Hi Fred,
The IT-18 will provide a high sensitivity test in that it will detect small amounts of leakage and I suggest doing it while you have those outputs disconnected. Be sure to have the amp disconnected from the line with caps discharged.
The high voltage test, first involves having the output go to each rail with no load and measuring the drop on the .82 R emitter resistors. 75V max on the rails.
If any transistor on the off side has leakage you will measure a drop on the .82R. Try it with the pair that are still connected, then connect an additional pair and test at each rail as you go.
We could then do the same with a dummy load if you have one, might want to limit the supply to 50V max or whatever power level your dummy load can handle - not a speaker of course. Then the reading on the .82R resistors on the side supplying current will tell you how much (i = V/R) each transistor is handling - if one or more are damaged there will be an imbalance. Please let us know how you do on the no load tests, and what you have for a dummy load.
Hi Fred,
First of all, use your IT-18 to test beta and both leakage modes on all your output transistors and drivers. If they pass that test, confidence is pretty high that all those parts are fine. Can you still have problems? Yes, but those are rare. I have found signal transistors that have gone open when warm, so with the IT-18 connected as if to read beta, set the pointer cal further towards the middle. When you warm transistors up, you will see the pointer move towards the right smoothly. If you see any discontinuity in that movement, the transistor has trouble, or your connections are not solid.
To test further, connect the boards up as for normal operation. Using only one pair of outputs is fine as long as the bias is set for the one pair only, so the same voltage across the emitter resistor, or about 1/4 the current measured at the fuse holder (using an ammeter or low value resistor. This means that you also have the differential pair installed properly. Simply installing a matched pair is fine as that's what you want to end up with anyhow.
If the thermal switch is open, the tail current for the diff pair will be cut off, as will the Vas and output bias current. It's a thermally operated switch, so take care not to break a solder connection or the leads where they enter the white rectangular casing. This does not have to be in position on the heat sink for troubleshooting and testing.
-Chris
First of all, use your IT-18 to test beta and both leakage modes on all your output transistors and drivers. If they pass that test, confidence is pretty high that all those parts are fine. Can you still have problems? Yes, but those are rare. I have found signal transistors that have gone open when warm, so with the IT-18 connected as if to read beta, set the pointer cal further towards the middle. When you warm transistors up, you will see the pointer move towards the right smoothly. If you see any discontinuity in that movement, the transistor has trouble, or your connections are not solid.
To test further, connect the boards up as for normal operation. Using only one pair of outputs is fine as long as the bias is set for the one pair only, so the same voltage across the emitter resistor, or about 1/4 the current measured at the fuse holder (using an ammeter or low value resistor. This means that you also have the differential pair installed properly. Simply installing a matched pair is fine as that's what you want to end up with anyhow.
If the thermal switch is open, the tail current for the diff pair will be cut off, as will the Vas and output bias current. It's a thermally operated switch, so take care not to break a solder connection or the leads where they enter the white rectangular casing. This does not have to be in position on the heat sink for troubleshooting and testing.
-Chris
Hi Pete,
So I think I understand the concept, what I don't confidently know is how to actually connect things up to perform the test, what state the input board (should it even be connected?) should be in, etc.
It seems like you described some of this at some point, but I can't seem to find now.
I'm currently in the process of putting the input board all back together with all new semis and caps (though I'm not replacing anything else, in keeping with Chris's advice to minimize technician-induced error).
Nice to be back!
I'll do that first and post the results.
NB that I have to use my DVM for beta testing since the IT-18's gain test still isn't working right. Just haven't had time to investigate it further...
Hi Fred,
If the beta test does give an indication that varies with beta and temperature, it will work for the heat test just fine. You really do need to get that fixed. The manual contains a troubleshooting section. Given how simple this circuit is, my bet is with switch contact problems. You may also have an open in the leads from the meter to the test clips. Ohms function on your ohmmeter should make that a quick test.
What Pete's leakage test is all about is applying a voltage from collector to emitter and measuring the current flow. It should be absolutely zero with a clean, operating transistor. If you measure any current flow, you either have C-E leakage or C-B leakage. Either one condemns the transistor.
-Chris
If the beta test does give an indication that varies with beta and temperature, it will work for the heat test just fine. You really do need to get that fixed. The manual contains a troubleshooting section. Given how simple this circuit is, my bet is with switch contact problems. You may also have an open in the leads from the meter to the test clips. Ohms function on your ohmmeter should make that a quick test.
