Classic powers up, but dead short at input

[Vic]

The problem with running MPPT charge controllers in this way is that using the grid and especially with a step down transformer,
the MPP voltage can be too close to the Voc and this is where the CC has trouble....  It expects the input source to be
a limited source but to the Classic (or other MPP CC's), it can seem like Hoover Dam   ...   

Thanks boB for the detailed description of how MPPT CCs work in this situation.
73,  Vic

[boB]

During a brief curiosity thing, I was able to pull about 375W out of it, around 90-95V, with the Voc being somewhere in the 110V range... both fluctuated a little because of line variations (I'm at the end of a 100 ft extension), as well as some shifting that appeared to be associated with temp. The latter isn't entirely clear.

Steve

This amount of "cushion" should be fine.  Not sure why it died ?

[MadScientist267]

I'm hoping some answers come out of the bench... It's not something that I'd like to see repeated. There's a few plans that hinge around the hydro mode for this project, and of course I'd like to implement them safely. Any insight from the autopsy would be greatly appreciated; hopefully there's something to hear in the next few days, as it's on its way there as I write this.

Steve

[MadScientist267]

Ok, I've made some changes, done a lot of experimenting with trying different things...

I received it back from repair, and was charged for them because the unit had gone overvolt.

I have some proof that the previous high voltage input involved, while not helping anything by giving margins, was not the cause.

The classic is generating these dangerously high voltages (to the unit), internally.

My findings are reproducible (at least on this unit), and I have discovered that if the classic goes from absorb to float, either by force or automatically, the voltage skyrockets into potentially destructive levels.

If the voltage reaches 150, it goes into resting, and the voltage then begins to drop. If it does not, more of an RC type curve is revealed; the rate of climb slowing as it nears its highest, then will begin to fall. All of this occurs over the course of several seconds.

In addition, I decided to put some resistance across the input to see how hard and for how long the classic was driving power into it. It peaked out at ~35V into a 150R 10W. Remember, this was with *NO external input power*, just the classic still connected to the battery. This persisted for 15 or 20 seconds or so, and the resistor was getting very hot. Finally, it began to decline and settle in closer to the battery voltage, where it then let go upon entering resting mode.

I have video that captured a 129.3V excursion... Tho it should be done over as I was clearly a little irate while recording it.

Someone let me know if you really need to see it happen, I'll film it again, with and without the resistor, for your viewing pleasure.

I don't know if it's just this unit and there's something that is supposed to be inside protecting against this and it has failed, or if it is software causing a problem, but there's *absolutely* no doubt in my mind at this point that the same phenomena is what killed the MOSFETs the first go around.

At least the resistor can protect it, albeit not an ideal method.

Steve

[Vic]

Mad..,

Am not an MPPT CC Designer,

But would ask you to NOT place any "load" on the input of any Buck MPPT CC.

At some point,  seems to me,  that this will become no different than shorting the PV+ input to the PV -- input,  and you might be back in the territory of having damaged your CC.  At least,  it seems quite possible to me.

You mentioned some behavior of the Classic when it transitions from Absorb to Float.  What is the source of input power to the CC when it was in Absorb?

Opinions,   FWIW,   Vic

[MadScientist267]

First, the resistor was eventually placed there to keep the voltages from skyrocketing. I paid a fair chunk of change to get it fixed and returned over something "I caused". I'm calling it "insurance".

Second, a full wave rectified transformer, Voc ~22V, Vmp ~15V. It has nothing to do with the voltage shooting off scale.

The classic is effectively operating as a boost converter for about 15 or 20 seconds after going into float mode. I either A, soak it up somewhere or B, let it run off the chart again and get stuck with the next repair bill.

Which do you think I'm going to choose, hmm?

[Vic]

OK,  fine...

But,   would seem that it would be safe to either use an AC-operated battery charger (and not involve an MPPT CC),   or perhaps use real PVs into the Classic,  and perhaps not run the Home Brew DC PS,  that you appear to still be using.

However,  you do get to choose.  FWIW,   GOOD LUCK,   Vic

[MadScientist267]

There will be real PV.

And I bet it does it with that too.

Been there done that with the "AC powered battery charger"... I brought it back because it would have caused problems that don't exist.

This entire project requires thinking outside the box. I expect cookie cutter thinking from experience with cookie cutter installations.

What I need is for those involved in designing the unit to bring their expertise to the table so that this can be hashed out and resolved.

[dgd]

So the Classic was damaged by over voltage 
Just looking at the photo of your transformer and diode rectifier, I can take a good guess why you are having problems.

First the recification is leaving a good ac ripple in place, this is usually smoothed out somewhat by using some suitable power caps plus a few noise suppression caps. This is probably not effecting the Classic too much but nevertheless should not be there.

