Home
JAQForum Ver 24.01
Log In or Join  
Active Topics
Local Time 13:00 28 Nov 2024 Privacy Policy
Jump to

Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me.

Forum Index : Electronics : 8 KW Inverter Build

     Page 8 of 22    
Author Message
oztules

Guru

Joined: 26/07/2007
Location: Australia
Posts: 1686
Posted: 12:59pm 26 Nov 2016
Copy link to clipboard 
Print this post

For a very simple controller, that works very nicely for me. Most times the fan is running so slowly, you cannot hear it, but it still moves enough air for sub 2kw loads down here.


........oztules
Village idiot...or... just another hack out of his depth
 
Madness

Guru

Joined: 08/10/2011
Location: Australia
Posts: 2498
Posted: 03:55pm 26 Nov 2016
Copy link to clipboard 
Print this post

Hooked up to the house and ran for about 40 minutes and bang! All 12 fets on the side that switches at 50HZ are full of craters!Edited by Madness 2016-11-28
There are only 10 types of people in the world: those who understand binary, and those who don't.
 
Clockmanfr

Guru

Joined: 23/10/2015
Location: France
Posts: 429
Posted: 11:20pm 26 Nov 2016
Copy link to clipboard 
Print this post

Hi Madness,

Flippin heck! you are not having much success with this.

I am at a loss to put my finger on one particular area of your build that is giving problems.

I re-read your build, and yes you have a different toroid build, and power board to us/me standards, and your blows are when things get warm.

Could it be your toroid? faulty/fractured wire wound switching in and out, or shorting a secondary wind occasionally and therefore giving the control board hard work adjusting?.

Even my windings with New wire, I still come across fractures. I am one of those guys that likes definite Mylar and resin wrap separation, and not relying on the copper enamel alone.

Power board,.... I hate mechanical connections, straps etc, and nuts & bolts. Where I use them on the FET's I use star/grip washer to ensure biteing contact, so I ensure good track size and separation distances, and have tried to ensure a good big surface area contact for soldering.
I don't trust via's on my double sided double copper Power boards, so I double up and normally allow for a pins through and solder top and bottom.
There are other things that I consider, ground plane loops etc, but they get real technical and that's not my field, but with vast amounts of POWER on this board using the PJ board as a basis and then vastly improving it, seemed a good way forward.

If you were a bit closer, I would come round with a few bits, swap stuff about and wade through every thing step by step over a couple of days.


Everything is possible, just give me time.

3 HughP's 3.7m Wind T's (14 years). 5kW PV on 3 Trackers, (10 yrs). 21kW PV AC coupled SH GTI's. OzInverter created Grid. 1300ah 48v.
 
oztules

Guru

Joined: 26/07/2007
Location: Australia
Posts: 1686
Posted: 12:02am 27 Nov 2016
Copy link to clipboard 
Print this post

I just don't understand this at all.

Maybe time to get a PJ 15kw pair and start from there.
Just don't know why.

.........oztules
Village idiot...or... just another hack out of his depth
 
Madness

Guru

Joined: 08/10/2011
Location: Australia
Posts: 2498
Posted: 12:13am 27 Nov 2016
Copy link to clipboard 
Print this post

  oztules said   I just don't understand this at all.

Maybe time to get a PJ 15kw pair and start from there.
Just don't know why.

.........oztules


Won't be doing that.
There are only 10 types of people in the world: those who understand binary, and those who don't.
 
mason

Regular Member

Joined: 07/11/2015
Location: Canada
Posts: 86
Posted: 02:45am 27 Nov 2016
Copy link to clipboard 
Print this post

Madness,


I feel your pain

Edited by mason 2016-11-28
 
Madness

Guru

Joined: 08/10/2011
Location: Australia
Posts: 2498
Posted: 03:07am 27 Nov 2016
Copy link to clipboard 
Print this post

Hi Mason,

Thanks for your input, I get the same waveform on a single core toroid that is not bolted in. I am very suspicious of the MOSFETs. A really crappy board I have been using for testing control boards gives a perfect sinewave. Currently, it has IRF640 FETs, I have a new batch of MOSFETs that arrived a few days ago, I am going to try them and see if that makes a difference.

This thing is not going to beat me.
There are only 10 types of people in the world: those who understand binary, and those who don't.
 