What Pete's leakage test is all about is applying a voltage from collector to emitter and measuring the current flow. It should be absolutely zero with a clean, operating transistor. If you measure any current flow, you either have C-E leakage or C-B leakage. Either one condemns the transistor.
-Chris
The input board needs to be disabled if you want to do this test, if you don't want to, just let me know. I know you are getting input from a lot of different directions so it must be confusing. Like I said, I like to leave things out and test as much as possible before proceeding. This goes completely against what some others have said here, but I was debugging these sorts of things when I was 10 years old - less actually. I do it my way.
It is completely reasonable to replace all the input semis since they are so inexpensive, but we might have wanted to probe something to determine if it was the part that failed.
So have you disconnected the input board? That was just a thought on my part to use molex connectors. The connectors were to allow you to use either input section which also caused alarm here. I'm guessing you won't have to do this once we finish with the output stage. I also have an intense dislike for the cheap Molex connectors but I was not going to have you leave them in.
I'd suggest board in, diff pair out, or T1 open, drive to the output drivers needs to be disabled in one of these ways. 1K resistor across VAS C to rail for pos output test, 1K
resistor C of Q4 to neg rail for neg test - never both 1K's in at the same time.
Dummy load?
"I like to leave things out and test as much as possible before proceeding. This goes completely against what some others have said here.".
It doesn't go against my methodology. I like to test every possible component before firing things up. There's nothing wrong with changing a handful of cheep parts. A set of 8 outputs isn't so cheep so it pays off to be careful. It's very disheartening to have to do a repair twice and not very profitable if you do it in a shop.
David.
It doesn't go against my methodology. I like to test every possible component before firing things up. There's nothing wrong with changing a handful of cheep parts. A set of 8 outputs isn't so cheep so it pays off to be careful. It's very disheartening to have to do a repair twice and not very profitable if you do it in a shop.
David.
I agree, not an easy amp to learn on given all those outputs if something goes wrong.
I am a bit nervous to work on mine.
I am a bit nervous to work on mine.
Likew I said before ... it's good to use single pairs of outputs and light loads or lower power. It doesn't effect stabilty to any great degree!
If you think this is bad, try working on one of those old Nikko Beta series. Talk about over kill in design. I've never seen so many transistors in a power amp and if you missed one thing....
There was also about 20 1/4W and 1/2W flame proof resistors and every one them had to be tested.
David.
Hi Pete,
Okay, the output boards are completely disconnected. I'll do IT-18 tests first and post those results.
The input board is all clean and ready for me to start re-populating the stuff I pulled. I'll leave out the diff. pair and jumper T1 such that I can easily open it.
I'll see if I can find some connectors I can use to temporarily re-connected things, but I will pull them for the final build-up/re-wire.
Re: dummy load, I have an 8-ohm 100 watt resistor attached to a heat sink. See post 171. Set between two ice-packs, it was pretty unfazed by previous tesing.
Yes, the IT-18 will work for the heat test. It calibrates fine. Just the test itself doesn't work right. I'll give it another look, but I haven't seen anything obvious yet and I've cleaned everything. I'll post whatever I find over in that IT-18 thread we set up.
cheers.
Just to let you know, don't power it up with the input board disconnected.
The drivers are not fully biased without the input board, and the outputs
could cross conduct.
The drivers are not fully biased without the input board, and the outputs
could cross conduct.
But don't I need to use the PS to test the output transistors? I'm a little confused. I've got the Elenco, but it only puts out about 17 volts.
The good channel is still connected to the input board, but the two output boards/heat sinks from the bad channel are now completely disconnected from everything.
Last edited:
I simply mean that for the high voltage leakage test you need to hook them back up, they can be disconnected for any IT-18 or meter testing. If your IT-18 has beta test problems we can do a basic test when/if you hook up the dummy load. If they share current evenly then they are reasonably well matched on beta.
You can just make a loop around the post and solder the wires if you have trouble finding connectors.
Output transistor IT-18 test results
Got my IT-18 working correctly. Here are the results for the 8 output transistors:
Q13: no leakage, 82hfe
Q14: no leakage, 71hfe
Q15: no leakage, 88hfe
Q16: no leakage, 80hfe
Q17: 10 microamps C-E leakage (!), no C-B leakage, 44hfe
Q13: no leakage, 52hfe
Q13: no leakage, 42hfe
Q13: no leakage, 45hfe
Recall that earlier (post 282) I measured high voltage across Q17's emitter resistor.
