The No.1 Website for Pro Audio
Roland JC120-H Output Transistors
Old 8th November 2016
  #1
Gear Head
 

Roland JC120-H Output Transistors

So my buddy's Roland Jazz Chorus suddenly quit working. After a look under the hood and finding both fuses in the power section blown, I tested the big filter caps in the power section and they were off spec so I replaced those.. Replaced fuses...and pop, again blown. The main fuse of the amp did not blow however...

So I think I have it deduced to the two output transistors for the right side power amp of the head. Both tested completely shorted. The pair on the left side power amp are perfectly fine.

Here is the dilemma.. Where in the FRIG do I find a couple of genuine Toshiba 2SD845 transistors?? After doing some research, I am extremely cautious of buying 'fake' transistors on the internet. Apparently there are some companies that are putting out 'rare' or 'hard to find' or 'defunct' products like this for sale and they just either fail, blow up, don't work, etc. All of course adding up to no money back/no return to emails/calls.

I've seen on some forums about a replacement transistor but it isn't in a mt-200 housing? I'd love to buy a couple of Toshibas off of eBay...but yeah, does ANYONE know of a reputable seller of some older solid state amp parts like this?

Also, does anyone know about the extra code at the bottom of these transistors? 3 of them say O 5B and one says O 9F but they both test exactly the same. Is this a date code?

Thanks for any and all help!!!
Attached Thumbnails
Roland JC120-H Output Transistors-img_0556.jpg   Roland JC120-H Output Transistors-img_0554.jpg  
Old 9th November 2016
  #2
Gear Addict
Google says that you can replace them with Sanken 2SC3264. Good brand, also mt-200 and available from e.g. Farnell.

However it's very likely that whatever caused the old ones to blow up will do the same to the new ones. For starters quite often if the power transistors are dead so are the drivers, and the "domino" can go back to (or come from) the Vbe multiplier, VAS, VAS current source... Unless you manage to identify where the fault started, throwing new transistors at it isn't likely to solve the problem.

Do you at least have a service manual or schematic so you can try to figure out where the problem is?

Cheers,

Cabirio
Old 11th November 2016
  #3
Gear Head
 

Quote:
Originally Posted by Cabirio View Post
Google says that you can replace them with Sanken 2SC3264. Good brand, also mt-200 and available from e.g. Farnell.

However it's very likely that whatever caused the old ones to blow up will do the same to the new ones. For starters quite often if the power transistors are dead so are the drivers, and the "domino" can go back to (or come from) the Vbe multiplier, VAS, VAS current source... Unless you manage to identify where the fault started, throwing new transistors at it isn't likely to solve the problem.

Do you at least have a service manual or schematic so you can try to figure out where the problem is?

Cheers,

Cabirio
Thanks! I do have a schematic after some digging on the net. I am really enjoying learning a lot about solid state amps so please don't tell me to give up and take it to a shop!!

Here is a picture I've taken of part of the schematic. I haven't had a class yet on anything to do with transistors/diodes yet(that's this spring), so yes, I am ignorant to a lot that is going on.. but all of this research I feel is helping me learn how to test components and more easily scan over complicated schematics to quickly find different 'sections' of the amplifier.

So are the transistors I have labeled as driver correct? If so, I checked them this morning and both test fine on the right power amp with the shorted output transistors.. They are Hitachi brand. Q29 is a 2SD669A NPN and Q30 is a 2SB649A PNP. Any other ideas of components to test? The diode connected with the driver transistors? Could any resistor failing in this area just wreak havoc? I'll gladly pull and test any and every component. I just see it as honing my soldering/testing skills.

I can send any picture you might need. Full schematic, actual pics of the board, etc. Thanks for any and all help!
Attached Thumbnails
Roland JC120-H Output Transistors-img_0585.jpg  
Old 11th November 2016
  #4
Gear Addict
How are you testing the transistors?

I've found the full schematic that corresponds to the one you posted (attached), but from what I've read there are many versions of the JC 120 so job #1 is to make sure that it's the right one. Compare the schematic to the PCB starting with the differential pair (Q23-Q24) and follow it through the VAS (Q22), Vbe multiplier (Q32-Q33), drivers (you got those right, Q29-Q30) and surrounding components. Make sure that the resistors and caps are exactly of the values shown in the schematic and make a note of the type of each of those transistors.

