Relay Bypass – do I really need one?

[aquataur]

A look at claims to improve reliability and performance using a Relay Bypass.

I recently felt impelled to haul out my knowledge on programming microcontrollers and try that on a relay bypass. Invariably, I tripped over several statements on web pages that turn out to be untenable claims it you look at them closely. However, I do not assume that there is negative intention or self benefit behind it.
The subsequent sales arguments are found on the following source. The writings appear to be educative, but turn out to be sales arguments. As such they may serve well as a foundation for thinking.

• Typical 3PDT latching mechanical switches are noisy, hard to press, and prone to mechanical failure after repeated use. Source

Noisy may imply electric noise, but indeed what is meant is mechanical noise that becomes audible by switch activation. While this may be of concern in a recording studio (where you usually do not use a performance switch), nobody will hear it during a loud live performance.

Hard to press; sounds ridiculous to me as an argument. Usually you benefit from a certain tactile feed-back in order to know that the switch reacted rather than have to look at some visual indicator.

Prone to mechanical failure; while I concede that many (even expensive boutique units) are equipped with the lousiest switches you can imagine, a quality switch of sorts will last 30k activations.

30.000 is a huge number. For your beloved stomp-box, you are willing to change such a switch if it dies after years, but honestly, most of the stomp-boxes I have (made) abide in dormancy and never saw even 100 activations. A cheap momentary switch may not even see 10k activations if it does not die because of its lousy construction before that.

So for your touted 1 million relay activations, what diamond-draped momentary footswitch can you obtain that outlives lives that?

It will be gone, you guessed it, at 30k activations maximum. And you will have to buy another one, if you can get one that fits your assembly mechanically. And it will cost you as much as a 3p3t if you want a good one.

• 3PDT switches are the main cause of guitar pedal failure. These switches are not particularly though, and they often break, especially as we smash it continuously with our feet on stage.

And the momentary footswitch is better?

• A classical high quality 3PDT switch is rated for 30,000 activation cycles. With relay bypass, we use a relay that will play the mechanical role of connecting ins and outs. Relays are usually rated between 10 and 100 millions cycles! Thus, this system is much more reliable. Source

I have yet to see the 100 millions on a datasheet. How can it be more reliable overall in the lights of the aforementioned?

You also have to take into account that a micro-controlled relay bypass actually means a huge increase in complexity. If a mechanical switch fails, big deal. It is a bit cumbersome to replace it, but everybody can do it. Everybody in this forum at least. But you likely have to replace the whole board. Microcode? Proprietary. Gone. Relay? Ever tried to unsolder even a two pin component like a resistor from a PCB with plated-through holes? Lucky if it worked. A transistor? A nightmare. A relay with 8 legs? Impossible without damage of some tracks or pads.

If a commercial unit incorporates such a circuit part, then, if you are lucky and warranty is still intact, you get a new one unless they claim you have abused it, but if warranty is over – that’s it. There is nothing to repair.

Somewhere in the dark past I have read in an industrial design paper that the failure rate rises with the square of the number of components on the PCB. If industry would truly follow their own dogma, they would advise against installing such a unit. Look at a 60 year old valve amplifier – you can get that working again like new. Look at a contemporary „professional“ device – none of that will work in ten years, corrosion will have eaten away all the delicate tracks irreversibly, and your eeproms will have failed. Clearly, all of that is for the benefit of cheap and fast production. It is not for your benefit in terms of longevity and repair-ability. Provided you want an effect that long…

• Replacing a switch is easy but sometimes quite annoying, as you have to rewire everything. To make replacement easier, you can use 3PDT miniboards and ribbon cables. This system is used in many commercial pedals. Source

This is even worse. You can easily replace individual wires, but try to get a double-sided PCB with plated-through holes out of 6 or even 9 contacts. A nightmare. Been there, done that.

• When wiring the pedal, cables are not always really fixed, and the solder can break and cause a disconnection in your wiring. This can happen quite easily on the 3PDT jack where there is a lot of wiring going on, or even on input jacks. (…) But the best way to prevent it is to include the wiring directly on the PCB, (…)

Again only good for production and a nightmare for repair. The most common source of failure on various „modern“ units I have encountered are the jacks or control units where they are soldered directly to the PCB. And if you are that bad at soldering, you won´t be able to solder the direct connection either.

