Sorry I should've been more specific on this test. Leave Q116 - Q119 out and power the amp up via DBT. Then measure the voltage on the incoming traces to the bases of Q117 & Q116 to see if the full neg rail is still present. This will isolate the output from the input circuitry and will point to whether the issue is up front or in the back half of the amp.
You'll have to first find the source of the short on B to C of the right ch Q117, 119 though. Perhaps and errant solder splash.
A base to collector short on those would certainly put full neg rail in the wrong place.
You may want to lift one side of R125R to see if it measures properly too.
You'll have to first find the source of the short on B to C of the right ch Q117, 119 though. Perhaps and errant solder splash.
A base to collector short on those would certainly put full neg rail in the wrong place.
You may want to lift one side of R125R to see if it measures properly too.
I found the short on Q117R and Q119R. It was actually from B to E, and I traced it to a shorted C3856 in Q129R. No idea how that shorted, but it was totally blown.
I removed a leg of R125R and it was 3.2ohms, so ok.
I have now removed Q116 - Q119 L&R and powered up the amp through the DBT. For the first time, the overload LEDs turned on, then off after a few seconds, and I heard relays click on the power amp PCBs.
Measuring the bases of Q116 and Q117, I get:
Q116L B: -78.5v
Q117L B: -78.4v
Q116R B: -78.7v
Q117R B: -78.6v
I removed a leg of R125R and it was 3.2ohms, so ok.
I have now removed Q116 - Q119 L&R and powered up the amp through the DBT. For the first time, the overload LEDs turned on, then off after a few seconds, and I heard relays click on the power amp PCBs.
Measuring the bases of Q116 and Q117, I get:
Q116L B: -78.5v
Q117L B: -78.4v
Q116R B: -78.7v
Q117R B: -78.6v
Thats very odd that shorting the inputs dropped the neg rail voltage. The input resistor R101 should be enough to keep the input pulled down low enough to keep that high of a voltage from coming up. The Q116/117 voltages are about right although the -0.4V at Q117 is a bit lower than we would like. I think you still need to measure around a bit though as the offset of .9V vs -.4V is a bit high. What is the voltage at the base of Q114 & Q115 now with the input shorted? Also what is the voltage now at the collector of Q106 with the input shorted? Also what do you measure at the feedback side of Q101 between R120 and the backside gate of Q101. It should be 0V but it looks like its going to be slightly positive based on the readings to Q117/116 being imbalanced.
When it comes time to begin bringing the output stage up what I would do to protect the output stage (and your pocket book) at this point is to only install one N & P output (say Q120 & Q121) and one each of the drivers say Q116 & 117. This should allow the output to come up but will only destroy one transistor if a fatal problem occurs. The single output transistors will be fine with no signal or even a very low signal into an 8 ohm load.
When it comes time to begin bringing the output stage up what I would do to protect the output stage (and your pocket book) at this point is to only install one N & P output (say Q120 & Q121) and one each of the drivers say Q116 & 117. This should allow the output to come up but will only destroy one transistor if a fatal problem occurs. The single output transistors will be fine with no signal or even a very low signal into an 8 ohm load.
I think essentially R101 wasn't pulling things down because without the RCA connected, there's no ground in that small part of the circuit. I confirmed that with some measurements on R101.
expected values in parens:
Q114L
B: 1.99v (2.0v)
C: 73.5v (74.0v)
E: 1.37v (1.4v)
Q115L:
B: -1.48v (-2.0v)
C: -73.5v (-74.0v)
E: -0.91v (-1.4v)
Q114R
B: 2.63v (2.0v)
C: 73.5v (74.0v)
E: 2.1v (1.4v)
Q115R:
B: -2.1v (-2.0v)
C: -73.4v (-74.0v)
E: -1.54v (-1.4v)
Q106L
C: -78.3v (0.3v)
Q106R
C: -79.0v (0.3v)
For the Q101 measurements, did you mean R102?