Madness

Guru

Joined: 08/10/2011
Location: Australia
Posts: 2498
Posted: 03:37am 27 Nov 2016
Copy link to clipboard 
Print this post

  Clockmanfr said   Hi Madness,

Flippin heck! you are not having much success with this.

I am at a loss to put my finger on one particular area of your build that is giving problems.

I re-read your build, and yes you have a different toroid build, and power board to us/me standards, and your blows are when things get warm.

Could it be your toroid? faulty/fractured wire wound switching in and out, or shorting a secondary wind occasionally and therefore giving the control board hard work adjusting?.

Even my windings with New wire, I still come across fractures. I am one of those guys that likes definite Mylar and resin wrap separation, and not relying on the copper enamel alone.

Power board,.... I hate mechanical connections, straps etc, and nuts & bolts. Where I use them on the FET's I use star/grip washer to ensure biteing contact, so I ensure good track size and separation distances, and have tried to ensure a good big surface area contact for soldering.
I don't trust via's on my double sided double copper Power boards, so I double up and normally allow for a pins through and solder top and bottom.
There are other things that I consider, ground plane loops etc, but they get real technical and that's not my field, but with vast amounts of POWER on this board using the PJ board as a basis and then vastly improving it, seemed a good way forward.

If you were a bit closer, I would come round with a few bits, swap stuff about and wade through every thing step by step over a couple of days.



Hi Clockman,

My toroid now is very much the same as yours, I ripped off the primary and changed it to 70mm square welding wire, you can see it the photo's from a few days ago. That had no effect apart from making it a little quieter. Apart from that I have 2 other transformers that give the same result so I would be very surprised if that was the problem. In the future I plan to build another the same as my original one but with more wire on it. Probably 6 in hand for the secondary and 80 - 100 mm square on the primary, but that is for the future.

While I had it running at 4.5 KW I checked the temperature of all the FETs, connections, choke etc. Everything was running cool, the hottest part was the choke at around 50 degrees, I have a Ferrite ring coming which I will put much heavier wire on to reduce the resistance. When it failed it was running at around 500 watts and had been for around 15 minutes.

All my PCBs are homemade, so no plate through vias, mine are all wire with 3-4 mm soldered on each side. Interestingly a really badly made board from one of my earlier attempts gives me a perfect sinewave, the big difference here is that it has just 4 FETs on it and are IRF640's. I am very suspicious of the FETs I have in the powerboard that failed. As they are Chinese cheapies perhaps they are rejects with some sort of fault, I see where some people have fits about MOSFETs not having matching date codes when running parallel in a bank as we are doing. I can't fault them with my multimeter, but I know there is more elaborate equipment required to test them properly.

If you are passing by you are more than welcome to come and visit. Perhaps one day I might be in France and can see what you are doing there in the home country of my ancestors. (They escaped to Prussia save their necks during the French revolution.)
There are only 10 types of people in the world: those who understand binary, and those who don't.
 
Madness

Guru

Joined: 08/10/2011
Location: Australia
Posts: 2498
Posted: 10:14pm 28 Nov 2016
Copy link to clipboard 
Print this post

After etching a new PCB with all tracks and pads connected to the gate drives reduced to 2mm wide I have got this. 300ns dead time, slightly lower idle current and runs very quite.

No load


9.5A fan heater


Before would have had sharp spikes around 0 crossing.

Hopefully, the curse is lifted, been nearly 6 months since I started this.

There are only 10 types of people in the world: those who understand binary, and those who don't.
 
oztules

Guru

Joined: 26/07/2007
Location: Australia
Posts: 1686
Posted: 10:48pm 28 Nov 2016
Copy link to clipboard 
Print this post

I'm too frightened to say anything until it has run the house for a week

........oztules
Village idiot...or... just another hack out of his depth
 
Madness

Guru

Joined: 08/10/2011
Location: Australia
Posts: 2498
Posted: 11:13pm 28 Nov 2016
Copy link to clipboard 
Print this post

Yes, I should just shut up and keep everything crossed.

