Nakamichi PA-5 — Bias pot questions

northpaw

Super Member
I have a couple of questions about bias pots on a Nakamichi PA-5 amp.

This PA-5 is in daily use, and I pull it once or twice a year to check it over and, in particular, to check and reset the bias if it has drifted. It uses single-turn 150Ω bias pots, and as is common, they are very touchy. I have cleaned them and they behave smoothly and their resistance varies smoothly, but it is just that even very small rotation has a too-large effect for my tastes. I’ve decided to replace them with a multi-turn.

First issue I ran into is that multiturn pots in 150Ω are hard to find, much less one with an appropriate configuration. But after looking at the schematic, it really doesn’t need to be 150Ω, as the bias circuit connects the wiper to one end of the track, effectively placing the “unused” part of the track in parallel with 0Ω wire (see partial schematic image below; bias pot is VR102, near the screw head symbol). I’ve checked the resistance of the “used” part of the track for each channel at their present settings, and I get 39Ω for one channel, 65Ω for the other. So it seems I can safely go with 100Ω pots, which are widely available.

bias area marked.png

First question: Is the fact that I get a sizable difference between the ohm settings on the bias pots for the two channels (39Ω for one, vs. 65Ω for the other) a problem? That is, is this within the expected range of variation for an amplifier circuit (such as due to typical sample-to-sample performance variations among the output transistors - 5 in each channel on the PA-5), or does it indicate there is some issue I should look into?

Second question: Is a Bourns 3296 series 25-turn pot a good choice for a replacement, or are there others that might be better?
https://www.mouser.com/ProductDetail/Bourns/3296W-1-101LF?qs=sGAEpiMZZMvygUB3GLcD7n4W/LGLppVQEdi8TtOM2aU=

The 3296 a cermet pot with good TC (100ppm/°) and is available in a range of configurations, but none are a perfect match for the vertical board arrangement in the PA-5. I need a pot with “face-mounted” pins in an offset pattern, and a top adjust screw. The 3296P is that, but due to its particular offset pattern, it would have to be mounted with the adjust screw pointing downward in the PA-5. The 3296W (link above) seems the best option: it has 3 in-line pins mounted on an edge and a “top adjust” on the opposing edge, and I will need to bend the pins to (1) match the offset pattern and (2) tilt the pot such that adjust screw points upward.

Anyone aware of an alternative or better pot I should consider? A 25-turn pot isn’t necessary; a 5-turn or 10-turn would be more than adequate.
 
Can't answer your more technical questions. But I have used that pot to replace an older offset pin pot. It had enough lead length to bend the center pin to fit.
 
First question: Is the fact that I get a sizable difference between the ohm settings on the bias pots for the two channels (39Ω for one, vs. 65Ω for the other) a problem?
I wouldn't worry about it, difference would be mostly due to the difference in Vbe of the bias transistors Q112, with other
channel. Maybe check Vdc at Q113base and Q114b, they should be close to equal but opposite sign, this would
rule out downstream issues.
Second question: Is a Bourns 3296 series 25-turn pot a good choice for a replacement
Probably others out there but it's the (only) one I'd use.
I need a pot with “face-mounted” pins in an offset pattern, and a top adjust screw. The 3296P is that, but due to its particular offset pattern, it would have to be mounted with the adjust screw pointing downward in the PA-5.
Your looking in the right place, need to select best option. Often the wiper and one of the legs are connected by a short/trackwork
see pic, in this case you can either feed the 3 legs through 2 holes or join 2 legs then feed them through 2 holes if you follow...
upload_2019-11-15_19-0-29.png

A 25-turn pot isn’t necessary; a 5-turn or 10-turn would be more than adequate.
Doubt there would be a significant cost difference though go with what's easily available.
 
Thanks for these insights and ideas, mbz. I'll check the voltages on the bases next time I have it open.

And yes, the wiper on the bias pot is connected to one end of the track (by a trace on the PCB, as shown in the image in the my first post), and I do follow what you are saying. So the options you mention are good alternatives. I'll sort out which option provides the best geometry, with an eye to sturdiness as the vertical orientation of the PCB means that applying a screwdriver to the adjust screw on the pot will always put some sideways force on the pins and their solder joints.

For reference, here is what the present pot looks like (labeled "151" or "1S1", and located opposite to the small transistor that is reverse mounted on the vertical PCB).

