Uncle Paul's B-2 Rehab

Uncle Paul

Super Member
Mr. Yamaha's B-2 restoration thread has been both informative and inspiring. As a result my first B-2 has been stripped down to the chassis and is undergoing rehabilitation from the ground up. Rather than dogpile on that thread I am starting my own to help keep that thread focused on his restoration effort.

Coming on six years ago I stumbled across a B-2 that had just been listed locally on Craigslist. I set up an appointment to audition it that afternoon and it was love at first listen. Unfortunately, after I drove it home I found that one channel was out. Fearing the worst I went to remove and test the VFETS only to find that one of the connector boards between driver and power FET board had snapped in half. Fortunately the VFETS were undamaged and after I replicated the connector board all was well

Note: My B-2's are far from museum pieces, but they aren't overly rough, either. I'm not trying to do a historical restoration, but rather the goal is reliability, sonics, and above all protect those VFETS! All original parts, however, are being documented and saved if they will be needed for some future reason.

The electrical portion of the rehab starts as always with the B-2's heart - it's power supply. I've made some capacitor changes based on insights from Mr. Yamaha's thread. While I'm waiting on some parts I started work on the Electrolytic Capacitor board and that brought up a couple of points I'd like to get some feedback on.

When I planned some of this rehab a few years ago I made a few decisions that I am rethinking about now. The first involves the main capacitors. At the time I decided to go with Mundorf M-Lytic 2200 uF 63V caps over the more typical Nichicon KG, but today would probably go with the Nichicon KG's and keep it simple.

They have been used in B-2's before
upload_2017-3-17_19-3-6.png

But when I look at them side by side with the original Nichicons it gives me doubts that they are truly electrically equivalent.
upload_2017-3-17_19-11-8.png

My first question is given that I have the Mundorf's in hand is there a compelling reason to sell them and get the Nichicon? I can say that if I had to do it all over again I'd have gotten the Nichicon.

My second concern is that I also bought 30A 600V Fairchild ISL9R3060G2 "Stealth: diodes AKA FREDs to replace the main PS diodes. These aren't expensive and it isn't a huge loss if I decide to stick with the factory diode bridges, but they are more efficient and less noisy, so should be an improvement. I was working on the Electrolytic Capacitor board this afternoon and decided to try to see how best to mount the FREDs and finally came up with the following:
upload_2017-3-17_19-20-24.png

This is just a test fit and function check even though they are soldered in. The complicating factor is that the anode of the FRED on the left does not go to a solder pad, it's just a convenient hole in the PCB that was there to accommodate the frame of the original Toshiba bridge, and there is no pad to solder to. It turns out that the cathode is enough to provide solid enough support. The question here again is that is there a compelling reason to stick with the original bridges over the FREDs?

In any event, they and the remaining bridges are coming out to accommodate a proper cleaning since Mr. Yamaha has raised the standard for PCB cosmetics. Just make sure when you go to remove the filter caps from this board to have plenty of braid on hand:
upload_2017-3-17_19-44-34.png
 
Last edited:
Mr. Yamaha's B-2 restoration thread has been both informative and inspiring. As a result my first B-2 has been stripped down to the chassis and is undergoing rehabilitation from the ground up. Rather than dogpile on that thread I am starting my own to help keep that thread focused on his restoration effort.

Coming on six years ago I stumbled across a B-2 that had just been listed locally on Craigslist. I set up an appointment to audition it that afternoon and it was love at first listen. Unfortunately, after I drove it home I found that one channel was out. Fearing the worst I went to remove and test the VFETS only to find that one of the connector boards between driver and power FET board had snapped in half. Fortunately the VFETS were undamaged and after I replicated the connector board all was well

Note: My B-2's are far from museum pieces, but they aren't overly rough, either. I'm not trying to do a historical restoration, but rather the goal is reliability, sonics, and above all protect those VFETS! All original parts, however, are being documented and saved if they will be needed for some future reason.

The electrical portion of the rehab starts as always with the B-2's heart - it's power supply. I've made some capacitor changes based on insights from Mr. Yamaha's thread. While I'm waiting on some parts I started work on the Electrolytic Capacitor board and that brought up a couple of points I'd like to get some feedback on.