While you're at it, check that there are no broken / burnt traces in the PCB and no burnt components. One possible culprit is something wrong with the Vbe multiplier biasing the output transistors too high, so pay special attention to resistors R145, R146 and R147: even if they look alright, lift one of their legs and measure their values to make sure.

A picture of the top and bottom of that part of the PCB would help too.

Cheers,

Cabirio
Attached Thumbnails
Roland JC120-H Output Transistors-roland_jazz-chorus2.jpg  
Old 16th November 2016
  #5
Gear Head
 

Quote:
Originally Posted by Cabirio View Post
How are you testing the transistors?

I've found the full schematic that corresponds to the one you posted (attached), but from what I've read there are many versions of the JC 120 so job #1 is to make sure that it's the right one. Compare the schematic to the PCB starting with the differential pair (Q23-Q24) and follow it through the VAS (Q22), Vbe multiplier (Q32-Q33), drivers (you got those right, Q29-Q30) and surrounding components. Make sure that the resistors and caps are exactly of the values shown in the schematic and make a note of the type of each of those transistors.

While you're at it, check that there are no broken / burnt traces in the PCB and no burnt components. One possible culprit is something wrong with the Vbe multiplier biasing the output transistors too high, so pay special attention to resistors R145, R146 and R147: even if they look alright, lift one of their legs and measure their values to make sure.

A picture of the top and bottom of that part of the PCB would help too.

Cheers,

Cabirio
Depending on if NPN or PNP, and the pinout from the data sheet for specific part, I have tested using the ohms range and also on the diode setting of my multi-meter.

The 2sd845 is an NPN and actually goes BCE, according to the data sheet.

Right Power Amp:
Far Right Transistor: Pos Base, Neg Emitter= .402 Kohm
Left Side Right Power Amp: Pos Base, Neg Emitter= 00.3 ohm

Reverse the leads: Far Right: Neg Base, Pos Emitter= .402 Kohm
Left Side Right Power Amp: Neg Base, Pos Emitter= 00.6 ohm

Left Power Amp:
Right: Pos Base, Neg Emitter= 12.70 Mohm
Left: Pos Base, Neg Emitter= 12.55 Mohm

Reverse the Leads: Right: Neg Base, Pos Emitter= OL
Left Side: Neg Base, Pos Emitter=OL

I've been following the same kind of testing for all other transistors so far. Looking for shorts.

R144: 10K Test: 9.91K
R145: 4.7K, Test: 4.7K
R146: 15K, Test: 14.93K
R147: 120K, Test: 120.2K

C80: 150pf, Test: 240pf
C81: 150pf, Test: 240 pf
These capacitors say B151, and online research shows tolerance +or- 10%

So yeah, the only thing I've really been testing for, as far as the transistors go is shorts. The driver's tested that they had none. Those tiny caps seemed suspect. But would such a tiny cap have that big of an impact on the entire circuit?? Damn. Is there anything I'm missing on testing all these guys? I'll happily de solder and re-test if need be. I had just heard that testing for shorts was a nice rule of thumb for transistors.

The right side power amp 2sd845 actually test with different numbers no matter what pin you put any lead on, so I am assuming they are done-zo...

I've included pictures of the top and bottom of the PCB board.
I had a random question about some random thread I've seen regarding Roland corp and their glue that they use on some older components. Some people have posted in forums that this old yellow glue becomes conductive over a period of time. Would this create problems with parts being glued together? I've included some pictures that show several components glued together.. Also there is a picture of perhaps a suspect leaky capacitor. All the way on the other side of the board than the problematic output transistors, but still. Anyways, thanks for any advice and sorry for being so long winded!
Attached Thumbnails
Roland JC120-H Output Transistors-img_0596.jpg   Roland JC120-H Output Transistors-img_0597.jpg   Roland JC120-H Output Transistors-img_0598.jpg   Roland JC120-H Output Transistors-img_0599.jpg   Roland JC120-H Output Transistors-img_0600.jpg  

Roland JC120-H Output Transistors-img_0601.jpg   Roland JC120-H Output Transistors-img_0602.jpg   Roland JC120-H Output Transistors-img_0603.jpg  
Old 16th November 2016
  #6
Gear Head
 

