Relay Switching Attenuator - Is it safe

[Littlewyan]

I'm building an attenuator which is going to have a foot switchable solo boost. Basically its an Airbrake with two rotary switches that I'll use a relay to switch between. Now when the relay is switching due to the design there won't be anything connected to the Amp as the relay switches the connection from the amp. Will this cause an issue? I mean I guess we're talking milliseconds here but I just want to make sure.

Here is the relay

http://uk.mouser.com/ProductDetail/TE-C ... PleQ%3d%3d

[dkochan]

Can you alter your design so that you're only switching a part of the load, like switching between one of two parallel (with the main, always connected) resistance loops?

[Littlewyan]

I could possibly look at somehow leaving the parallel resistor to stay connected so there is some load on the amp, however for a split second it will be a much lower load than normal. I.e. on the last attenuator setting it will be roughly 2.4Ohms and the amp will be set to 4Ohms.

I know valve amps can take no load for a very short amount of time, however this is only if the amp is turned on without one. Removing a load whilst the amp is on usually creates a flyback voltage I think.

[dkochan]

I really don't know.

4 ohms into a 2.4 ohm load for milliseconds is nothing. I don't know if or how fast fly back voltages will develop under those conditions, though.

Probably of equal or more concern is what happens if the relay just fails.

Have you looked at amp head switcher schematics and design notes? Similar issues switching from a speaker to a dummy load. I know I had seen some good info on that recently but couldn't find it again for you.

[Firestorm]

I'd be more concerned about the 150mW rating of the switch. Why not switch the output with Power Scaling (K O'C) or VVR (Dana).

[Littlewyan]

Ah good idea Dave, I'll look up a amp head switcher. Colossal on another forum suggested putting a 220Ohm resistor in parallel with the amp which is a good idea.

Firestorm - 150mW is the power consumption of the relay. It can actually take 3A 250Vac on the contacts. I would use power scaling but that means rebuilding the amp a bit :/.

[Firestorm]

I see. But I read 3A @ 5V. Might be okay. I should read more closely.

[Littlewyan]

Its not very clear on the Mouser site and the datasheet is rubbish. Went to the TE site in the end and read the spec on there.

[Littlewyan]

I looked at the schematic for Aiken's Head Switcher and he does exactly what Colossal suggested. He has a 470Ohm resistor in parallel with the amp.

I also found a beefier relay as that was another concern of Colossal's. The one I specified should have been fine but when I set about choosing the parts for this attenuator I wanted it to be overrated for what I'm using it for.

Here is the relay I plan on using:
http://uk.mouser.com/ProductDetail/TE-C ... tCk2CjU%3d

[Firestorm]

The beefier relays I've seen are the ones used to switch B+ voltages in old TVs (I assume). I "inherited" a box of them from an old repair guy. Some of them run the coils at 6-12VAC, some at 117. Contacts appear to be rated up to 20A (based on my box of junk). They're octal, so probably take up too much room, but often are DPDT or 3PDT. Still made as far as I can tell.

[Littlewyan]

Are those the ones that actually make a loud mechanical clunk when you turn them on? Those are the bad boys you want .

The only concern I have now is whether the Lorlin Switch is the best to use. Its rated for 5A at 250VAC. So as I'm probably going to put 20VAC on it at the most do I have a bit of breathing space with the current? Its my understanding that with a lower voltage you can go slightly higher with the current rating, but obviously you have to be careful. I doubt I could halve the voltage and double the current or anything that extreme.

[Littlewyan]

Once I've built the attenuator how can I determine if its safe when switching? As the relay switches in 6-8ms I'm probably not going to see it on a multimeter. I just want to know if its causing the amp to spike at all.