When I planned some of this rehab a few years ago I made a few decisions that I am rethinking about now. The first involves the main capacitors. At the time I decided to go with Mundorf M-Lytic 2200 uF 63V caps over the more typical Panasonic KG, but today would probably go with the Panasonic KG's and keep it simple.

They have been used in B-2's before
View attachment 890416

But when I look at them side by side with the original Nichicons it gives me doubts that they are truly electrically equivalent.
View attachment 890425

My first question is given that I have the Mundorf's in hand is there a compelling reason to sell them and get the Panasonics? I can say that if I had to do it all over again I'd have gotten the Panasonics.

My second concern is that I also bought 30A 600V Fairchild ISL9R3060G2 "Stealth: diodes AKA FREDs to replace the main PS diodes. These aren't expensive and it isn't a huge loss if I decide to stick with the factory diode bridges, but they are more efficient and less noisy, so should be an improvement. I was working on the Electrolytic Capacitor board this afternoon and decided to try to see how best to mount the FREDs and finally came up with the following:
View attachment 890433

This is just a test fit and function check even though they are soldered in. The complicating factor is that the anode of the FRED on the left does not go to a solder pad, it's just a convenient hole in the PCB that was there to accommodate the frame of the original Toshiba bridge, and there is no pad to solder to. It turns out that the cathode is enough to provide solid enough support. The question here again is that is there a compelling reason to stick with the original bridges over the FREDs?

In any event, they and the remaining bridges are coming out to accommodate a proper cleaning since Mr. Yamaha has raised the standard for PCB cosmetics. Just make sure when you go to remove the filter caps from this board to have plenty of braid on hand:
View attachment 890444
Good luck Uncle Paul!

Try using a sucker, you will never look back. Braid is oke for final cleanup, but is a waist of time otherwise...Plus the risk of damaging the pads is higher.

Make her purrty!
 
Good luck Uncle Paul!

Try using a sucker, you will never look back. Braid is oke for final cleanup, but is a waist of time otherwise...Plus the risk of damaging the pads is higher.

Make her purrty!

I have a Hakko 808, and it makes short work of most components, but the size of the pads for the big caps made it impractical. Instead, I heated the solder and literally drug the braid through it. It removed the majority in just a few seconds.
 
I have a Hakko 808, and it makes short work of most components, but the size of the pads for the big caps made it impractical. Instead, I heated the solder and literally drug the braid through it. It removed the majority in just a few seconds.
I am surprised to hear that..I have an ayoue rework station...I think 203...It only takes a few seconds to get rid of those blobs. Once everything is out, I sometimes go back with the wick to purdy em up
 
When I planned some of this rehab a few years ago I made a few decisions that I am rethinking about now. The first involves the main capacitors. At the time I decided to go with Mundorf M-Lytic 2200 uF 63V caps over the more typical Nichicon KG, but today would probably go with the Nichicon KG's and keep it simple.

They have been used in B-2's before
View attachment 890416

But when I look at them side by side with the original Nichicons it gives me doubts that they are truly electrically equivalent.
View attachment 890425

My first question is given that I have the Mundorf's in hand is there a compelling reason to sell them and get the Nichicon? I can say that if I had to do it all over again I'd have gotten the Nichicon.

Well, it's nice to know there are other options for those caps. I didn't realize Mundorf made those.

Mundorf makes high quality stuff. I use their polyprophylene caps (Mcap Supreme) in crossovers, and they're superb.

Nichicon is probably better known for their electrolytics, but if I had the Mundorfs already, I certainly wouldn't go and buy different ones. And after cramming Nichicon Goldtunes in there, smaller is just fine with me!
 
Last edited:
At the time I decided to go with Mundorf M-Lytic 2200 uF 63V caps over the more typical Nichicon KG, but today would probably go with the Nichicon KG's and keep it simple.

But when I look at them side by side with the original Nichicons it gives me doubts that they are truly electrically equivalent.

My first question is given that I have the Mundorf's in hand is there a compelling reason to sell them and get the Nichicon?

Size has no bearing on the quality of a capacitor in my experience, manufacturing technologies and etching processes have progressed vastly over the years.

The spec sheets on the other hand do, so I think it's important to check ripple current values, along with other things like ESR, DF, etc depending on the application. These things have no correlation to size in my experience, but they do definitely vary between brands/models.