• Input jacks and the 3PDT can sometimes become a bit loose, (…). It is really easy to fix, (...) either your tight it a lot with a pair of pliers, either you can include jacks directly on the PCB.

As above. Just the producer’s joy.

• However, it has one downside: if the jack input is not working anymore, it is quite boring to replace because desoldering and resoldering of the jack is not easy. ). (...) I plan to include the jacks directly on the PCB. It avoids a lot of wiring (reason 2), and prevents loose jacks.

Only a production merit. Bad service-ability, no tinkering for proto-types, no loose jacks if no exposure to repeated mechanical forces only.

• My recommendation would be to use PCB mounted potentiometers, which make the PCB fixed in the enclosure.

Production benefits only. No yanking in and out during protoyping.

So far the misleading information which I consider marketing rap. I don´t mind people claiming such things, but those pages are opinion shaping, and the claims are brass neck.

However, for the perfect momentary switch there is a solution: use an actuator (the mechanical part of a switch that uses an external switch) in conjunction with a micro switch rated for a million cycles. Omron B3FS is the first choice. Those are easily available, but this creates another mechanical problem.

Beyond all that, there are some never mentioned potential other issues that may be introduced with the installation of a relay bypass:

• Pops. Mr. Black explains in his great article on „What causes switches to pop“

what are the sources of this unpleasant effect. If those things are not fixed, even the best relay bypass will not rescue things, although the marketing promises exactly that.

Relay contacts are contacts too, although they may exhibit those problems on a smaller scale.

• Current spikes: Huge current flows are activated upon relay turn-on, worse if an LED is wired in parallel for indication. You cannot turn on a relay gradually or it will chatter, so this is inevitable. Is your pedal’s power supply management fit to handle that momentary surge without audible interference? It certainly was on prepared for that. If you can hear a LED’s inrush current pop, you can hear the relay too.
• Capacitive and/or inductive coupling: I have a wah pedal that picked up the switching spike by coupling. This is a valid concern.
• Mechanical issues with relays: some relays are found to be microphonic.

  True bypass with latching relay – Stompville

  stompville.co.uk

There are of course some genuine criteria that speak for a microcontrolled relay-switch, mostly if you want to achieve secondary or complex switching actions like:

• Radio buttons; Specialty product for proprietary use. If you were going that way, you would want to go about it differently.
• Latching relay for low power consumption (who uses battery operated stomp boxes any longer?)
• Remote control . Specialty product for proprietary use.
• Toggle action vs. momentary action upon button activation. This will not be the case very often.
• Mute-ing. This can make an already well behaved unit totally quiet (see Mr. Black above), which is of particular interest when on the beginning of a long gain chain.
• Secondary switching, like additional ground switching with an auxiliary open collector transistor. Specialty product for proprietary use.

So, do you still think you need a relay bypass? Opt for the least hassle and get a high quality 3p3t toggle switch.

 

[jwin615]

I think the biggest benefit of relay switching, which you didn't cover, is when it is implemented near the IO. This avoids the extra capacitance of the internal wiring or omits it completely, as well as possible noise picked up on those runs.
Signal running from a jack, down a short trace through a relay to another short trace to a jack beats any 3pdt from that standpoint.

 

[mckillacuddy]

All hail the mechanical footswitch. I am not a fan of all the extra parts and extra cost of relays, controllers, optocouplers, &c. And to be honest I am a fan of the snap of a good stomp. It’s a sound we can all recognize in our sleep. AND the more we use them the better coat and quality can get. I member when a 3pdt was like nine dollars! Nowadays you can get a heat resistant 3pdt rated for 50000 cycles for a fraction of that in whatever color you want. Hell yes

 

[aquataur]

The extra capacitance (against ground?) for this small piece of cable can not be more than a few pF. Where should noise come from? External noise is completely shielded by the metal case. The switching pops come from either inrush currents, or from voltage steps caused by the contacts itself and/or the huge DC blocking caps. I do not agree with you for the sources.

 

[jwin615]

The extra capacitance (against ground?) for this small piece of cable can not be more than a few pF. Where should noise come from? External noise is completely shielded by the metal case. The switching pops come from either inrush currents, or from voltage steps caused by the contacts itself and/or the huge DC blocking caps. I do not agree with you for the sources.

