Author Topic: PowerJack LF-10k 48v issues.  (Read 3634 times)

0 Members and 1 Guest are viewing this topic.

Offline Superscan811

  • Newbie
  • *
  • Posts: 8
  • Karma: +0/-0
PowerJack LF-10k 48v issues.
« on: June 13, 2017, 03:07:29 am »
First of all, I purchased this unit in 2014 but it sat in a corner up until a few months ago.

The Setup..
7kw of solar panels (120v).
External charger
48v Flooded cell Ni-cad battery pack  (1.2v per cell) which require 64v for boost, 60v equalization, 58v float).

My main goal for this unit is to make it bullet proof.

Problem 1.
Main power board blew.

Problem 2.
Inverter trips out when the voltage goes over 59v.

Problem 3.
Wanting to make a balance charger to correctly charge each cell.



Solution 1.
Rebuilding the power board.
Replacing the 80v caps with 100v and adding 2 more because there is room on the board.
Replacing the IRFB4310 FET's(100v 140a 5.6m?) with IRFP4568 (150v 171a 4.8m?) and adding 4 more, again there are extra slots on the board.

Replacing the controller board.
As I wasn't using the unit as a charger, there may be something faulty with the controller board. There is nothing obvious but I'd rather be sure.
The EBAY Seller jungleislandpower sells the new control boards. Don't know if they are any better than the older control boards but the new ones are modular, ie: there are separate daughter boards for the FET drivers, CPU, charger, a/c input and output detect, etc....
Purchased the whole set for $100 delivered.

Solution 2.
As I will be using the newer controller board, locating the new position for "R14" and apply the appropriate zener will be first on the list.

Solution 3.
Work in progress.. :-\



Modifications.
Apart from the fets and caps,

Reducing idle current
Viewed a post by oztules on how to reduce the idle current when using toroidal transformers.
Had an old 600va UPS with a large EI core transformer. Stripped it down so it's ready to go..

Extra fans.
May transplant  the guts into a different case to allow for better cooling.

Heatsink Insulation.
Noticed the inadequate insulation between the power board and the heatsinks. Plan on putting some thick plastic there instead of the heatsink mat.

PCB track improvement.
Will be running some 1mm dia copper wire from the Source and Drain to their respective inputs/outputs.



Any and all suggestions gratefully accepted.


Cheers.


Offline oztules

  • Forum Advisors
  • Hero Member
  • ****
  • Posts: 1177
  • Karma: +105/-8
  • Village idiot
Re: PowerJack LF-10k 48v issues.
« Reply #1 on: June 14, 2017, 01:31:17 pm »
It depends on the output your after.

The PJ is sensitive in the upper volt region, and will need to be modded with the zener or similar.

The power handling of the power board will probably be ok as is. It is not the weak link... the transformer is.

More caps are always better.... share the ripple.

More fets may not help much unless your driving it very hard for a long time.... once again, transformer will likely over heat  not the fets so much.

I don't know much about the newer models, and don;t use the PJ boards anymore.... so can't advise.

EI transformer is fine, except the power handling is half the torroid for the same weight. Losses will be a higher, but 600va is a bit too small to contemplate.....


........... oztules

.
Flinders Island...... Australia

Offline Superscan811

  • Newbie
  • *
  • Posts: 8
  • Karma: +0/-0
Re: PowerJack LF-10k 48v issues.
« Reply #2 on: June 14, 2017, 03:49:11 pm »
Thanks for the reply oztules.

Power usage will only be around 3-4kva normally but with an occasional peak when the washing machine and dishwasher are being used, so I'm hoping the 2 original toroidals in the 10kva unit will be fine.

as for the old 600va UPS EI transformer, I stripped the windings out of it to use the core for the inductor/choke you specified in one of your posts (at least I hope I read that part correctly...)

This setup is for my mother who is on a farm about 900 kilometres away, so that's why I want it to be very reliable.

Also planning to use the excess power to divert into the water heater if possible.
May end up doing a bit of trickery with an Arduino and 2 current shunts, 1 on the water heater and the other just before the battery pack.
If the differential is less than the normal usage, power on the water heater, otherwise turn it off.
Mum's hot water serviced is solar but with electricity as backup.
Some of her vacuum tubes have been damaged, so often the water ends up being heated via the grid.

I have lots of projects planned for the farm, including a water wheel of sorts for the river that runs behind the property.



Cheers.






Offline oztules

  • Forum Advisors
  • Hero Member
  • ****
  • Posts: 1177
  • Karma: +105/-8
  • Village idiot
Re: PowerJack LF-10k 48v issues.
« Reply #3 on: June 16, 2017, 03:38:35 am »
Be interested in how the transformer goes as a inductor in the filter position. It is very high frequency we are trying to damp..... probably up in the <100mhz range... not the 20khz range of the switchers.....not sure if the iron will do any good here..... never know..... but ferrite works.

If your using solar, then a grid tie will do well to help with the 4kw constant loads during the day.... less losses and built in mppt, and will take the pressure off the tranny. The two trannies will not like 4kw for more than 20 mins I would think... they are too small really. Good for a few kw continuous, but even then operate quite warm.... so if you can get a few kw running the house from a mppt grid tie, you will do much better. If you do a number on it ( rectify the genny output, and inject into the solar inputs ( may need a few caps to help regulate the input), you can use it for battery charging backwards through the PJ.

It  is my isolated position that is the reason i designed my own inverter using the egs002 and then the 8010 chip alone.... no hope of help here.

Edit. Re read your reply... you have grid available..... use your first monies to get grid tied.... thats the best for a start, then o/night power system after that.


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

Offline welshman

  • Full Member
  • ***
  • Posts: 136
  • Karma: +6/-1
Re: PowerJack LF-10k 48v issues.
« Reply #4 on: June 16, 2017, 04:26:54 am »
get a newer pj mainboard, mine works fine with the higher input voltages. the fets on my 15k run about 30 deg while the toroids get up to 90 at the same time, this seems to be only on charging the batteries though as taking power from the batteries, the toroids run much cooler around 50 at the same load.

the powerjack i have has been flawless, it's switched itself between charge and inverting mode up to 5 times daily with a generator providing the power, without a problem not blown a fet at all.

there is one slight issue, there is an inductor on the pj mainboard, when i run 12-13kw through the bypass it starts to get hot and the laquer is peeling now. no chance of a short with it as the polarities are isolated, but i think people should be aware that things can get hot on the power board before the fuses trip.

the only issues with the software of the device is when say it overtemps or undervoltages etc and shutdown the alarm comes on and even if the generator kicks in to charge later after there is no problem the alarm persists and can only be turned off by cycling the inverter off and on and if it overtemps it has to be manually cycled.

i;ve found the additional coil added to later pj's gets quite hot when doing nothing and i plan to swap to an e core variety hopefully using the advice of oztules here to get those idle watts down a bit further.

if you're interested in ardunio and current shunts check out my arduino companion thread, ill update the code on the page soon with the latest changes, but it might save you some time with the pitfalls of getting reliable data out of an arduino being able to measure teh current travelling at both directions on the shunt.

with regard to equalizing the charge on the batteries, check out my thread about modding my powerjack, i include the connection schematic for the battery bank, this setup i have used ensures batteries get equal charge and no slow to charge battery delays the others and at idle the entire bank equalizes.

now for cooling, even my 4 x 240v server fans are inadequate with ambient temps above 25 deg c. go big on cooling at a cost, or rewire those two toroids into one. which is what i plan to do in the future, hopefully with the help of oztules expertise.

Offline Superscan811

  • Newbie
  • *
  • Posts: 8
  • Karma: +0/-0
Re: PowerJack LF-10k 48v issues.
« Reply #5 on: June 21, 2017, 07:23:20 am »
Be interested in how the transformer goes as a inductor in the filter position. It is very high frequency we are trying to damp..... probably up in the <100mhz range... not the 20khz range of the switchers.....not sure if the iron will do any good here..... never know..... but ferrite works.
OK, didn't have the knowledge to realise that it was running in that range. Will "acquire" a ferrite core and proceed..

The two trannies will not like 4kw for more than 20 mins I would think... they are too small really.
Along this line, apart from physical space, what is/are the disadvantages of paralleling 2 or more separate transformers, as opposed to using just 1?


if you're interested in ardunio and current shunts check out my arduino companion thread, ill update the code on the page soon with the latest changes, but it might save you some time with the pitfalls of getting reliable data out of an arduino being able to measure the current travelling at both directions on the shunt.
Sounds good. I'll have a look..


now for cooling, even my 4 x 240v server fans are inadequate with ambient temps above 25 deg c. go big on cooling at a cost, or rewire those two toroids into one. which is what i plan to do in the future, hopefully with the help of oztules expertise.
Size is not an issue, noise is, as it is on the side of the house.. Too much fan noise, especially at night, could possibly be an issue. I prefer passive cooling anyway, as this eliminates one point of failure.

Cheers..


Offline oztules

  • Forum Advisors
  • Hero Member
  • ****
  • Posts: 1177
  • Karma: +105/-8
  • Village idiot
Re: PowerJack LF-10k 48v issues.
« Reply #6 on: June 22, 2017, 04:38:02 pm »
"OK, didn't have the knowledge to realise that it was running in that range. Will "acquire" a ferrite core and proceed.."........
It looks like about 14-25uh of inductance is in the right range to pull the idle current down...... This is far too small to have any effect on a 20khz signal, never mind a huge1000% reduction in losses from 5kwh/day to .5kwh/day and less.

It appears that the switch turn on is very very fast, the transformer impedance is very low... so the current rises very fast. With only a few uh of coil in there, it is enough to damp that inrush. The pulses are at 20khz, but the wave width is quite small, and may be only 5-10% duty cycle... so that may mean the effective wave length is 400khz. Now we have to switch on and off in that time frame.... we can start to see how fast the rising square wave wave front must look.... think about the wave width's dynamics..... we can go from idle to 20kw.... so there is a fair wave width differential to deal with, but you get the idea of how much time the fets is actually fully on at idle.... not much....

At idle,  these figures may be even worse, and the majority of time, the fet is between fully on and fully off in very very short times as the wave duty cycle gets less at idle.... If we can effect these wave fronts and take some of the peak power out of them, then the proportion of time the fet is turning on, versus fully on won't change, but the peak currents won't be there when it is in it's linear stage, where the losses are brutal.... thats my theory anyway why it works ;D  ..... code for " I don't have the knowledge either"

"Along this line, apart from physical space, what is/are the disadvantages of paralleling 2 or more separate transformers, as opposed to using just 1?"
Two transformers will have 25% more copper length in there as compared to a single one..... ie if you were to stack two cores together and then wind your half the number of turns at twice the cross section, you will have the same current carrying cross section as two, but you eliminate two sides of copper out of it ( where the cores face each other there is no copper ), so your winding length is less than the two cores separate.... thats another I^2R gain for free.

Those transformers are so  small, that you need any gain you can get for continuous power ( >20 mins). They are the heat source you will have problems with, the fets should be easy to keep cool..... passive wont do it at high power..... unless you develop a  chimney effect somehow.


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

Offline Superscan811

  • Newbie
  • *
  • Posts: 8
  • Karma: +0/-0
Re: PowerJack LF-10k 48v issues.
« Reply #7 on: June 23, 2017, 02:03:08 am »
If I understand things correctly, the Inductors act like a current limiter only when the fets are first switched on (high inrush current), but is relatively passive during the rest of the on cycle, so the time it takes to switch the fet from full OFF to full  ON, is the frequency we need to account for, NOT the frequency we are switching the fet.

As for the transformers,
less copper length = less resistence = higher current.

Head successfully wrapped around both concepts (I hope).

Thank you..



Cheers.

Offline oztules

  • Forum Advisors
  • Hero Member
  • ****
  • Posts: 1177
  • Karma: +105/-8
  • Village idiot
Re: PowerJack LF-10k 48v issues.
« Reply #8 on: June 23, 2017, 05:25:05 am »
Thats how i see it.... remember I am a hack, not an EE.

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

Offline Superscan811

  • Newbie
  • *
  • Posts: 8
  • Karma: +0/-0
Re: PowerJack LF-10k 48v issues.
« Reply #9 on: July 24, 2017, 05:09:50 am »
If you see these caps on Ebay etc.. DO NOT BUY THEM


Purchased 10 of these caps. Supposed to be 10,000uF at 125v but the tester told the true story.
Sent an abusive email along with pictures of my tests and they finally refunded me the full cost.


Then purchased 4 of these which were on spec..



Would definitely recommend testing all your components as Oztules said..

Building a test rig with a 150w 12v globe to test the RDS on resistance of the Mosfets.

Those test results to come later..



Cheers.







Offline lighthunter

  • Sr. Member
  • ****
  • Posts: 323
  • Karma: +13/-0
Re: PowerJack LF-10k 48v issues.
« Reply #10 on: July 25, 2017, 06:19:20 pm »
Lithium battery jumper packs same way. I purchased a pack 2 yrs ago boasting 12 AH capacity. I got it home and charged up, threw it on msw inverter running a 100w incandescant. turned out to be a 4 AH capacity. I kept it as there were no comparable units for $30 so i figured i still got a good deal. Ive used it a lot and though still going strong i thought it b great to buy a backup lipoly battery for it to b ready when it goes. I was surprised to find given the dimension specs of original, 130mm x 45mm x 25mm. I could not find a replacement inside those dimensions with greater than 4 AH capacity and i was looking among the high quality choices to fill the spot.

That said, the 4AH actal capacity of a labeled 12000mah battery of that physical size was not the fault of an underperforming/defective, or inferior battery. It was simply the fault of the sales literature.
Its unfortunate that merchants do that and weve seen it among many electric devices including the pure sine inverter in this thread. If you know about these shortcomings up front and usually by experience you do, one can still come away with a decent purchase just not one that matches the label. Of course if you have a critical application where spec must be accurate then do as you said.   :)
Health Warning: May contain traces of nut!
LH

Offline DJ

  • Full Member
  • ***
  • Posts: 82
  • Karma: +5/-1
Re: PowerJack LF-10k 48v issues.
« Reply #11 on: July 26, 2017, 10:27:13 am »

I love the ratings on some of those packs. Some are supposed to be over 50 Ah!
Sellers, particularly ebay merchants just seem to put whatever impressive number they dream up on things and any relevance to reality is out the window.

I have bought 3 Different types of Light bars now and they are all ridiculous ratings.  The last lot was supposed to be 210W and turned out to be 32.
The sellers always try to bull$#!+ and bluff their way out of a refund or return demanding to know what your qualifications are but I mearely ask what theirs are in order to be able to judge whether mine are adequate or not. Soon shuts them up. I got the feeling the last one really didn't know his arse from his elbow and the fact it had 210W on the box made it 210W and he had no clue as to how it could be different.

I have emailed every seller before purchase to confirm the lights are rated at what they say. Some I don't hear back from so I know what that means but the others have assured me they are as rated on the box which is why I purchased them.

So far all have refunded either partially or in full.  One guy went on and on about me trying to rip him off by doing an ebay complaint who charged the item back but no matter how many times I pointed out he was ripping people off by knowingly still selling scam products, I was still of course the bad guy.
Should have bought 20 of the things and had them all charged back.

Another one I see is fantasy ratings on 18650 batteries. Some of these sellers have batteries with capacity's 4-5 times what the top manufacturers are claiming.  Best I have seen so far is a 1500Mah 18650. I'm led to believe the best " genuine" cells atm are under 3000 mah. 

Over rating goes a long way back. I remember asking an audiophile acquaintance what the PMPO rating on sound equipment was about 20 years ago. I said sarcastically that  given some of the numbers I had seen on small portable units, it must be a measure of what the board will take when hooked to a HV transmission line before it vaporises a millisecond later.  He looked at me and he said that's pretty much about what it is!
I also asked why my Harmon Kardon Amplifier was only rated at 35W but would near take the windows out the house but friends other brand units were rated at 100W and were no where near as powerful. He said it depended what they measured the total harmonic distortion at. He Said HK measured theirs very low where others might be at 10%. If rated to the same scale as the HK, their rating might only be 10W.

I also got caught with inverter. Totally fictitious ratings on those as well.  One real giveaway is look at the fuse rating. On 12V units I take the rounded number as 10 for the multiplication factor for the output for 12/ 240V units.  This allows for different efficiency, different battery voltages, output voltages ( 220-250)  etc.  Therefore if the unit is 1000W output, it's going to be ( at least) 100A in. If the thing only has 2 x20A fuses, you know it's BS and you are going to be lucky to get 500W out the thing.  I see a lot like that. Huge ratings but the 12V in fuses are a fraction of the multiplied output.

All depends how you want to screw and distort the numbers.  :0(

Offline Norm

  • Sr. Member
  • ****
  • Posts: 475
  • Karma: +26/-0
  • Today is the day you worried about yesterday.
Re: PowerJack LF-10k 48v issues.
« Reply #12 on: July 28, 2017, 08:01:22 am »
Yeah I run across these guys that profess to know it all .
I had  a neighbor that has a motorcycle every day he would start it up
v room v room  on and on for about an hour until satisfied that it was tuned up
another neighbor wanted me to check his car battery I checked it and told him
it was only 12 volts and needed to be about 13volts to be charged up
this cycle neighbor pipes in "Why its Only supposed to be 12 volts that's
why they call it a 12 volt battery .....all these years I've never seen any other kind of battery!"
 ::)  .....so I turned around and went home ...let HIM fix the guy's battery ,,,,right ?

Norm  :)

Offline bj

  • Hero Member
  • *****
  • Posts: 735
  • Karma: +23/-0
  • Lamont, Alberta, Canada
Re: PowerJack LF-10k 48v issues.
« Reply #13 on: August 02, 2017, 06:52:38 am »
  Hey Norm.  A lot of miles between where we live, but I think we have the
same neighbor.  ;D
"Even a blind squirrel will find an acorn once in a while"
bj