BA-5000 Rework

:crazy::crazy::crazy:

I continue to be mystified...

Nothing has jumped out at me yet to be the culprit in why this thing is idling hot. I haven't really tried to be precise yet in setting the bias since I am assuming that something is amiss, just getting it in a ballpark of 2 or so mv. The bias transistors appear to be doing their job since the readings immediately drop several millivolts when the cover goes on. With the bias approximately correct, I'm still getting the same 112-115° readings today.

In looking at the possibility of oscillation, I tested the output with a dummy load and a 150mv 1khz signal. Both channels were putting out just at 1 watt and around 8.2V. This is the output on the scope, which looks very clean to me:

IMG_1561.JPG

And the same output expanded to get a better look at the peak:

IMG_1562.JPG

Didn't grab a picture, but the square wave version of the same signal looked great.

Running this signal into the dummy load for 15 minutes or so at the 1 watt level slowly raised the heatsink temp up to around 124°, which is troubling.

I am still getting dc offset readings in the 60mv range, but am unclear as to whether the source of those high levels might also be at the root of the temperature issue.

Looking at the parade of outputs, I checked them all with the temp probe and found them to all be within a degree or so of each other. Next was looking at individual bias readings for the outputs (the ones I can actually get to) and found one thing that I am curious about. Assuming that the manual's specified 180ma also includes the bias current for the driver transistor, it made sense to me that individual outputs would have a correct bias current of 36ma and testing directly across several of the emitter resistors gave appropriate millivolt readings. I did find however, on the left channel, that the 2 of the NPN outputs were reading 2mv lower than the other, so I am concerned that the necessary current sharing is not happening. The way these are wired, the emitter from the driver is split into two and two on the outputs, with a 4.7ohm resistor in line to each pair. My first thought was that I had mismatched these resistors and the two pairs were getting driven at different levels but all the resistors match. I am afraid of gain mismatch on the outputs, despite all of my efforts to keep them consistent, but the fact that the millivolt readings are consistent on either side of the split from the driver makes me think that maybe there is another explanation. Also makes my head hurt, so that was where we call it a day and hope that inspiration strikes or a helpful reader can point out where I am being really dense.
 
Your idling temperature is not too far away from the heatsink idling temps I was seeing on the BA-5000 I looked at IIRC - I seem to remember something around 40 degrees C. This temperature was important to me because I had lowered the acting temperature of the fan switch, and also I was keen to see the bias stabilise and clearly demonstrate that it was in a thermal loop as regards stability, your 114 F = 45 C.
 
The bias and temperature readings have continued to be troublesome. The bias, although not changing rapidly, does not want to stabilize on either channel. The temperature continues to creep up at idle.

So, next comes the step of pulling the driver boards from my working amp and installing them in the rebuild. I have been reluctant to do this, not only because it is a bit of a pain to get the boards in and out, but also because I am very concerned about the extremely delicate leads on the 2sc984 bias transistors. But I hoped to eliminate the output devices as the culprit, so I did the swap.

With these driver boards (which have been recapped) I tweaked in the bias as it idled for a couple of hours and it holds quite steady around 180 and the heatsink temp settled in around 104°. Also, the DC offset is considerably improved over the other boards: 600mv and 1.2mv.

On the boards from the rebuild, I had replaced the dual transistor 2sa798 with a pair of KSA992's, so figuring that this is the culprit for the high offset I put 798's back in and I also double check the bias transistors, which are fine. These boards go back in and I try the bias again. Still not too stable, but better than before, and the temp still rises up to that 112-114° zone, but a much slower rise this time. Offset has improved by app. 30mv per channel.

At least I can be pretty comfortable that the outputs are fine, and that the issue is isolated to the driver boards, which is a big relief. I guess these boards will come back out and I will start going over them with a fine tooth comb and see what turns up.

And on one positive note, I did dial in the upper and lower calibrations on the power meters and they are both working fine. It's a little scary feeding 49v into those meter boards to set the upper level...
 
Try exchange the bias transistors just to see what happens.

Well, I did that, and more, and have spent countless hours this past week becoming completely flummoxed by this project...

I have finally pulled it off the bench, onto the back burner for a while, admitting at least short term defeat. Close, but no cigar. Hit the wall. Over my depth. All the cliches apply:crazy:

I had gone back to the 798 inputs, and still bias drift and too much heat, then replaced the bias transistors (NOS 984's). Same troubles, so then I tried new pairs of 992's for the input differentials, stringently matching for Vbe and hfe. Same troubles, so out with the boards again, and remove all of the new transistors (being contemporary, generally agreed upon substitutes for the originals) that had been used during refurbishing the driver boards and returned the original transistors after testing them all, including the 798's again.

