While testing my newly purchased Classic 150 I have observed some unexpected behavior.
As an input source I have been using an adjustable power supply with foldback current limiting at about 25A. The battery bank being charged is ~2400 Ahr and 12V. The battery is in pristine condition and typically only requires about 0.3A to hold it at it's temperature compensated float voltage at it's current temperature of about 11C. It therefore takes a long while to drift down to nominal float voltage when transitioning from absorb to float when the battery is truly at 100% SOC.
I have been testing the CC to ensure that all functions of interest to me perform as expected.
I noticed that if I remove the input from the CC [input is truly seeing an open circuit] several weird things happen. At that moment I was trying to verify the transistion to resting which I think I remember seeing in some thread is triggered by 90 seconds without input. I removed the input immediately after the MNGP showed transition to FLOAT from ABSORB. Mine doesn't go to resting for as much as 6 hours without input. This is not a one time fluke but is reproducible. The MNGP status display will show FLOAT up until a couple of minutes before finally transitioning to resting. The last couple of minutes it will show MPPT FLOAT. It typically makes the transition to resting once the battery voltage has finally drifted down 0.2 to 0.3V below the FLOAT voltage setpoint. During this period of non-resting without input, the MNGP display shows an input voltage well above the battery voltage, initially almost double when the battery is still up near the absorb voltage and drifting downwards from there spending most of the time 0.4 to 0.6V above battery. I checked with a meter and this is real, not a measurement artifact. My first guess was that the converter MOSFETs were still switching and the various GD,GS, and DS capacitances were somehow forming a voltage doubler charge pump feeding voltage back to the input. Alternately, since highside MOSFETs likely would be N-Ch devices the required highside gate drive voltages was by design being derived by using the main switches to drive a charge pump and somehow this highside gate drive voltage was finding it's way to the input causing the controller to be fooled into thinking it had input power. Investigating further, I hooked up an adjustable constant current load to the input and found that it could source 2A while still maintaining the input a few tenths of a volt above battery. 2A was the limit of my constant current load so I could not test beyond that.
Possibly relevant configuration items:
Solar mode
VariMax disabled
BTS not connected
LoMax on
No loads on battery other than CC
CC resting power draw 4.5W at ~13.2 Vbatt
CC not resting power draw 7.5W at ~13.2 Vbatt
At random intervals MNGP will display Got Comm? briefly
Up to six hours to rest seems not normal. The question is what's going on here?
Interesting testing procedure you have.
However I always thought that the Classic MNGP displayed what it was doing, FLOAT, BULK MPPT, ABSORB etc., depending on the SOC of the battery.
So the Classic is completing an ABSORB time and falling back to the FLOAT stage. It detects the Batt V is still above the set Float V so will stay at the FLOAT setting. Only when the batt V drops below the float setting does the Classic look to the input to send power to the battery. It detects no input so transits to RESTING.
Perhaps that 90 second to RESTING only commences when the Classic needs input.
Maybe do the same testing again with some loads included, eg an inverter and an electric jug boiling some water. I bet it goes to RESTING a lot
sooner.
dgd
What firmware version is this Classic running? boB will be by in a bit to answer the more technical questions about the FET's and stuff.
Ryan
@ dgd. Your bet is correct. Any variation in the test procedure whereby the battery voltage reaches a few tenths below the CC float setpoint will hasten the transition to resting.
@HC. MNGP FW 1042, Classic FW 1181 which are the lastest in both cases AFAIK.
I to would say if the Vbat is above the set point than that is likely the cause. boB will be a better judge.
Ryan
Quote from: Sweet Taterman on January 27, 2013, 06:06:57 PM
@ dgd. Your bet is correct. Any variation in the test procedure whereby the battery voltage reaches a few tenths below the CC float setpoint will hasten the transition to resting.
So it looks like your C150 is behaving normally :)
I hope you get your system producing power soon and reduce that grid supply bill...
dgd
Hi guys...
While the Classic is running (not resting), the switching will create the input voltage from the battery side.
That is an artifact of the buck circuit topology (boost in the opposite direction)
The Classic is supposed to go to resting when three (3) conditions are satisfied... 1, when the battery
voltage is below the set point voltage or a couple or 3 tenths of a volt below the set point,
2, after 90 seconds and 3, if the wattage is below a certain amount. That wattage threshold
should be around 20 watts by default.
If all 3 of those conditions are correct, then it may be that your Classic needs to have
a "Vulcan Mind Meld" (VMM) initialization. But normally, the Classic should come to
you with the VMM already cocked and ready to fire when you first fire it up. i.e., when you
first powered it up, it should ask you what mode (SOLAR--WIND--HYDRO), then
the voltages and time set then it should send a bunch of data to the Classic itself
from the remote (MNGP) LCD panel.
boB
Thanks boB. All observed behavior is consistant with the algorithm you have provided. May I presume that the occasional random Got Comm? is of no consequence?
Quote from: Sweet Taterman on January 28, 2013, 07:55:26 AM
Thanks boB. All observed behavior is consistant with the algorithm you have provided. May I presume that the occasional random Got Comm? is of no consequence?
Yes, sorry. That is what I have seen somewhat. We are working on that though.
BTW, what software revision or date is in your Classic and MNGP (Remote) ??
You can find this by pressing the STATUS button a few times to see that
screen which also shows what voltage of Classic you are connected to
and its serial number in the lower right corner of that revision screen.
Another thing to check on is to see if the MNGP to Classic communications is
talking over address 10. In the main status screen, (showing input output
voltages, watts, current etc), HOLD down the LEFT-Arrow key and "tap" the
UP or DOWN arrow key. If you are on address 10, it will change to 11 or 9
(or 12 or 8 or whatever) and say "Unused Address". Just put it back
to 10 if that was the case. Let us know if it is something else, like, 255.
Address 255 will also work but might give you the Got Comm ? message
more often.
Otherwise, if it is indeed address 10, just accept the annoying message for
now because we are speeding up communications and should hopefully have
an update for that soon.
Thanks !
boB
@boB
Thanks for the reassurance r.e., Got Comm?
MNGP FW 1042, Classic FW 1181 which are the latest in both cases AFAIK.
Verified unit is on address 10.
Quote from: Sweet Taterman on January 28, 2013, 04:33:18 PM
@boB
Thanks for the reassurance r.e., Got Comm?
MNGP FW 1042, Classic FW 1181 which are the latest in both cases AFAIK.
Verified unit is on address 10.
Yep. You're up to date.
Working on it...
boB
Sweet Taterman, on the got comm ? thing...
Do you have Follow Me enabled or do you only
have one Classic ?
boB
@boB
I have only one classic.