Yamaha B-2 restoration journey

Small update here. I replaced the single turns with 500 mW multi turns and it's MUCH better. Bias spot on at 40 mV and DC offset spot on at 0 mV. Verry happy.

I went from 60 mV bias to 40 mV and it sounds much better. Less boomy, more seperation between instruments. But I still wonder whether imaging can be improved. The M-2 I had spreads the instruments all across the room, but with the B-2 the instruments are very near the speakers.

IMG_4257.JPG
IMG_4258.JPG
IMG_4260.JPG

I also changed the LED's to amber and made the lamp circuit series instead of parallel. I removed the jumpers / resistors and connected the LED's in series with wires and without mutilating the board. This was possible by removing 3 lamp holder pins and use the holes isolated to put the wires through. I added one multi turn pot to have the ability to change brightness. At day light it looked pretty close to original, but at night it was much too orange.

So, back to the drawing board.

IMG_4253.JPG
IMG_4254.JPG
IMG_4255.JPG
IMG_4265.JPG
IMG_4271.JPG
 
Last edited:
Small update here. I replaced the single turns with 500 mW multi turns and it's MUCH better. Bias spot on at 40 mV and DC offset spot on at 0 mV. Verry happy.

I went from 60 mV bias to 40 mV and it sounds much better. Less boomy, more seperation between instruments. But I still wonder whether imaging can be improved. The M-2 I had spreads the instruments all across the room, but with the B-2 the instruments are very near the speakers.

Interesting claim.

Why did you originally set it at 60mV?
 
Small update here. I replaced the single turns with 500 mW multi turns and it's MUCH better. Bias spot on at 40 mV and DC offset spot on at 0 mV. Verry happy.

I went from 60 mV bias to 40 mV and it sounds much better. Less boomy, more seperation between instruments. But I still wonder whether imaging can be improved. The M-2 I had spreads the instruments all across the room, but with the B-2 the instruments are very near the speakers.

Interesting in that my experience is much in reverse - with my B-2's the speakers just disappear, where the M-2, while very good just doesn't have the magic. Maybe I'll experiment and use my second B-2 to vertically biamp one while using the M-2 on the second to compare.

I also changed the LED's to amber and made the lamp circuit series instead of parallel. I removed the jumpers / resistors and connected the LED's in series with wires and without mutilating the board. This was possible by removing 3 lamp holder pins and use the holes isolated to put the wires through. I added one multi turn pot to have the ability to change brightness. At day light it looked pretty close to original, but at night it was much too orange.

So, back to the drawing board.

You could try getting clear LED's then use a yellow dye meant for clothes like RIT. If you want a close match toss a bunch in the dye then take out two at a time over intervals so you wind up with several different shades to try. If you have a proto board handy you could set up the original bulbs along side and see which shade matches the best, day and night. I'm likely keeping the bulbs standard on mine - all my originals are still working well and I have new replacements as well.

BTW, I like that trimmer on the LED's - I hadn't thought about doing them in series to limit the current. In parallel I calculated about 3/4 of a Watt and would have used a 2-3 Watt resistor, so series is much more efficient.
 
Interesting in that my experience is much in reverse - with my B-2's the speakers just disappear, where the M-2, while very good just doesn't have the magic.
That was my experience too. A couple of years ago I traded a B-2 for an M-2 because I really enjoyed my M-4 while I had it and since then I always wanted to check out the M-2. It's a wonderful sounding amp but it just doesn't have the magic that the B-2 has and it didn't take very long before the B-2 was back place.

I also tried different bias settings with the B-2 and prefer the higher ~60 mV setting to the 40 mV setting. To my ears the sonic difference is subtle but noticeable so that's what I set for and leave it at. To each their own ... :)
 
but it just doesn't have the magic
Do you have a B-1 as well?
Any comparisons on sound and soundstaging between the two?
All the comparisons I've read suggest the B-1 sounds better, but in which respects I've wondered.

Small update here. I replaced the single turns with 500 mW multi turns and it's MUCH better. Bias spot on at 40 mV and DC offset spot on at 0 mV. Verry happy.
That's been my experience too, multiturns are far better / easier for setting bias accurately.
 
