PCB problem with voltage

sparky 1

Joined Nov 3, 2018
756
I think isolating the oscillation could be done with an SDR (software defined radio 100kHz and up) taking advantage of the frequency domain and FFT for low frequency band. This setup might use DIY probe made from a coax connected to the antenna port terminated with a small loop that you direct at specific locations on the pcb while viewing the spectrum display. This would pick up the oscillation selectively at the same time harmonics that may be related. This I think gives big picture of spectral charecter of the board than what might not be as evident in the time domain. Using a spec analyzer with a low frequency range may not be available so I mentioned that SDR now goes down to 100 kHz giving frequency domain.

The capacitors are more specified for a given range of ESR than being specifically tantalum also nothing said about XR7 for that reason I would expect some history with oscillation mitigation. There has been some discussion that some intuitively prefer the lower portion of the ESR range could mean that careful filtering has been a concideration. I would choose which ESR is better by evaluating different ESR values to see if any improvement can be made. I did not see any oscillation during the inrush of current.
 
Last edited:

Thread Starter

camerart

Joined Feb 25, 2013
3,724
Hi S and J,
I have:
1/ An experimental field signal meter, that has a range somewhere like: 100kHz-3gHz, but no filters, so I can't tell what frequency it's picking up (yet).

2/ A Tiny-SA specrum analyser.

3. A SDR dongle RTL2832U, so I can open a computer SDR.

I programmed the 4431 to switch ON/OFF all PORTS so the SPI was ON/OFF, perhaps that's the problem.

Is the 18F4431 playing any part of this? I fear that it's last VDD PIN will soon burn out, with a few more quick tests, and it's getting expensive. Can I remove it and still test the above?

NOTE: The 5V regulator is steady, and it doen't have a 100nF CAP, could there be a problem with the 3.3V Caps?
C
 

jjw

Joined Dec 24, 2013
823
Make a simple test program to blink the LEDs and make all other ports as inputs ( might the default ?) and measure the 12V/3.3V
 

Thread Starter

camerart

Joined Feb 25, 2013
3,724
Make a simple test program to blink the LEDs and make all other ports as inputs ( might the default ?) and measure the 12V/3.3V
Hi J,
I don't think there's enough time to program the PIC before it burns out. I've just ordered more, and have 1x left (unless I rob an old PCB).

If I remove the PIC, will I still be able to test for oscillations.
C.
 

jjw

Joined Dec 24, 2013
823
You can cut the trace from 3.3V to PIC to see what happens.
The frequency of the ripple voltage is near 125kHz.
Might be the SPI clock, 8MHz/64 ?
 

Thread Starter

camerart

Joined Feb 25, 2013
3,724
You can cut the trace from 3.3V to PIC to see what happens.
The frequency of the ripple voltage is near 125kHz.
Might be the SPI clock, 8MHz/64 ?
Hi J,
I'll try your idea first, but I think the PIC is no more really, so after a test, I'll remove it.

I can program the 4431 on PCB4, with PCB6 LED flash only, then remove it and replace the conked PCB6 one with it.

EDIT: There are only 2x Caps on the 4431 PCB4, they both seem like ceramic, all calm?
C.
 
Last edited:

Thread Starter

camerart

Joined Feb 25, 2013
3,724
You can cut the trace from 3.3V to PIC to see what happens.
The frequency of the ripple voltage is near 125kHz.
Might be the SPI clock, 8MHz/64 ?
Hi J,
Cut the 4431 power track. Here is the 10V battery at start-up, then 1 sec later.
C
 

Attachments

Last edited:

Thread Starter

camerart

Joined Feb 25, 2013
3,724
Check the 3.3V also after cutting the trace.
Hi J,
I missed the above comment.

I programmed 4431 on PCB4 to only flash the LEDs, then swapped the faulty 4431 with this one and voila :) the LEDs are flashing merrily, and no smoke.

Here is the OSC at the 10V battery:
The is no Cap at the battery input, but will add one and see if this wave straightens out.

Must be more carefull regarding PINs ON/OFF in the future.

Thanks J, and thanks to all for your suggestions. My oscilloscope has now been calibrated and I now know how to START/STOP.

C.
 

Attachments

BobaMosfet

Joined Jul 1, 2009
2,110
Hi B,
How do I calibrate the probes?
EDIT: Found and did a calibration.

For the images in #10, I stopped the Osc, then it automatically started when I switched ON the voltage.

I don't think mine has the add on, for printing, as there are no ports.
3.5" floppies! You remind me of loading the 16? floppies for Windows 3.1 and looking forward to seeing the paper clip :)
EDIT: Mine doesn't have the output extension, so no prinint etc

C.
GOOD FOR YOU for figuring out calibration! :)

1615481291679.png
Your "STOP/RUN" button is at the top right- press this and it will freeze the screen. Press it again, and the scope will scan again. You appear to be missing the extension that would allow printing using the HARDCOPY button to RS-232 interface- You might look around and see if you can find one for reasonable price and install it.

Here is a Beginner's Guide I found for your scope- it will help you tremendously:

https://www.physics.queensu.ca/~phys352/BMScope_intro.pdf

Your scope is a fantastic scope- it is fast enough to handle any project you will do. Take good care of it and be careful with it, and it will serve you well. The probes are probably more expensive than the scope itself at this point, so if you only have 2, see if you can obtain a 3rd one as a spare for a reasonable price before the become impossible to find.
 

Thread Starter

camerart

Joined Feb 25, 2013
3,724
GOOD FOR YOU for figuring out calibration! :)

View attachment 232498
Your "STOP/RUN" button is at the top right- press this and it will freeze the screen. Press it again, and the scope will scan again. You appear to be missing the extension that would allow printing using the HARDCOPY button to RS-232 interface- You might look around and see if you can find one for reasonable price and install it.

Here is a Beginner's Guide I found for your scope- it will help you tremendously:

https://www.physics.queensu.ca/~phys352/BMScope_intro.pdf

Your scope is a fantastic scope- it is fast enough to handle any project you will do. Take good care of it and be careful with it, and it will serve you well. The probes are probably more expensive than the scope itself at this point, so if you only have 2, see if you can obtain a 3rd one as a spare for a reasonable price before the become impossible to find.
Hi B,
How often does it need calibrating?

START/STOP: I found this button, and have been using it for my images, thanks.

Thanks for the manual, I'll give it a good coat of looking at :)

There are the printout modules for sale at £100+, but I wonder how useful it woud be?

Yes, I could do with a spare probe, especially as one of mine doesn't have the clip end fitting. What do you think of these? https://www.ebay.co.uk/itm/TDS210-T...159173?hash=item3b4e812dc5:g:xpIAAOSwvvhfYXQi

C.
 

Thread Starter

camerart

Joined Feb 25, 2013
3,724
looks like the power supply oscillation issue will get solution, interesting project.
There is datasheet that discusses input output how to derate and which ceramic capacitors. A ceramic output cap derated = 20uF
https://e2e.ti.com/support/power-ma...tps7a91evm-831-input-output-capacitors-values
If there is concern on the uC board, the input decoupling cap on the input pin uC is important.
Hi S,
I don't understand derate!

Can I clarify? I only use batteries. At the moment, I test with an 8x NI-MH pack, 10V, but eventually, LI-ON 3xpack 12V will be used.
There is 1x 5V reg, which seems happy with a ceramic polarised 100uF cap after it.
2x 3.3V reg.

Earlier you mention, that the 3.3V should have 5V inputs, is this necessary? I can route the 2x 3.3V through the 5V is necessary.

Can you recommend your layout for this set-up please?

The original question, didn't give some of you the advantage that there are connections between the 2x PIC SPI etc, and they both should never be ON at the same time.

C.
 

Thread Starter

camerart

Joined Feb 25, 2013
3,724
Hi,
I spoke too soon! :(

There is now a 100uF polarised electrolytic cap before and after the 3.3V 4431 REG.

EDIT: The 4431 REG input and output show correct voltages and smooth.

After a while the LEDS started errors. The 3.3REG got hot and the VDD PINs are tarnishing, so I pressume hot.

The voltage at the LEDs is app 1V.

All PINs of the PIC show sensible resistance.

What to test? I'll try cutting the PIC VDD first, and see what happens.
C.
 
Last edited:

jjw

Joined Dec 24, 2013
823
Hi,
I spoke too soon! :(

There is now a 100uF polarised electrolytic cap before and after the 3.3V 4431 REG.

The before and after votages show smooth.

After a while the LEDS started errors. The 3.3REG got hot and the VDD PINs are tarnishing, so I pressume hot.

The voltage at the LEDs is app 1V.

All PINs of the PIC show sensible resistance.

What to test? I'll try cutting the PIC VDD first, and see what happens.
C.
Is the PIC the old one ( faulty )?
Have you looked at 3.3V regulator d/s about thermal requirements?
It needs a copper pad below. The bigger the better.
 
Last edited:
Top