The other thing I've found is that the actual (measured) capacitance of new capacitors is consistently lower than spec by as much as 10-15%, and while this is within the +/-20% spec, it's mildly annoying. Nichicon is definitely guilty here (Chemicon and Panasonic too, in liquid electrolytics). I still think those brands make good capacitors, but I've actually started buying higher capacitances in some of their product lines so that I can get something that's +5% of intended value, rather than -15%.

In any case I don't think you've made a mistake as long as the Mundorfs have equal or better specs than the originals. Have you measured the capacitance out of curiosity?
 
Last edited:
Size has no bearing on the quality of a capacitor in my experience, manufacturing technologies and etching processes have progressed vastly over the years.

The spec sheets on the other hand do, so I think it's important to check ripple current values, along with other things like ESR, DF, etc depending on the application. These things have no correlation to size in my experience, but they do definitely vary between brands/models.

The other thing I've found is that the actual (measured) capacitance of new capacitors is consistently lower than spec by as much as 10-15%, and while this is within the +/-20% spec, it's mildly annoying. Nichicon is definitely guilty here (Chemicon and Panasonic too, in liquid electrolytics). I still think those brands make good capacitors, but I've actually started buying higher capacitances in some of their product lines so that I can get something that's +5% of intended value, rather than -15%.

In any case I don't think you've made a mistake as long as the Mundorfs have equal or better specs than the originals. Have you measured the capacitance out of curiosity?

I can't believe I didn't go to the spec sheets first
upload_2017-3-18_11-19-19.png

I expect modern caps to be smaller, but when you hold the two the size of the Mundorf just seem wrong - it's like a house that's 3,000 sq feet in the inside and 1,500 on the outside.

Here is a comparison from what I could find. I was surprised that Nichicon didn't didn't list ESR and simply gave a formula to calculate leakage. It looks like Mundorf has the better specs, so since I have them on hand, Mundorf it is. When I'm finished with this B-2 I'm going to compare the "Fred Mundorf" PSU to the original in my second B-2.
upload_2017-3-18_11-13-8.png

http://www.nichicon.co.jp/english/products/pdf/e-kg.pdf
http://www.mundorf.com/english 1.1/Broschuere Einzelseiten/MLGO.pdf
 
Nice work.
I think ripple current is the more important spec for PSU caps.
For what it's worth, another guy on here "MiamiBoy" used those M-Lytic caps in his M-40 and was pleased with the results so I can't imagine them being bad.

If you have a chance please do measure their actual capacitance (and ESR if you have a tester), I'd be curious what the results are.
The "actual" capacitance is something the datasheets don't help with, so I've been relying on my tester to get some practical datapoints.
 
I don't want to disturb your nice project, but the capacitor hook-up with additional wires between main capacitors and their PCB is bad practice: you add inductive elements in the system which may cause nasty harmonics/resonances at higher frequencies (beyond 50 kHz, but the closed loop bandwidth is around 150 kHz, so it counts). Idem ditto for the additional wiring around the diodes. The noise (as far as relevant) of the original diodes can be suppressed with 5~10nF film caps (>100V rated).
Keeping diodes and power caps very tide and very close together with as little as possible inductive elements is a basic design rule for any power amp.

The 'bridge' of the original diodes acts like a heatsink, also directing heat to the large PCB tracks for more heat dissipation, and provides stiffness to relatively large parts which may suffer from damage by inertia during transport.
New quality diodes are indeed more efficient and dissipate less heat; probably not needing any heatsink for normal operating cases.
In any case you probably won't be running 100W output power for hours and hours; even the original diodes don't get top loaded.

Regarding capacitance value of new caps: their properties change after a number of hours, including improved capacitance.
Also, most of us measure capacitance with a very low voltage (~1V) and very little current.
When I measured Nichicon KG's with a 35Vdc bias, they measured well around 100% of their rated capacitance (the Cornell Dubilier even slightly above)
 
Last edited:
Regarding capacitance value of new caps: their properties change after a number of hours, including improved capacitance.
I have wondered this, and have measured before and after (various time periods) but haven't seen any improvement in capacitance where the new value was below labeled value.

Would it really be the case that capacitance would have improved when tested with the DC bias, but not improved (at all) without? Could you suggest a simple way to generate the DC bias you mention, or does it require expensive equipment?