IMG_0197a.jpg
 
wouldn't worry about it, difference would be mostly due to the difference in Vbe of the bias transistors Q112, with other
channel. Maybe check Vdc at Q113base and Q114b, they should be close to equal but opposite sign, this would
rule out downstream issues.

I received the replacement bias pots today, so I pulled the unit and before replacing those, I first checked the voltages on the bases of Q113/114 and Q213/214. They were 0.81V and-0.78V, and 0.80V and -0.78V, respectively.

So they are close to equal and opposite in sign, and as suggested by mbz, all seems fine with the downstream circuit, and the differences are the normal variations in the bias transistors Thanks for this guidance, mbz.
 
Bias pot replacement went smoothly. I used Bourns 3296W (25-turn sealed pots with 3 in-line pins) and just made little bends to match the offset pattern on the board. The pins were plenty soft to bend easily, and as Scott indicated above, are long enough to permit some bending and still reach through the board..

Here is a pic of the original and new pots compared. The original is 150Ω, the Bourns are 100Ω.

P1030793a.jpg

Here is the new right channel pot installed, with a tilted up orientation so the adjust screw would be easy to get to.

P1030795a.jpg

Adjustment of bias is now very easy and precise, with lots of control provided by the much higher resolution of the 25 turns. The largest source of change now comes from the temperature effects of taking the cover off.

For checking the bias, I use a set of mini alligator cables between my DMM leads and a short cable with a JST connector on one end, which fits directly into the bias Test Point fittings on the PA-5. I made the error of not being careful enough one time, and a small patch of metal exposed beneath one of the alligator clip covers contacted the case of the PA-5 while I was taking a measurement. The result was a quick spark. No immediate damage observed, and the amp did not shut down or go into protection, but then the bias was wonky on that channel. A careful inspection of the amp board revealed about a cm of tattered trace on the PCB between TP102 and R155 (seen as a white trace instead of a dark one in the photo below). I cleaned that up, and soldered in a jumper for the repair. Everything was fine after this. I got off easy, I think.

P1030797a.jpg

Amp sounds great. Did before too, but I know that bias drift will not be due to resistance changes.
 
20200206_215442.jpg I guess this is an old thread,but anyway I just picked up one of these and recapped it I also installed a 5k pot and a 10uf cap in the spots already on the pcb. This allowed me to adjust offset to 0 mv. I tried to replace the pot your talking about and wasn't successful it wouldn't give the range required to get to 40mv only 8mv. I have a feeling the install of offset pot maybe responsible but it works with the original,odd. Did you bend the middle pin down.? I guess I got a different one to yours bur similar ohms and appears same as your choice,you see something I don't? cheers
 
Last edited:
Yes, I bent the pins relative to each other so they would match the holes in the PCB. Not at the right angles you have, but it is all the same.

I have a feeling the install of offset pot maybe responsible but it works with the original,odd.

I think it is possible this is the issue. If you do a careful comparison of this part of the schematics for the PA-5 and PA-7 (which is what you are trying to emulate), you will see that the addition of the 5KΩ pot and the extra 10µF cap are not the only changes. There are 2 resistors on each board -- R114 and R115 in the case of the L channel -- that differ. They are 10KΩ on the PA-7, but 1.8KΩ on the PA-5. This difference is clearly in compensation for the insertion of the 4.7KΩ DC offset pot, as roughly half of the resistance of that pot is in parallel with each of the R114 and R115 resistors, whereas in the PA-5 there isn't any additional resistance in parallel with the R114/R115 resistors. I haven't thought through the consequences of your change for the bias part of the circuit, but with your modification of the DC offset circuit, you have about 1KΩ on each branch instead of 1.8KΩ.

You state that "it works with the original", and if by that you mean that after the addition of the DC offset pots, the bias is adjustable in range only with the original bias pots in, not the new ones, then I don't understand. I did check the actual ohms that my original bias pots were set at, to make sure that going with a 100Ω pot was adequate; but if your original bias pots were set between 100 and 150Ω, then you would need to get a 200Ω replacement pot for it to work (these differences would be attributable to vagaries in the performance of the bias transistors, as discussed above). That is the only thing I can think of.

N.B.: If you do modify the resistors to match the PA-7 bias and offset circuit, and that is the cause of the discrepancy, remember to minimize the bias pot settings before powering up. You don't want to risk blowing your outputs.
 
M2M - As Northpaw suggested its primarily the lower value bias pot that affected your adjustment range on the bias. I would go with the 200ohm to ensure there will be more than enough range available.

