2 head 2 speaker switcher

[andresbianco]

Hello all,

I'm designing a 2 amps to 2 speakers switcher (to select one amp with one or the other speaker). I'm doing all the logic with relays and precise switching (time on and off ) in order to avoid no-load situation when switching between (tube) amps.




Anyway, ground loops may occur of course. I suppose when sending to amps and also in the returning path from amp's speaker outputs (if some of the poles are connected to ground).
Hope you could help me to figure this out. what I mean is, of course I can switch both , hot and grounds, to avoid interconnection between amp's grounds but that doubles relay count.

Isolation transformer could solve the sending to amp in, but what about amp out?

Thanks!

PS: R.G. I was watching your splitter design, I have a doubt about frequency response of 42TM018, from 300Hz to 3,4kHz?? do you compensate that in any way? thanks

[R.G.]

I'm designing a 2 amps to 2 speakers switcher (to select one amp with one or the other speaker). I'm doing all the logic with relays and precise switching (time on and off ) in order to avoid no-load situation when switching between (tube) amps.

Anyway, ground loops may occur of course. I suppose when sending to amps and also in the returning path from amp's speaker outputs (if some of the poles are connected to ground).
Hope you could help me to figure this out. what I mean is, of course I can switch both , hot and grounds, to avoid interconnection between amp's grounds but that doubles relay count.

You are correct that you will have ground issues in the general case of any amp, any speaker. There will be some combination of amps (and their internal AC power leakages) that will cause hum and possibly other noise issues. Like so many things in musical electronics, you can get away with not doing the right thing, but not always. For general use, put in the ground switching relay contacts. Suitable relays are available for $3 to $5 each, so it's not a huge cost difference.

As to the no-load situation, you can get away with (there's that term again) just hanging a 47 to 120 ohm resistor on the output of most tube amps and have them not go into destructive runaway. But with today's boutique amp builders, there's no way to really tell what you'll get. I guess it depends on whether you want to sell these things or not. If you sell them, there's almost certain to be some people coming back and telling you that you destroyed their amp. Actually, that will probably happen anyway, but at least you'd know you didn't kill the amp.

Isolation transformer could solve the sending to amp in, but what about amp out?

IMHO, use the extra relays and separate the grounds in all cases. You can use transformer isolation on the front end, or you can relay isolate it, too.

PS: R.G. I was watching your splitter design, I have a doubt about frequency response of 42TM018, from 300Hz to 3,4kHz?? do you compensate that in any way?

The article at geofex goes over that point. The 300 to 3.4k spec is the minimum the transformer maker promises. You'll get --at least -- that. In actuality, the high end runs out to over 20kHz on all the samples I tested. The low end peters out somewhere around 200Hz for the typical ones, but I used a very low impedance opamp driver to drive the primary with a lot of current available. The low end of a transformer response is often determined by the primary inductance stealing signal current. If there is a low impedance source driving it, the source can simply pour in the extra current, and the combination of driver and transformer has an extended bass response. In the samples I messed with, I got tested response down to ~ 60Hz.

[andresbianco]

Thanks R.G.!

Probably i'll go with relays.
I also wonder if i will need some muting circuitry in order to avoid noise when switching guitar from amp 1 to amp 2 Inputs (maybe transistor or LDR mutes)?

I'm not aiming to sell this, so I agree with you on putting those resistor in the amp out but I'm going rely on relays.

Regarding transformers, if you want to make a passive isolated splitter you will need to move to an expensive transformer, right?

Thanks again!
AB

[MetalGuy]

I did something similar years ago but never finished it completely but at least now I know the problems.
What I would do is the following sequence:
1/ mute amp's input
2/ switch OT to a dummy load, say 8-10 Ohms
3/ AFTER that switch amps output
That means there should be few milisec delay between 1/ and 3/ just to be on the safe side. How to provide that delay to the relays maybe someone else would advise.

Using relays to switch amps' inputs to ground is noisy especially if it's a high gain amp so think about an LDR solution or use two FETs in series and parallel or something similar but bear in mind that it won't be 100% switched to ground thing because those have some resistance left.
Use one of the R.G. splitters with transformers - for two inputs you'll need only one trafo.

[andresbianco]

Thanks MetalGuy

The delay is accomplished using different RC times (charging and discharging a capacitor in the transistor's base), i.e. using 10uF with 22k for "slow" on time and 4k7 for "fast" off (See R1, R2 and C1 in the schematic).
I used this method to avoid no-load when switching speakers, see oscilloscope image included where both signals overlap so both speakers are on for a while when switching (blue trace correspond to one channel and yellow to the other one).  Relay on means speaker is off, due speakers are connected to N.C. relay's pin in this case.

Thanks,
AB