Author Topic: Overvoltage shutdown issue  (Read 8252 times)

0 Members and 1 Guest are viewing this topic.

Offline dochubert

  • Sr. Member
  • ****
  • Posts: 265
  • Karma: +4/-0
  • 11 powerjacks, 9 that run, 3 actually used daily
Overvoltage shutdown issue
« on: December 11, 2015, 07:16:15 pm »
Hello All,
I've read through many posts here and learned a ton about Powerjack inverters already.  My search for good solutions to the overvoltage shutdown problem that seems common to PJ inverters is what led me to this site, and the best answers I've seen so far anywhere.  Thanks, thanks, thanks!  Someone commented previously that Oztules was a genius.  If not I'd have to say that he's at least a reasonable facsimile!

Ok
The last and apparently generally accepted fix is the 1 meg resistor soldered in parallel to the 10 meg resistor R14 (says cut if low next to it) on the control board, with the alternate being a 5.6v zener diode instead of the 1 meg resistor.
I do have some questions, though:

1.  Does anyone know what the 1 meg resistor does to the setpoint?  At what voltage will it shutdown now?
     Does it not shutdown on high voltage ever now?
2.  Has anyone tried a potentiometer at R14 to try to set the setpoint to where it should  have been set?
3.  Is a 1 meg resistor used on all voltage models? 12v 24v 48v?  Other values?
4.  If the value isn't important in reducing the resistance, why not just short the resistor?  (Maybe the ic it feeds doesn't like zero  resistance?)
5.  I'm assuming the use of the zener instead is intended to prevent ANY overvoltage shutdowns, yes?
6.  Does this change have any effect on the low voltage shutdown setpoints?
7.  Could have sworn I had a couple more....

I'm not a fan of disabling safety functions, but correcting design errors is always good.  A 12v "name brand" inverter generally has its high voltage shutdown at at least 16v, and as high as 19v.  I would guess that PJ setting it at 15v on 12v units (30v on 24v and 60v for 48v units) is either poor understanding of the world market's intended usage, or more likely their attempt to safeguard substandard parts/poor quality assembly and limit burned units that get returned to them.  Maybe both
If it sounds like I'm disparaging Powerjack, sorry. Not my intent.  I actually agree completely with Oztules that for the money they are GREAT inverters, especially since he and others here have solved most of the design shortcomings.
   
So, with that all said,  I would like to nail down a resistor value that sets the alarm/shutdown to say 17v (thats 68v on my 48v unit)
I can now prevent the annoying midday low load full sun high voltage shutdowns using that zener or resistor, but it would be nice to know that if something really DID go south, that there is still some protection in there.
So that's the current quest.
More to follow
We're all going to DIE!  (eventually)

USA = Communist former republic
---  dochubert --

Offline oztules

  • Forum Advisors
  • Hero Member
  • ****
  • Posts: 1177
  • Karma: +105/-8
  • Village idiot
Re: Overvoltage shutdown issue
« Reply #1 on: December 11, 2015, 09:26:31 pm »
There is a reason I can't help you with this... poor memory... been a long time since I tracked through the circuit board to find out that the big resistor was the way to stop the problem.... can't remember how the voltage gets to there.... but the 1m simply loads the voltage divider being used so that more voltage is needed on the high side arm to pull the reference point up.

I don't want the inverter to drop out ever from over voltage, as it would only exacerbate the problem causing it.. I choose panels that are 60 cell, and build my own controllers for over 100 amps... I don't expect failure, but if it did, then the voltage would only rise up a few more volts, and the silicon panels would not have the emf to push it much further anyway.... so I won't see any high side damage.

Low side, will never happen, as there is 3 days usage in the battery bank if the sun did not turn up ....if it puts in an appearance behind clouds or whatever, the battery bank will end the day fully charged or near to it, as I have a mountain of solar there, and a dull day still fills the bank... just how I designed it to do...
....still have 10kw spare... in the shed.

So, I have not had the reasons to find out the answer for you, nor will I ever need to,

Wish you well in your trevails...


