In other words you are human. I am not sure what else to check at this point. So this is a schematic that has been converted to a PCB or just a protoboard?

It's nearly all 0402.
Shamefully, considering I was the go to EMC circuit design guy, I left the bandwidth limiting/ suppression caps out and decided to stack them as there is very little space and there has to be a clearance around the board.
This sits in one of those car cigarette lighter usb charger cases. It's ground flooded, with Vcc running along a thick track the other side, which is also ground flooded.

 

It's nearly all 0402.
Shamefully, considering I was the go to EMC circuit design guy, I left the bandwidth limiting/ suppression caps out and decided to stack them as there is very little space and there has to be a clearance around the board.
This sits in one of those car cigarette lighter usb charger cases. It's ground flooded, with Vcc running along a thick track the other side, which is also ground flooded.

If you can hand solder 0402 I am impressed! I have had difficulties hand soldering 0603. On all my hobby projects I tend to stick to 0806. I think I can leave you with a good piece of advice moving forward. Anytime you need an OpAmp always use a part that is designed for that task. When you need a comparator only use parts that are designed to be comparators. I personally feel this will lead to significantly less headaches for you on all future designs. I realize your current design is fairly far along now that you have a PCB. I am not saying you must redesign your schematic and PCB. But if you keep having issues trying to use the opamp as a comparator you may have little choice in the matter. Much success! This board should 'in theory' work, hopefully you can figure out a solution and not have to redesign. By the way, I like your board design for cigarette lighter looks nice

 

In other words you are human. I am not sure what else to check at this point. So this is a schematic that has been converted to a PCB or just a protoboard?

Well I've been running on the battery, connecting disconnecting for the last..... probably 8 hours and not a dead chip in sight!!
So your analysis was spot on.
Something is going on with my PSU that is not cricket.
I'm loathe to get rid of it as it's built like a brick and not a switch mode in sight. They even issue all the schematics and calibration procedure!!! Don't get that today.
I will stick my scope on it and set it to trigger on over-voltage. For the time being it's a search for a new bench switch mode PSU.
Thank you for your time, your simulations. This wasn't a quirky ST opamp, it was it's Vcc as you suspected!

 

If you can hand solder 0402 I am impressed! I have had difficulties hand soldering 0603. On all my hobby projects I tend to stick to 0806. I think I can leave you with a good piece of advice moving forward. Anytime you need an OpAmp always use a part that is designed for that task. When you need a comparator only use parts that are designed to be comparators. I personally feel this will lead to significantly less headaches for you on all future designs. I realize your current design is fairly far along now that you have a PCB. I am not saying you must redesign your schematic and PCB. But if you keep having issues trying to use the opamp as a comparator you may have little choice in the matter. Much success! This board should 'in theory' work, hopefully you can figure out a solution and not have to redesign. By the way, I like your board design for cigarette lighter looks nice

[/QUOTE]

Thank you!

Yes I have a stereo microscope so 0402 is not a problem, plus I was used to soldering them when I was on the Ericsson Baseband team 20 years ago!!!!! My hands have a little bit more of a tremor. I tried a 0201 recently and had a bit of ado with it!!!

Yes, thank you for the advice. It was a matter of space and laziness!

One thing I learned in this project was running a DC line under a surface mount capacitor that is used for slow charging is a very bad idea!! I've had to add loads of conformal coating to this rogue track!

I send you my regards and many many thanks!

 

Well I've been running on the battery, connecting disconnecting for the last..... probably 8 hours and not a dead chip in sight!!
So your analysis was spot on.
Something is going on with my PSU that is not cricket.
I'm loathe to get rid of it as it's built like a brick and not a switch mode in sight. They even issue all the schematics and calibration procedure!!! Don't get that today.
I will stick my scope on it and set it to trigger on over-voltage. For the time being it's a search for a new bench switch mode PSU.
Thank you for your time, your simulations. This wasn't a quirky ST opamp, it was it's Vcc as you suspected!

Thanks! I am glad you got it figured out. Happy to help.

 

As you see this is powered by 13V. It is meant to be used in a car. It detects when the car engine is turning by the spikes appearing on the 13V line.

This is far too unreliable. Detecting alternator voltage rise is better but still not that reliable as engine might not give much of a rise at idle when heater, lights, etc are on. The only really reliable method is detecting engine revs. Depending on the actual car this can be incredibly easy or may need some additional hardware. Some cars may have a 'running' output from the ECU or cluster but all cars have a timing signal based on crankshaft or camshaft rotation for ignition and/or rev counter. Converting the latter to a 'running' signal can be done with a simple 555 timer circuit and some input conditioning.

Worst case, a simple detector reflecting of a crankshaft or camshaft marker is an easy is an easy add-on

What car is it?

 

Hello there

It's nearly all 0402.
Shamefully, considering I was the go to EMC circuit design guy, I left the bandwidth limiting/ suppression caps out and decided to stack them as there is very little space and there has to be a clearance around the board.
This sits in one of those car cigarette lighter usb charger cases. It's ground flooded, with Vcc running along a thick track the other side, which is also ground flooded.

Just as a supplement for those following this thread (Metro code) (inch code)
https://forum.allaboutcircuits.com/...-size-and-color-standards.163700/post-1438251
Pardon the intrusion.

 

Thurlby and Thandar PSU

If it's the 0 - 30v, 3 or 6A, 1 - 4 channel linear type then it's almost certainly one of the clones of a old Bosch design - my Tenma, GW-Instek, and many others sold near identical units from the '70s through to around 2005 before most went over to a digital SMPS design though there were hybrids with digital controls but linear backend.
Anyway, to keep the linear pass transistor from being over stressed the design for the main output channel(s) has a transformer secondary with multiple taps that are selected by relays to attempt to keep the raw DC input to the regulator within a few volts of the desired output. You can sometimes hear them clicking over as you vary the output voltage, depending on the quality of the relay. These PSU feature an output on/off button to turn all outputs on or off simultaneously. Adjusting the output voltage while 'on' leads to wild spikey voltage changes as the relays click over and is not recommended. Noisy carbon track pots don't help either. My user manual does warn not to change the output voltage while it is 'live'!

 

This is far too unreliable. Detecting alternator voltage rise is better but still not that reliable as engine might not give much of a rise at idle when heater, lights, etc are on. The only really reliable method is detecting engine revs. Depending on the actual car this can be incredibly easy or may need some additional hardware. Some cars may have a 'running' output from the ECU or cluster but all cars have a timing signal based on crankshaft or camshaft rotation for ignition and/or rev counter. Converting the latter to a 'running' signal can be done with a simple 555 timer circuit and some input conditioning.

Worst case, a simple detector reflecting of a crankshaft or camshaft marker is an easy is an easy add-on

What car is it?
Hello there

Just as a supplement for those following this thread (Metro code) (inch code)
https://forum.allaboutcircuits.com/...-size-and-color-standards.163700/post-1438251
Pardon the intrusion.

This is far too unreliable. Detecting alternator voltage rise is better but still not that reliable as engine might not give much of a rise at idle when heater, lights, etc are on. The only really reliable method is detecting engine revs. Depending on the actual car this can be incredibly easy or may need some additional hardware. Some cars may have a 'running' output from the ECU or cluster but all cars have a timing signal based on crankshaft or camshaft rotation for ignition and/or rev counter. Converting the latter to a 'running' signal can be done with a simple 555 timer circuit and some input conditioning.

Worst case, a simple detector reflecting of a crankshaft or camshaft marker is an easy is an easy add-on

What car is it?

Hello Irving
Skoda Octavia2
I haven't shown the rest of the circuit!
If the signal is present for 3-4 seconds it turns the 5V reg on.
If it's not then it turns off after 20mins.
It's pretty constant and I can just plug it into my cigarette lighter, no wiring required.
This was just a "experimental" circuit, to get my hand in again and go through the formal process of PCB manufacture. It was never going to be a Bob Pease job!!!!
As for the PSU, it does not use any relays.
It's a straight up linear regulator circuit with a fixed 40V+ supply.
I would download the manual to this site, but I don't want to clog up this site's arteries!
It's a T & T PL-320. The pdf includes schematics calibration procedure etc.
Most impressive.
Yes cranking the voltage up is not a great idea, just an act of desperation!! It's pot is heavily low pass filtered with a -3dB point at 21Hz
I'm more interested in its response with my circuit, it murdered, attached.
All the PSU's capacitors on the regulation board were replaced at "regular" intervals as were the large reservoir caps on the transformer bridge rectifier PCB and all their ESR's are fine.
The one thing that stops all of this dead is a massive reservoir cap on the output!!!!
The graph is the output at a PSU setting of 12V. Ignore the tail off after it "reaches" 12V!

 

This is far too unreliable. Detecting alternator voltage rise is better but still not that reliable as engine might not give much of a rise at idle when heater, lights, etc are on. The only really reliable method is detecting engine revs. Depending on the actual car this can be incredibly easy or may need some additional hardware. Some cars may have a 'running' output from the ECU or cluster but all cars have a timing signal based on crankshaft or camshaft rotation for ignition and/or rev counter. Converting the latter to a 'running' signal can be done with a simple 555 timer circuit and some input conditioning.

Worst case, a simple detector reflecting of a crankshaft or camshaft marker is an easy is an easy add-on

What car is it?

I think one of his goals was a simple device that plugs into a cigarette lighter without adding additional stuff under the hood. But you are right, I am not sure how reliable this is.

 

I think one of his goals was a simple device that plugs into a cigarette lighter without adding additional stuff under the hood. But you are right, I am not sure how reliable this is.

So you won't be using it in your life support machine then? :0D