Author Topic: guess who bought a power jack inverter  (Read 196571 times)

0 Members and 1 Guest are viewing this topic.

Offline James96

  • Newbie
  • *
  • Posts: 6
  • Karma: +0/-0
Re: guess who bought a power jack inverter
« Reply #495 on: November 03, 2016, 08:39:15 pm »
alright, here's the feed out;

what had been happening was that the system was actually working for 40 seconds backfeeding and all, but the grid tie inverter would at that point show an islanding fault (basically it thought that it was feeding onto a downed power grid and as a safety measure it is designed to shut off) I watched the backfeed dc amps go above 100 before the gti threw the fault, but the voltage didn't pile up that much, the inverter started at 232vac and it went up to 236, I was expecting it to go up at least 15vac, and I think this is what made my picky gti think it was islanding, the manufacturer says that some new software may fix the issue, but I was wondering if there was a way to tweak the powerjack to pile up the voltage a little more

Offline oztules

  • Moderator
  • Hero Member
  • *****
  • Posts: 1177
  • Karma: +105/-8
  • Village idiot
Re: guess who bought a power jack inverter
« Reply #496 on: November 04, 2016, 12:21:36 am »
Start with less solar panel, and work your way up.... some is better than none while testing.... more time to find out what is really causing it.


............ozatules
Flinders Island...... Australia

Offline edgolla

  • Newbie
  • *
  • Posts: 3
  • Karma: +0/-0
Re: guess who bought a power jack inverter
« Reply #497 on: January 23, 2017, 08:44:44 pm »
I do not have a powerjack myself.  But, a friend of mine just install one that is rated something over 10 KW.  It seems to work ok until the battery voltage gets to about 58 volts and then cuts off.  The manual indicates that over voltage should be 4 x 15.5 or 62 volts.  Is there any adjustment or way to adjust the high voltage cutoff.  The unit works fine at night and on cloudy days.  But, cuts off on bright sunny days.  This is of course a 48 volt system which will charge up to about 58 volts for normal use.   Ed

Offline oztules

  • Moderator
  • Hero Member
  • *****
  • Posts: 1177
  • Karma: +105/-8
  • Village idiot
Re: guess who bought a power jack inverter
« Reply #498 on: January 25, 2017, 05:10:02 am »
Long time since I used them now, but  5v6 zenner across the 10meg resistor next to the main chip works. There are pics here and elsewhere depicting the orientation and position. I no longer recall it properly.

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

Offline lighthunter

  • Sr. Member
  • ****
  • Posts: 323
  • Karma: +13/-0
Re: guess who bought a power jack inverter
« Reply #499 on: January 25, 2017, 07:06:54 am »
Hi oztules!  I'm trying to understand toroid transformers. Just by the power ratings of PJ 24v vs 48v pure sine inverters it seems they think a similar toroid can double the power output by Increasing primary voltage from 24 to 48. At first glance this makes sense as the primary current would only be 1/2 to get the same work done.

Recently i modified an original pj 24/110 toroid by changing the 24v primary 12-in-hand to 6 in hand series to 6-in-hand. This changed the primary to work as a 48v unit but since i changed a 12 strand parallel to a 6-strand series my copper resistance will have doubled. So end result I havent changed the KVA rating at all.

 Therefore i dont see a gain as far as physical size and kva rating of raising input voltage. Am i missing something??

 I'm not arguing 24v vs 48 as its no comparison with power handling.  Just trying to understand limitations of toroids. You have mentioned saturation is not the result of driving a transformer current too high but a result of overdriving volts for a given winding/turns. Since that is a fact, can I effectively increase the KVA rating by simply decreasing the primary copper resistance by removing 12-strand  primary and replacing it with. 24-strands keeping turns the same of course. If this decrease of primary copper resistance increases kva rating, what role does the physical size of the core play?

Is a larger core necessary other than more real-estate for wire?

Thank you so much for your knowledge and teaching abilities!
Here are 2 photos, the toroid got an additional 120v winding as well. This pj is now 48v 120/240.
It works great, measured values are 122/117/239. The values are offset intentionally for 240 at grid tie and 122 for house loads.
Health Warning: May contain traces of nut!
LH

Offline lighthunter

  • Sr. Member
  • ****
  • Posts: 323
  • Karma: +13/-0
Re: guess who bought a power jack inverter
« Reply #500 on: January 29, 2017, 09:32:05 am »
Health Warning: May contain traces of nut!
LH

Offline oztules

  • Moderator
  • Hero Member
  • *****
  • Posts: 1177
  • Karma: +105/-8
  • Village idiot
Re: guess who bought a power jack inverter
« Reply #501 on: February 03, 2017, 04:01:23 pm »
"Is a larger core necessary other than more real-estate for wire?"

Really it is only copper  resistance that dictates the core size, as once we have magnetising current out of the way,, no amount of power throughput will change the magnetics, as the core will not saturate from heavier current, only from lower frequency or higher voltage, but not more current..... even though it feels like a ampere turns thing, it is not really that as far as the core saturation is concerned.