Interesting claim.

Why did you originally set it at 60mV?

After reading a lot opinions / experiences about it.

Interesting in that my experience is much in reverse - with my B-2's the speakers just disappear, where the M-2, while very good just doesn't have the magic. Maybe I'll experiment and use my second B-2 to vertically biamp one while using the M-2 on the second to compare.

You could try getting clear LED's then use a yellow dye meant for clothes like RIT. If you want a close match toss a bunch in the dye then take out two at a time over intervals so you wind up with several different shades to try. If you have a proto board handy you could set up the original bulbs along side and see which shade matches the best, day and night. I'm likely keeping the bulbs standard on mine - all my originals are still working well and I have new replacements as well.

BTW, I like that trimmer on the LED's - I hadn't thought about doing them in series to limit the current. In parallel I calculated about 3/4 of a Watt and would have used a 2-3 Watt resistor, so series is much more efficient.

Yeah I could try that, but series won't happen with 4 white LED's and 12V supply. They have 3.2V forward voltage.

And what is RIT?

That was my experience too. A couple of years ago I traded a B-2 for an M-2 because I really enjoyed my M-4 while I had it and since then I always wanted to check out the M-2. It's a wonderful sounding amp but it just doesn't have the magic that the B-2 has and it didn't take very long before the B-2 was back place.

I also tried different bias settings with the B-2 and prefer the higher ~60 mV setting to the 40 mV setting. To my ears the sonic difference is subtle but noticeable so that's what I set for and leave it at. To each their own ... :)

True, the B-2 has more magic than the M-2 and sounds definitely more warm. Btw there's also a preamp and speakers in the chain, so the magic have to be made with all components. Maybe the C-4 is not fit for the B-2.
 
Yeah I could try that, but series won't happen with 4 white LED's and 12V supply. They have 3.2V forward voltage.

And what is RIT?

I didn't think about series summed voltages exceeding the supply Voltage. I keep tripping on fundamentals... grrrr...

I made an assumption about series since in parallel the pull up resistor would calculate to 120 Ohms at .645W, so you would likely want a 1 or 2 Watt resistor as a safety factor and that trimmer is 0.25W. If they are in a 2x2 array then it's 150 Ohms with a 0.5 or 1 Watt also as a safety factor. That Bournes 3266 is 0.25W, so it is just above the calculated value of 0.21W. Now I'm going to make the assumption you are using a 2x2 array (no time to sketch from your photos) - it that trimmer getting hot?


Calculations:
4 LEDS in parallel:
E = 12V - 3.2V = 8.8V
I = 20 mA * 4 = 80mA = 0.08 A
Since R=E/I the resistor would be 8.8V/0.08A = 110 Ohms, so I would go with the next highest standard value of 120 Ohms.
Power is E^2/R, so (8.8*8.8)/120 = .645W, so a minimum of 1W and 2W preferred.
So the pull up resistor would be 120 Ohms 1 or 2 Watts.

2X2 Array:
E = 12V - (2*3.2V) = 5.6V
I = 20mA * 2 = 40mA = 0.04mA
Since R=E/I the resistor would be 5.6V/0.04A = 140 Ohms, so I would go with the next highest standard value of 150 Ohms.
Power is E^2/R, so (5.6*5.6)/150 = .21W, so a minimum of 0.5W and 1W preferred as a safety factor.
So the pull up resistor would be 150 Ohms 0.5 or 1 Watt.

RIT is a common fabric dye sold here in the US and in almost every store. It can be used to dye all sorts of plastics as well and is very stable after it dries. IIRC you are located in France and I'm certain you have an equivalent product.

Here is there web page: https://www.ritstudio.com/
 
I didn't think about series summed voltages exceeding the supply Voltage. I keep tripping on fundamentals... grrrr...