Q101L
G1: -1.0v
G2: -22.45v
Q101R
G1: 0.2v
G2: -7.47v
Good ideas on bringing this up gently.
expected values in parens:
Q114L
B: 1.99v (2.0v)
C: 73.5v (74.0v)
E: 1.37v (1.4v)
Q115L:
B: -1.48v (-2.0v)
C: -73.5v (-74.0v)
E: -0.91v (-1.4v)
Q114R
B: 2.63v (2.0v)
C: 73.5v (74.0v)
E: 2.1v (1.4v)
Q115R:
B: -2.1v (-2.0v)
C: -73.4v (-74.0v)
E: -1.54v (-1.4v)
Q106L
C: -78.3v (0.3v)
Q106R
C: -79.0v (0.3v)
For the Q101 measurements, did you mean R102?
Q101L
G1: -1.0v
G2: -22.45v
Q101R
G1: 0.2v
G2: -7.47v
Good ideas on bringing this up gently.
Yeah Now that I'm thinking more about it I believe that it was latched up to the negative rail due to the disconnected ground reference. I've heard of amps doing this, but never personally seen it. So the ground is connected back to the chassis on the input RCA now?
The feedback side of Q101 comes in from R120. It should be zero or very close. I see that is still quite a bit off from 0.
What do you measure at the point between R118 & R119, that should be close to zero or slightly positive based on your prior measurements.
The feedback side of Q101 comes in from R120. It should be zero or very close. I see that is still quite a bit off from 0.
What do you measure at the point between R118 & R119, that should be close to zero or slightly positive based on your prior measurements.
I have the two input wires clipped together with one end of a alligator cable, then the other end of the alligator clipped to the chassis in an area where the paint is removed (presumably at the factory).
I forgot to mention that the DBT triggered when I was measuring Q106R, so it's possible I damaged that part. Hopefully not but I want to note it so I remember to check it again.
The point between R118 and R119 measures:
L: -72.6v
R: -71.4
On the Q101 side of R120:
L: -22.75v
R: -7.53v
I forgot to mention that the DBT triggered when I was measuring Q106R, so it's possible I damaged that part. Hopefully not but I want to note it so I remember to check it again.
The point between R118 and R119 measures:
L: -72.6v
R: -71.4
On the Q101 side of R120:
L: -22.75v
R: -7.53v
So are you still at < 1V on the bases of Q116/117? I'm guessing the answer to that is no not with that measurement in the center of R118 & R119.
Do you still have Q116 - 119 pulled out of the circuit?
If you still have < 1V at the Q116 & 117 bases then power down and drain the main caps and measure the resistance on R116,117,118,119. I think you measured one or two of those prior.
Do you still have Q116 - 119 pulled out of the circuit?
If you still have < 1V at the Q116 & 117 bases then power down and drain the main caps and measure the resistance on R116,117,118,119. I think you measured one or two of those prior.
We really need to figure out whats up with the input LTP. The current source and +10V cascode voltage reference are way out of spec and I do think this is whats causing most of the issues. I'm really beginning to think that the input Q101 may actually be bad. You should pull it again and use your new tester on it. You should see some amount of drain current. I believe mine showed about 0.5mA which is not the Idss you'll measure if you connect a higher voltage source to it as I'll show below but it'll give you an idea of whther or not they're working and how closely the pairs match.
You could also test the Idss of the Q101 matched pairs by using a simple 9V battery test the same as the image in this first post of this thread. The Idss of the eahc of the matched pairs should be very close probably within 5% of each other for Idss.
You could also test the Idss of the Q101 matched pairs by using a simple 9V battery test the same as the image in this first post of this thread. The Idss of the eahc of the matched pairs should be very close probably within 5% of each other for Idss.
I thought I'd expand on why I think you should look more closely at Q101. There's only about a 0.1V drop across R103/104 which says there's next to nothing flowing through the LTP input. The voltage drops you have recorded look correct on the LTP from top to bottom and there's a 10V drop across Q101 leading to the Q106 current source. So current should be flowing through Q106 but its not even when you had near 0V at the bases of the Q116/117 output drivers.
The DMM readings we took a few nights back are suspect to me, even though they were similar. I'd like to see how much Idss current these devices are actually flowing.
The DMM readings we took a few nights back are suspect to me, even though they were similar. I'd like to see how much Idss current these devices are actually flowing.
Q101 L&R pulled and tested with transistor tester:
Q101L:
side 1: N-JFET, Id = 0.57mA@Vg=0.40v
side 2: N-JFET, Id = 0.56mA@Vg=0.39v
Q101R:
side 1: N-JFET, Id = 0.49mA@Vg=0.34v
side 2: N-JFET, Id = 0.49mA@Vg=0.34v
So, this seem close to what you measured with your similar transistor. I didn't try to 9v battery test.