I've also thought about installing the DC offset components into the board of my PA-5 as well. However to get around the issue Northpaw mentioned about installing higher value resistors at the R114/115 locations I was going to use a 20k or possibly even a 50k potentiometer. This would minimize the amount of resistance drop from paralleling those pots into the circuit. If DC offset is low to begin with then a small amount of resistance imbalance is all that will be needed to correct the DC offset. My PA-5 has about 20mV of offset on both channels once its fully warmed up, so not a huge amount.
 
Thanks for your responses. I attempted the install late last night and this time it worked. I think the test wire/pins were the issue . I can dial into the single digits and up into the 80 mV, right now at 40 mV .Still takes a while to get both almost the same even with the 25 turn pots.

I am interested in other tweaks you have performed . I see some caps and a resistor matbe on the large caps in your picture,northpaw? What are those and what did they do ? thanks
 
Those are all film bypass caps of a few different values spanning 10µF to 0.01µF.

I have not recapped my PA-5, including the large PS caps. I naturally had some concern the large ones might have deteriorated through age. I did measure them since they are easy to take out of circuit; they checked out fine, so I did not feel the need to replace them, especially given the signifiant expense for caps of that size. Nonetheless, out of concern that their ESR might now be high at higher frequencies, I added those film bypass caps. I don't know if they made any difference in the sound (nothing stood out to me either better or worse, but I don't hear much over 10KHz). There are differing opinions on the efficacy of using such bypass caps in the PS, but the arguments presented by some of the experienced techs on AK (e.g., EchoWars) for bypassing the PS caps with small film caps were sensible if not persuasive to me. And it was easy (did not even solder them, just wrapped the leads around the screw posts on the large cans) and cheap to do.
 
I can't see from pic ,is it a combination of a 10uf and a .01uf film per large cap? I didn't replace those just the amp ones ,not many to replace . I added a 220uf BP nich UES across c102/c202 its sounding very good . It had a fluttering relay on a channel . A few pics of what i did.20200207_102311.jpg 20200207_102340.jpg 20200207_102401.jpg
 
Nice work. Did you pull the boards to do the cap work? It seemed to me to be tight to get at several of the caps without getting the boards out, and I didn't relish unmounting all the output transistors on the heat sinks.

On the PS caps, I used 3 per large cap: 2 Panasonic polyester film caps, one 10µF, one 1µF, and the third is a Dayton 0.01µF. Here is a better photo of what they look like.

P1030799a.jpg
 
Amendment to the above: the photo has 2.2µF Panasonic polyester caps installed instead of the 10µF I mentioned.

Shortly after this photo was taken (late last year), I switched in the 10µF caps for the 2.2µF, which I hadn't intended to use. The 10µF caps are a fair bit bulkier than the 2.2µF ones, but still fit in space available. I don't have a photo of that, and the unit is presently in use.
 
Thanks , and for the picture too. Its very easy to do I did it as it sits be careful the tracers are fragile. I would suggest cut the old leads if possible then heat and use tweezers to remove remains. To remove the whole board just remove the center row of screws in the middle of heat sinks and the screw on standoff lifts up very easy amp to work on. Cover top of the big caps and it will sit on there, i had to address a fluttering relay. I thing just changing the cap responsible for delay 10uf c110 would have don it but i opened up relay and cleaned contacts since i had it apart. They looked perfect anyway,its worked fine since. Would you have part numbers for those caps if you don't mind? thanks
 
Watch out for the thermal white paste, i had it everywhere.:eek:
The piece of aluminum that all is bolted to comes out as one.
 
The caps I used, with the supplier part numbers, are given below. But please be aware that these are not necessarily the best caps for the job. I used what I happened to have handy.


Parts Express 027-450 Dayton Audio DFFC-0.01 0.01µF 400V

Mouser 667-ECQ-E2106JF Panasonic 10µF/250V

Mouser 667-ECQ-E2105JB Panasonic 1µF/250V
 
Thanks ,gives me a base to bring up product on Mouser, with 6 million plus parts it can be frustrating.I will do some research on the bypassing since I haven't payed much attention to the process.
 
20200131_130930.jpg Here's a few pics of it dismantled for relay issue20200131_130405.jpg [This is where i figure out you dont need to remove the heat sink just the screws in the middle row. cheersATTACH=full]1755925[/ATTACH] 20200131_132426.jpg
 
Back
Top Bottom