Heres the pdf of the schematic that I am using
Attached Files
File Type: pdf Roland JC-120H Jazz Chorus (84).pdf (402.3 KB, 338 views)
Old 16th November 2016
  #7
Lives for gear
i would check the transformer and other parts of the power supply before running ahead and replacing the transistors... like the person above said... likely doing this will result in blowing your new transistors as well - though it doesn't sound like you're replacing them on anything other than a hunch?
Old 16th November 2016
  #8
Gear Addict
Although from your measurements it does look like those transistors are shorted, the proper way to test transistors is with the diode setting. Measure all junctions (BE, BC, CE) in both directions, all should be open except the BE and BC ones when they are forward-biased, which should read a voltage somewhere between 0.5 and 0.8V (maybe lower or higher depending on transistor type, for these it will probably be closer to 0.6-0.7V). Do this for all transistors in the faulty channel and the output transistors in the healthy one, and just to be sure measure diode D22 too.

Those caps may measure higher than they should because of the capacitance between the test leads, in any case even if they were off by that much, that wouldn't have caused the problem. The ones in the last picture do look like they have leaked, but from their values they seem to be the ones in the +/- 15V regulator feeding the preamp (C6 and C7). Not a great idea to use 16V caps there, make a note to replace them and go with 25V ones, like these (it's ok to use 220u caps in both positions):

C6, C7: EEU-FC1E221 - PANASONIC ELECTRONIC COMPONENTS - Electrolytic Capacitor, 220 µF, 25 V, FC Series, ± 20%, Radial Leaded, 8 mm | Newark element14

Also, even if the electrolytic caps in the power amps look fine, I'd replace them too in both channels just to be sure, it isn't always visible when electrolytic caps go bad. You can use these:

C35, C45, C68, C86: EEU-FC1H100L - PANASONIC ELECTRONIC COMPONENTS - Electrolytic Capacitor, 10 µF, 50 V, FC Series, ± 20%, Radial Leaded, 5 mm | Newark element14

C28, C59: ECE-A1HN100U - PANASONIC ELECTRONIC COMPONENTS - Electrolytic Capacitor, 10 µF, 50 V, SU(Bi-polar) Series, ± 20%, Radial Leaded, 6.3 mm | Newark element14

For these last two it's better to use a bi-polar type with a rating of at least the rail voltage, otherwise if for some reason the amp output fails and sticks to one rail, you may get fireworks.

I only see glue between those two ceramic resistors, even if it had become conductive it wouldn't be a problem.

I've run some LTSpice simulations and, we'll see what the measurements show, but I bet what happened was that overheating or a surge in mains voltage caused thermal runaway of the output transistors. That Vbe multiplier implementation is, er, questionable, but it's very easy to make it much more stable with just an extra cap and resistor.

Anyways, the cap replacements aren't urgent and we'll see about the Vbe multiplier, if we find that it's fixable you can wait and order the caps together with the new output transistors, so for now make sure all other transistors in the faulty channel are ok and report back. Then we'll do some voltage measurements, both in the healthy channel and the faulty one without output transistors, and take it from there.

Cheers,

Cabirio
Old 16th November 2016
  #9
Gear Addict
Quote:
Originally Posted by Bitsmith View Post
i would check the transformer and other parts of the power supply before running ahead and replacing the transistors...
You're right about this, the power supply is the first thing to check and we will in the next set of measurements.

Cheers,

Cabirio
Old 21st November 2016
  #10
Gear Head
 

So here are results from correct diode setting testing of all the transistors in the faulty channel. Also included testing of left power amp output transistors to compare.

Ran into a problem... So the glue I talked about happened to be all over Q22, connected to C62. I was attempting to remove Q22 and a leg snapped off of it like a brittle piece of hay. Hardly any pressure or pulling on it and snap...

The actual marking on the component itself says MMPS U06. Color code on back(from top to bottom)reads Orange, Yellow, Red. Now, the schematic says Q22 is a 2SC2229-Y (or 2SB649A-C). Does this mean my schematic could be the wrong one? Can't find a MMPS U06 anywhere on the net besides a mega old data sheet. The transistors the schematic states don't seem extremely similar when looking at their data sheets. This is the ONLY component that I've come across that isn't exactly what the schematic states. If you have any suggestions or links for an equivalent that would be awesome because it is going to have to be replaced. I can take it to my University parts guy tomorrow and see if he can possibly match it..

