Which relays for switching?

[chipaudette]

Hi,

I've been modding my Fender DRRI by cascading the "normal" channel's pre-amp stages into the "vibrato" channel. This is giving some nice additional overdrive. But, I'd like to insert some attenuating resistors or other tone-shaping elements later in the signal chain.

Currently, I have a DPDT switch that brings the gain stages into the beginning of the circuit. If I want to bring in the new elements, I'll either need a 2nd switch (not very elegant) or I'll need to swap my DPDT switch for a 3PDT or 4PDT (requiring long leads and a lot of room for the switch itself).

Alternatively, I could switch all of these elements using a single switch connected to a bunch of relays. But, I don't know anything about working with relays. I'd really appreciate any guidance on the following questions:

What relays should I use?
Electro-mechanical or solid-state? (It's a tube amp...I've got 100s of volts)
What part number is recommended?
Where can I get them?

What kind of control signal do I use?
Can I use the low voltage line off my transformer (12ax7 heater voltage)?
Do I need a regulator or anything?

Thanks a lot for any help you can give!

Chip

[Don Moose]

You want 'low signal' mechanical relays.
- The Omron D20G5V-2 are nice. You might pick the 'sensitive coil' version to keep your coil currents down.
Then you need these de-popping steps:
- Definitely put a diode across the coil (leads as short as possible, cathode to (+), anode to (-))
- Add a large resistor (10M or so) from each signal switching terminal (NC, NO, COM) to GND.

Hope this helps!

[chipaudette]

Low signal?

So, I want a "low signal" relay even though my voltages could be very high in my tube-driven circuitry (100s of volts)?

Thanks for the de-popping recommendations!

Chip

[Enzo]

Tube amps run on high voltages, but the signal path where switching usuall;y occurs is generally not at B+ potentials. Look at the circuits you wish to switch. How many instances are you switching at B+ levels? Furthermore, the voltage rating will be for the open contacts. A separate voltage rating might be for absolute voltage to frame. Point being that switching two different plate loads might be possible even though they sit at a coupl ehundred volts because the voltage across the contacts is not that high.

To switch in the signal path, you want low current contacts because they will be low resistance, and made for signals.

Remember that even though the plate of a 12AX7 might be at 200 volts, there is still only a milliamp or two of current flowing.

If I wanted to diable a whole gain stage, I would not be switching the B+ off and on, I would just leave the tube running and ground off the output of the stage after the coupling cap.

[chipaudette]

AC signals could be 100s of volts?

Hi,

My plan was to switch gain stages into and out of the signal path. I planned on cascading the two halves of a 12AX7 (with some inter-stage attenuation) into the front of another 12AX7.

The location of one of my hypothetical relays would be after the coupling cap of the 2nd gain stage and before the grid of the 3rd stage. Since I'm likely to be getting some saturation of the 2nd gain stage, I figured that the pre-coupling-cap AC voltage might be swinging between zero and the plate voltage. For my Fender DRRI, I think that that would be something like 0-260V. After the coupling cap (which is the signal that would be hitting my theoretical relay), the voltage might be swinging between -130 to +130V.

So no, I'm not switching my B+ in and out. But, I still might have high AC voltages because my signal might be swinging rail-to-rail. Since it's a pre-amp stage, though, I do agree that the current flowing through the relay would be small.

So, a "low signal" relay was recommended. My worry is that I don't know what "low signal" means. If it means "low current", then I'm pretty psyched because I'm imagine that these low signal relays are reasonably small. Alternatively, if "low signal" it means "low voltage", I'm worried that my -130 to +130V AC signal might NOT be considered low voltage. True, it has zero DC voltage, but +/- 130V is pretty good sized. I don't want any arc'ing.

Do you have any experience to relieve my fears?

Thanks,

Chip

[pdman]

You want 'low signal' mechanical relays.
- The Omron D20G5V-2 are nice. You might pick the 'sensitive coil' version to keep your coil currents down.
Then you need these de-popping steps:
- Definitely put a diode across the coil (leads as short as possible, cathode to (+), anode to (-))
- Add a large resistor (10M or so) from each signal switching terminal (NC, NO, COM) to GND.

