iSpindel - digital WiFi hydrometer

[kelper]

Is anybody finding that Ubidots is missing reports? I only graphed 11 points in the last four hours. Should have been at least 15.

 

[Harry Bloomfield]

you mean it's harder to set up?

I had no issues building or setting up my iSpindel, I will admit I struggled a little getting Ubidots working. No doubt it does need a bit of savvy to get it working, but it doesn't need bluetooth device to be dedicated to it and its data is accessible anywhere you happen to be.

 

[Harry Bloomfield]

Is anybody finding that Ubidots is missing reports? I only graphed 11 points in the last four hours. Should have been at least 15.

I have nothing brewing at the moment, but I have noticed Ubidots miss the odd report. I have just assumed that the report was missed by my router.

 

[Bonley]

This thread is gold. I just built an iSpindel from an OSD kit off ebay (The Jeffrey V2.1). Calibration went fine. It's position in the tube is repeatable enough. Linked up to Ubidots. It only lasts about 8 days though?

That was with a Samsung 25R 2500mAh 18650 (43g), which I verified the capacity of. It stopped reporting when voltage dropped from 4.2v to 3.1v after 8 days. 15 minute interval. I changed it to 1hr interval and it lasted 8 days, again.

I had trouble with WiFi signal being <-90dB due to it being in a SS brew bucket, in a fridge, in the garage. Thought this might be part of the battery problem, but a WiFi extender and a passive antenna got that up to ~-45dB and still the same 8 day life.

Bought a 3500mAh Panasonic NCR18650B (47g), verified the capacity, and it's currently dropping voltage at the same rate as the 2500mAh...

I'm running out of ideas. Anyone got a clue as to what the problem could be?

 

[The-Engineer-That-Brews]

This thread is gold. I just built an iSpindel from an OSD kit off ebay (The Jeffrey V2.1). Calibration went fine. It's position in the tube is repeatable enough. Linked up to Ubidots. It only lasts about 8 days though?

That was with a Samsung 25R 2500mAh 18650 (43g), which I verified the capacity of. It stopped reporting when voltage dropped from 4.2v to 3.1v after 8 days. 15 minute interval. I changed it to 1hr interval and it lasted 8 days, again.

I had trouble with WiFi signal being <-90dB due to it being in a SS brew bucket, in a fridge, in the garage. Thought this might be part of the battery problem, but a WiFi extender and a passive antenna got that up to ~-45dB and still the same 8 day life.

Bought a 3500mAh Panasonic NCR18650B (47g), verified the capacity, and it's currently dropping voltage at the same rate as the 2500mAh...

I'm running out of ideas. Anyone got a clue as to what the problem could be?

Hmmm... I'd have thought that a 2,500 mAh battery should power something like that for at more like a couple of weeks.

sounds to me that your unit is taking a bit too much power, or waking up too often. With an embedded controller design like that should be asleep most of the time (consuming basically zero current), and then waking itself up every hour or so to take and report a reading.

Afraid I'm not familiar with the software you're using, but is there a setting on it controlling how frequently it takes a reading? If so, then you might need to back it off a bit.
Other possibility would be a duff board - if you have a multimeter you can use to measure the current consumption then perhaps one of the other users on here could compare with theirs.
Are there any extraneous components on the board (power LEDs) that needed to be removed?

 

[kelper]

My iSpindel has a Samsung R25 battery and goes two weeks easily. I'm logging every 15 minutes. The voltage drops very slowly. In two weeks, voltage dropped from 4.08 to 3.80.

 

[gingerneil]

Have you checked the values of your resistors ?

 

[Bonley]

Hmmm... I'd have thought that a 2,500 mAh battery should power something like that for at more like a couple of weeks.

sounds to me that your unit is taking a bit too much power, or waking up too often. With an embedded controller design like that should be asleep most of the time (consuming basically zero current), and then waking itself up every hour or so to take and report a reading.

Afraid I'm not familiar with the software you're using, but is there a setting on it controlling how frequently it takes a reading? If so, then you might need to back it off a bit.
Other possibility would be a duff board - if you have a multimeter you can use to measure the current consumption then perhaps one of the other users on here could compare with theirs.
Are there any extraneous components on the board (power LEDs) that needed to be removed?

I took the LED off the Gyro, left the one on the Wemos D1 mini. Most of the time there are no lit LEDs on the board when it's in use. I've tried 900s interval and 3600s interval, and the voltage drop is the same. The LED on the Wemos is off most of the time, and flashes briefly at the update interval before going dormant again. I've estimated the current consumption to average 13mA, but I'll get out a multimeter and let you know.

 

[The-Engineer-That-Brews]

13mA sounds plausible for when the thing is active, but in between measurements/reports it should be taking almost nothing, assuming the software has been decently written.
Are you arriving at 13mA by dividing the capacity of the battery by how long it lasts? If so, that's not telling you much: you need to take a measurement while it's transmitting, and another while it's quiescent :-)

 

[pottsworth]

Do you have 3 resistors or 2 resistors and a bat43 diode?

I believe that connection from d0 to rst is what controls the board going into deep sleep.

