Power supply issue by Vin on Sipy + Deep Sleep Shield

[nespressif]

@robert-hh Last night I tried to execute without the deep sleep shield and the behavior is the same, it shuts down in the first instruction.

I want to try with other programs (a simpler program that does not read sensors, just send a byte via Sigfox) and also with an external source that provides 5v stable.

Let's see if anyone has achieved to use Sigfox with a Sipy fed it directly through the Vin.

[robert-hh]

@nespressif There is no reason why you should not be able to supply a SiPy by connecting power to Vin and GND. I'm doing that in my own expansion board with Lopy, LoPy4 and Fipy, where the power from a separate USB/UART module is routed to Vin and GND, and I did that several times on the bench from a lab power supply in the absence of just another expansion board.

[nespressif]

@robert-hh And you achieve to send data via Sigfox by feeding the module with an external power supply through the Vin? With which voltage do you feed it 3.3v or 5v?

Because I already tell the code that they don't use Sigfox or ADC work fine, but my code with ADC and Sigfox is turned off. I have to try several codes, to be able to discard more things.

[robert-hh]

@nespressif I feed it with ~5V though Vin. I do not use SigFox, but LoRa. But bot use the same Modem chip, and the power pattern should be the same. The SiPy has no SPIRAM, so the Size of Python code it can run is limited. As a rule of thumb, about 500 Python lines of code. If your code needs more RAM for it's data, then less.

[nespressif]

@robert-hh My code works fine when I feed the module with the expansion board or pysense by USB or with the lipo battery in the JST connector.

This is why I think it's not a code problem, in addition, all libraries are integrated into the firmware as frozen code, as you suggested, so most of the code does not count for those 500 lines, only the main.py and little more.

I will try with other codes and also try with 5v stable through Vin.

Do you think that if I connect the 5.72v that my ATX power supply gives to the Vin, will I damage the module?
Since the data sheet indicates a range of 3.5-5.5v, yesterday I didn't dare to try, I don't want to break another module.

[nespressif]

@robert-hh @dan I've already achieved to isolate the problem,this function crashes the program only when I feed the module from the Vin, with or without ds shield. Everything else works fine, it sleeps and wakes up as expected, I still have to measure the consumption.

wake_s = ds.get_wake_status()

In this function of the deepsleep.py library, the program breaks, when the usb cable is not connected in the expansion board.

def get_wake_status(self):
        # bits as they are returned from PIC:
        #   0: PIN 0 value after awake
        #   1: PIN 1 value after awake
        #   2: PIN 2 value after awake
        #   3: PIN 3 value after awake
        #   4: TIMEOUT
        #   5: POWER ON

        wake_r = self.peek(WAKE_STATUS_ADDR)
        return {'wake': wake_r & (TIMER_WAKE | POWER_ON_WAKE),
                'P10': wake_r & 0x01, 'P17': (wake_r & 0x02) >> 1,
                'P18': (wake_r & 0x08) >> 3}

I don't know if it's a bug, it's strange that no one's ever reported it before.

So if I feed it with a battery I can't know the reason wake up, which for me is interesting. Nor can I know what error reports, because I do not have connected the cable uart.

[nespressif]

Please, Can anyone replicate my issue?

[reidfo]

I'm not sure if this will help or not, but I use the FiPy with expansion board 3.0. If I try to boot off just the LiPo battery it will not boot. The solution to that is to tie P8 to high (3V3).

[nespressif]

@reidfo Thanks for your input, that problem is already reported for the expansion board 3.0, but it is not the same.

The problem here is that when you use only Sipy+ds shield feeding it from the Vin, if you use the mentioned function, the program breaks.

[robert-hh]

@nespressif At which point do you connect Vin. At the SiPy or at the DS-Shield? As far as I understand it, it has to be at the DS Shield's bottom side pins.

[nespressif]

@robert-hh I usually connect it through the pins at the bottom of the deep sleep shield. But no matter where I connect it, if I don't use the deepsleep.py library, the Sipy works perfectly feeding it through the Vin. If I use Sipy+ds shield with the deepsleep.py library and I don't use the ds.get_wake_status() function, everything works perfectly.

The problem occurs when I use Sipy+ds shield+deepsleep.py+ds.get_wake_status() and feed it through Vin (without expansion board). If I add the expansion board and feed it with USB, everything works perfectly.

[robert-hh]

@nespressif That sounds as if some other Pins that are connected with the expansion board interfere with the DS shield behaviour. You mentioned to use Expansion Board 2.
Does the behaviour change if you remove all jumpers from the expansion board, and if yes, which jumper makes a difference?

[nespressif]

@robert-hh No, I really only use the expansion board to program the Sipy, once programmed I no longer use it, I only use the Sipy+DS Shield+3.7v Lipo Battery, only these three elements and nothing else, connecting the positive to the Vin of the Shield and the gnd to the gnd of the shield, if I include in my program the get_wake_status() function does not work and without the function, it works fine.

I think it is a problem of the deepsleep.py library, the get_wake_status() function uses the functions self.peek() and self.magic() and within this self._send() function that uses self.uart.write(bytes(data)). I don't know how it works, but the error may be due to the fact that you want to write data to Uart and it doesn't have the expansion board connected, you don't have uart. But this is just a supposition of mine, I don't know how the library works.

[robert-hh]

@nespressif That was nto my intention. You said it works with Expansion boartd and does not w/o. So I suggested to get with the expansion board the electrical status w/o by opening all connections besides GND and Vin.

[nespressif]

@robert-hh Ok, I didn't make myself clear. What I mean is that if I use the expansion board to feed the Sipy+ds shield, the program with the function get_wake_status() works fine and doesn't break.

But I just want to use the Sipy+ds shield+3.7v lipo battery.

Please, read my previous comment, I think I explain myself better. I will also upload some photos soon.

[robert-hh]

@nespressif I understood your statements well. I suggested to test, whether it fails with expansion board, but all jumpers from the expansion board removed.

[nespressif]

@robert-hh Ok, as soon as I can try to remove the jumpers from the expansion board and try.

[nespressif]

@robert-hh I've already tried it, with the expansion board it works well with the jumpers as well as without them.

[robert-hh]

@nespressif That's .. interesting. I assume it is hardware related. And without the jumpers and without a SD card inserted, at least according to the 1.2 schematics of the expansion board, nothing stays connected but the button at P10, which is open, and Vin, GND and 3.3V. There are additional capacitors at 3.3V and Vin (10µ + 0.1µ). Of course, the wires on the board add a small capcitive load, but that should not change so much.

[nespressif]

@robert-hh Like I said before, I don't know much about electronics. I think it's a problem with the deepsleep.py library, but it strikes me that no one has reported the error before, because if you feed it with batteries through Vin, you can't use the get_wake_status() function and therefore, you can't know why he woke up. But let's see if someone from Pycom can solve something for us.

I have two questions about Deep Sleep Shield, to see if you can help me:

1.- What is the Deep Sleep Onewire Comm Pin (P10) for?
2.- Looking at the diagram I see that there are some Pins with voltage dividers. Can I use some pin of the Shield, to measure the charge of the battery that feeds it through Vin?

Thanks for your help