Looking at the datasheet for the TPS65217 power management chip, a 6uA current draw seems to be a feature of the full-off state. I did a quick scan for any way to pull it to a no-draw state and didn't find anything, but a more full read might.
Based on your results, I'll be implementing a low-voltage monitoring circuit that disconnects the battery when it hits a lower limit (3.3v, likely). I did a quick Google and found several candidates along with a number of model-car and model-aircraft devices that could likely be adapted.
Greg in Mississippi
Based on your results, I'll be implementing a low-voltage monitoring circuit that disconnects the battery when it hits a lower limit (3.3v, likely). I did a quick Google and found several candidates along with a number of model-car and model-aircraft devices that could likely be adapted.
Greg in Mississippi
Why not add an output that is fired when it is powered off by the BBB cpu?
That way if the 5V goes out the BBB will still run until battery dead.
If shutdown drops off the batt relay output and isolates the battery by the relay contacts.
Just a thought.
Laters,
Drew.
That way if the 5V goes out the BBB will still run until battery dead.
If shutdown drops off the batt relay output and isolates the battery by the relay contacts.
Just a thought.
Laters,
Drew.
6uA is not going to drain a battery in any short period of time. I've had flashlights with way more than that last a month or two off, with nominally 500mAh RCR123s. Something else is causing a parasitic drain. But, it would surely be easier to physically cut power than find it. With such a low draw, it's doubtful better can be done, being all solid state.
I haven't implemented it yet, but my plan is to use an Arduino Nano clone, and one or more relays (and transistor, and diode, and maybe resistor), to cut power, once the BBB is safely off, and to also make sure there's enough reserve power to be turning it back on.
I'm not able to fulfill this promise - so no release next week - moving to January 25.
3V3 after shutdown
My attention is not building more Hardware around the cape, less is more in high end Audio. I wonder why are no others have the same problem. Let us think about a simple solution for everybody.
My attention is not building more Hardware around the cape, less is more in high end Audio. I wonder why are no others have the same problem. Let us think about a simple solution for everybody.
Heya miero,
Thanks for the update but I'm pretty sure no one will complain, your efforts with the driver etc are very much appreciated!
Thanks again, make sure you have some.time away from a pc!
Laters,
Drew.
Quick pic of the all in one BBB-DSP-DAC box I have been working on and a close up of the nano-cape with the NDK NZ2520SD 24.576 oscillator (gee they are small - fun to solder!) 3.3v power from the TPS7A4700 on the board at the top. Upsampling via SOX to 96k which is the frequency the miniSHARC runs at. Because of this no need for dual oscillators.
All working as intended - just need to tune the DSP parameters. Only problem I have not is there is no way to feed the analog test tone into the system so I can do frequency response measurements!
Not mainstream in terms of what most others are doing I know but running the miero Botic code so here it is...
All working as intended - just need to tune the DSP parameters. Only problem I have not is there is no way to feed the analog test tone into the system so I can do frequency response measurements!
Not mainstream in terms of what most others are doing I know but running the miero Botic code so here it is...
Attachments
I remember reading about this issue, maybe here -
https://groups.google.com/forum/#!topic/beagleboard/ogyyRufIVVs
You will see in the last post of that link that the author's BBB draws 25mA when "off". This would drain an 800mAH battery in 32 hours.
It seems to me that the only comprehensive solution, as I mentioned last September, is a fully external 5VDC UPS ...
... although this sounds like a good idea -
Q181699, how is your relay activated? Is it a logic-activated relay? And how is the relay powered?
Agreed. I was just pointing out that the TPS chip does continues some activity even in the 'off' mode. Doing a quick scan through the BBB SRM, it appears there is other activity that is likely adding to the current draw.
EDIT: I was posting as Linuxfan was posting just ahead of mine. Yup 25mA on a mostly empty LiPo will result in a fully dead cell fairly quickly!
Also it depends on where in the cell's discharge curve the system is set to 'off'. If it is already at or past the minimum, it will not take much current draw to take it down the rest of the way.
LOL... I agree on that too in theory, but that BBB is likely a few orders of magnitude more complex than all of the rest of my audio system and possibly yours too. A simple voltage detection circuit to control your relay sounds like the best 'Save the battery' solution, but it does not really deal with a controlled UPS-like shutdown if the BBB is operating on the battery. I glanced at the battery connection article referenced earlier in this or the HW thread:
BeagleBone Black: BBB - Rechargeable on-board b... | element14
I didn't see anything in there or the SRM about using the TPS chip and its capabilities to gracefully shut down the BBB as the battery acting as a UPS get's close to its low-voltage limit. Anyone have any thoughts or pointers to more info on this? I'd look deeper, but have to do some real work stuff instead tonight.
Sounds like a good, configurable solution. Please share when you have it going.
Greg in Mississippi
Last edited:
The answer is NO: the TPS65217C is not smart enough to do anything, itself, about a dying battery input. See shabaz's reply on this subject here -
BeagleBone Black: BBB - Rechargeable on-board b... | element14
Very cool Chris. That you are exercising Miero's driver and getting expected results is VERY relevant to what the rest of us are doing! Thanks for sharing!
Can you use the speaker test function in Alsa to produce test tones? Miero may have to provide a correction for how to make it work with his BBB driver, but it goes something like this:
speaker-test -Dhw:Botic -c 2 -f 1000 -r 96000
BTW, got this from Miero the man himself. Very generous & productive dude!
Thx, interesting set of comments, sorry I missed them at first.
Later!
Greg in Mississippi
P.S. Chris, I just soldered up 9 of those tiny NDK on an un-related, but not too dissimilar project. I could not have imagined doing it without my trusty hot-air rework station!