If you have 3 resistors, maybe try the diode, or check the diode is the right way around.

 

[Bonley]

Have you checked the values of your resistors ?

From the colours on them I have a 220K, 4.8K, and a BAT diode. Supposedly a BAT43 from what the kit said, but not sure how to check now it's soldered under the Wemos D1 mini.

 

[Bonley]

13mA sounds plausible for when the thing is active, but in between measurements/reports it should be taking almost nothing, assuming the software has been decently written.
Are you arriving at 13mA by dividing the capacity of the battery by how long it lasts? If so, that's not telling you much: you need to take a measurement while it's transmitting, and another while it's quiescent :-)

Yes, it was an average derived as you guessed. I'm also on the latest firmware (6.3.1) from the github. Nothing custom.

Turns out all my digital multimeters have blown fuses on the mA scale. Ordered more and tried my "ornamental" Avometer. Calibration seems questionable but at least good for indication. Not certain I'm reading the scale right either. The scale has 100 and 25 range options. I'm guessing you read the 100 scale for 1/10/100/1000 settings, and 25 scale for 2.5/25/250/2500 settings.

On the 1mA setting the meter reads a hair over 5 on the 100 scale while the iSpindel switch is in the off position. So this is 0.05mA? one the 250mA setting I read ~6.4 on the 25 scale, so 0.064mA? Switching it on pegs the meter until you change the sensitivity to 100mA. At which point it shoots up to ~90(mA?) on the scale for ~2.5 seconds as the LED is flashing, then settles at 80(mA?) for a further ~6 seconds, after which it drops down to a stable 10(mA?) and if you swap to the 10mA setting, it pegs the meter.

All of which sounds too high to me. But it being analogue makes it very non-ambiguous as to when current is flowing, when it isn't, and the relative difference between them. I'm asking around my mates to borrow a better meter. Not sure I have anything that draws constant current to calibrate it against... Maybe my battery charger/discharger? I'll let you know if/when I learn more.

 

[The-Engineer-That-Brews]

Yes, it was an average derived as you guessed. I'm also on the latest firmware (6.2.0) from the github. Nothing custom.

Turns out all my digital multimeters have blown fuses on the mA scale. Ordered more and tried my "ornamental" Avometer. Calibration seems questionable but at least good for indication. Not certain I'm reading the scale right either. The scale has 100 and 25 range options. I'm guessing you read the 100 scale for 1/10/100/1000 settings, and 25 scale for 2.5/25/250/2500 settings.

On the 1mA setting the meter reads a hair over 5 on the 100 scale while the iSpindel switch is in the off position. So this is 0.05mA? one the 250mA setting I read ~6.4 on the 25 scale, so 0.064mA? Switching it on pegs the meter until you change the sensitivity to 100mA. At which point it shoots up to ~90(mA?) on the scale for ~2.5 seconds as the LED is flashing, then settles at 80(mA?) for a further ~6 seconds, after which it drops down to a stable 10(mA?) and if you swap to the 10mA setting, it pegs the meter.

All of which sounds too high to me. But it being analogue makes it very non-ambiguous as to when current is flowing, when it isn't, and the relative difference between them. I'm asking around my mates to borrow a better meter. Not sure I have anything that draws constant current to calibrate it against... Maybe my battery charger/discharger? I'll let you know if/when I learn more.

Nice!
Heh ok it's been soooooo long since I used an analogue AVO, so I can't rely on my memory so far as the switch settings go - however if it's just the fuse that's gone, on your mA input for the digital meter, you'll be safe just to bypass it (the fuse) with a bit of wire temporarily.

Is your AVO one of these? Takes me back... AVO model 8 universal multimeter

 

[Bonley]

Nice!
Heh ok it's been soooooo long since I used an analogue AVO, so I can't rely on my memory so far as the switch settings go - however if it's just the fuse that's gone, on your mA input for the digital meter, you'll be safe just to bypass it (the fuse) with a bit of wire temporarily.

Is your AVO one of these? Takes me back... AVO model 8 universal multimeter

Yep, it's a Model 8 MKII my grandad left me. I've actually ended up using it quite often because it's rather robust and has a very large operational range.

I've borrowed my mates Fluke 115 which shows 0.000A when the iSpindel switch is in the off position, but 0.064mA is below the meters minimum detection limit. Turning it on shoots up to ~87mA for the aforementioned ~2s, then ~80mA for another ~6s, then sits at 11mA until the next update. So the AVO meter was shockingly pretty accurate!

Looking online Deep sleep should be more like 77uA, so it seems like my Wemos isn't going into deep sleep like it should. It shouldn't be losing 0.064mA when the power switch is in the off position either. I've left the Fluke logging and it's gotten 0.011A min, 0.099A max, 0.012A average. Which is roughly what I estimated from the battery capacity and run time.

