Pi and voltage regulator fail - overheat?

All things antweight

Moderators: BeligerAnt, petec, administrator

Post Reply
User avatar
MarkR
Posts: 375
Joined: Mon Dec 18, 2017 12:46 pm
Location: Reading Hackspace
Contact:

Pi and voltage regulator fail - overheat?

Post by MarkR »

I thought it was all working.

I have a Raspberry Pi Zero - this draws about 100-200mA

I built a custom power supply / regulator board which also controls the flipper weapon. I used a "Chinese special" KIA1117-5.0 voltage regulator ( see https://datasheet.lcsc.com/szlcsc/KIA-S ... 116068.pdf )

Now I had it running nicely yesterday. Now I'm testing it again, but with a *fully charged* 2S lipo.

It's crashing out and going wrong. What I *THINK* is happening is that the voltage regulator is overheating due to the battery voltage being somewhat (about 1V) higher than it was yesterday.

The old version of Betsie used a similar configuration, but with a much bigger LM7805 regulator, which I guess can sink more heat.

If the pack is at 8.4v , the regulator needs to drop 3.4v, 3.4v * 150mA = 0.51W - so approximately 1/2 watt. Is this too much for such a small surface-mount package?

NB: I don't drive any motors through this regulator.

What can I do to fix it?

* glue some kind of heat sink? (there is not much space)
* put some heat transfer compound or glue or something, under the package?
* Add a diode at the GND pin, to bias the voltage up a bit, so the regualtor needs to drop less voltage (con: output voltage goes higher, I think the Pi Zero can take ~ 6v)

This is my first designed PCB and it's already showing several design flaws :)

Mark
Robots: Betsie - RaspberryPi controlled flipper bot with gyro stablisation - too clever for her own good?
Stacie - tidy flipper; 4wd driven by hair bands
Paulmchurd
Posts: 227
Joined: Sun Dec 10, 2017 4:39 pm

Re: Pi and voltage regulator fail - overheat?

Post by Paulmchurd »

Under load regulation it says VIN-VOUT =3V. I don’t really know what “load” is but if it is this then your load if is 3.4 volts which is a little too high and could be the reason? I feel like load is all about current though and maybe is connected.

Also have you put the recommended capicitors on that it shows on the data sheet?

This is the voltage regulator that I see on most Chinese boards. It has a higher voltage rating. I have a few I could send in the post. All depends if it fits on your pcb.

https://uk.farnell.com/texas-instrument ... 7456876370
User avatar
MarkR
Posts: 375
Joined: Mon Dec 18, 2017 12:46 pm
Location: Reading Hackspace
Contact:

Re: Pi and voltage regulator fail - overheat?

Post by MarkR »

I have connected capacitors.

Considering the situation again, it seems possible that the regulator has developed a fault caused by an earlier overheat condition. I have some spares so I will replace and test it today to check that it can give stable 5v.

The problem I have with Antweight electronics is that stuff which works on the bench quite often seems to fail in the robot, occasionally spectacularly.

I already discovered that one of my PCB traces was too thin. That was exciting.
Robots: Betsie - RaspberryPi controlled flipper bot with gyro stablisation - too clever for her own good?
Stacie - tidy flipper; 4wd driven by hair bands
User avatar
MarkR
Posts: 375
Joined: Mon Dec 18, 2017 12:46 pm
Location: Reading Hackspace
Contact:

Re: Pi and voltage regulator fail - overheat?

Post by MarkR »

Replaced the voltage regulator last night, now it seems much better. The old one was giving overvolt conditions, I'm quite (happily) surprised that the Raspberry Pi came to no har.

I suspect that some earlier short or something, had fried the regulator and it was going haywire.

More testing definitely needed. If I can get Betsie running reliably, should be able to make it to at least one of next month's events.
Robots: Betsie - RaspberryPi controlled flipper bot with gyro stablisation - too clever for her own good?
Stacie - tidy flipper; 4wd driven by hair bands
Post Reply