Unfortunately, now, when I power it up there is an immediate high pitched whine when I throw the switch. After nearly jumping out of my skin, I manage to get it turned off. :eek::eek::eek: Okay, what new manner of hell is this? So, out go those boards, and in with the boards from my working amp. Same screeching sound. :yikes: Okay, I know that those boards were fine and the the amp dialed right in with them just a couple of procedures earlier so I put the DBT and variac back in line and the bulb burns bright. Tried all manner of bulb wattages, still glowing bright, and I don't even try to get the variac over 80%. The DBT seems to burn moderately bright up through the soft start relay clicking in and then pops up to almost full brightness.

At this point, I assume the worst and figure that I have screwed something else up during all of the R&R on the boards. Thus begins a new round of troubleshooting, starting from the mains coming in the chassis. I bypass the power switch in case it is shorting...nope. I pull the soft start relay assembly...all working fine. Check the primary on the transformer...no short. Disconnect the power from the driver boards...powers up fine, bulb dims out, voltages from the power supply boards are fine. All fuses are intact. These driver boards are known to be good, so I have to consider that maybe I have bunged up the wiring of the output terminal boards, or somehow managed to blow an output device.

This means that the heatsink assembly must come back out, since otherwise half of the wiring is completely inaccessible. Oh joy:(. This is a very disheartening process considering how difficult it is to extricate that thing. But, out it came. All the wiring looks fine, no shorts that I could discern. Well, maybe I've blown an output. Go to put the curve tracer on them, but with them still in place I can only get a correct pattern off of the driver TO-3's, not the outputs. Knowing that it is not good policy to re-use the silpads, the last thing I want to do is dismount all of these transistors. But I have to test them, so very very very carefully I pull them one at a time and very very very carefully replace them as they are tested. Since I have drivers out, just in case there is a problem using the GAS 118/119 badged versions of the 747's and 1116's, I test the 747/1116 pairs I have and pick 2 pairs to install. All of the outputs test fine, the silpads appear to be intact and all 20 devices are still electrically isolated from the heatsink, so back in it goes.

Still no joy. And to further cloud the issue, the startup symptoms are now a little different. As I bring up the variac and the DBT starts to glow, it begins to pulsate at rate of about twice a second, and when I get up to around 60-70% on the variac I start hearing a clacking noise from the chassis in time with the pulsating. First I think it is the soft-start relay, but no action there. Even though it seems to be too loud for the reed relays on the buffer board, I check those and no vibration and the noise doesn't seem to be coming from there. Then I realize that the noise is from the power meters, which are jumping in time to the pulsating. :eek2: I am only dialing the power up for 10-15 seconds at a time to hopefully prevent anything catastrophic, but even that amount of time is nerve wracking with these symptoms. When I put a meter on the speaker terminals and bring up the power again the meter is showing approximately 6vac bursts in time with the pulsating DBT and jumping level meters. When I pull the signal leads off of the driver boards the same thing happens.

At this point I am assuming that it is just possessed.

So, that's where things stand. Completely flummoxed, and certainly not too proud to be the first one to call myself an amateur moron if some astute viewer can point out some dumb thing that I am missing...
 
The only thing I can suggest is check all the PSU voltages, including the voltage at the ZD01's on the F-2518 boards - this is with driver boards installed of course, so to achieve this try backing the bias down as low as it will go, have you tried this? I have a recollection that you have been having problems with bias adjustment and range, with this in mind I am a bit suspicious of the NOS bias transistors you have been using.

The only other thing I can think of to try - needs careful consideration - is to try the 'suspect' driver boards in your working unit - on DBT (and variac?) of course.

This is a tough one. :(
 
Last edited:
I am confident that your attention to detail and tenacity will pay off for you. Don't lose the faith - you will figure this out.
 
Thanks for the responses...

bit suspicious of the NOS bias transistors you have been using.

I am bit suspicious of everything at this point...:rolleyes:

While it's possible that the bias transistors are causing some problem, the fact that the unit won't power up under the DBT with either the refurbished driver boards or the boards from the working amp would indicate that the problem is elsewhere. The boards from the working amp are back in that unit and working fine.

I have tried dialing the bias trimpots to their lowest settings, but the same DBT symptoms of too much current draw and the strange pulsing glow and bursts of AC voltage on the speaker terminals still persists.

