Yamaha B2 input switch/attenuator bypass

belgianbrain

Super Member
I had touched on this subject in another thread, but I think it belongs in its own thread. Maybe the moderators can help sort that out, if needed.

Recap:

Does anyone think there is much to be gained by wiring one of the inputs directly to the driver boards therefore bypassing numerous selector switches, volume pots, etc.? I never use those anyway. I just use one set of inputs in DC mode with volume unattenuated.

I'm not a big fan of making modifications to heritage equipment, but this one is easily reversible.
 
There is an ENORMOUS pitfall in doing that in the B-2...... the 0 Vdc track of the driver board won't be connected to the 0 Vdc point in the power supply anymore. That is, if your new RCA connectors are not grounded to chassis.

What Yamaha did in the B-2, i.e. routing the driver PCB 0 Vdc back to PSU via the screens of the input signal all the way back to chassis from RCA connector, was a very bad earthing/wring principle in the first place, and worth modifying anyway, even without a direct signal cable.

Hence, if you intend to install chassis-insulated RCA connectors as well, run a dedicated black 0 Vdc return wire from driver PCB "E" solder point to the central 0Vdc point on the capacitor board.
If you keep the original RCA connectors in place, do NOT cut the wire to chassis, en ensure that the RCA connectors you use are indeed connected to chassis. While modding anyway, I would have the RCA connector floating from chassis and run that dedicated earthing wire from driver board to capacitor board.

Thank you, Oilmaster!

I haven't tried it yet, largely for that reason! I looked at the schematic and the grounding methodology had me scratching my head and I was chewing on exactly that aspect. I appreciate your response, because I certainly didn't understand it to that depth.
 
I'm also considering replacing the electrical cable when I have the B2 open. Wondering about just installing an IEC connector since all my other gear uses this.

The B2 appears to be a floating ground design since it has no ground prong on the cable. Are there pitfalls to installing an IEC connector and connecting the chassis to earth ground through the 3rd prong? No need to muck up this thread if anyone can link me to such an article or thread elsewhere.
 
The B-2 has originally no grounded chassis (grounded to mains earth/ground that is), like the majority of 70's gear.
Depending on your entire system, and mains grounding layout in your home, it is in most cases not recommended to connect the B-2 chassis to mains ground so easily.
Also, remind that the mains ground (earth) is nowadays also the 'garbage bin' of many IEC filters in many household equipment running more and more with switched mode power supplies.
In other words, that "ground" is merely a landfill full of noise and transients (it's a not a dead quite stable reference level anymore)

A lot of 70's vintage gear has the outer bushing of the RCA connector connected to chassis, reminding that the shield of each interconnect is the signal return path.
Connecting that chassis to a noisy mains ground injects that noise in all the inputs and outputs (elevates the reference level to be more exact), and in addition reference level 'levelling' between multiple interconnected audio units may occur (i.e. ground loops).

Hence, if you may want to install an IEC connector, with a ground pin, only ground the circuit use a "ground loop breaker" circuit (aka safety loop) as in figure 4 of this ESP tutorial.

ATTENTION: in that case you have make sure that all possible B-2 PSU/signal connections to chassis are removed, such as the RCA connectors AND the 0Vdc wire on top of one of the heatsinks. It might be better to delete the mains ground connection to chassis as in figure 4 of the ESP tutorial, which will allow the existing grounding topology inside the B-2 to stay as is.
 
Thanks again, Oilmaster. I think I'll leave it with a two pronged cable. :idea:

Ok., I performed this modification. New, chassis insulated RCA connectors on the rear and cables straight to driver boards. Additional ground cables run from E connectors on driver board back to E on power board between +85/-85V rails. Old front selector board disconnected from driver boards.

So, the sound now bypasses a bunch of dirty switches and POTS that used to be in the signal path. One can spray deoxit, but without complete disassembly and cleaning, I suspected those items are mucking up the sound at least a little.

There was a noticeable improvement, particularly on clarity of highs and instrument separation.

I can now comfortably say that this B2 surpasses my Bryston 3BSST in performance! I highly recommend this tweak. I'm sure those dirty old switches and pots were compromising the performance of this OUTSTANDING amp.
 
A knowledgeable member of this board commented to me that I have now modified the 25K input impedance of this Amp.

Good point.

Instead of passing through 1.5K and then seeing 25K to ground, the input signal now sees 390K to ground and then passes through 2.2K before arriving at the input FET TR104 (left channel).

It seems that I have increased the input impedance of the amp considerably. Comments? Concerns?
 
Last edited:
Interesting thoughts, Oilmaster.

I have been chewing on this for the past day as well and I have not been able to conceive of how the change I made could be bad. But like you say, there are many complicated interactions with the preamp happening.

My current preamp I use in tube mode, and the output tube is a 6sn7 buffer, balanced, which goes through a solid state JFET summer and then on to the B2 via coax.

Anecdotally, I do think the sound has noticeably "improved" with this mod. Both bass and treble are clearer.

I put on some Marcus Miller yesterday, and holy $hit, that bass guitar was never so real. My wife actually commented that it sounded like it was in the room. It takes a lot for her to comment about audio.
 
And what about some kind of hybrid solution? Bypass the volume pots, but keep the switches (after giving the switches a complete disassemble, deep clean, de-oxidize, sandpaper and faderlube treatment).

