BA-5000 Rework

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

Only because of the circuit design possibly not being happy with lowered voltages, in other words, a bulb wattage that will adequately protect the unit won't allow it to power up properly. I didn't find this with the one-and-only one that I looked at, but mine wasn't a basket case, and just needed some 'tweaking'.
 
Okay so you've replaced the emitter resistors, so what exactly are they? What I know about the McIntosh Autotransformer amplifiers of which I have worked on a few, they are very sensitive to the type of emitter resistor used. They must be inductive wirewound, if you use anything else the amplifier will likely oscillate and go unstable, it does happen, I have seen it.
This is for the earlier MC2505,2100, 2125, 2205 etc etc....From what I can see this pretty much what the BA5000 design is based on, so I would think the same rules apply here....
Uncanny how similar the design is, right down to the current limiting circuit....
 
Okay so you've replaced the emitter resistors, so what exactly are they?

The KOA BPR low inductance plate type current sense resistors pointed out in this post by Echowars:

http://www.diyaudio.com/forums/parts/269581-heads-up-koa-radial-emitter-resistors.html

Assuming any hypersensitivity of this design to the emitter resistor type, I would still be confused by the fact that the output section was operating initially with no apparent indication of oscillation (see post #102) and just a slight bias drift. If it comes down to these resistors being the culprit, so be it, but it would truly be unfortunate to have to do a wholesale replacement of all 16 of those very difficult to access resistors. Barring any inspiration that solve this, I will probably re-re-build the driver boards and see where that takes me next before I do anything as drastic as digging the output section more than I already have...
 
The distinction seems to be very fine - at least according to the linked document about Current Sense resistors and how they work. To be honest I am puzzled by the name - all resistors are by definition 'current sensing' if you are referring to a voltage drop. :dunno:

http://www.mouser.com/pdfDocs/Ohmite_CSResistors.pdf

However, I would be extremely wary of ultra low inductance resistors being used in conjunction with high speed transistors - where ordinary so called non-inductive resistors were used before.
 
Last edited:
However, I would be extremely wary of ultra low inductance resistors being used in conjunction with high speed transistors - where distinctly inductive resistors were used before.

Can we assume that the original Micron cement wirewounds are inductive? I assumed that they were non-inductive. And in a .33 ohm resistor what would be considered a low or high value of inductance? I will put some of the originals and some of the BPR plate resistors on the LCR meter and see what kind of readings I get...
 
Put it this way, and I can only comment with regards to older Solid state McIntosh amplifiers with Autotransformers, but I wouldn't be using those particular resistors.
They need to be wirewound.
As this amplifier seems to be very close in design, I feel that some of the same rules may apply.

Well, let us know how you go......Very curious as to what this issue may be...
 
I will put some of the originals and some of the BPR plate resistors on the LCR meter and see what kind of readings I get...

That would be interesting, I was under the impression that the Micron resistors were not low inductance, but even if they are low inductance (bifilar wound?) they won't be as low as the plate resistors - which now I think about it, might have a small but perhaps significant capacitive coupling effect?
 
Some quick test results:

Inductance at 1khz-

Original Micron 5w .33ohm- .6uH

BPR plate resistor 5w .33ohm- .3uH

So I would be curious about a couple of things. Would these measurements be considered low or high inductance in this circumstance? In other words, is the .6uH resistor low enough that it would indicate a bifilar type, or would the .6uH be considered a level of inductance typical of a standard wire wound and the plate type is low enough to be consider non-inductive? And could this difference in inductance between the two styles be significant in this circumstance?
 
This thread niggles at me,
R701 10 ohm 2watt and R705, R707 4.7ohm 1 watt new and correct value (same for other channel)
Ground verified from input jack through driver card to chassis ground? Many of the AU series amps used this grounding scheme and things go rapidly wrong if there is a break along the way.
 
Yes, all components on the heatsink/output board assembly are new. I rechecked values when I pulled the assembly back out (post #106) but all were fine. The 10ohm is the non-inductive Ohmite WN wire wound (http://www.mouser.com/ds/2/303/res_wh_wn-180230.pdf) and the 4.7 ohm is metal oxide.

I have checked grounds throughout, but will do again before I pull the driver boards yet again...
 
I am not going to a new thread but one of my BA-5000 bit the dust after a few years in hibernation. I recapped the board and found some rather questionable attempt at the 2520 board.

Darn thing does not come out of protection even after full recapped on all boards (except the reserviour caps) which seems to be OK. I will replace them too.