The circuit. Any reason why you would run shielded cable internally.
Clocks, buck/boosts, etc.
If you want to split hairs over 30k/1M cycles, why not split hairs over a few pF?

 

[aquataur]

Clocks... I have a circuit with an LFO. There is some evidence with feedthrough. The PCB layout, although stringent, inevitably is vulnerable, so the cable to the input... Point taken.

So a proprietary design, as you describe it above, will help. But not one of the retro-fit units that reside at the footswitch. For this a buffer right at the input may help.

 

[falco_femoralis]

I have an idea for a build that requires relay bypass simply because I want to mount the switch ontop of a PCB and there won't be room for a mechanical switch.

Other than that, it comes down to whether you care about cost or about theory. If you want to keep your cost down and make your builds more accessible to potential buyers, then go with the mechanical switch. If you are building for yourself, or are building a piece that is going to be expensive anyway and cost won't be as much of an issue for a buyer, then go with the relay switching. If you want to split hairs then go for it, but like anything there is a time and place.

So no, you don't really need one, but if you like the idea, then build a pedal with one and try it out

 

[PedalBuilder]

[With the caveat that the cost analysis predates the recent tariff nonsense]
Relay switching isn’t meaningfully more expensive than 3PDT switching if you are careful with your parts sourcing. I used JLCPCB to assemble my small relay bypass boards, with the exception of the ATTINY13A microcontroller, which I source, program, and assemble myself. The cost per build is a little under $4, including the SPST switch. A high quality 3PDT switch and a corresponding 3PDT breakout board cost about the same, maybe a little more. Cost aside, 3PDT hardware bypass does one thing—toggle between bypass and engage. Depending on how you program the microcontroller, you can use a smart relay bypass do the following and more: momentarily toggle states while the footswitch is held down; to always power up in a bypassed (or engaged) state; to selectively function as a momentary switch if the switch is held down when the effect is connected to power; and communicate with other circuits with any otherwise unused pins on the microcontroller.

I’ll also add that I have seen quite a few examples of pedals whose 3PDT switches have broken due to vigorous stomps, overheating during soldering, and/or poor manufacturing. When one of those fails, you are out $3 or so for a new switch, and you need to desolder at half a dozen wires. I haven’t seen a failed SPST switch, but if one failed, it’s only two short wires that need to be replaced. It also wouldn’t surprise me if a momentary SPST lasts longer in real world use, as they are mechanically much more simple than a latching switch.

So with cost being about equal, I’ll gladly take relay bypass, as it offers more capabilities, is more reliable, and easier to service. Also designing my own relay bypass boards and code was a fun challenge in a way that designing a simple 3PDT breakout board was not.

 

[aquataur]

Have you actually read my writing? I never questioned the cost of a relay bypass system.
A look at my local dealer tells me that a Gorva 3p2t switch, which I trust, is about 6 bucks, and one of the momentary ones I would trust like Alpha or APEM, are 9-10 bucks. I have no means of validating their respective quality. Nobody has, everything is subjective.

I don´t care about the money. I was writing above that you can find a lot of claims that may create the illusion that a relay bypass will solve all problems. It doesn't.
If one talks about reliability one has to look at the subject with a much elevated view. Behold the added complexity. There is much more to fail.
Yes there is all this fancy stuff you may do with a microcontrolled system (or even an analog equivalent). As I said, specialty.
For a normal stomp box, does it make my (say) fuzz face switching more quiet? My wah pedal? Does it live in ten years?

For me, it comes down to the muting function. That's a real difference to the analog counterpart. Is the added complexity (read: potential to fail and introduce other problems) worth it?

Everybody may proceed as they wish. Take those lines as an changed perspective on the subject.
And you, Pedalbuilder, talk from a perspective of somebody, who has to make a living. I don´t.

 

[PedalBuilder]

Have you actually read my writing? I never questioned the cost of a relay bypass system.

Did you read the post above mine? Other people have replied to this thread besides you.

A look at my local dealer tells me that a Gorva 3p2t switch, which I trust, is about 6 bucks, and one of the momentary ones I would trust like Alpha or APEM, are 9-10 bucks.

If those are the prices that you’re seeing, I’d recommend trying online retailers. You can get the same or comparable switches for a fraction of the cost.

I don´t care about the money.

