short on VBAT? another board another ailment 130mA currents

Post here first, or if you can't find a relevant section!
ag123
Posts: 798
Joined: Thu Jul 21, 2016 4:24 pm

short on VBAT? another board another ailment 130mA currents

Post by ag123 » Tue Sep 19, 2017 4:16 pm

recently i tried to meddle with RTC on my Baite MM clone, RTC 'simply works' thanks much for the really good effort here.
then i hooked up a CR3032 coin cell on VBAT pin on the MM, it seemed to run quite well initially keeping the clock going after removing USB power.
after less than a month the CR3032 drains from some 3+ volts to around 1.5v, i'm rather surprised as VBAT is intended for low power and basically keep the LSE crystal osc and RTC running?
i then tried to measure resistance and voltages, if the coin cell is connected, it reads zero volts across to ground. but if i measure resistance across with a multimeter (VBAT to ground removing the coin cell) i get some 200 ohms.
This is pretty much a short between VBAT and ground?
it seemed the specs gives typical currents of 1 uA, this would seem much greater than 1 uA
would there be any need to put MM into 'low power modes' rather than simply disconnecting USB power?
:?
Last edited by ag123 on Wed Nov 08, 2017 2:33 pm, edited 1 time in total.

User avatar
Pito
Posts: 1627
Joined: Sat Mar 26, 2016 3:26 pm
Location: Rapa Nui

Re: short on VBAT?

Post by Pito » Tue Sep 19, 2017 6:38 pm

You do something wrong with your measurement. 200ohm would drain the cr2032 in 10hours completely.
3V/200ohm = 15mA. Cr2032's capacity is 150mAh.
According to the schematics I saw the Vbat is connected directly to pin Vbat.
If you measure with a multimeter (plus on Vbat, minus on GND), which uses several volts for measuring the resistance, it may happen you open the upper clamp diode (ESD protection at every pin) at the VBat pin and you may measure the forward resistance of the upper clamping diode against VCC/GND (could be 200ohm).
https://training.ti.com/clamp-diodes

With 20uA (doublecheck the datasheet plz) when RTC is running you drain the cr2032 in 150000/20 hours.
Pukao Hats Cleaning Services Ltd.

ag123
Posts: 798
Joined: Thu Jul 21, 2016 4:24 pm

Re: short on VBAT?

Post by ag123 » Thu Sep 21, 2017 12:29 am

thanks pito ! it seemed the coin cell which has already drained to 1.5v is possibly completely drained, as measuring it across when connected to VBAT - GND gives 0 v, i think you are right about the clamping diodes. i'd use a new coin cell and possibly connect a small series resistor and measure voltages across again, that would tell if after all VBAT is literally shorted, i think there are also some other posts about BP, MM pins shorting across not just VBAT.
it is not easy to identify the tiny features on the boards simply examining them with the naked eye, there might be some shorts at the solder joints etc

actually i'm starting to wonder if disconnecting the usb cable supplying power could have forward biased the clamping diodes, that would increase the currents from microamps to milliamps literally a thousand times higher. if that's true for the general case it would be bad news as it may mean coin cells won't last on VBAT pin on MM board designs when power is removed

User avatar
Pito
Posts: 1627
Joined: Sat Mar 26, 2016 3:26 pm
Location: Rapa Nui

Re: short on VBAT?

Post by Pito » Thu Sep 21, 2017 1:48 am

The Vbat current when RTC is on should be 1.2uA according to the datasheet.
Simply insert an ammeter between Cr2032+ and Vbat (or Cr2032- and GND) and you may see the current.
I do not have MMini Baite schematics handy (do we have it somewhere?).
The Vbat pin could be setup differently (on silicon) so the clamp diodes (if any) could be arranged such they will not be fw biased when Vcc off and Vbat on. You are right the 3V at Vbat with Vcc at 0V could mess with the upper clamp diode.

PS: My 2 MMini clones when not powered
First shows 7Mohm with multimeter's + at Vbat pin (and 2Mohms with - at Vbat pin) against GND
Second shows 5Mohm with multimeter's + at Vbat pin (and 2Mohms with - at Vbat pin) against GND
That would be 3V/6Mohm = ~0.5uA, plus add current of the running RTC, it may get 1.2uA in total.

The Baite's Vbat goes directly to stm's pin 1 here.

Btw, MMinis do not have 32kHz RTC crystal mounted, how did you wire it?
Pukao Hats Cleaning Services Ltd.

