PCB: low voltage On-Off switch drives AC mains relay \ includes soft start .. H9KPXG

Member
Joined 2017
Paid Member
Thanks, @RickRay , but this/my board seems to be malfuncioning, the bulb is fully lit from the moment on I plug it in, and the led-on is on regardless of the switche‘s position. Led-off does not light up…

I believe I‘m in for a fiddling (debugging) session, probably starting with
  • a general inspection for lousy soldering
  • misplaced parts
  • broken parts according to @Mark Johnson ‘s list mentioned recently, this (#699)

The thing that baffles me is that the board consistently behaved like this and that there is some functionality (there’s a relay clicking when a power-cut occurs/the rear switch is activated…
 
Last edited:
Member
Joined 2017
Paid Member
It wont matter while using a momentary switch. But sins a toggle switch doesn't produce a positive going edge while switching it off, the D-flip flop has to be disengaged by pulling P6. R9 now becomes the data path and the circuit will, with some delay, follow the state of the toggle switch.
I am not sure about this, OmeEd, are you describing how it should work, or is this a debug procedure starting with P6?
 
It's an minimalist explanation intended to describe why the position of the front switch during power-on matters, (when using a toggle switch for the purpose). This as you observed the circuit automatically passing to the on-state, and imo that's not necessary a fault (when using a toggle switch).
 
When powering it up with the back-switch while the front-switch is in the off-state, the circuit should start in the in the correct state, namely OFF. Otherwise it would take an on-off cycle to synchronize. Two more remarks.

1. You mention using a bulb, I take it you use it as a load (As substitute for the main transformer). When using it in series with the circuit (some people do this while initial testing an amplifier), it might create problems with SMPS.

2. This circuit works way more elegant when used with a momentary push-button. Why do you insist on using a toggle-switch?
 
  • Like
Reactions: 1 user
Member
Joined 2017
Paid Member
Rhank you again, @OmeEd

When I switched it on the first and second time, I indeed used a dim-bulb device in series with the circuit). It went as erwartet: bright on start, dim down to invisible.

(And, to be frank, the first run was with the whole PSU connected, it works)

I then connected H9KPXG directly and used the bulb (40W) as load.

H9KPXG never showed me an off state.

I have the front-plate cut to fit this one form-factor of the switch, out of habit and because of its pleasant look. I‘m already looking for a momentary switch that would fit the slot… (but if it would be working with what I have, I would be happy with that)
 
The thing that baffles me is that the board consistently behaved like this and that there is some functionality (there’s a relay clicking when a power-cut occurs/the rear switch is activated…

If it were me... It seems the board may be working fine when using the rear switch - that is, relay clicking when that switch is used - then I would move on to working exclusively with the front switch. When the board is turned on, get your multimeter across the light bulb and just make sure you're not sending high current/voltage to the bulb. If all seems well, connect your transformer and try again.

(And, to be frank, the first run was with the whole PSU connected, it works)

Yes, this is encouraging!

Somewhere in these 39 pages there are two comments I seem to recall. If only I could find them... 1) There may be a minimum voltage needed to get the H9KPXG to work as intended. This is why adding only your transformer might be the right next step. 2) You may need to cycle the H9KPXG a few times before it works as intended. It probably has to do with the logic chips and a couple of capacitors...

EDIT: Found a reference:
https://www.diyaudio.com/community/...ncludes-soft-start-h9kpxg.354971/post-6229775
 
  • Like
Reactions: 1 user
H9KPXG never showed me an off state.
The picture of the assembled board you posted in #746 reveals that the CD40106 is not correctly inserted in her socket. Pins 1 to 7 are bend outside of the socket.


IMG_6377(1).jpeg
 
  • Like
Reactions: 1 user
I got a bit help. Rereading #766 my attention was caught by the 6 word sentence "H9KPXG never showed me an off state", this made the 40106 first in the chain of suspects. It was under active investigation when I spotted the problem. If it proofs anything it's the importance of factual fault reporting and the emergent power rising from a joint effort. Thank you all.
 
  • Like
  • Thank You
Reactions: 4 users
Member
Joined 2017
Paid Member
I would like to confirm the board is working.
Haven’t yet put any load on it, just switched it on and it instantly turned on the OFF-led. (With the switch obviously in the open position)
On is on as well, and the relay‘s clicking just as expected…

This was one of my more ridiculous (not hazardous) bungles. (The most idiotic one I keep secret
🤐
)

Again many thanks for all the help and encouraging and support and this project to Mark, OmeEd, RickRay, IAIMH, and the community!