I've checked and GPIO16/D0 is connected to RST via the BAT43. It "wakes up" relatively consistently and uploads data... Digging into the logs it shows the time between data points varies from 56mins 37s up to 57mins 22 seconds, thats on a 3600 interval. With a couple where it seems to have skipped an upload. Is this normal? With an interval of 900 it's similar, ranging from 14minutes 12 seconds up to 14 minutes 25 seconds, then the odd missing upload. I'm wondering what would stop it from going into deep sleep mode?

 

[The-Engineer-That-Brews]

Yep, it's a Model 8 MKII my grandad left me. I've actually ended up using it quite often because it's rather robust and has a very large operational range.

I've borrowed my mates Fluke 115 which shows 0.000A when the iSpindel switch is in the off position, but 0.064mA is below the meters minimum detection limit. Turning it on shoots up to ~87mA for the aforementioned ~2s, then ~80mA for another ~6s, then sits at 11mA until the next update. So the AVO meter was shockingly pretty accurate!

Looking online Deep sleep should be more like 77uA, so it seems like my Wemos isn't going into deep sleep like it should. It shouldn't be losing 0.064mA when the power switch is in the off position either. I've left the Fluke logging and it's gotten 0.011A min, 0.099A max, 0.012A average. Which is roughly what I estimated from the battery capacity and run time.

I've checked and GPIO16/D0 is connected to RST via the BAT43. It "wakes up" relatively consistently and uploads data... Digging into the logs it shows the time between data points varies from 56mins 37s up to 57mins 22 seconds, thats on a 3600 interval. With a couple where it seems to have skipped an upload. Is this normal? With an interval of 900 it's similar, ranging from 14minutes 12 seconds up to 14 minutes 25 seconds, then the odd missing upload. I'm wondering what would stop it from going into deep sleep mode?

Good sleuthing. I'm familiar with coding for several different types of microcontrollers, so I know generally how to put them into deep sleep mode; but unfortunately I know nothing about the specifics of the Spindle. Is there a user group or GitHub page where you could ask?

 

[kelper]

Good sleuthing. I'm familiar with coding for several different types of microcontrollers, so I know generally how to put them into deep sleep mode; but unfortunately I know nothing about the specifics of the Spindle. Is there a user group or GitHub page where you could ask?

The components are listed in post 1
The Github stuff is at GitHub - universam1/iSpindel: electronic Hydrometer

 

[pottsworth]

Mine occasionally misses uploads - I believe it is because the wemos only has a short window, so if it doesn't manage to connect to wifi it gives up for that upload.

There is a dedicated Facebook group for the ispindel, it might be worth asking in there.

One bought - if you're drawing current even with the switch off, could it be a dodgy tp4056 module?

Would unfortunately probably require de-soldering the 4056 or d1 to prove.

 

[Bonley]

I dug out a spare Wemos d1 Mini and wrote a simple little bit of code that just periodically connects to the wifi then goes into deep sleep. I measured the current consumption and it was as it should be: basically nothing in deep sleep mode.

So I loaded that same code onto the iSpindel and it wouldn't even turn on. Probed voltages and the TP4056 was dropping 3.3v of the available 3.9V, leaving only 0.6V for the Wemos. Ordered 5 new TP4056 on prime and swapped it out. Now it turns on, but It's getting the same ~80mA connected to wifi , and ~10mA in deep sleep.

Desoldered the TP4056 and jumpered it out. Still the same. Removed the D1 mini and replaced with the one I knew was consuming "the right" amount of current. Replaced the TP4056. Checked, now consuming 1mA or less in deep sleep!

So both a dodgy TP4056 and Wemos D1 mini.

Just re inputting all the settings, and re calibrating it in it's new form. Then I'll start an endurance stress test. Thanks for the help everyone!

 

[The-Engineer-That-Brews]

I dug out a spare Wemos d1 Mini and wrote a simple little bit of code that just periodically connects to the wifi then goes into deep sleep. I measured the current consumption and it was as it should be: basically nothing in deep sleep mode.

So I loaded that same code onto the iSpindel and it wouldn't even turn on. Probed voltages and the TP4056 was dropping 3.3v of the available 3.9V, leaving only 0.6V for the Wemos. Ordered 5 new TP4056 on prime and swapped it out. Now it turns on, but It's getting the same ~80mA connected to wifi , and ~10mA in deep sleep.

Desoldered the TP4056 and jumpered it out. Still the same. Removed the D1 mini and replaced with the one I knew was consuming "the right" amount of current. Replaced the TP4056. Checked, now consuming 1mA or less in deep sleep!

So both a dodgy TP4056 and Wemos D1 mini.

Just re inputting all the settings, and re calibrating it in it's new form. Then I'll start an endurance stress test. Thanks for the help everyone!

Well done! Fingers crossed you’ll get a better battery endurance now :)

 

[Mr Sworld]

Hi,

I'm hoping someone can help me with the Plato Formula. I don't have a full MS Excel copy so can't get the macro to work. Could someone plug the following measurements in and post the formula?

S.G. Tilt
1.000 24.8
1.010 25.7
1.012 26.37
1.016 27.6
1.018 29.3
1.022 30.9
1.026 33.1
1.032 35.8
1.038 39.2
1.046 42.8
1.068 56.1
1.090 64.3

Many thanks!