I made an assumption about series since in parallel the pull up resistor would calculate to 120 Ohms at .645W, so you would likely want a 1 or 2 Watt resistor as a safety factor and that trimmer is 0.25W. If they are in a 2x2 array then it's 150 Ohms with a 0.5 or 1 Watt also as a safety factor. That Bournes 3266 is 0.25W, so it is just above the calculated value of 0.21W. Now I'm going to make the assumption you are using a 2x2 array (no time to sketch from your photos) - it that trimmer getting hot?


Calculations:
4 LEDS in parallel:
E = 12V - 3.2V = 8.8V
I = 20 mA * 4 = 80mA = 0.08 A
Since R=E/I the resistor would be 8.8V/0.08A = 110 Ohms, so I would go with the next highest standard value of 120 Ohms.
Power is E^2/R, so (8.8*8.8)/120 = .645W, so a minimum of 1W and 2W preferred.
So the pull up resistor would be 120 Ohms 1 or 2 Watts.

2X2 Array:
E = 12V - (2*3.2V) = 5.6V
I = 20mA * 2 = 40mA = 0.04mA
Since R=E/I the resistor would be 5.6V/0.04A = 140 Ohms, so I would go with the next highest standard value of 150 Ohms.
Power is E^2/R, so (5.6*5.6)/150 = .21W, so a minimum of 0.5W and 1W preferred as a safety factor.
So the pull up resistor would be 150 Ohms 0.5 or 1 Watt.

RIT is a common fabric dye sold here in the US and in almost every store. It can be used to dye all sorts of plastics as well and is very stable after it dries. IIRC you are located in France and I'm certain you have an equivalent product.

Here is there web page: https://www.ritstudio.com/

France?!

The Netherlands :p

I aways use this calculator: http://led.linear1.org/led.wiz
 
This guy reckons he bought recently and not fake. Hard to imagine.

Maybe they are fake, but fake AND working?

I have the source, if not too expensive i will buy a set and then find some one to verify them.

11r6ups.png
 
Not fake
the 8 down on the right are coming from a B2 complete
Fake a not so well marked !
SONY trs look originals
 
I just got around to pulling the four power caps out of the second B2 I own and was surprised to find that all the caps measured very low at only about 50% or slightly less of their rated value of 18,000uF. I am leaning toward purchasing four Nichicon KG’s as replacements. Any evolution of thoughts if these are the best replacement caps of readily available options?

Also, it was mentioned earlier that the Nichicon caps are rated for a temp tolerance of only 85C and this might be an issue. I have another B2 hooked up to my primary system that I removed the cover so I could easily take laser temp readings. After I had it playing for a couple of hours the highest temp reading I could find was 65C and this was when my laser was right on one of the VFETs. The heat sinks that the VFETs are connected to were mostly in the 52C to 62C range depending on where I took the reading. But the area where the power caps are housed is relatively cool at no more than 42C. Given these measurements is the 85C temp rating of the Nichicons really an issue?
 
Last edited:
I just got around to pulling the four power caps out of the second B2 I own and was surprised to find that all the caps measured very low at only about 50% or slightly less of their rated value of 18,000uF. I am leaning toward purchasing four Nichicon KG’s as replacements. Any evolution of thoughts if these are the best replacement caps of readily available options?

Also, it was mentioned earlier that the Nichicon caps are rated for a temp tolerance of only 85C and this might be an issue. I have another B2 hooked up to my primary system that I removed the cover so I could easily take laser temp readings. After I had it playing for a couple of hours the highest temp reading I could find was 65C and this was when my laser was right on one of the VFETs. The heat sinks that the VFETs are connected to were mostly in the 52C to 62C range depending on where I took the reading. But the area where the power caps are housed is relatively cool at no more than 42C. Given these measurements is the 85C temp rating of the Nichicons really an issue?
I have yet to see a B-2 where the original PSU big cans that are not way out of spec or tottaly shot.
Move some of those power resistors on the PSU board circuit side and you should be OK.
 
Mine all measured 18 or 19 mF and had low ESR, but I replaced them anyway. I would go for the 22 mF KG's and not the CDE's like I did. I will swap them in the end for KG's. And what rottalpha said indeed, move the large 3 Watt resistors to the underside of the pcb.
 
Back
Top Bottom