Any thoughts on the feasibility of powering up the driver board from a bench supply? I understand that not all boards will function outside of the complete circuit of a given design.

I am taking this roadblock as an opportunity to step away from it for a while, and clean up some of the detritus of this very long term project. As someone humorously pointed out in an engine rebuilding manual I was using some years ago, "Do you really need all those tools you're standing on?" So, now is some time to tidy up and sort through the spare bits from rebuilding this unit and put my workbench back in order...

But I appreciate any input, and can hopefully get back into it soon with some fresh perspective.
 
Re: F2520 buffer board, are you positive that both legs of that diode assembly are passing voltage? If one side has failed increasing current draw will upset any number of apple carts including offset stability and potentially slow oscillation as the current draw pulls down the dc voltage as the smoothing caps fail to maintain current storage. Just had a receiver bite me this week with the same problem, its a bear to sort out as (was pointed out in a pm thanks rcs) a low voltage diode test will not duplicate full load conditions.
-Lee
 
Re: F2520 buffer board, are you positive that both legs of that diode assembly are passing voltage?

Thanks for pointing out an avenue of troubleshooting. Would the fact that the unit powers up normally, no DBT problems, and solid engagement of the soft start relay when the power leads to the driver boards are disconnected indicate that the buffer board is operational? But I will check the buffer board when I get it back up on the bench- that very long 3 ft. lift from the floor to the bench!
 
I read John's ba5000 thread again, not sure if he put that buffer board back to as schematic drawn or not. Anyway, another thought from the armchair crowd, are you sure you have not lost a ground off the driver board assemblies. If you have then trying to get offset stabilized will be a non starter as the board has no reference to earth potential. Been down that road a time or two as well.
 
with this in mind I am a bit suspicious of the NOS bias transistors you have been using.

Other than not being certain of the gain range of the NOS 984's that I have, I'm pretty confident that they are real. If they are counterfeit, and I'm not sure that the market for these would support the effort of counterfeiting, then they did a really good job of aging them down:
2sc984.JPG

As far as gain range, nothing on these or the ones that came out of the unit originally or are on the boards in my working unit seem to indicate gain range. The manual specifies (B,C), but these letters appear on none of the ones I have. While I didn't specifically look for a gain match between old and new (which I will do when I dive back into all of this) I did put the new ones on the curve tracer and they all looked good. But possibly their gain is the culprit in idling hot and idle current drift?

I read John's ba5000 thread again, not sure if he put that buffer board back to as schematic drawn or not.

He did in fact, and was successful. I attempted to put the board back to what the schematic indicates, but had no luck, and eventually returned it to the as-found state. Which did seem to work in terms of getting the amp up and running, it was just during my subsequent dealings with the hot idling that I began having this new trouble with powering the unit up. And thanks, I will do a thorough check on the grounds as I assemble my list of everyone's suggestions of items to look at.

And I also happened to notice this contemporary thread:

http://audiokarma.org/forums/index.php?threads/au-20000-power-up.767255/

where the description in #7 sounds suspiciously like the trouble that I am having with powering up :crazy:
 
Okay, 110lbs back on the bench :eek:

I initially just took a look around to see if there was some obvious mis-wiring or some such that would explain this unit's odd behavior. Everything seems to be correct, but again, with the variac and DBT, it exhibits the same pulsing on the DBT and meter jumping as before. Preparing to once again dive into the driver boards, I decided to pull the speaker output fuses and give it another try. So of course it powers right up, bulb dims out, relay snaps in and it sits there behaving itself.

This is a new development, and possibly a positive development, but raises a whole 'nother raft of questions...

Other than a short downstream of the fuses, this at least narrows things down a little bit. First I replace the fuses, and pull the leads to the meter drivers from the output posts in case there is a problem with those boards, but once again the pulsing is present. Then I pull the leg of the zoebel from the 8 ohm tap, but with the zoebel out of the circuit, the pulsing persists. Then I separate the output transformer from the circuit and it powers up normally.

Okay, maybe we're talking a bad transformer, but it appears that with either channel hooked up to an output transformer, the pulsing is present. I'm not certain what would be the official correct test for an output transformer, so I first put the meter on it and got an approximate 1.8 ohm reading between the common leg and any of the three taps. Next I hooked a variac up to it and fed it 14 volts into the common and 2ohm taps, and the 4 ohm tap put out app. 20v and the 8 ohm tap was showing app. 28v. This would indicate proper functioning with how I understand the transformer to work in the circuit.