That’s a nice luxury to have. For those of us who don’t have an endless supply of money, the cost of the hobby matters.

I was writing above that you can find a lot of claims that may create the illusion that a relay bypass will solve all problems. It doesn't.

I don’t think that anyone who patronizes this site actually thinks that a relay bypass will solve all problems.

If one talks about reliability one has to look at the subject with a much elevated view. Behold the added complexity. There is much more to fail.

A simple circuit of a voltage regulator, a microcontroller, a passives, and a relay is exceedingly unlikely to fail, especially if you incorporate reverse polarity protection. The bottom line is that a latching three pole switch is a lot more likely to break in real world use than a momentary single pole switch.

Yes there is all this fancy stuff you may do with a microcontrolled system (or even an analog equivalent). As I said, specialty.

You can implement these features in any pedal, from a basic fuzz face to an analog flanger. They add creative possibility without adding any cost or build difficulty, so why not include them?

For a normal stomp box, does it make my (say) fuzz face switching more quiet? My wah pedal? Does it live in ten years?

I’ve use my relay bypass in dozens of builds and haven’t noticed any change in switching noise. Others have done the same.

For me, it comes down to the muting function. That's a real difference to the analog counterpart. Is the added complexity (read: potential to fail and introduce other problems) worth it?

There’s a potential, sure, but if the real world likelihood of failure is less than the real world likelihood of a latching switch failing, then I’m not sure that the potential is worth getting hung up about.

Everybody may proceed as they wish. Take those lines as a changed perspective on the subject.

Honestly, you come across as somebody who is looking for an argument, not a conversation. I’m not sure if that was your intention, but it’s how you come across to me.

And you, Pedalbuilder, talk from a perspective of somebody, who has to make a living. I don´t.

Does this mean that you think that I build pedals for a living? I don’t. I build for fun, for my own use. But like most of the other people here, I’m also not Croesus or Elon Musk, so I do care how much money I spend on my hobbies.

 

[CR0SSBL4DE]

I'm an electronic engineer. A mechanical 3PDT switch hand soldered (specially if you are soldering individual cables by hand) is definitely way more prone to failure than a relay.
As a hardware product maker, when you think quality, you don't think a design that is easy to fix. You think a design that is robust enough to avoid the need of fixing. Take an iphone as an example.

 

[Gordo]

Wow, this got messy.

I was gonna go with "yes you do" but it seems a tad underthought at this point.

Honestly, I add it because it's cool and/or I can. In the truckload of pedals I've ever built I believe I may have broken one DPDT/3PDT and I can count on two fingers how often I've had to troubleshoot noisy connections. I've likely just jinxed myself...

 

[aquataur]

If those are the prices that you’re seeing, I’d recommend trying online retailers. You can get the same or comparable switches for a fraction of the cost.

That is the online prices.

That’s a nice luxury to have. For those of us who don’t have an endless supply of money, the cost of the hobby matters.

Not that I have money to burn, but if a switch each month costs three bucks more or not, that don't break my bank.

I’ve use my relay bypass in dozens of builds and haven’t noticed any change in switching noise. Others have done the same.

I am not sure what you are saying by "change", for better, for worse? In connection with my quote I guess you say the relay bypass does not make things quieter electrically. Which is what I expect.

Honestly, you come across as somebody who is looking for an argument, not a conversation. I’m not sure if that was your intention, but it’s how you come across to me.

Your perception is your perception. In your first reply, you did not address any of the points I was making to an extent that I was wondering, if you have read them. During dealing with the subject I found that many of the claims people make are standing on shaky grounds and should be taken with a grain of salt.

 

[aquataur]

As a hardware product maker, when you think quality, you don't think a design that is easy to fix. You think a design that is robust enough to avoid the need of fixing. Take an iphone as an example.

That is a valid approach. However it is the approach of a company, of industry. They have a certain life-cycle in mind, and when this is expired, the defective unit is discarded. Those units are not made to be repaired. Take the mobile phone. A typical wear thing is the battery. This could be made replaceable (and some phones had this), but this is not wanted. Of course a mobile phone is a bad example because they expire naturally by means of updates and incompatibilities caused by that.

But this is a different approach. Early hardware units used to come with a complete service and repair manual. They were designed to live long (-er at least). This is gone, the approach is gone.