Q101L:
side 1: N-JFET, Id = 0.57mA@Vg=0.40v
side 2: N-JFET, Id = 0.56mA@Vg=0.39v
Q101R:
side 1: N-JFET, Id = 0.49mA@Vg=0.34v
side 2: N-JFET, Id = 0.49mA@Vg=0.34v
So, this seem close to what you measured with your similar transistor. I didn't try to 9v battery test.
Okay those Q101's look okay. I would install one of the C3503's into the Q106 location. Also see what Q106 measures with the tester. See if the Hfe is abnormally low or anything like that.
One thing I just realized is that there's about 150V from C to E on Q107 and Q104/105 as well. These Toshiba transistors are only rated to 120V max and although there's probably a small amount of headroom above that voltage, 150V cannot be a good thing. I'd go ahead and pull Q107 and test it.
If any of these Q107,104,105 prove to be bad the KSA992/KSC1845 are usually used in their places. the KSA/KSC have a lower current rating but a higher power rating, but in these locations this shouldn't be a problem.
One thing I just realized is that there's about 150V from C to E on Q107 and Q104/105 as well. These Toshiba transistors are only rated to 120V max and although there's probably a small amount of headroom above that voltage, 150V cannot be a good thing. I'd go ahead and pull Q107 and test it.
If any of these Q107,104,105 prove to be bad the KSA992/KSC1845 are usually used in their places. the KSA/KSC have a lower current rating but a higher power rating, but in these locations this shouldn't be a problem.
Leaving Q101 out for the moment, I pulled Q104, Q105, Q106, and Q107. I'm using the transistor tester to start with. I can do DMM measurements later if you think I need them.
Q104L:
2* diode symbol. Vf=937mV 761mV
Q104R:
2* diode symbol. Vf=866mV 703mV
Q105L:
BJT-NPN
hFE=533
Ie=2.8mA
Vbe=606mV
Q105R:
BJT-NPN
hFE=371
Ie=3.4mA
Vbe=658mV
Q106L:
2* diode symbol
.59 ohms 7mV 692mV
Q106R:
2* diode symbol
1.30 ohms mV 697mV
Q107L:
BJT-PNP
hFE=375
Ic=3.5mA
Vbe=620mV
Q107R:
BJT-PNP
hFE=408
Ic=3.8mA
Vbe=635mV
So, strange readings on Q104 and Q106. Have you ever seen the transistor tester indicate 2*diodes rather than a transistor?
Q104L:
2* diode symbol. Vf=937mV 761mV
Q104R:
2* diode symbol. Vf=866mV 703mV
Q105L:
BJT-NPN
hFE=533
Ie=2.8mA
Vbe=606mV
Q105R:
BJT-NPN
hFE=371
Ie=3.4mA
Vbe=658mV
Q106L:
2* diode symbol
.59 ohms 7mV 692mV
Q106R:
2* diode symbol
1.30 ohms mV 697mV
Q107L:
BJT-PNP
hFE=375
Ic=3.5mA
Vbe=620mV
Q107R:
BJT-PNP
hFE=408
Ic=3.8mA
Vbe=635mV
So, strange readings on Q104 and Q106. Have you ever seen the transistor tester indicate 2*diodes rather than a transistor?
Another thing to watch for when using the tester is abnormal readings in Vbe or Hfe. I've seen some bad ones still show as transistors, but have an Hfe of like 4, or a Vbe of 300mV.
One of the junctions should be reading as a short on Q104/106 if you test them with your meter. Thats the low resistance measurement you're seeing.
Anyway this is a good thing as we knew something was bad in this area. Now you've identified them so you can get the front end LTP operating again. Once that begins working hopefully the rest of the amp responds however I'm still suspicious of Q110 too as the emitter voltage you measured 69V is not correct for that one with the base at 76v. Before you submit another parts order I would pull that one and put it on the tester too.
Like I was saying in my response last night, I'd buy a decent supply of KSA992's just in case one of the current mirror transistors (Q102/103) turns out to be bad you'll need to come up with some well matched pairs. I'd get at least 25 of the A992's to have a decent pool of them to pick from. With 10 - 12 of the C1845's you should be fine and have some spares if needed. These trans are cheap so shipping will probably cost more than the parts.
One of the junctions should be reading as a short on Q104/106 if you test them with your meter. Thats the low resistance measurement you're seeing.
Anyway this is a good thing as we knew something was bad in this area. Now you've identified them so you can get the front end LTP operating again. Once that begins working hopefully the rest of the amp responds however I'm still suspicious of Q110 too as the emitter voltage you measured 69V is not correct for that one with the base at 76v. Before you submit another parts order I would pull that one and put it on the tester too.
Like I was saying in my response last night, I'd buy a decent supply of KSA992's just in case one of the current mirror transistors (Q102/103) turns out to be bad you'll need to come up with some well matched pairs. I'd get at least 25 of the A992's to have a decent pool of them to pick from. With 10 - 12 of the C1845's you should be fine and have some spares if needed. These trans are cheap so shipping will probably cost more than the parts.
I pulled the Q110s and checked them. They seem ok, though the hFE and Ic is noticeably lower than the parts I measured yesterday:
Q110L:
BJT-PNP
hFE=126
Ic=1.2mA
Vbe=651mV
Q110R:
BJT-PNP
hFE=130
Ic=1.2mA
Vbe=668mV
Agreed on ordering lots of the cheap transistors. So you're saying I should order the KSA992 (PNP) to replace the 2SA970 (PNP) and the KSC1845 (NPN) to replace the 2SC2240 (NPN). To replace the 2SC2705 on Q106, can I use the KSC1845, or should I order more KSC3503?
Q110L:
BJT-PNP
hFE=126
Ic=1.2mA
Vbe=651mV
Q110R:
BJT-PNP
hFE=130
Ic=1.2mA
Vbe=668mV
Agreed on ordering lots of the cheap transistors. So you're saying I should order the KSA992 (PNP) to replace the 2SA970 (PNP) and the KSC1845 (NPN) to replace the 2SC2240 (NPN). To replace the 2SC2705 on Q106, can I use the KSC1845, or should I order more KSC3503?
The hFE seems fine for those 2sa1370's, the grading will determine which hFE is normal. There's four different grade ranges running from hFE 40 up to 320. Yours could fall in the E grade of hFE 100 - 200. You can look at the 1370's and you should see a letter designating the grade typically at the end of the part# Like A1370 E. Anyway just looking at the specs of the A1370 it looks like the KSC3503 would be a great pin for pin replacement if you wanted to replace them. However it looks like they're okay now based on your measurements.
For Q106 you will want to use a C3503. There'll be about 4mA flowing through the Q106 current source and dropping like 78V so the power dissipated will be >300mW which will be too much for a C1845 which is rated for 500mW max to take and be reliable in the long term. The C3503 (1.2W) will take that amount of power no problem. Thats why Nak put a 1W rated transistor in that location I'm sure.
I'd also replace the 10V zeners ZD11 that set the cascode reference in the front end. The Voltage delta you measured appeared to indicate they were working fine, but the front end took a surge of some sort. So they're a 30 cent part and if you're buying some parts get a pair of those too just to be safe.
You might want to think about installing the back to back 12V protection zeners ZD15, ZD16 on the input that is showing to be for the Canadian models of this amp. I can't see it being a bad thing having that protective voltage clamp on the input.
All of these zeners are nothing special and can be replaced with standard 1/2watt parts.
While you're waiting on your new parts to arrive you should go ahead and pull Q108/109 and verify that they test good on the tester. Those are BL graded 970's so the hFE should be between 350 - 700.
One thing to note is that when you go to order the A992/C1845's you'll see a variety of part numbers, its all packaging, bent legs, bulk, tape mounted, etc. I usually opt for the tape mounted ones because you can test them with your tester without removing them from the tape and write your readings down on the cardboard tape. Some people have reported getting closer hFE matched devices when they come on tape too.
Another great and cheap gadget to have in your testing arsenal is a killawatt meter. I like to plug my DBT into it and monitor Watts of power being pulled from the wall. It a great tool to have around the house too for appliances and other electronics. I use it all the time for my amps I'm working on. I am always watching my power draw when I first bring an amp up. Some of the amps I work on when bringing them up on the variac I can tell right away from the killawatt that something is wrong before the DBT even begins to glow. Its also nice because you can power the channels individually and see if one or the other has an abnormally high power draw, indicating a possible bias problem or other issues.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.