So, the mystery continues- powers up fine with no connection to the transformer regardless of meter circuit or zebec circuit being in play, but the transformer itself seems normal.:crazy:

Any inspiration out there amongst the collective wisdom of the forum?
 
Is there any possibility that the transformer windings are wired in anti-phase on the output compared to the input? - just a thought - and not sure if this would upset things that much. But I think maybe disconnecting the transformer is too drastic a measure to yield meaningful results?

When this pulsing behaviour has come up before it has been basically a bias issue, whether because the particular unit (not a BA-5000) would not tolerate the DBT, or because the bias itself was either set too high, or there were still some damaged components present (like for example the bias control transistor) - I do not remember which it was.

Try a temporary substitute (different type number) bias transistor? to see if there is any difference, as I wouldn't be too worried about dead nuts-on bias tracking at the moment.

Sorry if any of this is a repeat - I didn't re-read the entire thread.
 
Last edited:
I have flicked through the thread quickly so I apologies if I cover something which is already covered.

1, Are the emitter resistors original or have they been replaced?

2, R736, C714, which are across the 8Ω Tap and the Negative terminals on the output Autotransformer, are these still there, are they within Spec, hiZ or OC?

3, I would be looking at the emitter resistors on the driver transistors, make sure they are in spec.

4, Go back further into the driver circuit and look into any of the resistors which are of low value that are basically there for protection purposes and make sure they are all in spec.

5, Check D01 is good and replace with a pair of 1N4148 in series if you need to.

6, I would also make sure that Zener ZD01 is also functioning, but i don't think that wouls contribute to this issue....

Thats about all I can see for a start...

Wouldn't hurt to make sure those current limiter transistors and surrounding components are good too...
 
And a close up shot of the driver board along with your part choices/designations.
 
The first thing I do to any BA amplifier is change all those damn tinned pin headers and sockets on all the boards for gold plated ones. Yes, you can get them. The tinned ones absolutely suck.

This one makes sense to me, because the thing is acting different every time you fire it up... While it is acting up, wiggle some chit.
 
Last edited:
Thanks for everyone's replies and suggestions and I'll try to address most of them in regard to where things stand.

The two primary aspects of this that have me the most flummoxed are that 1) everything was up and running initially, but with a degree of bias and thermal drift that I tried to address by reverting back to the original semiconductors on the driver boards since the rebuilt unit performed perfectly with the stock driver boards from my working unit. It was when I re-installed these reverted boards that the pulsing DBT behavior began and it wouldn't power up properly. 2) If the output transformers, zoebels and meter driver connections are all correct, why does the unit power up fine when the fuses are pulled to disconnect them all from the output?

Is there any possibility that the transformer windings are wired in anti-phase on the output compared to the input? - just a thought - and not sure if this would upset things that much. But I think maybe disconnecting the transformer is too drastic a measure to yield meaningful results?

I did check that the transformers are properly in phase, and since utilizing the 2ohm taps for speakers essentially bypasses the transformer it doesn't seem that disconnecting the transformers really differs from that situation.


whether because the particular unit (not a BA-5000) would not tolerate the DBT,

What would cause a unit to "not tolerate the DBT?"

Try a temporary substitute (different type number) bias transistor?

Will do.

Are the emitter resistors original or have they been replaced?

They are all new, plate type current sense resistors, and all tested fine when I re-pulled the heat sink assembly after things went to hell...

R736, C714, which are across the 8Ω Tap and the Negative terminals on the output Autotransformer, are these still there, are they within Spec, hiZ or OC?

Zoebels are good.

4, Go back further into the driver circuit and look into any of the resistors which are of low value that are basically there for protection purposes and make sure they are all in spec.

5, Check D01 is good and replace with a pair of 1N4148 in series if you need to.

6, I would also make sure that Zener ZD01 is also functioning, but i don't think that wouls contribute to this issue....

Once I get the transformers wired back in I will once again dive back into the driver boards.

And a close up shot of the driver board along with your part choices/designations.

Will do. And while I utilized new terminals and terminal housings (which were of course previously butchered) I left the the original headers. And while upgrading them may be a good choice, I don't know that I suspect them since the unit was operational initially. It was only after taking the driver boards back to the original semiconductors (which were all checked on the curve tracer before going back in) that this weird pulsing failure to power up reared its ugly head.

Thanks agin for all the input, and I will report any progress.
 
Back
Top Bottom