A Forum run by Enthusiasts of MidNite Solar

The Open Source software/hardware corner => The Black Box project => Topic started by: zoneblue on March 10, 2014, 04:16:25 PM

Title: Classic Amps anomoly.
Post by: zoneblue on March 10, 2014, 04:16:25 PM
Ok so heres the clearest example to date of the Classic Amps register anomoly.

It was a clear blue day, and there was no change in the house loads all morning, due to the fact that we were all having a holiday lie in. In the graph below, something happened at 8:34:21am, that caused the classic amps register to drop by 1.7amps, for no apparent reason, mid bulk. This caused the apparent house loads to drop (below zero) when they did not.

What is going on there?

As before blue curve is classic amps, pink is WBjr.
Title: Re: Classic Amps anomoly.
Post by: zoneblue on March 10, 2014, 04:57:11 PM
This is an unfiltered, one hour detail. Blue curve drops to zero for 6 seconds, then resumes at 1.7amps too low, verified by WBJr. Different from other sweeps. FTR the classic didnt reboot.

The 1sec register data  record is here: www.zoneblue.org/files/2014-03-10.zip
Title: Re: Classic Amps anomoly.
Post by: boB on March 11, 2014, 03:20:13 AM

I'm not sure what is happening here BUT, another useful graph line might be the Classic input voltage.

This would help tell us if the Classic is trying to voltage regulate or whatever.

boB
Title: Re: Classic Amps anomoly.
Post by: Halfcrazy on March 11, 2014, 06:34:54 AM
It has to be a calibration issue. There is no way the classics Amps could be less than the WBjr amps. I sort of wonder if it could be something going on in the current sense? Maybe we want to swap it out and see if the problem goes away? How often does it do this?
Title: Re: Classic Amps anomoly.
Post by: zoneblue on March 11, 2014, 03:57:12 PM
Ryan, i noticed it a couple of times shortly after the WBJr was installed, where the condition persisted for an hour or more during late bulk and abosrb. Then only ocassional short spikes here and there til now, more likely attributable to the noise in the registers. Previously it occured on cloudy days and so was less easy to seperate out from the changing classic amps. This is the first time its occured in the middle of a clean bulk stage amps ramp up,  and makes it so easy to spot.

Blackbox curently collects two sets of raw data. The file above is an example of a daily record, which currently stores only registers: 4115,4116,4117,4120,4121,4132,4371 once per sec.
I have those files for each day back til 11th Jan this year.

The other data file is a rotating log file containing the whole range from 4101 to 4376, that gets overwritten from one second to the next. Blackbox reads this file each minute and stores the following registers in the database: 4101 4102 4103 16387 4127 4126 4120 4131 4132 4133 4134 4119 4115 4117 4371 4116 4121 4138 4118 4369, but they are processed into readable values.

We could add more (debugging) registers to either list to collect in future if that will help.  If so let me know which registers are of interest.  Heres your graph:


Title: Re: Classic Amps anomoly.
Post by: zoneblue on March 11, 2014, 04:51:27 PM
Btw the following day was good as gold. But it (also clear blue) has that same 6s bulk pause around the same time, 8:17am. Might be a clue.

Guys i wouldnt stress too much about it, maybe just keep an eye out for it in the future. Theres quite a lot of variables not the least of which is me doing modbus reads every second.