I've found on the basic tester that better capacitor types (solid polymer, polypropylene film etc) measure much closer to their labeled value (both when new and after hours of use). This has been my experience even with caps rated between 35V and 450V. Is this because liquid electrolytics have far more sensitivity to operational voltage/current?
 
I don't want to disturb your nice project, but the capacitor hook-up with additional wires between main capacitors and their PCB is bad practice: you add inductive elements in the system which may cause nasty harmonics/resonances at higher frequencies (beyond 50 kHz, but the closed loop bandwidth is around 150 kHz, so it counts). Idem ditto for the additional wiring around the diodes. The noise (as far as relevant) of the original diodes can be suppressed with 5~10nF film caps (>100V rated).
Keeping diodes and power caps very tide and very close together with as little as possible inductive elements is a basic design rule for any power amp.

The 'bridge' of the original diodes acts like a heatsink, also directing heat to the large PCB tracks for more heat dissipation, and provides stiffness to relatively large parts which may suffer from damage by inertia during transport.
New quality diodes are indeed more efficient and dissipate less heat; probably not needing any heatsink for normal operating cases.
In any case you probably won't be running 100W output power for hours and hours; even the original diodes don't get top loaded.

Regarding capacitance value of new caps: their properties change after a number of hours, including improved capacitance.
Also, most of us measure capacitance with a very low voltage (~1V) and very little current.
When I measured Nichicon KG's with a 35Vdc bias, they measured well around 100% of their rated capacitance (the Cornell Dubilier even slightly above)

Thanks for your input - very valuable things to consider on both fronts and your insights are appreciated.

I've never liked the method of attachment in the photo from my first post. If I used a wire attachment method I'd probably fab an angle brace to shorten the distance (The person that installed those caps mounted their tops to the PSU PCB support) to a minimum.

I've also considered modifying the PCB as below, but that depends upon being able to remove the epoxy over the traces without damaging them and them being clean enough to tin afterward. It looks like you can heat the epoxy with a heat gun to soften it and then scrape away with an exacto knife or scalpel. I'd need to find a similar donor board to test before I proceed. There is room for a 5mm pad (in green):
upload_2017-3-18_14-48-49.png

EDIT: If I went this route I'd move the caps as close to the rectifiers as possible to shorten the electrical distance, then add some closed cell foam or other to support them

I've even considered fabbing a new PCB that would also let me mount the FREDs using heatsinks (the TO-247 heat sinks I've found all have mounting posts). Cost would be about $70 each with a minumum of three. Structurally, the FREDs are more solid than I thought they would be and I purposely mounted them as far away from the PCB as possible to increase the leverage against the unsupported leg. I'm not sure why the wire jumper would cause more inductance than the anode/heat sink of the OEM bridge. The wire itself is 20 AWG mil spec silver plated copper with a teflon insulator.

The noise (as far as relevant) of the original diodes can be suppressed with 5~10nF film caps (>100V rated)

Where would you put these at? The filter caps? Note that each bridge pair has a .01 uF cap - is this the location?
 
Last edited:
Nice to see somebody doing some plotting works these days before attacking :)
Yes, many of us dream of making much better PCBs for the B-2 to obtain a design with modern parts an known improvements.

Given the fact that the original large solder pads are so nearby, I would consider to bend the legs of the capacitors under a 45 degrees angle towards the original solder pads, and than solder up the entire pad.
Adding 1 to 2.2uF film caps (Panasonic ECQW or equivalent) can be done bending the legs back up above the original solder pads, so you create a take-off point

The snubber caps for the diodes would across each diode (hence 8 pieces), between the leg of the diode and the pin of the heat sink next to it.
These would do.
 
Where would you put these at? The filter caps? Note that each bridge pair has a .01 uF cap - is this the location?

The existing 0.01uF caps (C251) are HF bypass caps across the secondary windings of the transformers.
The way they are shown and located in the circuit diagram does not really reflect their function; if only some engineers could have some better presentation skills....

They have nothing to do with the existing rectifier diodes, despite being shown like they do....

While you are on it, better replace those old C251 polyester caps by decent polyprop caps; same as proposed for the rectifier snubber caps (so 10 caps in total)
 
Nice to see somebody doing some plotting works these days before attacking :)
Yes, many of us dream of making much better PCBs for the B-2 to obtain a design with modern parts an known improvements.

