Heavily modified Power Jack

[noneyabussiness]

Hi all, new to posting but i just wanted to post up some details on the off grid setup ive been working on, and maybe some questions too boot.. oztules i have perused maybe 35-40 thousand of your 65 odd thousand words and u have not only helped greatly but have inspired to keep going... this system runs an average of 5 odd kw all day, 8 kw worth of panels and 1100 odd AH of 48v battery bank, we've pulled 9.5 kw for over two hours and had a 22°c (mosfet) temp rise... it runs 12 mosfets per leg (instead of 6, 4110's) the original twin transformers lasted 5 weeks!! Before melting themselves so we had dyne industries in Melbourne wind one up...  so far im very impressed... just a couple of pics of the mods ive done..and transformer.. more tomorrow but gotta get to bed..

[noneyabussiness]

Also, its been running for about 8 months now...

[oztules]

Can't wait to see the rest... thats some transformer you have there...
12 fets per leg..... now thats serious too.

tell us more pls.

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

[noneyabussiness]

the first pic is the finished product, and the second is a pic of the bank setup(sadly thats the only clear photo i have, after a early failure)..

we had some fun setting it up, ill post up the Arduino code for the fan control when i can, its pwm progressive with temp and works flawless.. even has lcd printout of temps and fan speed, uses a common 10k thermistor as sensor..

also im just coding a setup for all the info to be wirelessly transmitted to house, ill put that up too when im finished..

hope this helps..

[noneyabussiness]

this is arduino code for pwm fan control, i modified the timer registers on timer 2 to increase the frequency from 450 odd hz to 32 KHZ, because the fans whined like my mother in law at the lower speed, but you do lose a bit of resolution but in this case it didnt really matter.. there is a library called "thermistorfun" that u need to download but a simple google search will do.. hope it helps..

[noneyabussiness]

best pic ive got of it complete... before i tidied up wires etc..

[noneyabussiness]

and a pic of the 100watt (against a cheap 50w job) resistors I ended up putting inline with transformer.. 10 in total which is .001 ohm... which leads me to a question..

the original transformers were directly connected to mosfet banks (except iron core "inductor") and when we did the same on the larger (lol) transformer we went through about 3 banks of mosfets until i figured out putting the .001 ohm resistor in series stopped it from happening...  it would happen randomly, under load, idling etc. i tried putting a .01 ohm resistor in series with each mosfet (as advised by dyne) and it still did it, so i tried with the said resistor and never looked back, its fine being there i was just wondering if anyone has any ideas why?? ive spent MANY house researching it and the only thing i came up with is rise time of mosfets may be slightly (ns/us) different and the resistor is limiting surge at that moment..

[oztules]

Iron inductor? the inductor should be ferrite, and greater than 18uh.  What is your idle current .

Wondering of your fets have too much capacitance for the drivers. A bigger inductor ( ferrite) may help here. I'm thinking all these resistors are really having an inductive effect more than a resistive effect.

Is that a dual "C" core you had wound, and what is it's rating.

Typically the C core is better than the EI but worse than the torroid or R cores. It may have enough leakage to help keep inrush currents down, and it will also tolerate DC on the primary better ( the tiny air gap will act as a fuse "sort of" for the magnetic field that would saturate a torroid with DC component in the primary.

RossW would have a better understanding about this resistor problem than me, .......but we may need to prod him into action .......

I think the inductor I use is probably involved with the switching transitions more than filtering the 23khz.... it is to small to do that.

In your case a bigger one may have helped stop the blow ups... but thats just conjecture.

You really have taken this another stage.... but gee the drivers.... yet you say the cooling is good under load..... The 8010 ones I have built surprised me with the ability to switch very high  cap gates too.

[noneyabussiness]

Morning. .. you are right it is a duel  "c core".. afer discussing this with dyne thats what they recommend. .. funnily enough it was actually cheaper to go toroidal and i wanted to,  but like i said they suggested so not being a pro (tinkerer who reads alot), i relented. And she is rated for 10kw continuous 32v primary 240v secondary .. ferret makes sense, i used the original that came with the inverter, two green toroids wrapped once.. they get hot so there doing something but maybe bigger is the go. Idle current is hard to determine because of fans and where this thing lives but on the bench its about an amp or so.. and the drivers seem to be doing just fine, earlier i thought maybe it was them, so i put them on my old but reliable cro and under about 5kw load it looked great (no photos sorry), very clean... ..  i have to build a backup for it so it will be a 8010 modified like u said as it about the only complaint is lack of current limiting except for the sub board (has its own).. hasn't been a problem and has earth leakage switch etc.. all banks of batteries are individually fused etc.