That's something I would be interested is as I do not use the pots, but may use the switches in the future.

Is that even possible?
 
And what about some kind of hybrid solution? Bypass the volume pots, but keep the switches (after giving the switches a complete disassemble, deep clean, de-oxidize, sandpaper and faderlube treatment).

That's something I would be interested is as I do not use the pots, but may use the switches in the future.

Is that even possible?

Yes, you could simply jump the pot terminals from the wiper to the correct side turning it into a straight piece of wire. It could easily be undone later if desired

upload_2017-2-1_11-35-53.png
 
Yes, you could simply jump the pot terminals from the wiper to the correct side turning it into a straight piece of wire. It could easily be undone later if desired

View attachment 861309

Do I not need resistors which acts the same as the pots while fully opened? Or is the resistance minimal while fully open?

The switches operate only the power to the coil of output relays; no signal being handled by them.
By all means keep the switches alive. Just gently coil of the original signal wires from RCA connectors to pots, and from pots to driver board, so that the modification is reversible and the unit's originality contained.

This is for the pots bypass?
 
Do I not need resistors which acts the same as the pots while fully opened? Or is the resistance minimal while fully open?

Yes, you do. I pulled the trigger before I aimed on that one. The 25k pots are part of a voltage divider with a 1.5k resistor, and at full volume are set all the way to 25k. The problem is that you would need to unsolder the pots from their daughterb board and run a high quality low noise resistor replacing it entirely. I don't know how feasible that is mechanically. I'll look at mine (the meter/switch sub-chassis is off the one I'm working on) tonight when I get home.

terminal b on the daughter board is ground.

upload_2017-2-1_12-38-26.png
 
Anyone wanting to maintain the original resistance values and wire new RCA connectors directly to the driver boards could just put a 1.5k resistor in series with the wire connected to the IN line on the driver board and a 25k resistor from IN to E on the driver board.

The main question for debate is, is there really a benefit to this or a danger to not doing this? We're not sure there is.
 
Last edited:
Yes, I'm leaving it as it is with inputs wired directly to driver boards. This amp sounds so good now, it's stunning.

I finally understand when people say they've never heard another amp that sounds like the B2. :trebon:
 
Last edited:
Yes, I'm leaving it as it is with inputs wired directly to driver boards. This amp sounds so good now, it's stunning.

I finally understand when people say they've never heard another amp that sounds like the B2.

I've listened to and owned a number of great amps over the last 40 plus years. For the most part, as long as the amp is capable of driving the speakers to an acceptable level the differences between them are very subtle. Not so with the B-2. Listening to a B-2 is like hearing in color - it's hard to describe, but almost like I can see the music floating in the air. Other amps are enjoyable, but not captivating. I don't know if I could handle hearing a B-1 or B-3 given that they considered better by owners whose opinions I respect.
 
Alrighty...when I become fixated on a particular piece of gear, I end up studying the schematics and wondering how it could be improved.

Now that the inputs are wired directly to the driver boards, the signal path on this thing is pretty neat and simple. 85% of it is the driver boards.

So, I'm looking at all the resistors in the signal path and picturing shiny new metal film resistors.

Two possibilities:

1. New metal films in signal path only;
2. New metal films for all resistors on driver boards.

Any comments as to potential for improved sound or the quality of the original resistors?
 
Alrighty...when I become fixated on a particular piece of gear, I end up studying the schematics and wondering how it could be improved.

Now that the inputs are wired directly to the driver boards, the signal path on this thing is pretty neat and simple. 85% of it is the driver boards.

So, I'm looking at all the resistors in the signal path and picturing shiny new metal film resistors.

Two possibilities:

1. New metal films in signal path only;
2. New metal films for all resistors on driver boards.

Any comments as to potential for improved sound or the quality of the original resistors?

Just a thought... you could carefully remove and document original resistors on one channel, replace with metal film, then compare in mono using the balance control in your pre to pan between them. This way you know for sure that there was an improvement and either way you are half done
 
Any comments as to potential for improved sound or the quality of the original resistors?
The original resistors are carbon film and I'd expect them to still be fine but metal film is a better more stable resistor overall. I changed virtually every resistor and when I did I measured each one as I changed it and I didn't find a single one that was out of tolerance. It took hours and hours to do both boards. Could I hear a difference afterwards? No but it was a fun project at the time. Was it worth it? It was for me but YMMV ...
 
I am sure this will get done before too long, at least for the signal path. Time is always the bottleneck for me...

There are some 0.22 ohm resistors on the FET boards. Anyone have a comment about those? At that stage, those are kind of like resistors in the crossover network in the speakers, transferring the full power Of the amp.

I upgraded the signal path resistors in my Troels Gravesen crossovers for my NS-1000M to Duelund cast graphite, and shockingly, there was a noticeable improvement!

After a little research, it seems many people prefer carbon film resistors for sound to metal film. I'm leaning towards replacing the signal path resistors only - maybe with Riken Ohm. Comments?
 
Last edited:
There are some 0.22 ohm resistors on the FET boards. Anyone have a comment about those?
The OEMs are wire wounds and as such will have some inductance. I replaced them in one of my B-2's with some Ayrton-Perry (non inductive) types. I couldn't hear a difference but once again, YMMV.
 
Back
Top Bottom