• The move to the new server is done. There are some software and database maintenance updates in process. This has us passing the hat around to help out. We appreciate any donations. Seriously, even a dollar helps. The payment page may be found here - https://www.audiokarma.org/support.html

MC2105 Rebuild

I did some more voltage measuring today and I think the amp still has an issue.

DC offset is 9mv in the left, and 35mv in the right.

I also tried to take a look at the bias voltage. I measured around 45mv at the Emitter of Q124 and also 45mv at the Collector of Q126. That seems correct according to the schematic, but when I measured at the Emitter of Q123 and Collector of Q125 I only measure around 16mv. This seems too low.

I swapped output boards between channels to see if the bias measurements would follow the boards, but when I powered up on the DBT the bulb stays bright. If I swap the boards back to their original location the bulb goes dim and I measure the same bias voltages as I had earlier.

I also tried with one output board installed at a time, and the only time that the DBT stays bright is when I have the output board from the channel with proper bias readings installed into the channel that originally measured low.

In other words the output board that was repaired in 1997 keeps the DBT dim when installed in either channel, but I only measure 16mv at the E and 16mv at the C in whichever channel I install it in.

The output board that looks original/untouched measures 45mv in its original channel and keeps the DBT dim, but when I install it into the channel with the mj15003 output transistors my DBT stays bright.

Any suggestions?
 
Last edited:
I pulled all of the mj15003 transistors from the previously repaired channel and checked them. They all tested fine. I also measured all of the WWs in this channel. The .33ohms all measure .5ohm, and the .56ohms all measured .7ohm. I’m not sure how accurate my fluke meter is down that low however. I also removed the 5 extra micas that were installed. I only left one mica on Q125 in this channel.

I put the amp back together and have the same results as yesterday. Bias seems low in this channel and correct in the other channel, and if I put the output board from the good channel into this channel then the DBT stays bright. I measured the Collector of Q125 while the DBT was bright and measured around 1.0v.

Does this mean that this channel is going into oscillation when I install the other channels output board? If so, is it most likely being caused by the replacement WWs that were used?
 
Last edited:
Without a osciliscope and a proper distortion analyser how can you know.......It beats me. I am uncomfortable just guessing.

It is not uncommon for a fluke to measure the resistor at that low level high.
 
Following, mine just emitted a low, loud, constant hum. I don't think it has ever been gone through. I don't know if I can do it myself, but, I'm following along for now to get a feel for things.
 
I think I need a professional for mine. Somewhere I can drive it to. I don't think I have the patience for rebuilding mine right now. But I am glad this thread is here so I can see more of what it's about. Thanks.
 
I ordered some wirewounds from McIntosh parts dept. to replace the generics that were installed in the previously repaired channel. I’m planning to install them next and then reassess the situation with the bias measurements.

They also measure .5 ohm and .7 ohm with my fluke, but hopefully they are of the correct type for use in the 2105.
index.php
 
Last edited:
You can't reliably measure resistors with such low resistance with a DMM, there is always residual resistance in the cables and even the small bit of oxidation on the leads and contribute as well.
 
I have replaced all of the generic wirewound resistors with the special McIntosh wirewounds that I received last week. I have also checked all of the mj15003 transistors using the 6-way diode test and they all appear to be fine. I removed the mica insulator from all transistors except q125 in this channel. Originally all 6 transistors had a mica installed.

I powered up on the DBT and the bulb went dim with the output boards installed in their original locations. I swapped the output boards between channels and the DBT stays bright, same as before. So the generic wirewounds weren’t causing the problem. At this point I think the output section in this channel is good.

I’m thinking my next move will be to check all 6 output transistors as well as the wirewounds in the output section of the other channel. That channel looks like it has never been touched before. It has original type output transistors and wirewounds still.
 
Last edited:
Without a scope and distortion analyzer you are asking for pure speculation.

Is your problem following a driver board? Does the amp work with both boards in place if they are in the correct location?

If either channel works with the "good" board the " bad" board has a issue.

These amps do not ALWAYS go into oscilation but they CAN which is why we continually point out they need a proper bench test for the proper repair.
 
The amp works when both output boards are installed in their original locations. The DBT goes dim and I have listened to the amp this way. If I swap the output boards between channels then the DBT stays bright.

I have also tried installing one output board at a time. One output board will keep the DBT dim when it is installed in either channel. The other output board is fine when it is installed in its original channel, but keeps the DBT bright when I install it in the channel with the mj15003's.

It’s becoming clear to me that something is wrong on the one output board. But if that output board has something wrong shouldn’t it keep the bulb bright in either channel? Perhaps I have an issue on that output board and also in the output section as well.
 
Last edited:
I will speculate that the original output section is more tolerant than the new parts output section.

We know that while the MJ15003 devices have been the go to replacement output they do have a little bit more gain, are "faster" than the originals and so can be more prone to oscilation tan the originals.......bench testing with a scope should be mandantory.

So it appears the non repaired driver board will cause the updated output to go nuts but not the original.

Have you lifted a lead and tested the bias resistor? A few ohms high can cause excessive notch or crossover distortion that is quickly observable via a scope and distortion analyzer......