Perhaps this little guy that snapped so easily was the problem all along? I believe you identified it as the VAS.

Also, still not completely sure how to test these two JFET's that are thrown in the mix (Q20 and Q17) I included my testing with the diode setting and doing it how I did the other ones essentially.. Any help with how to properly test these would be awesome. I've watched some videos but it seems you have to have a very complete testing setup at your house to properly do so.

It still seems that all components are testing good? So narrowed down to output transformer?
Old 21st November 2016
  #11
Gear Head
 

Heres the test pdf.

Realized I messed up on first column of old output transistors. I put .162 where should say 2sd845
Attached Files
File Type: pdf Jc120 Transistor Test2.pdf (18.4 KB, 174 views)
Old 21st November 2016
  #12
Gear Head
 

Also the pictures of the transistor with the snapped leg
Attached Thumbnails
Roland JC120-H Output Transistors-img_0627.jpg   Roland JC120-H Output Transistors-img_0628.jpg  
Old 21st November 2016
  #13
Gear Addict
That's great, thanks for organizing the measurements so clearly. Yup, all the other transistors look good. Don't worry about the JFETs, they are used as switching transistors for muting and there's no way that a fault in the power amp would have propagated to them.

It's not unusual to find parts that don't match the schematics, maybe that's what they had in stock at the time and decided to use it... This is especially true for guitar amplifiers, for some reason manufacturers go through versions and revisions like crazy and they don't always produce schematics for each of them (*).

Fortunately the VAS isn't critical at all for the biasing of the output stage (which I still think is the best candidate for the cause of the problem), that's an ancient part and it's perfectly ok to replace it with a modern equivalent of the 2SC2229, this one is good for VAS duties: KSC2316YTA - FAIRCHILD SEMICONDUCTOR - Bipolar (BJT) Single Transistor, NPN, 120 V, 120 MHz, 900 mW, 800 mA, 60 | Newark element14. Get a couple to replace it too in the other channel, for the sake of symmetry, and the same goes for the output transistors, though you should take the measurements in the healthy channel with all the stock transistors first. It's past bedtime this side of the pond, tomorrow I'll tell you what you need to measure next.

(*) While you wait for your order, this free book is a very good read on solid state guitar amplifiers and it mentions this issue using precisely the JC-120 as an example (see page 40): http://www.thatraymond.com/downloads...ttala_v1.0.pdf
Old 21st November 2016
  #14
Gear Addict
Almost forgot: to fix that questionable Vbe multiplier, you're going to need two each of these:
3296Y-1-103LF - BOURNS - Trimmer, Multiturn, 10 kohm, 500 mW, ± 10%, Trimpot 3296 Series, 25 Turns, Through Hole | Newark element14
B32529C1223J000 - EPCOS - Film Capacitor, 0.022 µF, PET (Polyester), 100 V, B32529 Series, ± 5%, Radial Leaded | Newark element14

And a couple of 1.5k resistors.

Edit: it's advisable to put base stopper resistors on the new output transistors, so you'll also need 4 x 10 ohm resistors.
Old 21st November 2016
  #15
Gear Addict
Sorry, one more thing! From the picture here: 2SC3264 - ALLEGRO SANKEN - Bipolar (BJT) Single Transistor, NPN, 230 V, 60 MHz, 200 W, 17 A, 50 hFE | Newark element14 it looks like the transistor comes with an insulator, but it's hard to tell if it's mica or a Sil-pad. I'd give them a call to confirm because if it's mica you'll also need some thermal grease, like this one: 120-SA - WAKEFIELD SOLUTIONS - THERMAL GREASE, PACK, 4G | Newark element14 (clean well any old grease and apply a *very thin* layer of grease between heatsink and mica and between mica and transistor).
Old 28th November 2016
  #16
Gear Head
 

Waiting on parts to come in btw!
Old 28th November 2016
  #17
Gear Addict
Good, while you wait you could check a few more resistors, they are unlikely to be broken but it's better to make sure before powering it up for the measurements: R113, R114, R115 (the ones connected to the differential pair Q23-Q24), R101 (the one at the emitter of the VAS) and R117 (the feedback resistor from the output to the base of Q24).

