15kw PJ transformer

Firstly, PJ reverses "input" and "output" on the transformer...as far as I'm concerned, the "input" is the FET side, and the "output" is the AC high voltage side.  PJ nomenclature is reversed from this.

I am assuming from "15kw" that the battery voltage is 48v...all numbers from here on out assume the same.

2) PJ spec is 36vAC primary, 230vAC secondary.  This provides very little headroom for regulation before the wave starts to saturate.  GS units currently run 32vAC primary, 240vAC secondary; I'd still like to reduce the primary voltage for a bit more headroom.  HOWEVER, read farther on...

1) I am presuming this is a stock PJ control board (with upgraded CPU firmware)...if so, I would have to see a photo of the output board to know whether the center tap is required or not.  The large PJ inverters (in my experience) seem to regulate from one 120v phase...meaning that if you remove the center tap, the inverter will either give you 120vAC out (of your 240vAC transformer) OR fail to be able to regulate voltage (which may blow FETs).

One reason that PJ runs such a shallow transformer ratio, is because the LF Driver is mismatched and imbalanced.  The more of a square wave output the inverter produces, "surprisingly" the cooler the FETs are (because they aren't switching on/off at the SPWM carrier frequency).  Using a steeper transformer ratio (i.e. lower primary voltage) forces the FETs to run more of the wave in the SPWM region--and "strangely" they get a lot hotter.  This is because of the mismatched and imbalanced LF Driver.  Your mileage may vary.

Running a very low primary voltage starts to cause problems with EMI and resultant signal crosstalk on the FET control cable (rainbow ribbon cable).  Regardless of that, the signal switching crosstalk starts to reach fatal levels much past 6kw (at least in our experience with GS control boards with properly matched stronger drivers)...causing the FETs to run very hot.  In significantly higher loads, the FETs will blow immediately due to this crosstalk.

These issues are in part why we have decided to no longer promote/support upgrading PJ inverters.  Because by the time you replace everything that's being problematic, you basically end up at exactly the same place you'd be if you had started from the ground up with a completely different design.  Which is pretty much where we are at with Genetry Solar right now (prototype testing).

Thank you Sid for your quick reply, yes all off your assumptions where correct, I have come across a core 12" dia, 5" hole, 6" high, from memory I think that's about 10 turns primary 78-80 turns secondary, should be big  enough 🤣

Regards Alan

One reason that PJ runs such a shallow transformer ratio, is because the LF Driver is mismatched and imbalanced.     36v  primary transformer  ASL 9 .  

A   bad  LF Driver  on my  8kw PJ   need  a new  LF driver only  .    A  bad  LF driver on  15kw  20kw  25kw  30kw is very  expensive  .   I  will need  to replace  mosfets   100 dollars  and a control board with  new LF  driver  100 dollars .     Maybe now I   can not get parts . 

 Because by the time you replace everything that's being problematic, you basically end up at exactly the same place you'd be if you had started from the ground up with a completely different design.  Which is pretty much where we are at with Genetry Solar right now (prototype testing)  from  Sid  . 

That's an extremely high core Tesla (3 volts/turn), are you sure of these numbers?  The 12kw GS inverter runs a core just a bit smaller than that, and it's 1.5v/turn.

Technically, you might be able to run 3v/turn...but unless the core is designed for it, the transformer efficiency will be very low, and it'll put extreme strain on the FETs.  If you have access to a variac, you can test the number of turns vs no-load current off the AC line (magnetic excitation current)...add turns until the no-load current stops going down.  (Beyond this point, extra wire continues to reduce the turns/volt, yet without any improvement.)  No load current should be under 10W or so at idle (but I'm not a seasoned transformer winding expert like @ben).

Once again thank you Sid, no I'm not 100% sure of those figures, the math just drives me crazy, my core is 320mm diameter, 130mm hole, and 150mm high, so if anybody out there, can help me out that would be great, I'm looking for 36v primary and 240v secondary winding turns, every time I do the sums I seem to get something different, regards Fido 😥

 PJ spec is 36vAC primary, 230vAC secondary.  This provides very little headroom for regulation before the wave starts to saturate.  GS units currently run 32vAC primary, 240vAC secondary; I'd still like to reduce the primary voltage for a bit more headroom   from Sid .

From  other  youtube video  calculation  36vac  is too  high  for  240vac .   GS  32 vac primary  is better  but I  think  even lower  than 32 v  is better  for the FETS .  

That's a HUGE core!  (Bigger than the core PJ uses in their "30kw" inverters, which obviously can't do that continuously.)  Assuming it's a good core, you might be able to wind it for over 30kw continuous.  Can pretty much guarantee the PJ mainboard can't do 30kw continuous, but your transformer will run cold at a paltry 8kw load 😉.

It's also a good bit taller than the core we use in the GS 12kw inverter (100mm tall vs 150mm on yours).

Like I said above, you can trial-and-error to determine the turns/volt for best efficiency (lowest idle current).  Once you know the turns/volt, then it's a simple matter of calculating the desired ratio, then winding all the wire on it.

Hi Sid, thank you for all your help and advice, just reporting that at 7500w my transformer temp is 26°C, 5° above ambient, thank you so much 🙂

(Mistook you in brief for a PJ customer who'd just bought a 15kw PJ....)

That's good to hear.  Care to post any photos for the rest of us to drool over?

Sure Sid, may not look pretty but works so well, thanks for all your help, primary is wound 12 turns 4 in hand, secondary is 80 turns 12 in hand, center taped at 40, 2mm copper wire ( when I disassembled the pj transformer, it was copper coated aluminium wire 😡 once again thanks for your help 👍<a class="ipsAttachLink ipsAttachLink_image" href="/monthly_2021_07/20210712_094451.jpg.267162fb6c5e4bbf4c95be6d0616bf1e.jpg" data-fileid="607" data-fileext="jpg" rel=""><img class="ipsImage ipsImage_thumbnailed" data-fileid="607" data-ratio="178.15" width="421" alt="20210712_094451.thumb.jpg.0204c9ca5504f445039195368f807481.jpg" data-src="/monthly_2021_07/20210712_094451.thumb.jpg.0204c9ca5504f445039195368f807481.jpg" src="/applications/core/interface/js/spacer.png" /><a class="ipsAttachLink ipsAttachLink_image" href="/monthly_2021_07/20210712_094435.jpg.e216dbbcad799e2c90ff2ec15f6d0a84.jpg" data-fileid="608" data-fileext="jpg" rel=""><img class="ipsImage ipsImage_thumbnailed" data-fileid="608" data-ratio="56.20" width="1000" alt="20210712_094435.thumb.jpg.982c340235966854994b89496c0a2b16.jpg" data-src="/monthly_2021_07/20210712_094435.thumb.jpg.982c340235966854994b89496c0a2b16.jpg" src="/applications/core/interface/js/spacer.png" />

It's enameled aluminum wire, no copper to speak of.  Reduces cost and weight, at the expense of slightly thicker wire size for the same power output.

What's the no-load current of the inverter?