[noneyabussiness]

Just a note to show how tough they are, it was built for my brother, and he still has a connection to ergon (isolated 3 way switch," generator connection") so to  on the really cloudy days for a few months he would switch back to ergon.. we sorted it now but even though he was warned to only change at low loads he was switching at 5 + KW .. never bothered it... imagine the transients and spikes... i did install TVS diodes and MOV suppressors across mosfets too so i imagine they helped greatly. .

[oztules]

Excellent stuff.... my units have no problems switching across to mains and back at  5kw and more, so I think it is just the tough things that they are.

Interesting that dyne wanted the c cores.... did they say why.... can only think they were worried by the DC component.

You have certainly pushed the envelope.
Note the 8010 will need less primary turns... ie 26v:240v It lacks the dynamic head room of the PJ controller.... and change the gate resistors and install diodes..... otherwise the pj fet board is a good design.


.........oztules

[noneyabussiness]

Yeah, the person i spoke to was "vlad" and yes heavy Russian accent too boot... so i only really understood half of what he said, but yes the dc component and something about surge and tripping breakers.

Bugger about voltage change on primary for 8010 may have to stick with the PJ, cant (not impossible just what a job) change the primary .. its like a 3 inch "strap", i tried bending it slightly to fit it better where it lives and i couldn't even move it...

[oztules]

For H bridge switching, the DC component gets reversed every 1/2 cycle, so would not have been a problem.
The surge is very very real, and only soft starting and no....... and I mean no modification to the frequency while running.

This is also true of the small PJ torroids, they have very little leakage, and will cause he same fet explosions for the same reasons, the bigger transformers are just better at doing it.

That said, neither the PJ or the 8010 cause any problem.... except if you use the pj for charging without turning he unit off first before removing the grid... else as it reverts to 50hz, ... the change in Freq will most likely blow the fets up.

So he was right on that count.

The price you pay is...... twice the weight transformer for he same power, and more leakage under loads ( sag) so the fets have to work that bit harder.

The proof is in, the Dyne works very well, and thats what counts......but....next time get the torroid perhaps ....I know they work very very well too.

They are also easy to change the primary..

I think clockmans new transformer will weigh in over 50kgs, the equivalent C core will be 100kgs or over 220lbs in the old money.... there is the difference really for next time.

Yes, stick to the PJ for now.





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

[noneyabussiness]

Thx for the advice,  this thing weighs 98 odd kg so makes sense.. and the charge circuit isnt a problem  been disconnected... and i have to build another for my sister's inlaws so ill be going toroidal then...this time ill probably go the 8010 just for simplicity and sooooo much cheaper... with conversion the PJ cards are expensive now..

[solarnewbee]

Hey there noneyabussiness!

Having a bit of trouble with your arduino sketch. stops compiling at #include <thermistorfun.h>

Arduino: 1.6.8 (Windows 10), Board: "Arduino/Genuino Mega or Mega 2560, ATmega2560 (Mega 2560)"

C:\Users\Pancho\Documents\Arduino\libraries\sketch_jul29a_fan_control_works2_oldpwm\sketch_jul29a_fan_control_works2_oldpwm.ino:2:27: fatal error: thermistorFun.h: No such file or directory

 #include <thermistorFun.h>

                           ^

compilation terminated.

exit status 1
Error compiling for board Arduino/Genuino Mega or Mega 2560.
Invalid library found in C:\Users\Pancho\Documents\Arduino\libraries\sketch_jul29a_fan_control_works2_oldpwm: C:\Users\Pancho\Documents\Arduino\libraries\sketch_jul29a_fan_control_works2_oldpwm
Invalid library found in C:\Users\Pancho\Documents\Arduino\libraries\sketch_jul29a_fan_control_works2_oldpwm: C:\Users\Pancho\Documents\Arduino\libraries\sketch_jul29a_fan_control_works2_oldpwm

This report would have more information with
"Show verbose output during compilation"
option enabled in File -> Preferences.


 Any help would be great maybe direct me to another topic so we can discuss when you have time.
Still learning code at this point but thought your sketch would get me a head start on controlling my PJ fan since Oz says their analog is weird. don't want to burn one up for a simple problem since it will be 10,00 miles away from me.

Great read above, thoroughly enjoyed.

Regards,
SN