Second, the reason you are seeing some sharp rises in the voltage input to the Classic, as the Classic regulates input from Absorb to Float and Bulkmppt to Absorb is IMHO caused by Inductive Kickback. (google this) This will be your primitive transformer/rectifier design because of the lack of proper AC ripple removal and the complete absense of a snubber circuit.
This kickback is the quickly rising voltage and its likely the Classic has been exposed to over voltage before it's Hyper VOC detection could react.

Reading your recent postings it would appear you are definitely doing things to the Classic that it was not designed for, I suspect its on the way to being rubber-ducked again.
WHy not just get the solar panels and connect them then do your mad scientest stuff?

dgd

[boB]

FWIW, the Classic does allow some reverse (boost) current out of the PV input, but it is limited to around 9 amps or so.

This is common with MPPT charge controllers with synchronous buck converters.  Anything over that and it should
go to Resting immediately.  But you can draw some power from the input side for about 90 seconds. The voltage
that it goes up to should also be limited based on the previously measured Voc before it turned on the first time.
IIRC, that is around 10 volts over Voc.

Sorry Steve that we charged you for this.  I had forgotten all about this.
boB

[boB]

Steve, I have a question if you have the answer...  I'll also ask Jose in Warranty.
What was the high voltage reading of the input to your Classic ?  How high over
150V did it go ?

I saw your IRC post also and saw that you noticed it go to high voltage when transitioning
from Absorb to Float...   I will look into this right away.

The input voltage (PV) ~should~ be limiting itself to about 10 volts above Voc... But maybe
it did not see any reverse current because there was not PV connected to the input ?

Will do a sanity check today.  Or, insanity maybe.

boB

[RossW]

Just looking at the photo of your transformer and diode rectifier, I can take a good guess why you are having problems.

First the recification is leaving a good ac ripple in place, this is usually smoothed out somewhat by using some suitable power caps plus a few noise suppression caps.

I witnessed a great deal of discussion about this in IRC between Steve and Bob.

As I recall it, the recommendation was specifically *NOT* to put big filter caps on it, because this may make the supply too "stiff" and actually cause damage! I believe the original supply DID have filters (was a fairly grunty commercial supply) but deemed "unsuitable for this use".

So while your comments may be right in a general sense, please don't just assume that you're right and everyone else is wrong, without the benefit of hindsight and/or detailed discussion with the manufacturer about these quite specific issues.

[dgd]

Ross,

I did say that the ripple probably did not effect the Classic.
I most certainly did not assume everyone else was wrong and I was right, I have no idea who everyone else is anyway as I comment only on the forum postings NOT some previous off forum discussion I did not participate in.
Too 'stiff', yea right, got to be better than the existing non-stiff AC component on the DC being supplied to the Classic.

My main point was the rising voltage at Classic regulation steps. Unless the laws of Physics have changed that transformer/rectifier combo is
voltage spiking due to inductive kickback.  Is the design of the Classic capable of dealing with these? Methinks not hence the broken Classic sent back for repair.

dgd

[RossW]

Too 'stiff', yea right, got to be better than the existing non-stiff AC component on the DC being supplied to the Classic.

As I understand the discussion, the whole "stiff" issue was a serious consideration - the ability of the transformer / rectifier and any filter caps ability to deliver "substantially more current" than the classic would see from a normally fairly "soft" PV installation, including the comparatively long supply cables.

In discussion, Steve chose to use a modest transformer, with significant current limiting, to prevent the issue happening again.

Unless the laws of Physics have changed that transformer/rectifier combo is voltage spiking due to inductive kickback.

But riddle me this... with a bridge rectifier between the classic and the xfmr, and the primary of the xfmr unplugged, how is there going to be any inductive kickback? If the output of the classic is positive voltage, the diodes in the bridge will be reverse-biased, and no current could flow to the xfmr secondary, so there can be no inductive kickback, surely?

I'm talking out of turn, I'm not privy to the entire setup, but from what I have seen go past on the screen, it's quite clear that it isn't an "ordinary" situation. (There was also a video posted earlier today - volts were stable about 17V or so with the transformer connected. Then the classic was put to float mode and the volts rapidly climbed. At about 70V (WELL above the open-circuit output of the transformer), the input to the xfmr was removed, yet the volts CONTINUED to climb to almost 130V. I don't know what's going on, but I can't explain it.... but I think it's quite clear it ISN'T just inductive kickback from the transformer!)

[dgd]

OK, like you I'm not privvy to all of the previous discussions.
So the stiffness issue is when the Classic regulates then possible inrush of current due to power caps could damage the Classic.
I leave this discussion to those with full knowledge of the situation.
As for kickback and  the bridge rectifier diodes,  the issue here may be those rectifier diodes may be too slow to deal with fast rising spikes
and they may be rated for much lower peak voltage.
Anyway, don't know but I personally would not 'test' a Classic with that power source or connecting a load resistor between the PV inputs.
...back to interesting stuff now s my Cubie arrived today Yay!

dgd