Improvements on the B-2 are beyond me at this point, but PCB design is pretty straightforward if tedious. Costs have also come way down and OshPark will manufacture boards at $5 per square inch which includes 3 boards. For example, my replacement connector board was 4.84 square inches, cost a total of $24.20 and I received three boards for that total price. If I needed to replace a B-2 PCB and couldn't source one elsewhere I wouldn't hesitate to brew my own. I have to think that more than a few power supply boards have cracked due to bowing.

Given the fact that the original large solder pads are so nearby, I would consider to bend the legs of the capacitors under a 45 degrees angle towards the original solder pads, and than solder up the entire pad.

We have a winner! The caps are just tacked in right now as I'm going to add some 2.2 uF bypass caps, hopefully next week. I still had to chamfer the capacitor side of the PCB to get the caps to seat flush, but it worked out very well. THANK YOU for this very simple solution!

upload_2017-3-18_19-43-23.png

I hope this PCB is up to @Mr. Yamaha standards for cleaned up!
upload_2017-3-18_19-43-57.png

Chamfered areas are under the arrows on the capacitor side of the board.
upload_2017-3-18_19-45-14.png

The snubber caps for the diodes would across each diode (hence 8 pieces), between the leg of the diode and the pin of the heat sink next to it.
These would do.

Ordered, thanks!

The existing 0.01uF caps (C251) are HF bypass caps across the secondary windings of the transformers.
The way they are shown and located in the circuit diagram does not really reflect their function; if only some engineers could have some better presentation skills....

They have nothing to do with the existing rectifier diodes, despite being shown like they do....

While you are on it, better replace those old C251 polyester caps by decent polyprop caps; same as proposed for the rectifier snubber caps (so 10 caps in total)

Ordered as well - thank you for your input and suggestions!
 
BTW, those original pesky jumpers are laughable. There are at least a few ways this board could be improved.
 
ooh...:oops:
Although looking nice, I meant something else regarding bending of the caps' legs towards the solder pads
Sorry for not being clear enough :oops:

What I meant was: do drill the new holes as per your plot plan, mount the capacitor, and THEN bend the legs under 45 degrees towards the original solder pads.
Like that the caps will all be nicely aligned, and then nicely laying on the chassis.
It is important the PCB does not have any weight stress of the capacitors once back in position.
 
ooh...:oops:
Although looking nice, I meant something else regarding bending of the caps' legs towards the solder pads
Sorry for not being clear enough :oops:

What I meant was: do drill the new holes as per your plot plan, mount the capacitor, and THEN bend the legs under 45 degrees towards the original solder pads.
Like that the caps will all be nicely aligned, and then nicely laying on the chassis.
It is important the PCB does not have any weight stress of the capacitors once back in position.

No worries, and I may still go that direction - the caps can still be moved. If I leave as is I would need something to support the caps.

Now, if I change back to the original plan I can bend the two on the left below to reach the solder pads, but the two on the right would either need epoxy removal for solder pads or by extending the leads. The only downside to moving the caps at this point is that they come with an adhesive foam on the top of the cap, so they are now glued in place which helps to take the strain off the solder pad connection.

upload_2017-3-19_9-35-10.png

Something to ponder a bit - I will be pretty busy this week and hope to have all my commitments done by the time my next box of fun from Mouser arrives.
 
Nice project, Uncle Paul :thumbsup: Will follow this thread. I considered the Mundorfs also, but went for the CDE's because of the size which matched the originals.

About the standards, I have to give most credits to Sensei Oilmaster :D

I have very high standards in working neat, clean and consistent, plus I'm a bit of a perfectionist / autist, but most tips and tricks were provided by Oilmaster or at least veryfied with Oilmaster :) Really amazing help.

About your work so far, looks good!
 
Nice project, Uncle Paul :thumbsup: Will follow this thread. I considered the Mundorfs also, but went for the CDE's because of the size which matched the originals.

About the standards, I have to give most credits to Sensei Oilmaster :D

I have very high standards in working neat, clean and consistent, plus I'm a bit of a perfectionist / autist, but most tips and tricks were provided by Oilmaster or at least veryfied with Oilmaster :) Really amazing help.

About your work so far, looks good!

Well, I hope to be as good a student - so far just a white belt, and it's kind of dingy at that...

How is everybody cleaning their chassis? Seems like every B-2 thread is full of as new looking hardware.
 
Back
Top Bottom