What can I do to test it further, or is that leakage too small to worry about?
Got my IT-18 working correctly. Here are the results for the 8 output transistors:
Q13: no leakage, 82hfe
Q14: no leakage, 71hfe
Q15: no leakage, 88hfe
Q16: no leakage, 80hfe
Q17: 10 microamps C-E leakage (!), no C-B leakage, 44hfe
Q13: no leakage, 52hfe
Q13: no leakage, 42hfe
Q13: no leakage, 45hfe
Recall that earlier (post 282) I measured high voltage across Q17's emitter resistor.
What can I do to test it further, or is that leakage too small to worry about?
I knew that the high voltage across the Q17 emitter resistor was a bad sign
but was hoping that your output stage was OK. Are you interested in replacing
just the bad output or would you plan to change them all? I would say that
yes Q17 is bad. We can proceed if you want to understand how to do the
test, or else you can just rebuild the output stage. I would only use Toshiba
outputs if you want to stay with the original parts, Chris will offer an opinion
also I'm sure. I might have a few Toshibas that I can give you - I have to
look for them. I'll let others comment also.
If you want to continue:
It seems that we've gotten all the data we can from them disconnected so I
think we're ready to hook it back up to the driver board. Do you have the
new higher voltage output stage drivers Q11 and Q12? If you do not have
spare drivers you might want to continue testing with the lower voltage
parts. Just have one set of outputs connected to start.
This is the idea to get started, you can go to the full 81V when you have the
correct Q11 and Q12:
http://www.diyaudio.com/forums/soli...r-high-dc-adcom-gfa-555-a-41.html#post2134698
Fred if you want original outputs and you can't find them state side, one of my old suppliers in Vancouver BC has the outputs listed on their web page.
The 2SD424 list $3.99 CAD
The 2SB544 list $4.65 CAD
If you like you can contact Daniel at Main Electronics Vancouver BC by email for availability and current pricing. Be sure though make it clear you want original Toshiba parts. I have purchased some questionable part there.
Main Electronics Bulk Semis Page
David.
The 2SD424 list $3.99 CAD
The 2SB544 list $4.65 CAD
If you like you can contact Daniel at Main Electronics Vancouver BC by email for availability and current pricing. Be sure though make it clear you want original Toshiba parts. I have purchased some questionable part there.
Main Electronics Bulk Semis Page
David.
Hi David,
I still have three Alpha 220 and 230 amps sitting here. They even work! But they do sound odd when compared to a good amplifier. I have plans to try and fix that issue and may use them for testing speaker designs using active crossovers.
Hi Fred,
You only need to replace that one transistor. In the mean time, disconnect it and continue testing. Put a big, black X on it. Actually, before condemning this part, try to remove it and clean the part with Methyl Hydrate and a tooth brush. External leakage is just as destructive as leakage in the part.
Hi Pete,
Well, my opinion is identical to yours. I can't fault any of your suggestions in fact. My earlier comments were an attempt to keep things simple for Fred and any others just starting out. After all, suggestions were flying in all directions.
-Chris
Did warranty on these amps. The Alpha 220 / 230 and down were fine, The Alpha 440 / 450 was badly designed and hell to work on. All those filter caps up top used to vent huge amounts of fluid where it would run down and sit on the upside-down main PCB. Also, the heat would just sit there cooking everything.
I still have three Alpha 220 and 230 amps sitting here. They even work! But they do sound odd when compared to a good amplifier. I have plans to try and fix that issue and may use them for testing speaker designs using active crossovers.
Hi Fred,
You only need to replace that one transistor. In the mean time, disconnect it and continue testing. Put a big, black X on it. Actually, before condemning this part, try to remove it and clean the part with Methyl Hydrate and a tooth brush. External leakage is just as destructive as leakage in the part.
Hi Pete,
Am I really that bad?
Well, my opinion is identical to yours. I can't fault any of your suggestions in fact. My earlier comments were an attempt to keep things simple for Fred and any others just starting out. After all, suggestions were flying in all directions.
-Chris
Hi, Anatech
I did had one of those, don t remember the exact model, it was
a 2 X 100W Rms/8R with a pair of VU meter..;
The day it was blown, due to abuse as PA amplifier,
i did open it, and trust me , i didn t even tried to repair it !!
Just removed the two PCB and installed a design of mine
along with doubling the power devices..
Seems that it was the good decision , after all.