Once you get the parts, leave the healthy channel with all the stock components and in the faulty one for now install just the new VAS (careful with the pinout, the MPS-U06 is EBC and the KS2316 is ECB). Connect a loudspeaker to the output of the good channel only, nothing plugged into the input, turn down all controls, chorus off, then power up the amp.

Needless to say be very careful with those test probes: one little slip, you short something and innocent components die. Also I'm sure you're aware that there are lethal voltages in there, so even more careful with your fingers. Before powering up, do a test run to check that all measurement points are easily accessible with good clearance between the test probes and any adjacent component leads or PCB traces, rehearse how you're going to measure. If some points are easier to access from above and some from below, it's ok to do a first set of measurements, switch off, reposition the amp, switch on again and do a second set.

First of all measure the rails to make sure the power supply is in order and it gets to the power amps, so measure from ground (e.g. the mid point of the transformer secondary, point 39 in the schematic) to points 15, 17, 32 and 34 in the schematic. Then leave it on for about 10 minutes for the heatsink temperature to stabilize, then measure the following voltages:

Faulty channel:

- Ground to loudspeaker output (point 36)
- Between the bases of the drivers Q29 and Q30 (or the collector and emitter of Q32)
- Across the following resistors: R135 (330R), R118 (330R), R144 (10k), R14 (100k)

Healthy channel:

- Ground to loudspeaker output (point 21)
- Between the emitter of Q11 and collector of Q9 (points 18 and 14)
- Between the bases of the drivers Q12 and Q13 (or the collector and emitter of Q15)
- Across the following resistors: R54 (330R), R44 (330R), R69 (10k), R38 (100K)

The measurements in the healthy channel are to have a reference for the various biases before changing anything, and the ones in the faulty channel are to make sure that everything is in order, even if the transistors measure ok with the tester it's better to confirm that they're good with the actual biases of the circuit.
Old 16th December 2016
  #18
Gear Head
 

Last final to do and then update haha also my apartment flooded yesterday! Amps safe tho!
Old 16th December 2016
  #19
Gear Addict
 
samwinston123's Avatar
 

Weirdly enough I just fixed a JC120 that also had blown both of the internal fuses but not the main AC fuse. Turned out the wires on one of the speaker cones were shorting out to the speaker frame (which was grounded). Found this out because we saw the sparks! The suspect wires had been soldered too far up on the solder tabs so there wasn't enough slack to keep them clear. Not sure if that was a factory issue or a repair someone made. Probably unrelated to your problems, but it's something to check.
Old 18th December 2016
  #20
Gear Head
 

Quote:
Originally Posted by Cabirio View Post
Good, while you wait you could check a few more resistors, they are unlikely to be broken but it's better to make sure before powering it up for the measurements: R113, R114, R115 (the ones connected to the differential pair Q23-Q24), R101 (the one at the emitter of the VAS) and R117 (the feedback resistor from the output to the base of Q24).

Once you get the parts, leave the healthy channel with all the stock components and in the faulty one for now install just the new VAS (careful with the pinout, the MPS-U06 is EBC and the KS2316 is ECB). Connect a loudspeaker to the output of the good channel only, nothing plugged into the input, turn down all controls, chorus off, then power up the amp.

Needless to say be very careful with those test probes: one little slip, you short something and innocent components die. Also I'm sure you're aware that there are lethal voltages in there, so even more careful with your fingers. Before powering up, do a test run to check that all measurement points are easily accessible with good clearance between the test probes and any adjacent component leads or PCB traces, rehearse how you're going to measure. If some points are easier to access from above and some from below, it's ok to do a first set of measurements, switch off, reposition the amp, switch on again and do a second set.

First of all measure the rails to make sure the power supply is in order and it gets to the power amps, so measure from ground (e.g. the mid point of the transformer secondary, point 39 in the schematic) to points 15, 17, 32 and 34 in the schematic. Then leave it on for about 10 minutes for the heatsink temperature to stabilize, then measure the following voltages:

Faulty channel:

- Ground to loudspeaker output (point 36)
- Between the bases of the drivers Q29 and Q30 (or the collector and emitter of Q32)
- Across the following resistors: R135 (330R), R118 (330R), R144 (10k), R14 (100k)

Healthy channel:

- Ground to loudspeaker output (point 21)
- Between the emitter of Q11 and collector of Q9 (points 18 and 14)
- Between the bases of the drivers Q12 and Q13 (or the collector and emitter of Q15)
- Across the following resistors: R54 (330R), R44 (330R), R69 (10k), R38 (100K)

The measurements in the healthy channel are to have a reference for the various biases before changing anything, and the ones in the faulty channel are to make sure that everything is in order, even if the transistors measure ok with the tester it's better to confirm that they're good with the actual biases of the circuit.