I don´t live in the delusion that any pedal or any of my pedals is free of trouble. So my approach is a completely different one. It is that of an knowledgeabe amateur. A company can affort to have a huge number of design cycles on a product, or man-power, so that when it is released, there is little to fail.
Back to the main topic, I see that the motivation to use a relay bypass is more a practical one rather than an electrical one. A few good suggestions have been made.

 

[szukalski]

As a hobbyist, a relay based bypass with an MCU is appealing because it opens possibilities to do a multi-function IO board and be more efficient with my time.

For example, the below prototype has over-voltage protection (and sensing via MCU), polarity protection, 5V/9V and VCC out. Without footswitch and jacks, it will probably be €3-4 at hobby scale. SMD populated direct from the vendor. We're not doing mods on it, but it's a disposable module if repairs are needed. Lifetime wise, it's compatible with mechanical switching so could be replaced with whatever the tech has on hand.

Add the jacks/switch, connect 5 cables to the board with JST connectors and you're up and running. If you combine that with a pre-drilled enclosure then a full build can be pretty fast. This is a god-send if you have limited time (family commitments).

 

[Big Monk]

Ultimately, I ditched relay based switching due to cost, not reliability.

It's cool for onesey twosey DIY projects but is at least 3x the cost of mechanical switching.

 

[HamishR]

For me time is not too much of an issue, nor is money a major issue, but I don't spend a lot of money on my builds. For me simplicity is my guiding principle, so I just use a boring 3PDT switch with an LED ring. One less hole to drill! I'm not particularly electronically minded either, so some switch I can wire myself is fine. However I am suspicious of any extra electronics not related to the sound in a circuit - probably because of my lack of actual theoretical knowledge. But also because anytime I have used more complex gear which uses relays, especially amplifiers, I feel that the connection to the sound is affected. It could well be that the amps which have more switching (via relay) are built more cheaply or something, but I almost always prefer the sound of simpler amps.

Now all of my amps are built by me, so it's a moot point because they're all stoneage technology and have no channel switching or any relays anywhere.

But for every example where simplicity may be shown to be superior there is an example of where more complex technology is surprisingly resilient. I have a few old (1988) Ibanez delays like the PDM-1. The PDM-1 is a small, pocket calculator sized pedal with an LCD and push buttons for programming. It is housed in a plastic box which should be hopeless for a pedal. It also has a battery somewhere inside to power the memory as the pedal has 19 memory patches. This battery was supposed to run out years ago, but each of my pedals still works perfectly nearly 40 years after they were produced! My main one has seen many gigs and is a little worse for wear but works just fine. The only reason I don't use it any more is because it's not TBP and the buffers slightly affect the signal in ways I don't like. I'm sure players less anal than me wouldn't care!

So I have no real point to add except that maybe there is no answer to this question which will work for everyone. I'm suspicious of any extra electronics other than strictly necessary but that doesn't make me right. It just guides how I like to make stuff.

 

[aquataur]

example of where more complex technology is surprisingly resilient. I have a few old (1988) Ibanez delays like the PDM-1.

Yes things can be done resilient. I have a 40 year old Philips CD player, actually two of them, that run like a breeze.
At these ages a Sony will be ashes.

Well, I see that there are as many opinions as people.

 

[CR0SSBL4DE]

Also, p

That is a valid approach. However it is the approach of a company, of industry. They have a certain life-cycle in mind, and when this is expired, the defective unit is discarded. Those units are not made to be repaired. Take the mobile phone. A typical wear thing is the battery. This could be made replaceable (and some phones had this), but this is not wanted. Of course a mobile phone is a bad example because they expire naturally by means of updates and incompatibilities caused by that.

But this is a different approach. Early hardware units used to come with a complete service and repair manual. They were designed to live long (-er at least). This is gone, the approach is gone.

I don´t live in the delusion that any pedal or any of my pedals is free of trouble. So my approach is a completely different one. It is that of an knowledgeabe amateur. A company can affort to have a huge number of design cycles on a product, or man-power, so that when it is released, there is little to fail.
Back to the main topic, I see that the motivation to use a relay bypass is more a practical one rather than an electrical one. A few good suggestions have been made.

There are noise and signal integrity reasons too.

 

[aquataur]

There are noise and signal integrity reasons too.

Please elaborate. I am all ears.