Mictester's CMOS Relay Bypass Module

[Silver Blues]

A little while ago I posted a layout for this relay bypass circuit, and yesterday I had a chance to test it out. It works really well, but as it turns out I made a few minor mistakes in the original, including mixing up my energized and de-energized states and listing one row too many. So here is the corrected and verified layout.



It will fit inside a 1590B on top of the switch itself, but be sure to pick the lowest profile parts that are humanly possible. I put one into my Blueberry build and I could not fit another wire fragment in there if I tried. The Takamisawa is the best relay for this job IMO, but be sure to use the thinnest wire you can get your hands on because the pins are quite short and if your link is too fat it will prevent you from being able to solder it down. Sockets are the order of the day for most semiconductors in our line of work but I straight-soldered everything without issue. It's a nice circuit, I didn't experience any popping and the actuation was smooth and defect-free. Thanks to Mictester and mmolterax from FSB for developing this.

If you want to avoid the one standing resistor, you can use an eighth-watt device and solder it diagonally from the edge of the board to right above Q2.

UPDATE Sept 2022 - included an updated layout that removes the connections of the unused CMOS outputs to ground - should make the circuit more stable.

[bart]

This is awesome Silver_Blues. I tried this one, both Paul and Mark's versions and with Geezer's suggestion for a flyback diode but it still has an audible pop.

will be trying your layout in the next couple of days. Thanks a lot for this one!

[GeGhostRider]

thanks a lot! a great contribution.
I've ordered a few relays from the bay but I'd like to know if there's a reliable source for suitable switches.
the Carling ones are quite expensive and I only come across momentary cheaper ones.
where did you get yours Silver?
thanx again.

[Silver Blues]

I got mine from Tayda, these (though I got them off eBay and not their web store. I imagine they're the same). Best price to quality ratio I've yet seen, local stores were trying to milk $13 out of me for one of these. I've used these in three pedals now and haven't had an issue.

[GeGhostRider]

well, after some reading on the subject on FSB I thought latching ones were mandatory...but I may have confused things ( currently quiting smoking so...). I'll give it a try asap.
thanx again Silver

[Silver Blues]

Well, the guy who developed these originally developed a very simple circuit that used a SPST or SPDT on/off switch, however this version uses Schmitt triggers to toggle the coil in a latching relay with a momentary switch. There are a few user-submitted versions in the FSB thread but many are overly complicated - the best two IMO are the original latching one here and the non-latching one in my original post.

You do however need a latching relay. Any of the ones listed in the layout will work, I'm partial to the Takamisawas. Be careful though if you try and order NEC relays from eBay that you are getting specifically EA2-5SNJ and not anything else.

[GeGhostRider]

I've ordered the Takamisawas and the Tayda switches.
been wanting to implement this on a few builds for so long.
again many thanx for your contribution (and to the guys involved in development) and for taking time to answer.
I feel dead-silent switching comin'! ( obviously you can't hear it!)

[gul]

Hey! Which particular Takamisawa relay did you use? I'm excited to try your layout :)
Thanks, Aisha

[Silver Blues]

AL5WN-K. Here's the eBay seller I use, they have lots of 5, 10 or 20.

[gul]

Awesome, thank you!

On Apr 4, 2015, at 11:14 AM, "Silver Blues [via Guitar FX Layouts]" <[hidden email]> wrote:

AL5WN-K. Here's the eBay seller I use, he has lots of 5, 10 or 20.

---

If you reply to this email, your message will be added to the discussion below:

http://guitar-fx-layouts.42897.x6.nabble.com/Mictester-s-CMOS-Relay-Bypass-Module-tp19021p19162.html

To unsubscribe from Mictester's CMOS Relay Bypass Module, click here.
NAML

[purity]

Just wanted to add that there may be some errors with this layout. The unused inverter gates (on the right side of the IC) have inputs and outputs tied together. Only the inputs should be tied to ground or supply - the links under the IC need to be cut so the outputs dont feed back to the inputs.
On my build I remember I made alternating cuts to the links as shown in the layout. I think the ground connections are still ok, as the thing worked afterwards. The top right pin is tied to supply, the one below an input (tied to gnd) and the one below its output (and linked, so cut it), and so on. Follow that to find out which links connect outputs to their own inputs and cut them.

[Freppo]

There's a number of issues with that schematic, so I wouldn't recommend using it.

First, like purity mentions. The unused inverter outputs should not be connected to ground. Since it's an inverter, if the input is low (at ground) it will try to pull the outputs high, so it will basically cause a short between v+ and gnd (or very close to a short, which will draw alot of current) and it can damage the chip.

Second big issue is that it's missing a flyback diode between the relay positive and negative. It needs a diode (1N4148) with the cathode pointing at the positive side, to protect the other components from the discharge when the relay magnetic field collapses.

It's also way more complicated than it needs to be. You can also remove the whole transistor section which is unnessecery. A CD40106 inverter can sink enough current for most small relays. So a better way would be to keep the relay + side to v+ at all times and toggle the ground side insted directly from an inverter.

So keep the IC1A / IC1B section (this is the flip flop toggle) but remove C1 (not needed) and remove R1 (put a link there insted). Then connect the output of IC1B directly to the relay negative side. That's really all that you need. Don't forget the flyback diode, and you could use another inverter to have the LED isolated from the relay.

Cheers / Fredrik

[purity]

Thanks for the addition Freppo.
Am I wrong in stating that a flyback diode as youve described is not desired for this latching relay as the current to trigger the relay must flow both ways (or actually the flyback voktage should be blocked both ways)? Two zeners rated at 5.1V with the anodes tied together, and the cathodes across the coil are more (?) suitable for this application.

Also, the transistor part is to convert the constant high / low output of the inverters to a single bidirectional pulse towards the relay coil, which only needs a pulse and not a constant current. Minute current consumption is the design idea here in case you dont know the reasoning behind this circuit. Your alternative could use a more common and cheaper nonlatching relay though.

[Freppo]

purity wrote

Thanks for the addition Freppo.
Am I wrong in stating that a flyback diode as youve described is not desired for this latching relay as the current to trigger the relay must flow both ways (or actually the flyback voktage should be blocked both ways)? Two zeners rated at 5.1V with the anodes tied together, and the cathodes across the coil are more (?) suitable for this application.

Also, the transistor part is to convert the constant high / low output of the inverters to a single bidirectional pulse towards the relay coil, which only needs a pulse and not a constant current. Minute current consumption is the design idea here in case you dont know the reasoning behind this circuit. Your alternative could use a more common and cheaper nonlatching relay though.

I didn't know that this layout was for a latching relay. In that case you are right. My bad. :)
Two zeners would be the way to go for overvoltage protection (which the schematic lacks).
My version (which I have used alot with good results) uses a common nonlatching relay.

Aha, I see the point of the transistor section now. However, it would be possible to omit a few components and shrink the layout a bit by using a pair of the unused inverters as the edge to pulse converters insted of the additional transistors, and C1 and R1 are still not needed for the flip flop.

[Silver Blues]

Sorry for the necro-bump, but I've added an updated layout for this circuit that floats the unused CMOS outputs and leaves the inputs connected to ground for better results.