Sooooooo yeah real dumb question but how does this work(testing it without the fuses on the pcb board??) How can I test voltages with no fuses and if I put new ones in there they just pop? I am getting pretty much 0v for all of these tests. Also when hooked up to ground(point 39) my multimeter just doesn't stop beeping incessantly..
Attached Files
File Type: pdf JC-120 Testing 1.pdf (17.7 KB, 141 views) File Type: pdf JC-120 Testing.pdf (17.7 KB, 144 views)
Old 18th December 2016
  #21
Gear Addict
Oops, forgot to mention: put the fuses back. Unless when you do it, with the rest of the circuit like that (the bad output transistors removed), they do keep blowing up, in which case there's a problem elsewhere... Is that what's happening?

Edit: I reread my instructions and I didn't mention to remove the bad output transistors from the fautly channel. Sorry, my bad, I assumed you would but should have made it clear...
Old 18th December 2016
  #22
Gear Head
 

Quote:
Originally Posted by Cabirio View Post
Oops, forgot to mention: put the fuses back. Unless when you do it, with the rest of the circuit like that (the bad output transistors removed), they do keep blowing up, in which case there's a problem elsewhere... Is that what's happening?

Edit: I reread my instructions and I didn't mention to remove the bad output transistors from the fautly channel. Sorry, my bad, I assumed you would but should have made it clear...
I had them removed but put them back lol guess I misread what you said earlier. I'll try again today. Hope they don't pop again. Last two. I'll order a couple more of those as well. Thanks!
Old 18th December 2016
  #23
Gear Head
 

Interesting update! So removed the faulty output transistors, have good channel plugged into speaker cabinet, all knobs down, chorus off, good fuses in, make sure no wires touching anywhere, all good... power amp on. Terrible, loud, crazy sounds from speaker cabinet, fuses still fine but turned it off being scared I might damage a speaker(and I live in an apartment and it was crazy loud). Amp was on for about 30 seconds and fuses remained fine. Any ideas on what could be producing the noise? I do have all the news caps in that I could replace. Only thing I can see it being akin to sound wise was a really old crappy capacitor in a tube amp I had, or kind of sounded like a fender bandmaster I had when the screen grid resistors had blown. Loud, static, mixed with hum, but always changing, not just a constant sound.
Old 18th December 2016
  #24
Lives for gear
 

gbo backwards from the output transistors and test them. BTW, most transistors short from collector to emitter when the go bad.
Old 18th December 2016
  #25
Gear Addict
Huh, strange... First thing I would do: since you've pulled out and put back the transistors and maybe other components to test, double check that all solder joints are good (no spill over between adjacent ones), all transistors are in the right position, the right resistors in the right places... Bad caps it could be too, but it's strange that this would happen now if up to the point where one channel blew up, the amp was working fine, wasn't it?
Old 18th December 2016
  #26
Gear Addict
Quote:
Originally Posted by drtechno View Post
gbo backwards from the output transistors and test them. BTW, most transistors short from collector to emitter when the go bad.
Yeah, we know the output transistors in one channel are shorted, but all the other transistors tested ok. One VAS broke while removing it, but that was in the faulty channel...
Old 18th December 2016
  #27
Gear Head
 

Yes indeed it sounded fine before not functioning anymore but you could tell that it was on for sure, some noise from caps but definitely nothing 'glitchy' sounding. I'll meticulously go over every solder point and redo any that even look slightly suspect and try again.
Old 20th December 2016
  #28
Gear Head
 

Alright so turns out diode 27 has a non soldered leg. I have no idea why I would do that but hey, guess I did. I soldered that in, checked and redid about 3 spots that could've been cleaner and yes, no sound besides a tiny bit of those crackly leaky caps! Took those measurements. I took between Emitter of Q11 and collector of Q9 and I believe that was supposed to be point 13 instead of 14 on the schematic because those 2sd845 and the equivalent that I ordered's pinout is BCE. Anyways the only thing that seemed suspect to me was R14 didn't have a value that would settle at all. I convinced myself that I was crazy and tested it 4 different times at different angles and all that but yeah it would never stop except for like 1 second and that was on 0.001mv haha. Only thing that was a pain to get to really were the bases of the drivers.

