Visit our Popular Forums

conachair · 28-08-2018, 06:24

This looks promising >
https://www.adafruit.com/product/904

Has anyone else tried this?

An INO219 current sensor connected across the battery meter 50mV shunt. Just croc clipped at the moment but with some calibration it seemed to track pretty well, plus the INO sense resister is still in place so it will skew readings a bit on the battery meter as well - though looks like maybe decent readings accurate down to maybe 0.5A might be possible. Surprising it's s good as it is so far considering it's reading for to fractions of a mV. Connected to n ESP8266.

So for us lucky ones with signalk onboard it will be quite easy to connect get the ESP to send signalk over wifi to a Raspberry Pi of something running signalk. Then easy to record to a database and watch just what's going on when charging/discharging. Very

Already I've an esp sending barometer, battery voltage and some thermometers data over wifi straight into signalk. Works great, the database viewer can overlay graph with data from 24h before.

conachair · 28-08-2018, 07:33

Well that was easy, signalk is great! Tiny bit of averaging seemed to help stabalize the amps a bit, looks pretty close now but a way to go yet. Simple to copy across the code from the other ESP and make a new signalk Amps message, then the sigK server does the rest. Just turn off the mains charger and watch what happens

Fridge pulls around 2.5A, turned off just after 15:18 so looks in the right ballpark anyway, maybe with some proper wiring to the shunt and a bit more averaging it could get pretty accurate.

conachair · 28-08-2018, 14:10

And a bit of an equalize , votage just creeping up with current slowly dipping a tiny bit. Think the calibration is slightly high from checking with a DC clamp meter - but for a sensor costing less than a London beer this is gold dust

rgleason · 30-08-2018, 17:55

Conachair
This looks to me like a great start for a good home built battery meter for SOC and Capacity. Can you use rednode to do calcs and present the results to the display?
Perhaps through experience, measuring trailing amps you can show a graph of capacity? I suppose peukart's equation would be a part of it too.

Then perhaps when you are sure about the readings, you could disconnect for low voltage, high voltage, maybe even control some charge sources.

Have you seen the fellow who is making a very sophisticated arduino regulator?
VSR Alternator Regulator

john61ct · 30-08-2018, 19:44

From our own Al Thomason

http://www.cruisersforum.com/forums/...rt-195059.html

conachair · 31-08-2018, 03:11

Quote:

Originally Posted by rgleason

Conachair
This looks to me like a great start for a good home built battery meter for SOC and Capacity. Can you use rednode to do calcs and present the results to the display?
Perhaps through experience, measuring trailing amps you can show a graph of capacity? I suppose peukart's equation would be a part of it too.

Then perhaps when you are sure about the readings, you could disconnect for low voltage, high voltage, maybe even control some charge sources.

Have you seen the fellow who is making a very sophisticated arduino regulator?
VSR Alternator Regulator

You probably could come up with some sort of monitor for SOC with Peukert's law but doubt if I'll bother, happy batts is the goal ( and gathering data for no reason is a enough on it's own, never know what you might fine

). So seeing amps in plotted at the end of the day when still at absorption should give better idea when you might need an hour with the genny in the morning.

Capacity is much more exciting! Yet to try it but in principle this should work >

Some transistors on a heat sink connected to a load, then to do a load test the ESP can look at a set amps figure and control the battery draw to that - IN ADDITION TO THE HOUSE LOADS. Now how cool would that be, every once in a while discharge down to 12v or whatever without having to turn anything off. Then stick the plot over the last time and see if it's very different. Not sure what differences temperature would make but maybe that could be factored in a little somehow, thermometer recording battery temperature is easy. Chronograf seems like it should handle the graphing easily, I haven't dug past the defaults yet but it seems very powerful, or a few lines of python.

Another thought was sharing the code/design - something like an instructables for boats might be nice. Somewhere to store and code, components & PCB. Sites around now will make PCB's and ship them for a few dollars, other sites to get all the components in one place for next to nothing. I really think a lot of people could put something together like this much easier than they think but are put off.

What a little discharge/charge cycle looks like, I found another IO219 library with more config options, tracks the battery monitor and a DC clamp meter very well now, prob accurate to 100mA or maybe 200mA, plenty good enough for us lot out in the anchorage

rgleason · 31-08-2018, 04:19

Conachair, we have a supplimental hardware page at OpenCPN wiki with a special section on RPI devices. And if you would like I could make a separate page for you. Are you a manager / editor for the Opencpn wiki? I can do that too.

https://opencpn.org/wiki/dokuwiki/do...rdware:screens

conachair · 31-08-2018, 04:47

Quote:

Originally Posted by rgleason

Conachair, we have a supplimental hardware page at OpenCPN wiki with a special section on RPI devices. And if you would like I could make a separate page for you. Are you a manager / editor for the Opencpn wiki? I can do that too.

https://opencpn.org/wiki/dokuwiki/do...rdware:screens

Thanks, though not sure if that's the right place for something like this - bit more openplotter/signalk oriented. Initially I'll shove some bits I've done up on github at some point & post links here.

How this one is wired >