Cab switcher pedal

[Buffalo Tom]

Hello. I wanna build a pedal that can switch between two speaker cabinets using a single amplifier head. My idea is to get a make-before-break type of switching using two relays with a capacitor that keeps the voltage about 500ms after the power has been removed. So if properly match the tube amp output, say 8 ohm,  with two 8 ohm cabinets. The amp always has a 8 ohm load except during switching where the amp will experience a brief 4-ohm load for 500ms. This is when the overlapping happens and both speaker cabs are on for 500ms.

IS THIS MISSMATCH SAFE?

I have made two different schematics. Is version 1 OK or do I benefit to switch with a transistor like version 2.

How do I calculate the capacitor value to get a 500ms delay?

Im using 12V TE Connectivity K10P-11DT5-12 relays. According to the Datasheet the operate voltage is 9.0 VDC and release voltage 1.2 VDC. https://www.mouser.se/datasheet/2/418/4/NG_DS_1308242_K10_0413-918460.pdf

Thanks

Version 1.


Version 2.

[antonis]

For 500mS delay, you need to calculate capacitor voltage after one RC time constant, which should be 37% of Vcc (full charge capacitor voltage), according to formulae Vc = Vcc X e^-t/RC..
Of course, you can set any RC delay point (e.g. 0.7RC for 50% discharged capacitor) according to convenient R & C items values..
(500mS time constant can be set either by 500k & 1μF or 500R & 1000μF or any other combination of 10^x X 10^-x to set units for sec..)

BUT, in your first case, R can't be considered relay's resistance 'cause 1N4007 flyback diode is set in parallel with relay winding..
(and I shouldn't count on diode's incremental resistance for setting RC constant..) 

For BJT drive, set time constant (4XRC) for about 20% cap charge (80% discharged), if you trust on 1.2V release voltage..!!
(for more precise delay setting, measure particular relay contact release voltage and take into account electro cap tolerance..)
Of secondary significance is BJT relay driver for longer life mechanical switch contacts due to lesser current through them..

edit: Not enough coffee yet.. 
You can always implement first version without flyback diode and count on about 50R coil resistance..
LED may suffer from reverse voltage kick but we can't always get what we want..

[Buffalo Tom]

For 500mS delay, you need to calculate capacitor voltage after one RC time constant, which should be 37% of Vcc (full charge capacitor voltage), according to formulae Vc = Vcc X e^-t/RC..
Of course, you can set any RC delay point (e.g. 0.7RC for 50% discharged capacitor) according to convenient R & C items values..
(500mS time constant can be set either by 500k & 1μF or 500R & 1000μF or any other combination of 10^x X 10^-x to set units for sec..)

BUT, in your first case, R can't be considered relay's resistance 'cause 1N4007 flyback diode is set in parallel with relay winding..
(and I shouldn't count on diode's incremental resistance for setting RC constant..) 

For BJT drive, set time constant (4XRC) for about 20% cap charge (80% discharged), if you trust on 1.2V release voltage..!!
(for more precise delay setting, measure particular relay contact release voltage and take into account electro cap tolerance..)
Of secondary significance is BJT relay driver for longer life mechanical switch contacts due to lesser current through them..

edit: Not enough coffee yet.. 
You can always implement first version without flyback diode and count on about 50R coil resistance..
LED may suffer from reverse voltage kick but we can't always get what we want..

Thanks for explaining antonis. So version 2 is probably the one to go with. Just four more parts.. Will start with buying the relays then breadboard this thing 

[antonis]

 


As R can be considered total equivalent resistance of any possible discharge path..
For BJT drive version, ignore Emitter intrinsic resistance & Vbe drop and consider 10k Base resistor as directly connected to GND..

!!!BUT!!!

FLIP C1 & C2 POLARITY..!!!

P.S.
It should be a good idea to place another resistor between Cap & +12V for also setting a delay (smaller than 500mS, of course) on cap charging 'cause power supply may suffer instant short when ON..