Hope this helps!

Any idea on where to get these?
Searching google or yhoo shows nothing.
looking at the Newark catalog now. Any help appreciated.
I want to build an FX switching system and need good low signal relays.

thanks

[Enzo]

You really expect 260v p-p of signal in the preamp stages?

[chipaudette]

260 Vpp??

Am I expecting 260Vpp? Well, I've never measured the voltage being output by a saturating pre-amp tube, so I don't really know. I'll rely upon yours or others opinions on that matter.

But, since I'm looking to insert a relay into the circuit, I want to make sure that it is spec'd to the worst-case possible condition. In my amp, the worst case would be zero volts up to B+, which is 260V.

Now there's the question as to whether this is a "reasonable" worst case. Well, in my circuit, I'm purposely driving my pre-amp tube to saturation to create that overdriven guitar sound. If I'm heavily saturating the tube, then I must be approaching the voltage limits supplying the tube or else it wouldn't be saturating. Sure, I might not be hitting the full rail-to-rail 260Vpp, but a 150-200Vpp swing doesn't seem unreasonable to me. With a 100k plate resistor, a 200Vpp signal would require a 1mA swing either way around the quiescent current of the tube at rest. Seems plausible.

So, if I'm picking a relay for my circuit, I feel like I want one that could handle a couple hundred volts peak-to-peak because, in the output of a saturating tube, I feel that such levels are possible. Is this overkill?

Chip

[Enzo]

DO you have an existing tube amp of some sort that functions? If so, I suggest you apply a suitable signal, and see what amplitudes you find at each stage. Most amps are pretty much like the next, so it is not likely you will find one amp with 40v of signal at the. first stage plate and another with 200, and a third with 4. Most amps will be typical of all amps.

There is overdriving a preamp tube, and then there is... well, I'm not sure what.

To get the tube to turn off, your signal at the grid would have to be VERY strong, and to get the tube to conduct so hard the plate goes to zero will also require quite a signal at that grid.

Yes, the most voltage POSSIBLE on the plate would be the B+ voltage, and of course the least would be zero, but getting to those limits will be difficult and not to be expected in a guitar amp. Through the preamp, I expect to see a few tens of volts signal at most. I expect large signal levels at the output tube plates.

To get silly, if you fly somewhere on a plane, you could prepare for worst case - short of death - by wearing scuba gear AND a fur parka, in case the plane had to come down in the water or on a glacier. But chances are you'd just wear normal clothes because at worst, you might have to walk down to the pavement if the gate thing was broke, and it might be raining.

Look at the little rectangle pc mount DPDT relays that many amp makers use. This need not be overly complex. The relays are not that critical. SPend the time working on the design itself.

SOmething like those are compact, need relatively little coil current, and are easily sourced in a variety of coil voltages. What you don't need are huge clunking 20 amp plug-in type relays.

You don't want solid state relays, they are essentially triacs or similar inside.

Look in your Mouser catalog in the relay section, there are pages of "Low Signal" relays. REfer to their stock number:
653-G6A-234P-DC12
That is a typical relay. I picked 12v, but you could chose 5 or 24. The size and shape are common, 10x20mm, and most relay brands make them in similar fashion, so you can use the Omrons, or NEC, or whichever, and they will fit in the same holes.
Here (more Mouser stock numbers):
551-MR62-12USRY Same thing, different brand
817-RY-12W-K same again
655-V23105A5003A201 and again

And there are other brands. I buy stuff like this from Mouser, Allied, Digikey, etc.

COnsider your power supply. Though these relays come in various coil voltages, they also have a range of coil currents. SO within the selection of brands, and within a particular coil voltage, some will need more current than others. None will need a lot of current, so it is not probably much of an issue, unless you want ten relays all on at once.

No, the relay supply need not be regulated. A 12v coil will activate the relay over a reasonably broad range of voltages. Just don't have the relay wiring share a ground return with any signal circuitry.

[Don Moose]

Any idea on where to get these?
Searching google or yhoo shows nothing.
looking at the Newark catalog now. Any help appreciated.
I want to build an FX switching system and need good low signal relays.

thanks

Don't remember where I found them, but search Digikey for Omron low signal relay