quiet channel switching relays

[Mark Buckingham]

Are there any ways to really make channel switching relays quiet in a amp design?

I've gone the optoisolator route in the past, and that seemed to work, but was a bit pricy.

Any hints? Thanks in advance.

[Enzo]

Vactrols sell for $2.49 to $4.10 at Allied. I don't think of that as pricey, especially when compared to relays. And I like relays.

For more on that go to www.alliedelec.com and enter VACTROL in the search window.

JFETs are very simple to use, dirt cheap.

RElays work well too, really. There are a couple main sources of noise. If the relay is controlled by an actual switch, the little spark in the switch becomes a pop in the speaker. Proper snubbing of the relay coil and at least a cap across the switch helps. Using a transistor to control the relay eliminates that switch noise. It is then absolutely imperative there be a diode across the relay coil or the little transistor will become charcoal.

Once the coil noise is cured, simple enough, we consider the contacts themselves. They are not inherently noisy, but can make noise. The main trick is to not leave any contacts unterminated. For example if you are switching a cathode bypass cap in and out for gain adjustment, that cap has to charge when first switched in. This little current surge makes a pop in the speaker. By connecting a large value resistor across the switch, the cap remains charged, but the resistance makes it ineffective. When the switch shorts the resistor, then the cap enters the circuit, but doesn't have to charge first.

In situations where for example ther are two preamp channel outputs, and your relay selects which one to connect to the master volume control, those outputs need to be terminated, if nothing more than another high value resistor to ground.

Same deal selecting an input. You can't leave that input hanging on the relay depending upon the previous stage for anything. Terminate that grid.

[Steve Conner]

I second what Enzo said. Any time I had trouble with pops like that, it was down to the DC conditions of capacitors not being set properly (what he calls things being "unterminated" or "hanging".) It didn't make much difference what the actual switching device was.

I installed channel switching relays in an amp last weekend: one to switch the guitar input between the clean and dirty channel inputs (and ground the unused channel) and another to switch the output of the wanted channel through to the reverb recovery mixer etc. (and ground the output of the unwanted channel too.) I didn't need to add any of those high value resistors, because the grounding of the unused channel set up the DC conditions fine.

I put the relay coils in parallel with a catch diode across one of them, and put a 10uF cap across them too, to slow the rise and fall time of the switch signal. I then put a 4.7 ohm resistor in series with the foot switch to stop that cap getting discharged too hard.

Once I'd done all that there was just a little plink! noise when channel switching with the amp volume cranked and the guitar volume down to zero. I'm guessing that was contact bounce.

The relays I used were these guys: http://www.relays-r-us.co.uk/webpage.../nuova_cat.htm

Last night I was playing around with some H11F1 opto FET switches, and I managed to use them to make a footswitchable gain boost. They seem much like Vactrols except limited to 30V in the "off" position and 0.5mA in the "on" position. So they can't do everything LDRs can. They don't make any clicks or pops at all, and cost about $2 AFAIK.

[Mark Buckingham]

Thanks guys... I'll probably go with the vactrol or jfet route...

Just out of curiosity... Does anyone have an example circuit I can look at for the jfet method? I also need to know what a good jfet to use would be.

I'll probably have a switching arrangement where I need to turn 3 switches on, and turn 2 off, and vice-versa. I'm guessing I can just use a plain old transistor to handle turning on the other switches. Sound good?

Thanks.

[Enzo]

JFETs are easy that way, with the same gate voltage you can turn the P channel ones off and the N channel ones on, and vice versa, so an N and a P with gates wired together becomes a SPDT. JFETs are ON until turned off with a gate voltage.

VOltage is a consideration. 30-40 volt JFETs won't like 100 volts of signal.

[stevesamps]

getting rid of that POP ...

I would love to learn more about this topic, also
I have an amp with 2 channels that I hooked a manual
switch dpdt to ground one grid and manually switch
the channel and it makes a big ole POP.. they are completely
seperate channels so I imagine I need to look closer at grids to ground..
or in general the proper place to switch things..