The darlington pair can cause issues, the sentry monitor pair, the predrivers, the electrolytic caps, to properly restore these units you need to bench test first.

While the electrolytic caps are now 20+ years past their expected life time just assuming they cause circuit issues is shortsighted.

All the semiconductors do not have a end of life type expiration date like electrolytic caps. The bias resistors however are known to drift high.

You really need to find a test bench with the proper test gear.
 
I have not done anything at all to the driver boards. My next step was going to be to recap them, verify that all resistors were in spec, and implement the service update. That’s when I discovered this issue with swapping the output boards and decided to replace the generic wirewounds with the proper parts from McIntosh labs in order to rule them out.

I just went ahead and lifted a leg on the bias resistors. The reworked board measures 25.6 ohms, and the original board measures 26.0 ohms. So almost 20% high on both, and out of tolerance.

q109 and r135 look like the only parts that were replaced on the repaired output board. Q109 has an mpsa64 installed.

index.php
 
Tonight I pulled the output transistors from the other channel and tested them. q116, 118, 122, 124, and 126 test fine. It seems like something is wrong with q120 however.

This is what I can see with the diode function of my meter:
(B+,E-) = .559
(B+,C-) = .558
(B-,E+) = OL
(B-,C+) = OL
(E-,C+) = OL
(E+,C-) = .969 *this should also be OL.

I can also measure some resistance here.
(E+,C-) = 144k ohms

It looks to me that q120 is bad?

I also checked all of the wirewounds and they seem fine. I measured the wirewounds with all 6 output transistors pulled, but with the WWs still soldered in place. Should I lift a leg of each WW to be sure?

index.php
 
Last edited:
I’ve started going through the components on the output boards looking for any bad component. I am finding drifted resistors on both boards that should be replaced, but the same resistors are drifted on both boards. I’ve went through the transistors and diodes on the bad board and they all seem fine. I’d really hoped that I’d find something amiss here.

The only notable differences between the good output board and the bad output board is that R136 100ohm has drifted up to 110ohms on the bad board. On the good board this resistor has been replaced previously and measures 98.6ohms.

Also q109 on the good board has been replaced with MPSA64, and the original M100 is still installed in the bad board.

The bias diodes (d103/d104) show 1.136 on the bad board, and 1.121 on the good board when checking with the diode mode on my meter.

index.php
 
Last edited:
I'm inclined to say both bias diodes are high.

I measured several NOS 070-046s and NOS Motorola MZ2361s (the non-McIntosh branded/sorted) this evening and all of mine are consistent at ~ 1.1vdc. I would recommend replacing one (or both if you find that McIntosh is selling them separately). Whether it is the source of your problem / symptom I do not know. My 2105 is running MJ15003G outputs without any issues if that helps you any. The additional drop could be your test leads.

I do hope you're only rebuilding one board at a time, it will make troubleshooting much easier. Try the two series-diodes on the trouble board to see if that clears up the bright-bulb symptom (before making other changes). It doesn't have the same temperature coefficient as the original I believe, but does work for temporary and troubleshooting.
 
Appreciate the reply!

Right now I’m only lifting a leg on resistors, transistors, etc. to check for bad parts. Basically trying to find the difference between the two boards. I don’t want to rebuild the boards until I can figure out what the difference is between them as I know that will only lead to more headaches for me. So far it seem like either board should work in either channel, but obviously that’s not the case.

If I can’t locate any difference between the two output boards, then I think I will install two diodes in series in the bad output board as a test.

What about the original output that tested bad the other day, q120? Can someone confirm that it’s failed? I have never had an output test like that. Usually failed outputs have provided a clear short between junctions. I’m wondering if that output could be the reason why the DBT would stay dim when the bad output board was installed in its output section.
 
Last edited:
Of coarse no one can confirm a transistor has failed without it in hand on our own test bench, if there is a doubt, as there is a doubt, just replace all 6 and move on.

I would replace the drifted higher bias resistors and leave the bias diodes alone, might add them to a future Mac parts list if parts does have them again.
 
I’m going to work on putting an order together to replace the resistors and the capacitors on both output boards next. I’ll replace the resistors that I found drifted out of tolerance, replace the two caps on each board, and then re-evaluate. Are metal films the best choice for replacement?

I had planned on installing a fresh set of MJ15003g into both channels from the beginning. So I’ll add a set to the order so that I have them on hand.

I appreciate all of the input that you guys have given me.
 
Last edited:
You might want to add to your order a modern Mpsa14 darlington as well as the appropriate matched device (for the132- 090 and 100) as well as some low noise modern replacements for the darlington pair that you can select for matched HFE. Save the clips if you wish to track their heat together.

I just remove the old production solder tubes as DOB showed us to simplify the repair. I would struggle to get the transistor legs out of the tubes and he would just chuckle at us.
 
I’m glad that you mentioned the solder tubes as I was going to ask about removing those. Good to know that they can be removed. They sure seem like a pain to deal with!

Are ksa992f appropriate to use for the differential pair? I have lots of those on hand.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom