Author Topic: Need help finding high current drain on my PCB  (Read 7389 times)

0 Members and 1 Guest are viewing this topic.

Offline Psi

  • Super Contributor
  • ***
  • Posts: 9889
  • Country: nz
Re: Need help finding high current drain on my PCB
« Reply #75 on: October 24, 2021, 10:40:54 pm »
Yeah, I have also seen this.

All is fine until you get a batch of caps that is a bit more susceptible to cracking than all the batches you had used previously. Then the problem appears.

It can be a bit misleading too because the short is not always static, it comes and goes based on many factors. Handling, flexing, hot air rework, ambient temp etc.

It can also be a part mix up where you spec'ed 25V caps but got 16V version by mistake, and so are overvolting them a bit.
« Last Edit: October 24, 2021, 10:44:12 pm by Psi »
Greek letter 'Psi' (not Pounds per Square Inch)
 

Offline sailahTopic starter

  • Contributor
  • Posts: 46
  • Country: us
Re: Need help finding high current drain on my PCB
« Reply #76 on: October 25, 2021, 12:13:28 pm »
In this application the cap in question was a Taiyo Yuden TMK212BBJ106MGHT.  Will speak to my CM today about substituting in a flexible termination cap.  Might as well do the 10µF, 1µF & 100nF all the same.  Good idea or no?  So far I have not discovered an issue in any of the other caps in that array, just to 10µF bulk cap.

Was looking on Mouser and they have a few in stock in the qty I need.  Not looking for the cheapest option, my devices are installed in high ambient heat (up to 70C) and long term reliability in a high temp area is what I want.

100µF-> https://www.mouser.com/ProductDetail/810-CGA4J1X7S1E106K2
1µF-> https://www.mouser.com/ProductDetail/81-GCJ21BR71E105KA2L
100nF->  https://www.mouser.com/ProductDetail/581-08053C104K4Z2A

Regarding caps, and what little I know about them, and how little they cost in overall terms, is there a "best" cap for me to pick?  Is going with a X7R dielectric going to give me higher reliability over a lower temp dielectric?   These little $0.06 caps have cost me much frustration.  If there is a better cap for $0.25 for example, I want that.  Sorry for these most basic questions.
 

Offline Manul

  • Super Contributor
  • ***
  • Posts: 1105
  • Country: lt
Re: Need help finding high current drain on my PCB
« Reply #77 on: October 25, 2021, 07:01:18 pm »
I have no experience with Taiyo Yuden. Maybe they have quality problems and you were unlucky enough to use them. Known, reputable brands which come to my mind are AVX, Vishay, Kemet, TDK.

Normally, dielectric defines things like capacitance stability over temperature, decrease of capacitance due to DC bias, these sort of things. Not directly reliability. Reliability is more a function of manufacturing process quality, capacitor size (density) and, of course, usage.

Soft termination capacitors are a valuable improvement for applications, requiring high reliability under harsh conditions (especially flexing, vibration). They are widely used in military, medical, avionics, automotive applications. Not strictly necessary for your application, but would certainly eliminate some risk.

Hope you will not develop a capacitor phobia, likely you just got unlucky and it happens.
 

Offline sailahTopic starter

  • Contributor
  • Posts: 46
  • Country: us
Re: Need help finding high current drain on my PCB
« Reply #78 on: October 25, 2021, 07:51:18 pm »
I have no experience with Taiyo Yuden. Maybe they have quality problems and you were unlucky enough to use them. Known, reputable brands which come to my mind are AVX, Vishay, Kemet, TDK.

Normally, dielectric defines things like capacitance stability over temperature, decrease of capacitance due to DC bias, these sort of things. Not directly reliability. Reliability is more a function of manufacturing process quality, capacitor size (density) and, of course, usage.

Soft termination capacitors are a valuable improvement for applications, requiring high reliability under harsh conditions (especially flexing, vibration). They are widely used in military, medical, avionics, automotive applications. Not strictly necessary for your application, but would certainly eliminate some risk.

Hope you will not develop a capacitor phobia, likely you just got unlucky and it happens.

Thx I ordered a few reels of the soft termination caps from vendors I use regularly like TDK, Murata and AVX.  Have call with CM tomorrow and will discuss.  I don't know why we were using Taiyo Yuden other than it was on the BOM and was built that way.

So far I have performed the C13 cap reflow on about 20 devices and they all work perfectly.  I guess the CM can do an autopsy with Xray and see if cap broke or was soldering issue.  The solder looks OK in my magnifying glass but what do I know.

Now I just need to deal with fallout of all the devices that have failed in the field.  Hope my customers aren't too ticked off.

Trying to devise how I can do a current test on these to maybe catch the issue.  Problem is that the supercap masks the current draw of a couple mA pretty easily during that initial flash.  When I have been testing these, I simply unsolder one leg of the cap, test, and then resolder it.  Don't want to be doing that during production runs.  My cycle time on my programmer is about 20 secs.
 

Offline mikeselectricstuff

  • Super Contributor
  • ***
  • Posts: 13695
  • Country: gb
    • Mike's Electric Stuff
Re: Need help finding high current drain on my PCB
« Reply #79 on: October 25, 2021, 08:22:40 pm »
For production test, would it be easier to not connect the battery, charge the supercap and check that voltage was still above a certain value after a fixed tim, with the mcu in a test sleep mode?
Youtube channel:Taking wierd stuff apart. Very apart.
Mike's Electric Stuff: High voltage, vintage electronics etc.
Day Job: Mostly LEDs
 

Offline Kasper

  • Frequent Contributor
  • **
  • Posts: 737
  • Country: ca
Re: Need help finding high current drain on my PCB
« Reply #80 on: October 28, 2021, 05:22:56 pm »
Just came across this IP68 waterproof LoRa sensor that reminded me of your product.  I wouldn't expect seeed to be great quality but it could help you get some ideas for waterproof antennas.

https://www.seeedstudio.com/Dragino-LSN50-V2-Waterproof-Long-Range-Wireless-LoRa-Sensor-Node-Support-868MHz-Frequency-p-4474.html
 

Offline jmelson

  • Super Contributor
  • ***
  • Posts: 2758
  • Country: us
Re: Need help finding high current drain on my PCB
« Reply #81 on: October 28, 2021, 06:45:31 pm »
Interesting paper, thanks!
I am just wondering, what the HECK I have stumbled into doing right, so that I have almost NEVER run into this.
Now, I do my own reflow here, using a toaster oven with ramp-and-soak thermocouple controller, and reflow at the lowest temp I can get reliable melting at.  So, maybe the whole issue is crazy reflow profiles and very rapid heat-cool delta-T ramps.  And, maybe I'd never notice a capacitor leaking a few microamps, as I don't usually do battery-powered stuff.  But, I am just mystified as to why I have not run across this issue.  I have had exactly ONE capacitor that clearly was bad, and split into two pieces when replaced.

Jon
 

Offline Kasper

  • Frequent Contributor
  • **
  • Posts: 737
  • Country: ca
Re: Need help finding high current drain on my PCB
« Reply #82 on: October 28, 2021, 07:14:06 pm »
Interesting paper, thanks!
I am just wondering, what the HECK I have stumbled into doing right, so that I have almost NEVER run into this.
Now, I do my own reflow here, using a toaster oven with ramp-and-soak thermocouple controller, and reflow at the lowest temp I can get reliable melting at.  So, maybe the whole issue is crazy reflow profiles and very rapid heat-cool delta-T ramps.  And, maybe I'd never notice a capacitor leaking a few microamps, as I don't usually do battery-powered stuff.  But, I am just mystified as to why I have not run across this issue.  I have had exactly ONE capacitor that clearly was bad, and split into two pieces when replaced.

Jon


I too am surprised how supposedly fragile ceramic caps are.  I used to solder avionics comms equipment.  We would hand solder lots of ceramic caps. We'd put them between legs on ICs, solder 0805s ontop of 0603s, etc. We'd do all sorts of mods because it was easier to certify and perform mods than it was to update a PCB and certify it.
 

Offline sailahTopic starter

  • Contributor
  • Posts: 46
  • Country: us
Re: Need help finding high current drain on my PCB
« Reply #83 on: October 30, 2021, 11:57:17 am »
For production test, would it be easier to not connect the battery, charge the supercap and check that voltage was still above a certain value after a fixed tim, with the mcu in a test sleep mode?

Possible but that's not how we flash the firmware, not sure how we could calibrate that.  I'm sure that there's a calculation to figure out the voltage drop on a 0.47F cap over time given anything about say 10µA.  I suspect that 2.4mA is pretty simple to figure out, but what about 67µA?  Our cycle time on programming is about 20 secs.  Clearly the Keysight is the right tool for this job but I would need to get someone to program it to turn it on for a set amount of time and then switch off the outputs, measure.  Maybe not as big a deal but just one more thing...I need an easy button here :-DD

Just came across this IP68 waterproof LoRa sensor that reminded me of your product.  I wouldn't expect seeed to be great quality but it could help you get some ideas for waterproof antennas.

https://www.seeedstudio.com/Dragino-LSN50-V2-Waterproof-Long-Range-Wireless-LoRa-Sensor-Node-Support-868MHz-Frequency-p-4474.html


Thanks, we have our device certified to IP67, although generally speaking we are strictly in an industrial environment or maybe outside in a pipe rack.  I have had instances where customers broke the antenna but usually it was from a ham-fisted installer.  In thousands of deployed devices maybe only a couple have broken.

Here's the latest in this saga, again, massive thanks to mystery EEVblog member who walked me through numerous debug steps.  Thanks and thanks to everyone else for their kindness and interest with a complete stranger.

I spoke to my CM, who was also a bit mystified at this issue.  They place millions and millions of caps every year and never had an issue.  I sent them the faulty device with the traces cut to the bad cap but no other modifications were performed.

I had approx 55 devices left from that production run so rather than test them all by unsoldering the supercap, I decided to swap out C13 to a comparable soft termination X7R cap from TDK.  I also sent CM a reel of these for our run.  Probably not necessary but I also switched C14 and C15 to soft caps as well, belt and suspenders.  I had to upgrade my soldering situation and got a soldering smoke collector from Weller, nice 10" microscope to verify solder joins, new flux from Chipquik.  I don't do a lot of soldering so it was helpful practice.  Most of my joints were pretty good and if they weren't I did them again.

The CM is going to make 50 devices with no supercap and ship them to me.  I will test all 50, flashed, on Keysight to verify sleep current.  If all 50 pass, I'll greenlight the rest of the run and send them back to be finished.

Thanks, Peter
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf