Author Topic: power jack lf8000/48v blew up and smoked  (Read 19746 times)

0 Members and 1 Guest are viewing this topic.

Offline lighthunter

  • Sr. Member
  • ****
  • Posts: 323
  • Karma: +13/-0
Re: power jack lf8000/48v blew up and smoked
« Reply #45 on: October 20, 2017, 08:27:19 am »
Hi Sean :)  Yes, that looks like a good replacement. Im not sure it offers an advantage over irfb4110 though. Its very close to 4110 spec but seems more pricey. Then again quality and location can be factors too, You can spend $7.00 usd each on 4110 for genuine IR parts or 50¢ ea on ebay. Must use complete set of one type though.

 To compare devices look at::

Id (continuous drain current and pulsed drain current.) Same or greater than original.

Vdss (voltage ability)   same or greater than original

Qg (gate charge)   Less than or equal to  original

Package style, ex, to220, to247. same as original.

Internal circuit, same as original

Sorry for late post, info is probably not necessary at this point.

Just in case repairing drivers is too much of a headache, you can buy the driver board for $20 usd free shipping now. depending on which cntrl board you have it may be a simple swap an ready to go.

Good luck with that!



Health Warning: May contain traces of nut!
LH

Offline MadScientist267

  • Impossible Condition Curator
  • Hero Member
  • *****
  • Posts: 1514
  • Karma: +44/-4
  • Rules? What rules?
Re: power jack lf8000/48v blew up and smoked
« Reply #46 on: October 20, 2017, 07:17:01 pm »
Not that things can't get into splitting hairs when crossing a suitable substitute part when combing datasheets already lol... but I'd like to add "avalanche rated" to lighthunter's cheat sheet list... It's ok for it to show up in the substitute and not be in the original, but sometimes can "bite you in the heatsink" the other way around. (Ok that was bad I apologize lol)

This is a complex attribute of a particular device in terms of what is going on under the hood, but in a nutshell, it represents a device's ability to "stay in control" when Vds exceeds breakdown and how it handles the energy internally.

This is relevant to any switching design because the key offender is inductance, and while the characteristics of the nearby components can be compensated for (this is what "snubbers" do), there is a practical limit for what can be accounted for when external loads and their wiring get involved.

I see in this case that both IRF4110 and RU190N08 are indeed avalanche rated, so you're good there, but felt it was worth mentioning as one of the more important considerations in general.
Wanted: Schrödinger's cat, dead and alive.

Offline lighthunter

  • Sr. Member
  • ****
  • Posts: 323
  • Karma: +13/-0
Re: power jack lf8000/48v blew up and smoked
« Reply #47 on: October 25, 2017, 07:53:24 am »
Thanks Steve! i wasnt aware of the avalanche rating. Learned something new :)

Why dont we / or manufacturers put fuses on each FET (source) terminal , wouldnt this make repairs much more delightful for fellows or am i missing something? Fuse wire comes to mind.

Since i havent seen this in practice, im wondering if there is a pitfall in idea.
Health Warning: May contain traces of nut!
LH

Offline eidolon

  • Full Member
  • ***
  • Posts: 53
  • Karma: +2/-1
Re: power jack lf8000/48v blew up and smoked
« Reply #48 on: October 25, 2017, 11:03:52 am »
Fuses are resistors with much higher resistance than the FET on.  Source would be the worst place to place a fuse as any fuse failure would destroy the FET.  Commonly a failed FET will also short G-S.  If series gate resistors are used, you just look for the FET with the lowest gate  ohmage  to find the bad FET.

Offline rossw

  • Senior Moderator
  • Hero Member
  • *******
  • Posts: 879
  • Karma: +35/-0
  • Grumpy-old-Unix-Admin
Re: power jack lf8000/48v blew up and smoked
« Reply #49 on: October 25, 2017, 05:38:16 pm »
Also, fuses take A-G-E-S to blow, compared to the time it takes a FET to self-destroy.
Fuses may reduce the carnage, but they certainly won't save the FET in most instances.

Offline oztules

  • Forum Advisors
  • Hero Member
  • ****
  • Posts: 1177
  • Karma: +105/-8
  • Village idiot
Re: power jack lf8000/48v blew up and smoked
« Reply #50 on: October 26, 2017, 07:34:35 pm »
The best thing to do ( I do anyway) is to protect the input with say 120amps of breaker... standard 240v ones will do. ( yes it is DC and it is high current  I know that too) The 240v ones are typically $4.00 on ebay. They are not all the same, some weld shut on startup, most perform perfectly well. I use noark ones, they seem to run fine on <60vdc... DO NOT USE ON >60v DC though. I use 2 x 68amp ones on my inverters and battery strings for that matter.

EDIT: Further to the above. Clockman had a discussion with breaker manufacturers in Europe. They agreed that <60v, the 240vac ones are fine to use, but derate them 20%... thats it.

A big problem is the start current ( to charge the caps from cold at first turn on) This is in the order of over 400 amps, and is why ANY breaker will have a good excuse to weld up on start up splat. For some reason, most of the 240v ones seem to handle this ok, the noark ones from ebay do it with ease, and Ihave yet to see any untoward effects. Some of the other ones I have used... not so good. If you just splat the cable to the input bolts, you will so substantial damage to the lug or bolt threads... it is a big big surge..... especially on the 6 cap boards.

It is best to have a resistive charge up system, but my particular design will hang if the start up voltage rises too slowly, and requires a reset switch in the design. I think the PJ suffers this slow cap charging ok. Even a push button with 40 ohms in series to charge the caps will solve it all.
End Edit.

This protection may/may not stop fet destruction ( generally not)... but it will protect your circuit boards.

I do/did a lot of experimenting on these PJ and my own designs, and I don't care if the fets go bang,  but I hate the burnt resistors, burnt totem trannies, and subsequent board damage. It is stupid and unnecessary.

So input fuses will stop the cascading noises you normally hear on failure. It is just breaker goes, and the drivers and the rest of it survives, and you generally only need to replace the fets.... and they won;t have blown the pads away either.


.......oztules...... expert destroyer of fets and other tender electronic components.

Like this

7264-0

7266-1
Flinders Island...... Australia

Offline lighthunter

  • Sr. Member
  • ****
  • Posts: 323
  • Karma: +13/-0
Re: power jack lf8000/48v blew up and smoked
« Reply #51 on: November 03, 2017, 05:42:28 pm »
Some great information guys!! :) Makes sense, i knew there must be a reason not to fuse a fet since I never see it in industrial design. I still think I could pull it off by using some hall effect latches but it may require a re-layout of main FET board. I wouldnt even concern about it except mine did blow up a month or so ago and there wasnt a reason for it to have happend. As I went through the possibillities it did happen at a time when house loads could have been fairly high 3kw+ and then a hair dryer and well pump may have started on top of that. The surge might have required current limiting which I admit there is a problem.

When i changed this inverter from single phase to split phase it left one of the two without current limiting. Since the PJ limiting ct uses a physical wire connection rather than just magnetic, i cant add the two together by running two wires thru hole in opposite directions like i did for the face display meter. I may have to remove the original ct and change it to an inductive toroidal type. Anyone know what ratio ct i need to do this?

Very nice looking inverter in photo!!

BTW, i am using a 100A marine breaker on the 48 DC side, it never opened. So i will try the two 240 breakers, maybe they are quicker and of magnetic trip style although I may try to go with two 50A ones.

Again! Thanks for advice!!!
Health Warning: May contain traces of nut!
LH

Offline oztules

  • Forum Advisors
  • Hero Member
  • ****
  • Posts: 1177
  • Karma: +105/-8
  • Village idiot
Re: power jack lf8000/48v blew up and smoked
« Reply #52 on: November 04, 2017, 03:05:44 am »
"Anyone know what ratio ct i need to do this?"...... no but nearly anything will do. If you look at the circuit coming away from the original CT, it goes to a obviously added on resistor after the initial smd assembly.   By changing that value, you tune the new torroid CT ( plenty of 2000turn ones on ebay for peanuts).

If you put the torroid in series with the original CT (place the ac through the middle I mean) at some output, and read the ac rms  output of the 2000t torroid...  and the main CT......compare notes, and go from there.

In fact I use the 2000t ones in my 8010 inverters ( in pic). Just need to shunt it to give you the range your interested in. I generally cut out at 14kw ( about 60 amps@230v )


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

Offline rossw

  • Senior Moderator
  • Hero Member
  • *******
  • Posts: 879
  • Karma: +35/-0
  • Grumpy-old-Unix-Admin
Re: power jack lf8000/48v blew up and smoked
« Reply #53 on: November 04, 2017, 05:46:09 am »
If you put the torroid in series with the original CT (place the ac through the middle I mean) at some output, and read the ac rms  output of the 2000t torroid...  and the main CT......compare notes, and go from there.

Just a heads-up warning for the uninitiated.

CTs are not well understood by a great many people. They're a *CURRENT* transformer with a fixed ratio.
If you pass current through the primary of a CT and DON'T LOAD THE OUTPUT, you can get DANGEROUSLY high voltages.

Oztules suggestion is perfectly fine, but please don't just think you can stick a meter across the toroid output "safely".
Put some load or "burden" resistor across it BEFORE you power it up.

Offline oztules

  • Forum Advisors
  • Hero Member
  • ****
  • Posts: 1177
  • Karma: +105/-8
  • Village idiot
Re: power jack lf8000/48v blew up and smoked
« Reply #54 on: November 04, 2017, 07:49:29 pm »
Yes the typical CT runs into a very low impedance, and the open circuit equation looks very scary
Vs=Vp(Ns/Np) Vs= secondary volts, Vp=primary volts and Ns=turns of secondary, and Np== primary turns.
The info this comes from this site...... very informative, and nice pics.

http://www.electronics-tutorials.ws/transformer/current-transformer.html

They do the calcs as Vs=240(2000/1)=480000 volts.

Their particular example was on a 480v circuit, with 160 turns:1... they got 76kv... wow..

In the real world though, the 2000:1 torroid with 2500watts going through the hole ( 240v@10 amps or so), produces only 35v open circuit... which is a little bit less than the 480kv their equations  predict.... yes I went outside and tested it again to be sure.

I suspect there are differences I can't see, their diagrams are specific, but it don't add up.

Anyway, at 6kw you would feel a tingle, so stay under 5kw if you want to be pretty safe..

There are a lot of warnings regarding the average CT open circuit, so there is something in it for sure... I just haven't witnessed it first hand... all the testing I have done in this sphere is pretty woosy stuff.

So keep an eye out, where there's smoke there is fire, and Ross and other folks know lots more than me with this stuff .... but I haven't seen it first hand.... at all.



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

Offline rossw

  • Senior Moderator
  • Hero Member
  • *******
  • Posts: 879
  • Karma: +35/-0
  • Grumpy-old-Unix-Admin
Re: power jack lf8000/48v blew up and smoked
« Reply #55 on: November 05, 2017, 01:42:05 am »
In the real world though, the 2000:1 torroid with 2500watts going through the hole ( 240v@10 amps or so), produces only 35v open circuit... which is a little bit less than the 480kv their equations  predict.... yes I went outside and tested it again to be sure.

I suspect there are differences I can't see, their diagrams are specific, but it don't add up.

Your use of a 2K turn xfmr works in your favor here. A lot! And this is one of the things I hinted at - CTs are quite counter-intuitive. I don't mean to hijack the thread...

Most of us tend to think that a transformer with "lots of turns" is going to make "scary voltages" and that "less turns is more safe".  This is where the CT can bite. If someone were to take a transformer with only 20 turns (instead of the 2000 you tested), they may well be forgiven for thinking that it would be 1% of the output you have.... but it doesn't work that way! The transformer is going to try to generate 1/20 of the current in the primary, not 1/2000 of the current. So with your 2000 turn transformer, and a 10A current in the primary, it's going to "try to" push 5mA. With a 100R burden/load resistor, that'll make a whole 500mV. Without a resistor, it should still "try" to make 5mA, but as you found, transformer core losses, considerably less.

With our 20 turn transformer, it's going to try to push 500mA through the secondary and the secondary voltage (unloaded) will potentially be SIGNIFICANTLY more.

Even a conventional "off-the-shelf" CT to drive a 5A meter is designed to drive 5A into a meter with a coil resistance of 0.2 ohms (5 watts). If the meter is open-circuit for some reason (like someone working on it) .... well, lets just say, its the sort of mistake you only make once!