.................oztules
Flinders Island...... Australia

Offline dochubert

  • Sr. Member
  • ****
  • Posts: 265
  • Karma: +4/-0
  • 11 powerjacks, 9 that run, 3 actually used daily
Re: Overvoltage shutdown issue
« Reply #2 on: December 13, 2015, 04:47:35 pm »
Finally got some time to work on this a bit...

First I removed the 10 meg resistor and soldered a 10 meg trimmer pot in its place.

4938-04940-1

This control board is the one that came with my 15kw PJ   I had so much trouble with it that I bought a replacement and swapped it out so its now my spare.  The new one is more stable but still shuts down (of course).  I suspect fixing the high voltage issue will make it usable once more, so we're going to find out.  Anyway, next I have to swap it back in, which won't be happening until tomorrow when I hope to have more time for it.

In the meantime I want to post here some voltage drop readings across R14 at various battery voltages taken on my 48v inverter while running a light load over the last two days;
At 49.8v    3.27v
     50.2v   3.55v
     53.2v   3.88v
     54.5v   3.96v
     56.1v   4.06v
     56.7v   4.13v
It seems clear from these readings across the original 10 meg resistor that somewhere around the 5v mark would likely be when the alarm/shutdown happens.  It follows that reducing that resistance value would cause it to require a higher battery voltage to get to the assumed 5v shutdown.  And in fact the others that have added the parallel resister have proved that theory true.  Now we just need to determine what resistance will give us a 68v shutdown.

till next time...
We're all going to DIE!  (eventually)

USA = Communist former republic
---  dochubert --

Offline dochubert

  • Sr. Member
  • ****
  • Posts: 265
  • Karma: +4/-0
  • 11 powerjacks, 9 that run, 3 actually used daily
Re: Overvoltage shutdown issue
« Reply #3 on: December 13, 2015, 05:08:39 pm »
Looking like I put this post in the wrong place.
Just now figured that out.
Can it be fixed or wiped and i'll start over when i get smart enough to do it right.
My apologies
dochubert
We're all going to DIE!  (eventually)

USA = Communist former republic
---  dochubert --

Offline rossw

  • Senior Moderator
  • Hero Member
  • *******
  • Posts: 879
  • Karma: +35/-0
  • Grumpy-old-Unix-Admin
Re: Overvoltage shutdown issue
« Reply #4 on: December 13, 2015, 08:33:57 pm »
Looking like I put this post in the wrong place.
Just now figured that out.

Where do you want it moved, mate?

Offline dochubert

  • Sr. Member
  • ****
  • Posts: 265
  • Karma: +4/-0
  • 11 powerjacks, 9 that run, 3 actually used daily
Re: Overvoltage shutdown issue
« Reply #5 on: December 14, 2015, 12:22:02 am »
I had intended to put it in the controls, inverters section with the other discussions about the powerjack inverters, unless you think someplace else.
Thanks very much for your help
We're all going to DIE!  (eventually)

USA = Communist former republic
---  dochubert --

Offline dochubert

  • Sr. Member
  • ****
  • Posts: 265
  • Karma: +4/-0
  • 11 powerjacks, 9 that run, 3 actually used daily
Re: Overvoltage shutdown issue
« Reply #6 on: December 14, 2015, 03:18:00 pm »
Thanks Rossw for moving this where it belongs and for not blasting me for my new guy error.
I appreciate it.

Well, I put the modified control board back in and the unit runs fine but partly cloudy today so not able to get good readings. Battery voltage is bouncing around too much.
The trimmer pot was set to the 8.6 megohm max it would go to before i soldered it in, so its already a bit lower resistance than the original, but not too much.  The readings I could get indicate slightly lower voltage drop across the pot (R14) than with the original 10meg, which is what I expected.  So far so good.  Now I need some run time with good steady sun, but I don't think it is going to be today.  Mother nature not cooperating.
There have been no alarms or shutdowns so far even though I saw brief 59-60 volt transients.  A good sign but too early to get excited about.