I have avoided channel switching amps but its more about the
customer than what I want to deal with..

I guess I have to figure out what is making the pop
with the manual setup before I do anything else..

I have this amp as a prototype to play with...
I will try and post a rough drawing of what I tried so far..

I am not oppsed to jfets would love to see examples

[enash]

Relays

Are there any ways to really make channel switching relays quiet in a amp design?

I've gone the optoisolator route in the past, and that seemed to work, but was a bit pricy.

Any hints? Thanks in advance.

I have used 12VDC Telecom DPDT relays sucessfully for True bypass FX switching, input selection and other stuff. I have used a cicuit to control switch bounce from GEOFEX

There is a very small amount of switch noise which is no problem for practical purposes. I have considered Vactrols, Quad switch IC's and used relays because they are cheap, robust, flexible and when switched I know it is a metal on metal connection. I regularly play live in a few bands and I am very happy with the result.

I'm happy to give you more information is you are still interested

Cheers

[Tubis]

This is how i at the moment switch between to channels.
No switching noise or channel bleeding at all.
The optos are vactrol VTL5C1.
The vactrol to ground after the 470k is to take away bleeding from the highgain channel.
You could problably simplify the switching circuit some. But the components are so cheap so......

[MattT]

I've used vactrols and relays and neither makes any noise at all. The relays I used had the 5/6v coils and I put just a diode across each coil. In both cases I used the 5vac rectifier filament winding (both amps are SS rectified), a voltage doubler circuit, and a 5vdc voltage regulator. The whole switching circuit has it's own ground 'circuit'. In the case of relays it just required a few high-value resistors (usually 4.7M) in the right spots, as Enzo mentioned. It did require a bit of thought to have the LEDs work correctly (the different colors having the same brightness as well as the LEDs not interfering with the coils)...also accounting for the LED in the footswitch.

In both amps the switching is AFTER the first gain stage (or later). I've never done an amp that did the switching prior to the first gain stage but it seems that the earlier in the circuit the switch is...the more likely a 'pop' will be audible...just because of so much more gain happening after the switch. That first gain stage is usually taking a 100mVac signal and amplifying it to the 5vac ballpark...any pops get amplified right along with everything else.

One idea I used in my vactrol amp that I've never seen before (I'm not that well versed/read in switching circuits so that doesn't mean much) was to put a vactrol on the ground connection of a 'Master Volume'. The vactrol 'closes' (i.e. grounds the MV) when on the dirty channel but 'opens' (i.e. lifts MV) when on the clean channel. So this MV operates as a 'Lead Master' to allow you to balance the two channels. That leftover series resistance of the MV (when lifted) doesn't seem to make much of a difference in my amp.

[Tubis]

Oh, any led indicator in my circuit wouldnt be in series with the vactrol leds(or relays) because if the led indicator brakes the switching brakes.
I would put in separate mosfets to drive any led indicators.

[MattT]

Do LEDs fail that often? I've never had an LED fail...ever.

[Tubis]

Do LEDs fail that often? I've never had an LED fail...ever.

Sometimes but not very often ,no. But during a gig, it wont be to funny.

[enash]

I have used 12VDC Telecom DPDT relays sucessfully for True bypass FX switching, input selection and other stuff. I have used a cicuit to control switch bounce from GEOFEX

There is a very small amount of switch noise which is no problem for practical purposes. I have considered Vactrols, Quad switch IC's and used relays because they are cheap, robust, flexible and when switched I know it is a metal on metal connection. I regularly play live in a few bands and I am very happy with the result.

I'm happy to give you more information is you are still interested

Cheers

No Problem
First - check out the site below:
http://www.geofex.com/
One of the projects is : "A Remote Indicating Effects Bypass System" - lots of good info on bypass units. Within this project file is a layout of a bypass unit (I have attached a picture of this).

