relay bypass in inverse state

[Zilla]

hi all.

i recently built this relay bypass circuit using a verified OSH park board: 

it was working as expected at first, but during the course of debugging something on the effect board i think i touched something and now the switch is operating in the inverse state:  the LED is on when the effect is bypassed and is off when the effect is engaged.  I have no idea what happened.

is there any way to reset the CD chip so the LED is back in sync with the switching??

[duck_arse]

I'd guess you've pulsed the relay by accident. if it is unset, try just tapping the upper transistors base to supply via a 10k resistor.

[Zilla]

Thanks. I'll give this a shot tonight.

[Transmogrifox]

I can't imagine how pulsing the relay could result in a permanent reversal of operation.  When the CMOS inverter turns on the LED it turns on the transistor to the SET coil.  That will set the relay.

When it goes the other way it will reset the relay.  If you somehow pulse the relay you might find it momentarily out of sync for one cycle of the switch, but after one press of the switch it will drive the state according to the LED indicator.

The only thing I can guess is there's a damaged component, solder bridge, or something happened at the 1M and 0.1u cap to defeat switch debounce function.  If the switch was not debounced it may be able to land in a state that doesn't agree with the LED because it might be oscillating at a high frequency.

Shotgun approach is to replace CMOS inverter chip and transistors then double check all soldering work and components.  If the components are good and the soldering is correct it's not possible for this circuit to get permanently stuck in a reverse indication state.

Are you sure you didn't accidentally wire your FX connections in reverse on the relay contact pins?  It might be this circuit is working exactly as designed, but you inadvertently changed the send/return FX and dry paths around on the pins.  ??

[Zilla]

Nope. like i said it was working as expected but i had to do some debugging on the effect board and i think when i unplugged the power jack i might have shorted something with my hand or something on the relay board and now it's operating in the inverse state.

[Zilla]

so i tried tapping the transistors and it worked fine to swithc the state, but then when i click the stomp switch it does nothing the first time and then reverts back to the original state.

just so i'm not doing something stupid:



this is the layout i'm using.

Input - guitar in from input jack
Output guitar to output jack
send - send to effects input
return - from effects output

ill probably swap the cd4069 and see what happens.

[Zilla]

I took a brand new board. same behaviour - LED is on when the switch is in bypass and off when the effect is engaged.  am i missing something completely obvious here?

(yes i know the OSHpark layout has the 100k resistor on pin 14 oF the cd4069 labeled as 10k)

[R.G.]

Do you have the correct relay?

I have not traced out the circuit board you show, but the dual-coil latching relays have one pin for "set", one for "reset". They don't toggle, they just go where they're told.

The LED goes on because the input latch has driven the output inverters and transistors to pulse the "set" pin. So LED on, set pin gets pulled low then released.

The only way this would result in the wrong state is if the wires from the "set" and "reset" outputs of the transistors didn't go the the right pins on the relay (perhaps by using a pin-compatible but wrong-function relay) or if the circuit on the PCB doesn't route the signals to the right places on the relay.

I'm conjecturing, I haven't traced the PCB, as I noted.

[Zilla]

yes. it's a Panasonic TQ2-L2-5V

http://www.mouser.com/ds/2/316/mech_eng_tq-850574.pdf

[Zilla]

Could the relay be backwards?  I installed it with the bar in the same orientation as johnk's schematic.

Will try flipping the relay 180 degrees and see what happens

[R.G.]

Before you flip the relay 180 degrees, try re-routing the two 10K base drive resistors so they go to the opposite transistor's base. Clip them out at the body of the resistor and pull out the cut-off lead. Then install two new resistors, crossing them so pin 8 goes through a new 10K to the base of Q1, pin 10 to Q2.

That re-routes the set/reset pulses to the opposite side of the relay, and doesn't affect the switching. More importantly, you don't destroy either the relay or the PCB getting the relay out.

I still haven't churned through the whole PCB, but one possibility (which the above corrects) is that what was the "set" pin and the "reset" pin on the relay diagram was mis-interpreted in laying out the PCB. The Panasonic relay data sheet is pretty arcane about what pin is what for those. So just swap the sense of the relay, and if it does what you want, you're there.

[Zilla]

RG, this makes perfect sense. I think this will fix the problem and get the LED in sync with the relay.

Will report back on results soon.

[johnk]

to make the LED come on when the effect is engaged, the LED just needs to have the - go to the LED resistor, and the LED+ go to +9V.

[Zilla]

Thanks Johnk!