 mv
bad channel: 549 mv
R15:
good channel: .1mv
bad channel: 350mv
Q4 (2SC2912) and Q7 (2SA1210) In circuit diode test (using BK 2707A):
Q7:
Bad Channel:
B+/C- O/L
B-/C+ .580
B+/E- .000
B-/E+ .000
E+/C- O/L
E-/C+ .580
Good Channel:
B+/C- O/L
B-/C+ .607
B+/E- .530
B-/E+ O/L
E+/C- O/L
E-/C+ .625
Q4:
Bad Channel:
B+/C- .662
B-/C+ O/L
B+/E- .666
B-/E+ O/L
E+/C- .703
E-/C+ O/L
Good Channel:
B+/C- .656
B-/C+ O/L
B+/E- .661
B-/E+ O/L
E+/C- .733
E-/C+ O/L
Q8 out of circuit testing (tested with Q7 removed, R7 and R8 disconnected):
Bad Channel:
B+/C- O/L
B-/C+ .626
B+/E- O/L
B-/E+ .641
E+/C- 1.968
E-/C+ O/L
Work Done to date
Flushed Front End PCB with Techspray iso alcohol and Deoxit Contact Cleaner Wash
Replaced: Q4, Q7, Q8, R7, R8
Removed and reinstalled: Q11
Reflowed all above joints.
No work on the good (R) channel was done.
Available Test Equipment
Heathkit IT-18 Transistor Checker
BK 2707A DVM
Elenco XP-720 Power Supply
Velleman K7000 Signal Tracer/Injector
Velleman K8065 Audio signal generator
Tektronix 422 oscilloscope (soon)
Variac (soon)
Member
Joined 2009
Paid Member
Here are test results for the three original transistors that were replaced, taken with the "new" IT-18 (all tests out of circuit):
Q4 (2sc2912):
Beta: 24
Lceo: 0
Lcbo: 0
Q7 (2sa1210):
Beta: none (DVM reads E-B short)
Q8 (2sa970):
Beta: 40
Lceo: 0
Lcbo: 0
So if I'm interpreting the data correctly, Q8 and Q4 are okay, Q7 was shorted.
Next I will do in-circuit tests of these three transistors' replacements.
What else should I test?
Note to bigun: If I can't get this Adcom working again, building a new amp might be my next project!
Q4 (2sc2912):
Beta: 24
Lceo: 0
Lcbo: 0
Q7 (2sa1210):
Beta: none (DVM reads E-B short)
Q8 (2sa970):
Beta: 40
Lceo: 0
Lcbo: 0
So if I'm interpreting the data correctly, Q8 and Q4 are okay, Q7 was shorted.
Next I will do in-circuit tests of these three transistors' replacements.
What else should I test?
Note to bigun: If I can't get this Adcom working again, building a new amp might be my next project!
Hi Fred,
All right, you should probably replace both the 2SC2912 and the 2SA1210 as a pair. If you can't get the same part numbers, look at Fairchild and On-Semi. I have seen some Japanese part numbers cropping up in their product line. You can get this amp working properly! Also, Q7 may not have died from excessive heat. You would be surprised at how much thermal abuse a transistor could take over time. Until the amp is completely repaired, the initial failure will be a question mark.
The 2SC2912 and the 2SA1210 are mounted on "U" shaped heat sinks, so you do need to stick to using TO-126 (or whatever they call them these days) case styles. Do yourself a favor and buy 5 of each transistor so you can try to match the NPN with PNP beta. Also, to answer some earlier comments, these are not problematic devices at all. The issue was the driver transistors. They had a choice between higher current and higher voltage devices. At that time, you could only go one way or the other. The high current devices blew if the amp clipped with no load (Vce was exceeded), and the high voltage devices could blow into 4 ohm loads, and certainly a short (Ic was exceeded). These days you can use the On Semi parts and rest easy. Probably could back then too, but not with the Japanese devices in that case style.
I am a little concerned that your IT-18 might be reading low. The 2SA970 ought to have much higher gain, as should the 2SC2912. In the manual, there is a calibration procedure. It's pretty simple, all you need is a 180 ohm resistor (I think, the value may be wrong), and a fresh 1.5 V "D" cell battery. Use an Alkaline type to reduce leaking and pay attention to both the switch and battery contacts. The readings from a Heathkit IT-18 correlate very closely from readings made on a jig using power supplies to directly measure the base and emitter currents, so it's accurate. That even holds for power transistors! Yes, I was surprised. I didn't expect that kind of agreement. So, please clean things up and calibrate your tester and get back to us on that. Also, use 2SA970 for the old 2SA970s. They are available still and are an excellent transistor. Really good for audio, so buy extra for your experiments, also the compliment (2SC2240?).
When testing components, always either completely remove them from the circuit, or leave only one lead connected (easier to pull it for transistors, etc...).
-Chris
All right, you should probably replace both the 2SC2912 and the 2SA1210 as a pair. If you can't get the same part numbers, look at Fairchild and On-Semi. I have seen some Japanese part numbers cropping up in their product line. You can get this amp working properly! Also, Q7 may not have died from excessive heat. You would be surprised at how much thermal abuse a transistor could take over time. Until the amp is completely repaired, the initial failure will be a question mark.
The 2SC2912 and the 2SA1210 are mounted on "U" shaped heat sinks, so you do need to stick to using TO-126 (or whatever they call them these days) case styles. Do yourself a favor and buy 5 of each transistor so you can try to match the NPN with PNP beta. Also, to answer some earlier comments, these are not problematic devices at all. The issue was the driver transistors. They had a choice between higher current and higher voltage devices. At that time, you could only go one way or the other. The high current devices blew if the amp clipped with no load (Vce was exceeded), and the high voltage devices could blow into 4 ohm loads, and certainly a short (Ic was exceeded). These days you can use the On Semi parts and rest easy. Probably could back then too, but not with the Japanese devices in that case style.
I am a little concerned that your IT-18 might be reading low. The 2SA970 ought to have much higher gain, as should the 2SC2912. In the manual, there is a calibration procedure. It's pretty simple, all you need is a 180 ohm resistor (I think, the value may be wrong), and a fresh 1.5 V "D" cell battery. Use an Alkaline type to reduce leaking and pay attention to both the switch and battery contacts. The readings from a Heathkit IT-18 correlate very closely from readings made on a jig using power supplies to directly measure the base and emitter currents, so it's accurate. That even holds for power transistors! Yes, I was surprised. I didn't expect that kind of agreement. So, please clean things up and calibrate your tester and get back to us on that. Also, use 2SA970 for the old 2SA970s. They are available still and are an excellent transistor. Really good for audio, so buy extra for your experiments, also the compliment (2SC2240?).
When testing components, always either completely remove them from the circuit, or leave only one lead connected (easier to pull it for transistors, etc...).
-Chris
In my amp, these were not mounted to any heat-sinks (see the pic at post #73). Perhaps another tech removed them and forgot?
Anyway, I already replaced them with the same devices (from MCM), left the leads long and added the heat-sinks (following G2's advice). I did not match the devices' beta. Is that going to be an issue (at least for the purposes of the repair)?
I double-checked the calibration and it is correct. I cleaned all the contacts and switches with Deoxit contact cleaner. The battery reads a healthy 1.5v. I measured a new 2SC2912 with the gain test on my DVM and got 278, the IT-18 reading was 29 (x1 scale). I assume that's correct and that my DVM is reading x10, right?
I guess I am a little confused about the range switch on the IT-18...
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.