This unit allows for a True Bypass / FX switching. However what the circuitry does is control the speed of the relay switch to reduce the noise. You can use this relay control circuit to switch a relay however you like - True Bypass, Guitar Inputs, Chanel Switching, AMP switching.

Note :

1) Relays can be polarised or nonpolar and can have a variety of PIN layouts - you can get them in lots of eg 50 pretty cheap but may have to accomodate different pins
2) Power Supply - I have purchased 12VDC unregulated plug packs cheap (300ma to 700ma) cheap. Note - the parts list in the GeoFX site lists a transistor Q1 2N3904 625-2N3904. Beef this component up - use a arround 700ma rated transitor, this is a critical component which is subject to fail (from practical experience)
3) In addition - connect a Imeg resistor between input and output fro ground reference.
4) Make sure the whole unit is well shielded and play with earthing to get best result
5) WARNING - the applications I have used this for are all in the signal chain before the Amp input. I see no reason why this relay used in the AMP internal signal path will not work (would like to know if it does). As I have said there is still some switch noise but to me it is so minimal it does not matter.

Remote Switching: in this circuit the on/off only requires any positive VDC input to the NPN transitor. Therefore, a remote latch switch can operate on a 3VDC to 4VDC voltage which can drive the indicating LED. you can have as many switch points on the one circuit (eg on on a remote footswich and one on the amp or amp mounted unit). I have converted 19inch Rack power rail units to a footswitch with 12 switches. Use Carling switches. A good source is guitar techs who convert old Wah pedals to true bypass. They replace the SPDT switches with DPDT. You can get the used SPDT swiches cheap and they are top quality and reliable.

I have more info if you are interested

Cheers

[BackwardsBoB]

If you've got the right supply...

The old Carvin amps I play with have +/- 15V supplies for the opamps they use as buffers, tone controls, reverb drive and return circuits, etc. They use lots of JFET switching, mostly driven by a pack of CMOS inverters. Here's the circuit that was pretty much standard for about ten years:

http://www.carvinmuseum.com/pdf/amps..._schematic.pdf

If you look at Q4, right above the red "30-x100-4B" board number in red, you can see that its base is driven by a circuit that generates a pulse whenever the channel is switched. The inverters make two complementary channel-select signals at A8.2 and A8.12, and the capacitors, resistors and diodes generate a positive pulse at Q4 on both transitions of the signals. Q4 drives a signal that momentarily blocks input to the PI at Q13 and to the reverb and effects loop at Q11, killing switching noise dead. The pulse is very brief, and you don't really notice it.

Works good. Looking at the schematic, it looks like the switching circuitry is as big as the whole pre-amp, but the parts are small. The circuit would run well off a single 12V supply.

Obviously, if you can avoid the switching noise in the first place, you can avoid this part of the circuit, but if you're using relays or you are already doing JFET switching, the "glitch mute" itself doesn't add much circuitry. You could use the relay drive signal, a transistor to make an inversion of it, then the pulse generator network, another transitor, and the gating JFET(s). Total cost is under a buck.

While you're there, check out the "output power" switch over on the right that drops the supply to the PI, and the many examples of JFET switching circuits.

[GainFreak]

If you look at Q4, right above the red "30-x100-4B" board number in red, you can see that its base is driven by a circuit that generates a pulse whenever the channel is switched. The inverters make two complementary channel-select signals at A8.2 and A8.12, and the capacitors, resistors and diodes generate a positive pulse at Q4 on both transitions of the signals. Q4 drives a signal that momentarily blocks input to the PI at Q13 and to the reverb and effects loop at Q11, killing switching noise dead. The pulse is very brief, and you don't really notice it.

This is neat but works only if you have to switch between two channels. What is the situation if you have 3 separate channels with 3 separate switches + 2 more for your reverb and FX loop and you want that mute impulse for all 5 buttons?