[Buffalo Tom]

Fund this calculator that was useful.
http://www.ladyada.net/library/rccalc.html

So keep the resistor 10K and change to 470uF cap would give me 495mS delay. IF the release voltage 1.2 VDC according to the Datasheet is correct  Will try




Yes flipped the caps! And added two more resistors as you suggested. Correct?

Here is version 3

[antonis]

You're confused a bit..

For version 2, you don't actually need to know relay's release voltage..
Relay will stay ON till BJT is set in cut-off region, which is at about 700mV for cap positive plate..
(it's quite a brute approximation 'cause we consider BJT as sheer ON-OFF switch but its transistive period is much less than 500mS so we can safely ignore it..)
So, delay time is set by C1/R2 & C2/R4 for respective circuits..

And make R1 & R3 values lower than R2 & R4 respective ones, or else 500mS delay set for upper relay cut-off  will be diminished by almost(*) equal lower relay delay turn-on..

(*) Actually, lower cap should need slightly less than 500mS for charging due to about 700mV "remaining" voltage when BJT is set in cut-off region..

P.S.
Calculator posted above is for CHARGING delay use (Vc = V-(V*exp(-t/(R*C)))..!!
You need a calculator which uitilizes Vc = V*exp(-t/(R*C)) formula..!

[antonis]

Another issue could be the use of BD139 in conjunction with 12V supply and 10k Base resistor..

For BD139 class 6, minimun DC current gain is about 40, so a Base current of about 1mA (11.7V/10k) might NOT be sufficient for relay working current..
You can, of course, lower 10k value accordingly to relay current needed but you'll also have to raise cap value for the same RC time constant..
(or get a BD139 of 16 classification)

[Buffalo Tom]

You're confused a bit..

Yes sometimes

So this might be a working one... Lower values on base resistor and the "input resistor"..  470mS delay (I hope 
I found the TE Connectivity K10P-11DT5-12 relays in my Metropoulos head-switcher and got inspired to use the same relays in this switcher.

Version 4

[antonis]

So, you want us to make all the calculations DIRTY JOB...!!! 

[Buffalo Tom]

Found the Aiken cab switcher.. So that's another way to go. Much simpler.. Don't know if its better..
Think I got the schematic right. http://web.archive.org/web/20130709182001/http://www.aikenamps.com/CabSwitch.pdf

[antonis]

Fair enough, if you can leave without delay..

But you have to use a n-p-n BJT (which is correctly named as 2N3904 but wrongly drawn as p-n-p..)

[PRR]

Speakers are 2-terminal devices.

[Buffalo Tom]

Fair enough, if you can leave without delay..

But you have to use a n-p-n BJT (which is correctly named as 2N3904 but wrongly drawn as p-n-p..)

Good observation.. Thanks.
Now updated with correct symbol.

[Buffalo Tom]

Speakers are 2-terminal devices.

Yes they are. All negative wires (sleeve) on the speaker/amp connections are tied together. But Im not sure I understand the picture you posted  Whats going on there?

[antonis]

Paul pointed that you can effectivelly set out of order a speaker simply by shorting it, with a SPDT swich, like your relay..

[PRR]

> All negative wires (sleeve) on the speaker/amp connections are tied together.

Why? It's not needed.

Your initial assumptions lead you down this jungle path of timing-circuits and relays. It does not have to be this complicated.

[Buffalo Tom]

> All negative wires (sleeve) on the speaker/amp connections are tied together.
Why? It's not needed.

The two cab switch schematics I found online was made that way.. But I understand now that it can be done differently.

Your initial assumptions lead you down this jungle path of timing-circuits and relays. It does not have to be this complicated.

The idea you posted looks very interesting.. A completely different way to do it. But is this safe for the tube amp and the speakers? And still safe if cab 2 is a Torpedo Live or UA OX instead of a real speaker