User avatar
RogerClark
Posts: 7465
Joined: Mon Apr 27, 2015 10:36 am
Location: Melbourne, Australia
Contact:

Re: short on VBAT?

Post by RogerClark » Thu Sep 21, 2017 3:02 am

I looked on the original MM schematic and VBat is only supposed to be connected to pin 1 on the MCU
maplemini(1).pdf
(24.3 KiB) Downloaded 14 times
Looking at a MM clone, it also looks like VBat is just connected to pin 1

Of course your MM clone may be different

ag123
Posts: 798
Joined: Thu Jul 21, 2016 4:24 pm

Re: short on VBAT?

Post by ag123 » Thu Sep 21, 2017 12:31 pm

Pito wrote:
Thu Sep 21, 2017 1:48 am
Btw, MMinis do not have 32kHz RTC crystal mounted, how did you wire it?
thanks pito, this is tad ugly but it here is the hack, guess it's about time i solder a new MM :lol:
Image
looks ilke that huge blob of solder at the crystal joints looks pretty bad, would probably need to 'drain' all that solder off
VBAT is 3rd pin from the bottom

i think baite is basically MM clone hence presumably the specs would be same as/similar to the original MM

and the usual places to get those 32k crystals, it is a gamble that they even works, but for this piece it works (certainly not that epson 5pf high precision feather weight crystals) :lol:
https://www.ebay.com/sch/i.html?_nkw=32768&_sacat=0

ag123
Posts: 798
Joined: Thu Jul 21, 2016 4:24 pm

Re: short on VBAT? another board another ailment 130mA currents

Post by ag123 » Wed Nov 08, 2017 2:45 pm

well, i had gone ahead to solder up a new baite MM, this time round VBAT is pretty much high resistance, but on this new baite board the current consumption is 130mA ! measured at the usb supply using a usb current / voltage meter
https://www.ebay.com/sch/i.html?_nkw=us ... r&_sacat=0

i reused the old tactic of soldering a 32k crystal at pc14, pc15, well that actually works never mind no grounding caps
accordingly one should also have a series resistor with the crystal if one wants the crystal to last, otherwise it may be short lived, but i simply soldered it no resistors

the stm32f103 literally runs *hot*, it is pretty much very hot to touch
3.3v / 130ma ~ 25.3 ohms
and gives a power dissipation of some 0.43 watts
if you go by the specs
http://www.st.com/content/ccc/resource/ ... 161566.pdf
section 5.3.5 p42 table 13 max current consumption is a mere 50ma with all pheriperial clocks enabled, this is running more than double that pretty much close to 3x the normal currents

it is somewhat puzzling on the defect density of these 'cheap' boards & stm32f103 batches used in manufacture

but despite running hot, it is rather interesting that the chip actually runs the sketch fine driving ili9341 lcd etc

ag123
Posts: 798
Joined: Thu Jul 21, 2016 4:24 pm

Re: short on VBAT? another board another ailment 130mA currents

Post by ag123 » Sat Dec 09, 2017 6:22 am

now somewhat improved on a new baite mm, no shorts this time round
but the steal? no caps they are actually needed but for now i take the short cut as it seem to work
Image

User avatar
RogerClark
Posts: 7465
Joined: Mon Apr 27, 2015 10:36 am
Location: Melbourne, Australia
Contact:

Re: short on VBAT? another board another ailment 130mA currents

Post by RogerClark » Sat Dec 09, 2017 8:01 am

Cool.

I presume you will end up laying it flat and gluing the crystal to the underside of the board

ag123
Posts: 798
Joined: Thu Jul 21, 2016 4:24 pm

Re: short on VBAT? another board another ailment 130mA currents

Post by ag123 » Sat Dec 09, 2017 5:26 pm

well i simply soldered the 32k crystal at the pc14, pc15 pins that's intended for the crystal, but this time round i succeeded in using less solder. somehow by sheer chance it fills the hole nicely and keep the crystal contacts in place on the board

there is no load capacitor, it is simply the crystal. i've been trying to figure out how and where to put 2 load capacitors but i havn't thought out good solutions. hence i simply have the crystal that took a gamble, surprisingly it 'just works', time keeping seem rather accurate but is probably off, i've not actually measured that. i simply need the rtc time to timestamp the log files and i could make do with minutes of differences and perhaps sync the rtc say every once in a while (say a week to a month) when i hook it up via usb

Post Reply