Once the magetising current is out of the way, all other stuff will be work or heat... ie the magnetic fields of the primary will make an electromagnetic field around the coil of copper that will induce voltage into the coil of wire that is the secondary. The core just facilitates this magnetic interplay i guess is one way of looking at it

With that in mind, if you want more power from a given core, it means filling the thing right up with copper. If you can get the super conductor thing happening, then core size will be irrelevant.

So we use a bigger core so we can get a smaller number of turns of bigger wire in there.

Bigger cores allow bigger wire for two reasons... one more realestate, and two, we need less turns for the same number of volts ( or more volts per turn when we get a  big enough core).... so we can use even thicker wire before running out of room.

"Recently i modified an original pj 24/110 toroid by changing the 24v primary 12-in-hand to 6 in hand series to 6-in-hand. This changed the primary to work as a 48v unit but since i changed a 12 strand parallel to a 6-strand series my copper resistance will have doubled. So end result I havent changed the KVA rating at all."

Yes your resistance has doubled....... but your current has halved... and the losses will be an I^2 x R.... so if you had say 4 amps running through a 2 ohm core, we would have E=IR or 8v loss times 4 amps = 32 watts.... or 4^2 x 2 = 32 watts for I^2R.... done both ways.... for your next scenario, we would have 2 amps at 4 ohms for the same work....... so by E=IR we have a loss across the coil of  8 volts still....but losses will be 8 x 2 for  16 watts losses... or by I^2 R we have  2x2x4=16 watts losses.

So it looks to me you have done better.



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

Offline lighthunter

  • Sr. Member
  • ****
  • Posts: 323
  • Karma: +13/-0
Re: guess who bought a power jack inverter
« Reply #502 on: February 05, 2017, 09:31:40 am »
I feel dumb  ;D.  I never plugged the values in the formula and overlooked the obvious. Thanks so much for pointing that out!  I make the same mistake occasionally when I use a water heating element at a voltage increase thinking its a linear increase in power with volts and its not.
Health Warning: May contain traces of nut!
LH

Offline lighthunter

  • Sr. Member
  • ****
  • Posts: 323
  • Karma: +13/-0
Re: guess who bought a power jack inverter
« Reply #503 on: February 10, 2017, 08:14:27 pm »
 Ive had a rough start with my 24/ 48v pj inverter conversion. Does anyone else ever have frustrating moments with H-bridges?

 I wanna say i blew it up at least 4 times.  The very first run after the conversion ran 1/2 a day before vaporizing the bottom row of fets.  Each time i fixed it the inverter would run great for a while and then start to get noisy and then poof. Eventually watching the board powered on in the dark i could see fire flashing around the board in front of the fet bridges.  I cleaned board very well replaced  fets again and it ran for a few more minutes, poof.

After blowing up one more time i had the "thats it, i've had enough moment!!  I dug out an old dataq instruments computer scope "148u" I cut the ribbon cable leads for pin 1 and 2, connected 48v supply negative where it belongs and the +48v fused at 1 amp to the cut wires leading to control board. Then, i rigged a "110v actual voltage" to simulate the feedback.


Powered up board and repaired issues until i got identical pictures of all gate drives with no FETs in the main board. Then because i was totally out of 4110s and needed 7 more to finish the set,  I took a pile of used ones and put each on a 4 amp load with 8v on gate and measured voltage drain to source in this "on" state. Some measured as low as 24mv while some as high as 36mv. I choose those in the middle at 29 mv.  I then finished out the complete set of fets with "tested" components. Again i looked at all 4 drive outputs with fets in circuit.  When all looked perfect i started with a 24v control board and 24v supply.  It seems like the 48v circuit is more picky. One thing i want to point out, the idle current on 24v was 2.8A with no choke, i noticed the band around the toroid was warm in one spot after only a minute of on time. I took the metal band off toroid and watched idle current drop and again when i removed bolt from middle. Something someone here may want to consider. After switching to 48v setup i then installed a choke

Thank You Oztules, :)

The idle current went from 1.4A at 48v no choke all the way down to .4A for 19.6w :)

By the way the pics of top and bottom H drive reveals bipolar modulation on even the 2013/08 control board. The first pic is the bottom gate drives and the second is the top pair, 3rd pic is of a different time base and one top / one bottom;  this scope is very speed limited at 7000 samples/sec with 2 channels at same time.

One thing to note, since the 3rd pic is a top and bottom the timing should reveal a 180 degree offset. I was extremely puzzled about this, so much so that i thought the gate drive leads were switched around, i proved the timing was in fact correct by swapping in the 24v control board and it showed the same timing error and i knew it worked. Then i unplugged the feedback which causes the modulation to go 100% to 60hz square wave, then it shows the 180 offset. I think the problem is created because im only using 1 ground reference for scope and top/bottom drives do not have the same ground, in fact since i have the power supply disconnected on main board, the charge pump for upper drive cannot even function. Surprised it worked as well as it did.

This inverter has now run about 14 hours and ran only about 2.5kwh before it tripped on overvolt with dead battery. I did the R14 voltage mod,  Today may go better.
Health Warning: May contain traces of nut!
LH

Offline Madness

  • Newbie
  • *
  • Posts: 13
  • Karma: +0/-0
Re: guess who bought a power jack inverter
« Reply #504 on: February 11, 2017, 10:08:44 pm »
How is your Inverter now Light Hunter, I have been trying to get mine right for many months now. I keep getting nasty spikes in the output, works fine with only half the MOSFETs but is not right with more. One tip for testing I got from Oztules that is very helpful is to run it with no capacitors on the MOSFET board and put a 30-ohm resistor in series with the 48V supply. If anything shorts the MOSFETs don't go bang. If you have the CAPS on the board they can draw huge current just long enough to make the FETs explode.


BTW does your PC Scope have the ability to save an image?

Offline lighthunter

  • Sr. Member
  • ****
  • Posts: 323
  • Karma: +13/-0
Re: guess who bought a power jack inverter
« Reply #505 on: February 11, 2017, 11:05:27 pm »
Hi Madness :) there is 23.39 hrs on the inverter now since finishing it. It sounds and runs perfectly.
Today it was cloudy though and i only had 1 array hooked to it so it didnt get a workout, only 2kwh put into battery. Tomorrow i will add a 3kw array to it so it may be running at 4kw tomorrow fr the first time.  I think part of my problem is i buy cheap fets and i hadnt been testing them. From now on i will. I also did not clean board good enough earlier. Finally used gasoline and toothbrush then finished with alcohol. That cleaned it nicely,  Great idea on the series resistor and removing caps! I had thought of that but hoped i didnt need to go that far.

My pc scope is just a cheap 8 channel data aquisition unit. I dont know if i can save a screen shot but windows can since it runs on windows os. You can record the whole wave display for as long as you wish and view any time frame later as a still pic. You can view waveform with record mode on or off.  It cost $50 10 yrs ago made by dataq instruments in ohio. I'm sure they have better stuff now but it does help verify control board before you connect to fets.

Dont give up! you can do it! is it a 48v?
Health Warning: May contain traces of nut!
LH

Offline Madness

  • Newbie
  • *
  • Posts: 13
  • Karma: +0/-0
Re: guess who bought a power jack inverter
« Reply #506 on: February 11, 2017, 11:41:39 pm »
Yes I have had it running last time was for 2 weeks till it failed. Only way I have been successful though is to increase the dead time to the max but it is still quite there. Have had it running well over 6 KW thought for several hours, I have built it with extra windings in the toroid and big heatsinks to run at the higher power levels. Last failure though was after a couple days of very light running.

I will look for the arcing you mentioned next time I fire it up. My system is 48V I have 7.5KW of solar panels ATM which I am planning to add another 5 KW to shortly. I am running a 10KW AC that draws up to 3 KW plus a second AC that draws 660W and running a normal house. I have 2 4500W Trace inverters athat I am using ATM but they were built last century and I don't expect them to last forever.

Offline lighthunter

  • Sr. Member
  • ****
  • Posts: 323
  • Karma: +13/-0
Re: guess who bought a power jack inverter
« Reply #507 on: February 12, 2017, 09:18:04 pm »
I'm amazed, your equipment is in another category of power level. I may have to increase mine also though. Odd, ive got a 24v setup with a 200lb square transformer and it runs 4kw without an issue and very little heat, i turn fan on low just for a little protection.  Now this 48v setup is using 1 original pj toroid with modified windings and it only had to hit 1850w today and the former + heatsi ks got very warm. 3 times it shut off, guessing it was heat because it turned right back on.
It didnt display. Any red lights.
Health Warning: May contain traces of nut!
LH

Offline oztules

  • Moderator
  • Hero Member
  • *****
  • Posts: 1177
  • Karma: +105/-8
  • Village idiot
Re: guess who bought a power jack inverter
« Reply #508 on: February 12, 2017, 11:22:48 pm »
Lighthunter..... those PJ transformers are only 1kw continuous by my reading........ they do fine with huge loads for short periods, but continuous they are very much under cooked. You will need to wind a 100lb torroid to emulate the 200 lb square one...... and that PJ is not that.

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

Offline welshman

  • Full Member
  • ***
  • Posts: 136
  • Karma: +6/-1
Re: guess who bought a power jack inverter
« Reply #509 on: February 13, 2017, 07:09:01 am »
Lighthunter..... those PJ transformers are only 1kw continuous by my reading........ they do fine with huge loads for short periods, but continuous they are very much under cooked. You will need to wind a 100lb torroid to emulate the 200 lb square one...... and that PJ is not that.

..........oztules

What would be a safe max working temp for one of those PJ toroids as standard?

What seems to make my modified PJ get the hottest is backfeeding throught the toroid to charge the batteries. The other direction at the same draw produces about 1/3 of the heat.