Anyways, here are the measurements I got. Let me know if anything seems wonky
Attached Files
File Type: pdf JC-120 Testing2.pdf (17.9 KB, 92 views)
Old 20th December 2016
  #29
Gear Addict
I don't know how the schematic corresponds to the actual circuit, but the collector of Q9 is shown as point 14, note that in the schematic they are numbered 14-13-15 so if in the actual PCB they are ordered 13-14-15 that would be the one in the middle... Anyway, to avoid confusion, in the next test just measure directly across resistors R55 and R53 (the 0.33R ceramic ones), this is to know the quiescent current of the output stage, which is important to adjust it properly when the new output transistors are in. Other than that, all looks good in the healthy channel, except for a high-ish DC offset, but that could be just a slight mismatch between the input transistors and/or a leaky feedback cap (C28).

There's obviously something wrong in the faulty channel even without the output transistors, R114 (typo in the instructions where I called it R14 but I guess we're talking about the same 100k one, the tail of the differential input pair) should have around 33V across it, just like R38 in the healthy channel. The huge DC offset at the output also points to something wrong in the input stage. The currents through the drivers are also high (voltages across the 330R resistors) but this could be because of the input stage throwing things out of whack. The Vbe multiplier + VAS seem to be ok though, so we're narrowing things down.

So, the next step would be to investigate that suspect input stage. First of all take a good look at all the solder joints: R113, R114, R115, Q23, Q24, input resistor R105 (56k) and feedback resistor R111 (27k), follow the PCB traces connecting them to make sure there aren't any broken ones [* see note below] or any solder spill overs and make sure the transistors are in the right position.

Now, in the faulty channel, replace cap C68 with the new one, remove the driver transistors Q29 and Q30 (while you have them out you can test them again just in case) and remove the feedback and bootstrap caps (C59, C86) but don't replace them with the new ones. Short the output with a piece of wire from point 36 to ground (without drivers and output transistors all we're doing with this is grounding the base of Q24 through the feedback resistor to balance the input stage for the measurements).

In the healthy channel replace caps C28, C35 and C45. Again connect a speaker to the healthy channel, all pots down, switch on, wait about 10 minutes and then measure:

- Healthy channel: Voltage across R55 and R53 (the 0.33R ceramics) and, just to see if the new feedback cap improves the DC offset, from ground to loudspeaker output (point 21).

- Faulty channel: Voltage across R114 (100k), R113 (4.7k), R115 (4.7k), R144 (10k), C68 (10u/50V) and from ground to each of the bases of Q23 and Q24.

* Edit: PCB traces don't break spontaneously, but some times a solder pad comes off the PCB slightly and there's a (often hard to see) "hairline fracture" just at the edge of the pad where it joins the trace, that's the kind of thing to look for.
Old 21st December 2016
  #30
Gear Head
 

So first Test:
R55: 0V?
R53: 0V?
R114: 47.4V

Checked Solder Joints of:
R113, R114, R115, Q23, Q24, R105, R11: Redid them all, made sure perfect

Replaced cap 68- BTW, original wasn't fully connected when I checked. One solder joint on it had a fracture between the board.

Removed drivers Q29, Q30, caps C59, C86, shorted 36 to ground with a wire.

Replaced caps C28, C35, C45

Test 2:
R55: 3.9mV
R53: 6.8mV
R113: 17.1mV
R114: 45.6V
R115: 1.791V
R144: 100 mV
C68: 25.13V
Ground to Point 21: 32.3mV

Thats when I stopped because it seemed like with this 36 to ground setup, it was much easier to identify microphonic or noises components. It seemed strange. So I grabbed my trusty chopstick and started some gentle poking around. New caps obviously were much quieter. Visually leaky caps right after the input jacks were very noisy of course, and the main source of LOUD noise was around the R55, R53, Q15 and Q16 especially!! I'm going to check their connections next. If I poke gently on Q15, it will send the whole thing in a tailspin and I'll have to turn the amp off because its so loud, but I've also tapped once more and it stopped the noise back to dead quiet. Perhaps I've found a source of a problem?
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump