Output voltage adjustment

[lighthunter]

Hi guys, i'm wondering if the 120/240v split phase version of the pj board is different from the single phase version in some way causing your problem. Seems like you mentioned a loud hum when you tried it the first time too. Basically all you are doing is moving the the maintx-n to p4 connection off the board and replacing the maintx-n terminal voltage with a slightly adjusted one.

Your driver transistors should have a code on them like MY and NY which corresponds to NPN and PNP.

LH

[dochubert]

Hi lighthunter,
I don't understand it either, since it worked for you.  Was very very careful with hookup this time around, and first test had the toroid transformer output just where I wanted; 9-10 volts low.  Should have worked....
Well, guess I get to change my first set of mosfets

My spare control board's drivers didn't seem to have any markings, but scraping with thumbnail found very faint MY on the 2 nearest the control board edge.  Q08 and Q03 for the MY.  The other 2 don't seem to have any markings at all.  Duh, guess they're the NY

[lighthunter]

Hi Guys! yeah, i agree, MY is Q3,Q8 and NY is Q5,Q9 next to U1. I feel bad about your misfortune DocHubert. Pm me if there is any way i can help. If u send boards i will fix em for you. The worst part about a board failure is the downtime. If they can b fixed quick then one forgets about the bad experience quickly    Yeah it should have worked. By info it seems like it was running without feedback. One thing i liked to do was measure voltage between maintxn and maintxl before making change while the new feedback transformer is connected to the ac inverter output. This way the original and the (new) feedback voltage can be compared before change. I know you did this but somehow either an unexpected ground loop or something made the feedback voltage invisible to the control board.

Cheers,
LH

[dochubert]

Thanks lighthunter,
I appreciate the offer but will just suck it up and do the job myself when my mosfets and stuff arrive.  I knew the risks when I decided to try it. Nobody to blame but me.  Just a bad time (when is a good time?) as working on the house to list for sale in next couple of weeks.
 
If you decide to build an ozcontrol 8010 and have pc boards made, I would be grateful if you could make a couple extra for me as I have no experience with that.  I wiil pay for them of course.  I'm starting to gather parts for it in anticipation.  (Ah, an output voltage trimpot!)

Downtime does suck.  I keep wanting to go turn mine on and power the house on a nice sunny day like today and watch the power company meter spin madly backwards from 2000 watts of gridtie solar and no load.  (Yes, I kept my analog meter and told sdg&e to find something intelligent to do with their 'smart' meter.)

[lighthunter]

Hi Doc Hubert!

Yes, I kept my analog meter and told sdg&e to find something intelligent to do with their 'smart' meter.)

You animal!!! good job! am proud of you! Mine was switched before i got the place

In february of this year i was having an issue with my grid tie tripping off and i plugged it into the mains to test and seemed all good thinking the clothes dryer would eat up all the output. As luck would have it I smoked a belt on my snowblower and took a bit too long before checking on it. The dryer had finished the cycle and .... the smartmeter was reporting reverse power.. I never got a phon call but I'm pretty sure they knew it. I was worried cause a friend in texas had that experience as well and they came out to check things. Guess its not a big deal just another $400 that would take years to come back... They install another meter and all. Id just as soon keep em separate.

Back to your situation. Hopefully I will have a bare board done this week some time. Were you thinking you want a bare board or loaded and tested? I'm not a wizard with pc board making but its coming along ok. My those drill bits are tiny. I hope your repair is going well. Im still not sure where the lights were on your control board you spoke about but after measuring for shorted drivers with an ohmmeter i would remove any shorted ones and power it up without the ribbon cable plugged of course and see if your lights are back on. If not look for another fail part. Another way is to measure for the 5v on the 7805. Once you get power back you know replacing the drivers an optos an fets will restore normal use.

Cheers,
LH

[dochubert]

Hi lighthunter,
The lights I was talking about are the 4 leds on the main board. Each led indicates that section of fets is outputting properly.  On my pj, one led is completely out and another is dim.  They are the two sections that the big transformers connect to, the output half of the main power board.  No lights on the control board
Sorry, guess I wasn't clear about that.
I plan to just change all 4 drivers and both optos on the control, and all the fets on the main (and anything else that looks stressed)  Luckily, I do have a spare control board, so once the mainboard is repaired I can be back up and running.  The spare board is the one that has the added 470k ohm resistor in the resistor chain, and puts out 240v (mostly)  The 4110 fets haven't arrived yet.  Heck, I haven't even taken the boards out of the case yet

About those 8010 boards.  I will be happy and grateful to get a bare board and do the assembly.  I already have some of the parts.  I just read today that Oztules just made some changes; added a small transformer and changed some resistors.  Maybe we shouldn't be too quick to build....
Also, he recommends changing gate resistor values and adding diodes on the main pj board if changing to the 8010 control.  Don't know if you saw that one

That 'smart' meter thing was something.  Luckily, I'm home most of the time and was able to intercept 3 separate tries at them sneaking in and changing my meter.  I told each tech that I was refusing the changeover because the new meters were harmful to health and almost always resulted in higher electric bills.  They all acted shocked at such an idea, but were polite and left.  3 attempts says that sdg&e really didn't care about customer concerns (how shocking!).  Eventually, they decided to allow customers to keep their analog meters if they wished.  Of course this was long after 99% of them were already changed.  They would now agree to replace the 'smart' meter with what they were calling an analog meter, but which was actually a digital meter disguised as an analog meter. Of course the customer had to pay a service fee for changing it.  I have no direct knowledge, but would bet the 'new' analog meters are just as hazardous to health and transmit your electric usage same as the 'smart' meters.  Glad I was ornery and held onto my original.   And because of the analog meter, I can bank some KWs early in the month running my powerjack as long as I make sure I owe for 20-30 kwh by meter reading day.  I never let it be negative when read so nobody pays any further attention (hopefully).  I pay a small bill each month to avoid notice.  The price of renegade solar is constant vigilance! (and a few bucks)

[dochubert]

Might just as well tag on here;

After recently frying my powerjack while experimenting, I am still not back up and running.
My mainboard has the 4 small daughterboards each with 6 irfb4110 mosfets.  Tested new mosfets for function with dmm before installing them.  Isn't there a discussion on this site somewhere about testing these mosfets before using?  I couldn't find it.
Anyway, checked all the resistors, etc and each daughterboard got all new mosfets as well as new snubber resistors. The snubber resistors all looked fried but still read correct resistance.  Put new ones in anyway. 
Have a spare control card so used it and put unit together this morning. All connections good except 10 lead cable from main to control board not plugged in yet.  After connecting battery power, noted the leds on each daughterboard were lit, which is supposed to mean the mainboard is functioning and ok.  Plugged in the 10 lead cable to the control board.  Still good.  Turned on powerjack and got pop, pop, and smoke so turned it off.  Leds are not lit now on the 2 daughterboards on the +DC input side.  Can't yet tell how many exploded there.  On the output to transformer side, leds are both still lit even though I can see at least 2 fets have exploded.  (Doesn't say much for those leds as health indicators, does it?)
Funny, my experiment that originally fried the pj didn't physically destroy any mosfets.  In changing them all only found 1 that read bad out of 24.  My repair job results in several exploded fets.  How did it get worse?  Bad mosfets?  Did I do something stupid?

Questions:
Whats the best way to test 4110 mosfets before installing?
By blowing up more mosfets, have I also blown up my opto's and the 4 transistors on the spare control board?  Any easy way to test those in place, or should I just automatically replace?
Open to suggestions

[oztules]

replace the trannies and opto's, and mosfets and resistors again... and await the explosions.

Hopefully thats all that is wrong with it.. if it goes further than the totems and opto's your toast unless your very good at bush ranging.

If it is gone, perhaps do a 8010 board, and change the resistors and put diode across then... at least you know whats happening then.

Testing mosfet. Do an RDSon test. that will show that it turns on, off and carries current. ( I use 10 amps generally).


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

[dochubert]

Thanks Oztules,
I just have to face doing it all again. 
I did get one of Clockman's books with pc boards, but didn't plan to build just yet. 
Hate to waste all this sun so will probably reconfigure the battery bank for 24v and use my 8kw unit in the meantime until I get the 15kw working or build the new 8010 ozcontrol.  I have to order more mosfets so it will be awhile

[oztules]

I'm currently having a run of outs at the moment... on the 8010 board, trying to perfect the current system.... blew two lots of fets up today....sometimes it doesn't pay to play with something that works fine... trying to get that last bit out of it before it current locks.

At least the 8010 board does not seem to go with the fets so far.. it seems to have no problems, even though I designed it to just plug in the three chips and go again.



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

[dochubert]

sometimes it doesn't pay to play with something that works fine

Words to live by!

Glad to hear the 8010 control can survive the fets popping.  Definitely a plus!

On an almost unrelated topic, soldering that 8010 chip to its adapter board.  I'm old, hands not as steady as they used to be and eyesight not what it used to be either.  I'm waiting for a smaller tipped tip for my soldering iron and hoping that helps.  I keep bridging pins.  Frustrating.  Any suggestions?

[RFburns]

If the pins bridge you can use solder wick ( a desolder copper strip - could probably use the sheild off coax if you dont have any) to remove some of the solder  . Or drag solder see here http://www.instructables.com/id/How-to-Solder-SMD-ICs-the-easy-way/ .Or you could use a small toaster oven/ heat gun, proper surface mount solder paste / glue and bake it (Google will show you what I mean ;lots of home builders use surface mount) eg:http://www.starlino.com/surface_mount_reflow.html.   RF

[oztules]

I had the same problem originally.

I used a resin pen from when I did the solar panels.... simply pen some resin on the board... place 8010 on the right spot, and using no solder, and a smallish tip ( actually ended up using normal tip) and a hot iron touch each lead. The board is tinned already, and so the flux then allows the tiny amount to run nicely to the chip legs.

You cant get a run, as there is not enough solder there to do that..... and you may need to clean the tip with wet sponge and water first to get rid of the burnt resin/ rosin  and other oxides that may normally be on it.

It takes only 30 seconds or so to do each board, (I did a dozen at once, you get better after each, until it looks professional after about 3 )

It turns out to be simple, but the first ones I did using solder were a disaster as you have presupposed... I heat gunned them off, and started afresh.

RF, how is the egs002 coming along now??

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

[dochubert]

Thanks Oz and RF,
I've only tried one so far, and yeah, it looks pretty bad.  Except one side the row of pins looks like a machine job.  Must have gotten in the groove there. (Or an Elf snuck in and did that row.) Couldn't do it again tho.   

Checked those links RF, good info there.  Knew there had to be a better method.  Looks like there are several ways to do it.  I need to learn some new tricks, looks like.
Think I'll try your method, Oz.  Seemed the least likely to end up with a mess
I have i think 8 of the 8010 chips and 10 adapter boards so will be developing some new skills soon
Still looking for my solder wick.  Have 3 or 4 of them somewhere.  With the house for sale, my work area is now in the garage and my truck is in the driveway, and nothing else is where it should be.  Going to have a proper shop area at my next property
Thanks again

[RFburns]

Well Oz ,had a bit of fun with the EGS002 and have smoked a few mucking around with them -2110 drivers are pretty easy to smoke- (driving things that I shouldnt - BIG IGBT's - salvage from industrial VFD's) but havn't done much since the last one . But they perform bloodly well for the $$, been distracted with something else which ; if I get round to will post up eventually. RF

P.S Have ordered one of these to see what it like. http://www.ebay.com.au/itm/DC-AC-Converter-Pure-Sin-Wave-Inverter-H-Bridge-Drive-Board-TDS2285-/251800396314?hash=item3aa079221a:g:pNwAAOSwa39Uvil5