18Watt.com

I've been giving this some thought and come up with the attached.

At the moment I have a SPDT switch to ground the unused channel where it connects to the PI. This works well and doesn't pop.

What I want to do is have the option to use a foot switch, I didn't want to take signal grounds out to the switch as I would expect this to be problematic with noise. I'm hoping that replacing the SPST with a relay and switching it either via a foot switch or the panel switch will do the job. Can the more experienced eyes take a look at the schematic and highlight any potential problems?

Thanks as always

Martin

That looks fine.

The more important reason to keep the (-) rail separate from ground is that the center tap of the heater winding already is grounded (virtually or for real). Bad things will happen if you try to ground both the center and one end!

--mark

Thanks, just looking for a vote of confidence before I go messing about with this build. I'm split between building a footswitching cct or using an AB box. This is the last step of the build and I want to ensure I make the right decision to suit my needs, rather than have to keep re visiting it.

markh wrote: Bad things will happen if you try to ground both the center and one end!

Yup, if the smoke leaks out it won't work properly...

Cheers

Martin

Yeah, it's surprising the channel switching designs you see on otherwise reputable sites that show virtual earthing of the heater secondary with 100r resistors and an earthed -ve on the bridge rectifier, and think that's OK. It works (as different to if the heater CT is directly connected to earth), but still isn't good.

An AB box is pretty easy & cheap to build. You could try that first before further modding the amp. But either should do the job well.

I think you would still need a film cap, around 0.1uF or 0.68uF, between the virtual ground of this circuit, and your amp's ground. Otherwise, going from personal experience, it's likely that it will introduce buzz into the amp.

FYI, here's another relay supply circuit someone sent to me. I've never tried it, and I'm not quite sure how it works...

I'll take a good A/B box over channel switching, relays, LDRs, optos or FETs any day.

One can skip the cap and doing the virtual ground in favor for a real ground, but ONLY if the supply feeding it is isolated from the heaters.

ZP, what value of cap would you suggest to eliminate the potential noise problem?
I presume also, on that circuit you shared here, but never used, that the 5v supply is not powering a rectifier ('cept maybe a GZ34. Using a 5U4 would be a cinch to possibly letting out more than just smoke.)

WiseOwl, That will work, and as you noted, it cannot be direct to ground, at least using the heater taps that power heaters.

I have used the power supply that wiseowl shows with good results. The best results are had with a separate transformer tap. I have found using a 1000uf to 2200uf cap keeps things the quietest. I also find it useful to put the relay power supply/footswitch as far away from the input stages as possible.

I think channel switching has it's place and is useful. The question in a channel switching amp is how to keep as pure a signal path as possible. Obviously, the more simple the audio path the better, but a good sounding channel switching amp can be made. Of course, the best way to go are multiple optimized amps with good switching - but it's still switching.

I personally prefer the sound of tubes pushing tubes over a solid state boost, but that's just me.

Richard, I wonder what would happen, if instead of a bridge we used a full wave rectifier, where the 6.3V winding has a centre-tap. Then the relay's DC supply would be referenced to the same chassis ground as the heaters. I wonder if that would be better or worse for noise...

Shrapnel wrote:One can skip the cap and doing the virtual ground in favor for a real ground, but ONLY if the supply feeding it is isolated from the heaters.

Correct. But see my question to Richard about what happens in the case of a full-wave rectifier.

Shrapnel wrote:ZP, what value of cap would you suggest to eliminate the potential noise problem?

I've seen 0.68uF, and I've even used a 1uF film cap. But I'm not sure how much it improves noise in every case. When you start dealing with grounding issues, voodoo inevitably seems to creep in along with the science.

Shrapnel wrote: I presume also, on that circuit you shared here, but never used, that the 5v supply is not powering a rectifier ('cept maybe a GZ34. Using a 5U4 would be a cinch to possibly letting out more than just smoke.)

I wish I really understood how that circuit works, as it seems to form a kind of virtual centre tap with the pair of diodes and the pair of caps, and that's what goes to ground. But I wouldn't like to try it with a 5U4 or other directly-heated rectifier tube either.

zaphod_phil wrote:Richard, I wonder what would happen, if instead of a bridge we used a full wave rectifier, where the 6.3V winding has a centre-tap. Then the relay's DC supply would be referenced to the same chassis ground as the heaters. I wonder if that would be better or worse for noise...

The transformer in my case has to use the 2 100r resistors to form a virtual center tap, these would limit the current significantly if, I understand what you mean by full wave rectifier correctly.

Could you recommend a good AB box or point me to a schematic for one, as I see it the AB box with 2 leads forms a big earth loop so I assume some are designed to eliminate or minimise any effects?

rjgtr thanks for the input. I'm not going to make any mods until I'm satisfied I know what's best for me. I want to keep this amp tidy.

Cheers

Martin

zaphod_phil wrote:Richard, I wonder what would happen, if instead of a bridge we used a full wave rectifier, where the 6.3V winding has a centre-tap. Then the relay's DC supply would be referenced to the same chassis ground as the heaters. I wonder if that would be better or worse for noise...

Well, that method would give a 3v relay a possible chance? or w/ good filtering 3.15*1.4=4.41v might be enoughfor the 5v, depending on the coils.

Shrapnel wrote: w/ good filtering 3.15*1.4=4.41v might be enoughfor the 5v, depending on the coils.

That's more the way I was thinking of, especially as 5V relay's threshold will probably be somewhare around 3.5V. But I was wondering whether this way would be better wrt noise, or the bridge rectifier with the floating relay supply. I guess it's probably one of those things you don't know until you try.

Possibly so... But when I stop and think of it, if that voltage is too low, I'm not sure if that would induce noise via vibration in the switch contacts. (I reserve the right to be incorrect on this)

My experience with trying full wave is that it just doesn't produce enough voltage to reliably switch a 5V relay. You're better off with using the Bridge recto with a floating ground. If you get hum then using a 1uf to 4.7uf cap to ground will take care of it. I use small electrolytics for this.

wiseowl wrote:Could you recommend a good AB box or point me to a schematic for one, as I see it the AB box with 2 leads forms a big earth loop so I assume some are designed to eliminate or minimise any effects?

Never thought of the ground loop issue, but mine doesn't add any hum and I use 3m of shielded 4 core microphone cable split and heatshrunk at the ends with the std plugs to the switching pedal and another 5m lead to guitar. Schematic attached, but it's just the obvious.

Thanks,

got hung up on the whole ground loop thing.

Seems strange that after all the effort put in to eliminating them inside the amp that a great big one outside doesn't seem to hurt. Maybe it's all down to the proximity of the transformers.

I'll give external switching a try to start with, it'll be much simpler to implement. On the downside I cracked my faceplate drilling a second hole,:oops: so It looks like making a new one is going to be my Sunday job.

Cheers

Martin

zaphod_phil wrote: FYI, here's another relay supply circuit someone sent to me. I've never tried it, and I'm not quite sure how it works...

I was curious, so I scribbled it out, it looks like the full-wave doubler shown here:
http://www.play-hookey.com/ac_theory/ps ... liers.html
with some added decoupling/filtering,

Naturally, it has the same problem that all the other schemes have, It is only suitable if the rectifier winding (the one normally used for directly heated diodes) is unused.

At a wild guess (i.e without trying to remember how to apply the maths again ) it should supply 12V or so at a few 10's of mA from a 5V winding.

david

That's what I had a hunch it would be, but never got round to checking it out....

I reflote this old topic. Is there an schematic of this mod somewhere on the site?
Its seems it existed somewhere in time. Thanks!