An interesting (to me anyway) sidenote:
I have been messaging back and forth to Powerjack in Taiwan (or China)  since I bought this inverter in September.  They have always been very polite and prompt to answer, but generally consistently unhelpful.  It is possible they deliberately gave me wrong information.  Twice.  My most recent communication was on the 10 meg R14 resistor I'm playing with now.  When I stated that the 10 meg R14 resistor was for setting the high voltage shutdown, I was told by them the resistor is a 1.2 meg resistor and I should try reducing it to 1 meg to increase that setpoint.  That sort of goes against my own testing and the experiences of those people here who have played with it and found 10 meg resistors.

An earlier dicussion with them regarded the snubber circuit on the main board.  I asked for the value of a resistor there (it happens that it is also R14, but on the main board this time).  They told me that resistor was 47 ohms.  When I eventually had the main board out where I could read it, I found the resistor value was actually 220 ohms.
So, are they deliberately  putting out bad info?  Why would they do that?  No answers seem apparent.

It is interesting that the 1 meg value they claim for R14 on the control board is very close to the 900k value achieved by adding the 1 meg resistor in parallel with the 10 meg original. 
They concluded by saying that if I wasn't happy with the results of that change, I should look into reducing the values of R16 and R19 located on the underside of the control board.  I looked at these two resistors and noted that they are both the same surface mount type, with 153 written on them.  If R14 does the job I probably won't dig into these other two.  My eyes aren't good enough to trace those tiny traces out.  30 years ago maybe....
We're all going to DIE!  (eventually)

USA = Communist former republic
---  dochubert --

Offline rossw

  • Senior Moderator
  • Hero Member
  • *******
  • Posts: 879
  • Karma: +35/-0
  • Grumpy-old-Unix-Admin
Re: Overvoltage shutdown issue
« Reply #7 on: December 14, 2015, 03:35:27 pm »
Thanks Rossw for moving this where it belongs and for not blasting me for my new guy error.
I appreciate it.

Why would we blast you for a simple mistake? You asked - politely - if it could be moved. When asked, you answered politely where it should be. Took a couple of seconds to fix.

Honestly, NOBODY should feel scared to ask for assistance, or to request a post be moved, deleted or changed.
The board moderators and administrators are all nice, decent folk, who actually give a damn, and will help where we can.

Offline oztules

  • Forum Advisors
  • Hero Member
  • ****
  • Posts: 1177
  • Karma: +105/-8
  • Village idiot
Re: Overvoltage shutdown issue
« Reply #8 on: December 14, 2015, 04:05:59 pm »
The snubber resistors on all my boards have been the same as the gate resistors..... around the 47-51 ohm region... have not seen a 220 in this application

4944-0

...............oztules
Flinders Island...... Australia

Offline dochubert

  • Sr. Member
  • ****
  • Posts: 265
  • Karma: +4/-0
  • 11 powerjacks, 9 that run, 3 actually used daily
Re: Overvoltage shutdown issue
« Reply #9 on: December 14, 2015, 06:37:42 pm »
Thanks Rossw.  Guess I just feel siily for a dumb mistake.  Thanks for letting me be part of your group. 

Oztules, that's very interesting about the 47 ohm resistors.  I would much rather believe powerjack didn't lie to me than otherwise.  Now I'm going to have to look at those things again, and see if mine are really 220 or not.  If I remember correctly (about a 50/50 chance), my gate resistors are 47 ohm, so I figured at the time that powerjack just looked at the wrong one when I asked about the snubber resistor.  It was their assertion a few days ago that the 10 meg resister was actually 1.2 m that got me thinking 'deliberate' bad info.  That one I'm sure of because I read it with the ohmmeter after I removed it from the board, just yesterday.  Definitely 10 meg.

We're all going to DIE!  (eventually)

USA = Communist former republic
---  dochubert --

Offline dochubert

  • Sr. Member
  • ****
  • Posts: 265
  • Karma: +4/-0
  • 11 powerjacks, 9 that run, 3 actually used daily
Re: Overvoltage shutdown issue
« Reply #10 on: December 16, 2015, 02:20:30 pm »

Well, no shutdowns so far, but wasn't here all day yesterday to watch and see how high the battery voltage got.  Just started it up and have the freezer running on it as a light load, same as I did for previous testing with the 10 meg resistor.
Some readings:

the 10 meg                                                                   8.6 meg

At 49.8v    3.27v                                                   51.9v     3.63v
     50.2v   3.55v                                                   52.8v     3.74v
     53.2v   3.88v                                                   53.8v     3.80v
     54.5v   3.96v                                                   54.1v     3.82v
     56.1v   4.06v                                                   55.1v     3.93v
     56.7v   4.13v                                                   57.5v     4.07v
                                                                            58.0v     4.15v

Definitely going in the right direction.
Now I need to push the battery voltage up over 60v to see what happens in that range.  Maybe the sun will cooperate today.
My dump load is set to 60v so have to turn it off for now, and kick up my bulk charge  setting from 58 to maybe 61 or 62 for this testing.  Just don't want to leave it there too long....

Here are some pics of my current setup, in case anybody is interested:

4958-0
My 15kw 48v  inverter and my 180ah agm batteries.

4960-1
This shows my Xantrex c40 charge controller and my meters.  The dump load controller is to the left of the c40 and just turns on a 300 watt resistor (not shown)  The dump was to help keep the inverter online with the 58-60v shutdown settings.  I did have it set to 57.8 when the inverter had the 10 meg resistor and changed it to 60v when I went to the current 8,6 meg. Next it will be turned off entirely.
We're all going to DIE!  (eventually)

USA = Communist former republic
---  dochubert --

Offline dochubert

  • Sr. Member
  • ****
  • Posts: 265
  • Karma: +4/-0
  • 11 powerjacks, 9 that run, 3 actually used daily
Re: Overvoltage shutdown issue
« Reply #11 on: December 18, 2015, 12:07:20 am »
Interesting, if somewhat confusing observations today.  Turned the pj on today and switched the house onto it. 
Changed the float charge setting to 57.2v and the bulk charge to 60v which is as high as it will go.  Awhile later when the sun was up more and charging at around 9 amps, battery voltage was around the 59v mark and hit 60 briefly here and there.  No shutdowns/alarms.  Everything seemed to be stable and doing fine.  I went about my business, checking on it every few minutes.  Close to noon just as I happened to check, battery voltage at 58.6v the alarm started sounding.  I tweaked the R14 pot just a bit and the alarm stopped.
Why would it run at 59-60v earlier and then alarm later at the lower 58.6v?
The only thing that changed was the garage went from 50 degrees F up to low 60s as the day warmed up.

After tweaking the pot, the pj ran well the rest of the afternoon.  It again ran at 60v and 13-17 amps charging pretty steadily with about 400-600 watts load until the solar charging started dropping off as afternoon progressed.  No more alarms.
I don't know what the ohms value is now with the pot adjustment and can't read it in circuit.  The voltage drop across it at different voltages indicates a slight decrease in ohms from the 8.6 megs previous setting.
Problem is, barring transients, I don't see how I can raise the voltage higher with my charge controller already maxed out except for a direct connection controlled with the dump controller.  Not sure if I want to risk my batteries on that.

There is a voltage drop across R14 even with the unit turned off.  When I turned it off R14 dropped to 0.18v.  As the afternoon cooled, and presumably the inverter did too, the voltage across R14 climbed to 0.22v/  Now several hours later and down to 50 degrees again its reading 0.24v.  Probably means nothing but its interesting.

Another interesting note was the fan came on briefly twice today while I was there watching it run.  Each time the fan came on the R14 drop voltage decreased .2 to .3 volts but didn't return to previous level when fan stopped.  Then gradually increases back to previous level/ratio over several minutes.  Starting to wonder if R14 has another function besides high voltage shutdown/alarm.  Something temp related maybe. Sure hope we're not changing the hi temp setpoints here.
We're all going to DIE!  (eventually)

USA = Communist former republic
---  dochubert --

Offline oztules

  • Forum Advisors
  • Hero Member
  • ****
  • Posts: 1177
  • Karma: +105/-8
  • Village idiot
Re: Overvoltage shutdown issue
« Reply #12 on: December 18, 2015, 01:18:25 am »
The reason I went for the zener, is that the problem you get is generally at and around the float time when the battery is fully charged, and the choppers/regulators are running at very low/skinny pulse rates for minimal battery maintenance.

This will inject very short, but very high voltage spikes you won't see with the volt meter, but will see with the scope..... ie is is switching 70v or so into tiny time slices, which average only an amp or two... the PJ seems to have little filtering and absorbtion of these spikes, and sees them as true voltage... and shuts down accordingly.

If you drive the PJ with linear input voltage of 60v or so, the thing will not trigger (eg.. direct connect to array of a single panel that does not push the battery up higher than say 61v)... so it looks like it is seeing the pwm spikes ( peaks), not the just the  average voltage.

I don't think it is temperature sensitive, as much as it is sensitive to voltage peaking of the pwm waves.

............oztules
Flinders Island...... Australia

Offline dochubert

  • Sr. Member
  • ****
  • Posts: 265
  • Karma: +4/-0
  • 11 powerjacks, 9 that run, 3 actually used daily
Re: Overvoltage shutdown issue
« Reply #13 on: December 18, 2015, 04:14:13 pm »
Thanks Oztules.  That probably explains the alarm at 58v when it was running over 60 earlier without alarm.
So its looking as if even getting the setting where it ought to be won't stop all the shutdowns, which I guess is what you have been trying to tell me all along.  I'm a little slow sometimes...

I have a scopemeter.  Guess I need to start using it.

You know, I was thinking (before I heard about your success with R14) of putting a filter on the voltage going to the control board from the main board.  I thought about the 10 lead cable, and interrupting the + and - leads to insert a surge/spike suppressor filter between control and main boards to hopefully prevent those spikes from affecting control functions.  Does that sound like its worth trying, or am I missing something? 
Unfortunately, designing a filter for this purpose is beyond my capabilities.  Any ideas?  Or would something off the shelf work adequately well?
Again, I realize you are good with the zener stopping all shutdowns, and honestly, I am too in my own applications.  I just would like to really fix the issue rather than use a work around.  If no luck and I eventually have to give up the quest, I have zeners on hand....
We're all going to DIE!  (eventually)

USA = Communist former republic
---  dochubert --

Offline dochubert

  • Sr. Member
  • ****
  • Posts: 265
  • Karma: +4/-0
  • 11 powerjacks, 9 that run, 3 actually used daily
Re: Overvoltage shutdown issue
« Reply #14 on: December 23, 2015, 10:42:02 am »
Have been researching spike/surge filters last couple of days.  Just reread my last post.

"I just would like to really fix the issue rather than use a work around."

Open mouth, insert foot.

My new posting procedure needs to be:
 Write post as draft and save.
 Next day, read draft.
 If I can stop laughing long enough, go ahead and post.


I'm looking at trying this spike/surge protector (partially because I already have one on hand)

Translectric, Inc. - Noise Filter, 8AMP/6-48VDC
TESSCO SKU : 453000 Mfg Part #: SM480-8 Qty/UOM : 1 EACH UPC: 646444530002
SERVICEMATE noise filter and spike protector. 6-48 VDC input. 8A @ 100% duty cycle, 16A @ 50% duty cycle.

Its some amount of overkill with its 8a rating, but looks like it should do the job of smoothing those spikes before they reach the control board.  Also, should be able to use it on 12, 24 and 48v inverters.  Since probably everybody on this forum knows more about the subject than me, any comments would be appreciated.
Thanks
We're all going to DIE!  (eventually)

USA = Communist former republic
---  dochubert --