It gone into protect mode while idle without no burning up anything. No fuses blown before. But F-601 is blown after full recapped job. Needless to say, plenty of leaky caps on 2520 boards. What baffle me is that D606 has one leg cut and some monkey job added to circumvent the problem at hands with extras resistors and some traces were cut and new wires added to reroute for the monkey job. But it was working before so it was probably a temporary fix. So, instead of recapping and installing a proper D606, some monkey does what he did best. I think whatever was done needs to be undone first.

May be we can help each other out during this ordeal. Stay tuned and I will try to understand what you are dealing while attempting to troubleshoot mine at the same time.
 
Well, reading the thread slowly from Page1, I found your buffer board has the same monkey style job as in mine. I wonder whom really worked on them?

Was that the technical update from Sansui Tech bulletin back then? Cut one leg on D606, added resistors for mid point on the RL602, rewire the boards, cut traces; etc.

But not good. I am going to revert mine to original circuit.
 
I pulled my F-2520 to give a thorough check on all components and all traces again. The traces should be checked for high intermittent resistance near the wirewound power resistors which give some burnt on the PCB and substrate. I have replaced these 270ohm and 560ohm resistors with the new ones and mounted high from the PCB substrate to allow better air circulation. It's also probably safer to add a new trace wire even though they are thick traces.

I also have cleaned all the pins and sockets on all boards and connectors before. I pulled each one connector and carefully scraped off dust with a small sharp hobby knife. Looks like new.

The F-2520 board has the same mods on mine, Hype and you so it must be some kind of factory mod added or all of ours were worked by the same factory service Tech before.
 
I just fixed mine keeping the modified circuit which appears to be done by the Sansui factory service. I was a monkey who didn't realize what I did before and I am the reason why the F-601 fuse is blowing. Besides, getting old makes me a fool and being lazy to snap a few pictures before what I did makes this stupid mistake and cost me a day and puts unnecessary rework on my part. So, I pulled the board out last night and circuit is traced and compared what was being done by the mod and the original schematic.

OP, from your pictures, I think you may have made the same mistake as me by Connection the Green Jumper lead from the anode of the 1S1850 to the ground bus which prompts an immediate blowing of F-601. In fact, this wire needs to go to the +ve leg of 63V 4.7uF cap which will supply to the collector of the TR605 through the added 27K resistor at the back side of the PCB which would makes sense.

I think the ground bus is jumped /combined for L & R channel and also added two 56K resistors to both side of the input of the (may be to create a better common mode rejection and tame the ground noise when used in the mono mode).

For now, I am going to keep mine in this modified circuit and I think this circuit is done by Sansui to address the sonics issues rather than reliability issues.

Now, I am going to replace the power filter caps which is a challenge to get out and replace.

OP, I will keep checking your thread. If you need me to take pictures of the wires / connections and something else, let me know. I will help anyway I can.
 
What baffle me is that D606 has one leg cut and some monkey job added to circumvent the problem at hands with extras resistors and some traces were cut and new wires added to reroute for the monkey job.

Thanks for chiming in on my problem child...

I have not really had a chance to deal with mine much lately, but did want to respond. Yes, I also think that this was a mod from the factory. Both my basket case and my parts carcass had the exact same surgery applied to the 2520 board, while my working unit has the slightly different version of the board, as it seems most other BA5k's were produced. Unlike Hyperion's experience, I was unable to get my 2520 board to work with the mods removed and the board taken back to match the schematic; I ultimately kept it as found, which seemed to work okay- it has been other issues that have me grasping at straws.

At some point I will check serial numbers and see if possibly these "monkey style jobs" are early or late in production.
 
In regard to serial numbers and the quite possibly factory mod to the 2520 Buffer board, the working unit that does not have the mod was built in March of '76, and the rebuilt unit was built in August '75. The parts carcass I have did not include a back panel, so no serial number. So maybe they were altering the early units before they shipped them out?

In regard to the problem child, it powered up today when I re-installed the driver boards from my working unit. Of course, these are the same boards that I tried immediately after this phenomenon of pulsing dbt and refusing to power up occurred with the rebuilt set of boards. :dunno:

I had re-intalled the working unit boards yesterday as a last ditch effort before I started tearing things apart again, and a 100W bulb would not dim out. I walked away from it, and tried it again today with a 500W bulb and it powered right up.

That in itself is a relief, but I still have to try to track down the issue with the rebuilt boards, and will start by installing the NOS 984 bias transistors on the working boards and see what happens...