Last edited:
Thanks, I will read up on speaker-test, have used it before I recall. Problem is I am using REW which generates its own frequency sweep and records the mic output. may look at using the SPDIF input - not in place yet as I need to convert coax to TTL 3.3v, just waiting on some components to do that. Then find a laptop with spdif output or get HDMI to SPDIF converter - sure it can be done somehow!
A hot-air station would be useful for sure. I did manage to find a way to do it with a fine iron in the end involving soldering all surfaces first then placing the oscillator and reflowing.
The speaker-test tool is not able to generate sine sweep, but you can use play tool for this. The play is part of sox package.
$ play -r 96000 -c 2 -n synth 30 sin 500-5000 vol -3dB
This will play 30 seconds of sine sweep from 500Hz to 5kHz on two channels and at 96kHz sampling rate.
Enjoy! 🙂
$ play -r 96000 -c 2 -n synth 30 sin 500-5000 vol -3dB
This will play 30 seconds of sine sweep from 500Hz to 5kHz on two channels and at 96kHz sampling rate.
Enjoy! 🙂
... although this sounds like a good idea -
Q181699, how is your relay activated? Is it a logic-activated relay? And how is the relay powered?[/QUOTE]
@linuxfan: Thanks for the link, a lot of stuff, it seems there is a low current when BBB is in shutdown and PS comes from BAT.
@Miero: It isn't a bug in your patch or in Debian, it comes from PMIC in BBB layout.
So the right way for me (in my setup with permanently LiPo for safe shutdown) is to switch off the BAT line when BBB shutdown is complete (by 3V3 relay). This job will done by a simple 8Bit-controller, measuring voltage and current of BAT line. The controller is in sleep mode every time (while audio!), wake up comes from BBB PWR_EN goes off, relay is switched off by lowest current or critical low voltage. Controller PS comes from 3V3EXP BBB P9, it switches itself. PS comes back with normal BBB power on.
After more tests I'll post the schematic, maybe not in this thread, because here is commercial sector of Twisted Pear.
Q181699, how is your relay activated? Is it a logic-activated relay? And how is the relay powered?[/QUOTE]
@linuxfan: Thanks for the link, a lot of stuff, it seems there is a low current when BBB is in shutdown and PS comes from BAT.
@Miero: It isn't a bug in your patch or in Debian, it comes from PMIC in BBB layout.
So the right way for me (in my setup with permanently LiPo for safe shutdown) is to switch off the BAT line when BBB shutdown is complete (by 3V3 relay). This job will done by a simple 8Bit-controller, measuring voltage and current of BAT line. The controller is in sleep mode every time (while audio!), wake up comes from BBB PWR_EN goes off, relay is switched off by lowest current or critical low voltage. Controller PS comes from 3V3EXP BBB P9, it switches itself. PS comes back with normal BBB power on.
After more tests I'll post the schematic, maybe not in this thread, because here is commercial sector of Twisted Pear.
Thanks, will give that a try. Hopefully I can fool REW into thinking its the sweep it has generated and it will read the input from the mic.
Q181699: according the Figure 1 on page 16 of tps65217c.pdf after transition to OFF state everything is powered down.
However datasheet also states that in OFF mode TPS consumes 6uA current and that battery voltage always supplies the system(SYS pin).
The SYS_5V is in schematics connected to following components on the BBB board:
- TL5209 regulator ... this should be disabled because of low on EN pin
- USB HOST chip TPS2051 ... this should be disabled too
- 4 user leds ... these are OFF
- SYS_5V pins on expansion header P9 ... >>> power might be consumed by cape <<<
- HDMI connector ... >>> power might be consumed by connected device <<<
In OFF mode with AC adapter I measure 1.1V on SYS_V5 pin. What is the voltage there when battery is connected and AC adapter is not?
However datasheet also states that in OFF mode TPS consumes 6uA current and that battery voltage always supplies the system(SYS pin).
The SYS_5V is in schematics connected to following components on the BBB board:
- TL5209 regulator ... this should be disabled because of low on EN pin
- USB HOST chip TPS2051 ... this should be disabled too
- 4 user leds ... these are OFF
- SYS_5V pins on expansion header P9 ... >>> power might be consumed by cape <<<
- HDMI connector ... >>> power might be consumed by connected device <<<
In OFF mode with AC adapter I measure 1.1V on SYS_V5 pin. What is the voltage there when battery is connected and AC adapter is not?
@Miero: Please measure voltage on P9 pin 3 or 4 (VDD_3V3EXP) or put a LED on it. This line is still on after BBB shutdown with active BAT unless you disconnect the BAT line.
Q181699: according the schematics if there is voltage on P9 pin 3 or 4, then BBB is not powered down...
Please try holding power button for 8 seconds... the voltage must go down.
Additional question, does this really happen with Botic distribution too? There is explicit TPS chip power down enabled when CPU is powered down.
Please try holding power button for 8 seconds... the voltage must go down.
Additional question, does this really happen with Botic distribution too? There is explicit TPS chip power down enabled when CPU is powered down.
Last edited:
@Miero: It makes no difference, shutdown by SSH command, pressing button 8s or simple disconnect AC.
I use your distribution and tested the original Debian one, always the same.
Please simply try it out on your board.
I use your distribution and tested the original Debian one, always the same.
Please simply try it out on your board.
I thought about using a controller (and am open to adding that to the Hermes-BBB cape) but I wonder if there isn't a simpler approach. maybe just disable the battery via a normal GPIO or other already available signal after the sucessful shutdown?