More sleuthing in my future, but at least I can rest easy knowing that I'm not likely to have to start tearing apart the heatsink and connector boards.
 
In regard to serial numbers and the quite possibly factory mod to the 2520 Buffer board, the working unit that does not have the mod was built in March of '76, and the rebuilt unit was built in August '75. The parts carcass I have did not include a back panel, so no serial number. So maybe they were altering the early units before they shipped them out?

In regard to the problem child, it powered up today when I re-installed the driver boards from my working unit. Of course, these are the same boards that I tried immediately after this phenomenon of pulsing dbt and refusing to power up occurred with the rebuilt set of boards. :dunno:

I had re-intalled the working unit boards yesterday as a last ditch effort before I started tearing things apart again, and a 100W bulb would not dim out. I walked away from it, and tried it again today with a 500W bulb and it powered right up.

That in itself is a relief, but I still have to try to track down the issue with the rebuilt boards, and will start by installing the NOS 984 bias transistors on the working boards and see what happens...

More sleuthing in my future, but at least I can rest easy knowing that I'm not likely to have to start tearing apart the heatsink and connector boards.

Sorry I have been away for a few days to finish recapping my units. The Power Filters are not easy to replace and a hell of a job to get in and out.

I can't tell you where to look because you seem to have looked and replaced almost every active parts and some passive too. But here is what I do. When all else fails, I take each board and really clean the solder side with alcohol, windex, and rinse with purified water and immediately blow dry the board. When you clean the solder side with the denatured alcohol, really make sure you got all the nasty flux and sticky gunk out with an old tooth brush and soldering hand tools / picks. I then looked the board and components under the led magnifying lens very carefully. I usually found many problems when everything is clean which otherwise wouldn't be normally visible.

I am usually over confident enough not to use a DBT (blows a fuse more than a few times thanks to that ego) with my works so I wouldn't know what wattage bulb it takes to get the unit up and running. But I knew the inrush current is high enough to draw more than a few hundred watts at startup.

Good Luck with the problem child!.
 
If it takes a 500w bulb then perhaps quiescent bias is way too high.

Yes, one might think so, although I had the idle current ammeters in circuit when I was successful today and both channels were initially reading just over 100ma. Even accounting for the typical upward drift as it warms up, that's pretty safely under the 180ma spec.

It's quite possible that the amp may have fired up just by stepping up to a 150w lamp, but having been round and round with thing I was frustrated enough to jump up to the 500w photoflood. No hesitation, just dimmed right out and and the inrush relay clicked over. Lucy, you got some 'splainin' to do....

I usually found many problems when everything is clean which otherwise wouldn't be normally visible.

I will certainly be going over the rebuilt driver boards with a fine tooth comb and have to assume that something will show up to 'splain all of this curious behavior.
 
The problem child continues to perplex.

As noted, this unit has now been willing to power up with the driver boards from my working unit. No explanation, no change from the circumstances under which the unit previously refused to power up with these same boards after it started exhibiting the pulsing behavior on the dbt with either the working unit boards or the rebuilt boards (which had been reverted back to their original semiconductors).

Next, to confirm that the NOS 2sc984's I was using were suitable, I swapped them onto the working unit boards and the bias dialed in fine. Then, going back to the rebuilt boards, I once again built them up with the same new (commonly accepted substitute) transistors as before with the exception of keeping the original TR04 2sa818 associated with the 984 bias transistor, thinking that since my initial issue was in the bias setting possibly the previous substitute was not well suited in this position.

Also, in an effort to reduce the high DC offset readings that were present with the first testing with these rebuilt boards, prior to the crazy dbt performance, I have attempted a new round of scrupulously matching pairs ksa992's on the curve tracer. I happened to have a number of 992's marked as "fc" -not a gain grouping currently in the datasheet, with hfe readings around 550, 100 points higher than the batch of "fa" group that I used before. Remembering some reference to higher gain being good in this instance, I matched my 2 pairs from these "fc" devices.

Today, these re-re-rebuilt boards went back in the amp, and with only the briefest of glowing from the 500 watt bulb, it was up and running. I truly have no idea what might have been the source of all of the previous bad behavior. It's almost like wiggling the rabbit ears on the televions of yesteryear- things just start working. Did not take the time today to properly fine tune a hopefully stable bias setting, but was very happy to see that the 60mv offset readings from when things were previously running have now come down to 1.2 and 6mv. The newly matched 992 pairs seem to be the ticket.

Now, keeping my fingers crossed that a stable bias setting can be maintained along with reasonable temps on the outputs...
 
Back
Top Bottom