It is definitely very different, this latest PCB has TO220 MOSFETs. It is not the MOSFETs that are different as I have another board with same MOSFETs with previous design and it had the same nasty spikes at zero crossing. Now I am doing a further redesign to get gate resistors and diodes as close to the gate as practical. I can get it about 3mm closer, doesn't sound like much but the changes already made so far have had a dramatic effect. The fuzzy wave is about half the width now also. I am assuming less copper and not having tracks on both side is making the FETs turn on or off faster (or both). It is like it has a changed personality before it sounded angry, now it is quite and calm with just the slightest noise in the choke that was buzzing loud enough to be heard 10M away before.
There are only 10 types of people in the world: those who understand binary, and those who don't.
 
Clockmanfr

Guru

Joined: 23/10/2015
Location: France
Posts: 429
Posted: 10:59am 29 Nov 2016
Copy link to clipboard 
Print this post

Hi Madness,

Just sent you my Power Board Component list.

Normally when I rough test the Fets, 50 at a time, I get about 2 that are very different, so I reject them, the rest I try to match up.

My Gate resistors and fast diode piggyback are about 4mm from FET and all are about the same or as close as possible.

I remember Oztules and I discussing the Power board design it self, and learned folk advised about the tracks size, equal distances etc, fast diodes, and a few other alterations.

With the original PJ board it was the QC when they were being assembled that seemed to be the issue, mostly crap soldering. So PJ went to modular sub FET boards hoping QC would not be as much as an issue.

I will also keep my fingers crossed.
Everything is possible, just give me time.

3 HughP's 3.7m Wind T's (14 years). 5kW PV on 3 Trackers, (10 yrs). 21kW PV AC coupled SH GTI's. OzInverter created Grid. 1300ah 48v.
 
Madness

Guru

Joined: 08/10/2011
Location: Australia
Posts: 2498
Posted: 11:30am 29 Nov 2016
Copy link to clipboard 
Print this post

Thank you Clockman,

New PCB has ran overnight now with no load, I will be connecting it to my house shortly. It has had a few bursts at 9.5A connected to my test setup with a single core Toroid. To put it in the case with the big Toroid will take about 5 minutes, then I can give it a good workout. Here is an image of the revised board, I have reduced the tracks further on this than the one that I have just made. I will start making this new version after the current one has run for a few days successfully. In case you are wondering the extra holes near where the safety caps go is for MOV's for lightning protection.



There are only 10 types of people in the world: those who understand binary, and those who don't.
 
Madness

Guru

Joined: 08/10/2011
Location: Australia
Posts: 2498
Posted: 07:58pm 29 Nov 2016
Copy link to clipboard 
Print this post

The Inverter is now connected to the house for about 8 hours now. Did a load of washing in the Front Load washing machine and had HW on for about 30 minutes this morning, it ran for about 30 at 6 KW then. I gave it another workout this afternoon another 30 minutes with 34 degrees ambient, heatsinks got up to 48 and toroid is shown below.





Bottom of Toroid was much cooler where the fan is directly on it, I have a couple smaller fans that I will mount above it blowing air on the top from inside the case.
There are only 10 types of people in the world: those who understand binary, and those who don't.
 
Clockmanfr

Guru

Joined: 23/10/2015
Location: France
Posts: 429
Posted: 09:50pm 29 Nov 2016
Copy link to clipboard 
Print this post

Yes. that board looks better.

Although the FET legs look a little close together for me. On mine I have opened the legs so the solder pad is bigger and the space between the legs is wider. But to wide and the FET legs are fractured.

Why? because my installs are in a damp/moisture environment. Its amazing what humidity can do.

I trust these 2 pics help.?

Fingers still crossed.







60 for the toroid, mmmm !.

I like Oztules original toroid design with the big single primary. This arrangement allows a gap, 12mm to 14mm for the air to flow through the centre of the toroid.

Some consideration is required with the OzInverter regards air cooling, hot air rises, well it does here, so I have my base and top panel ventilated for natural air flow.
I also have 2off 120mm dia fans in a push pull arrangement on the toroid, and 2off for the PCB boards. I doubled up on the OzCooling circuit so toroid and PCB's can be independently cooled depending on their needs.




Here, the push fan is mounted on the side, but you can see the air gap between the primary cable and toroid secondary wrap, and a nice breeze is flowing up the centre of the toroid. That's one of the toroid two heat sensor probes that you can see at the front, just cable tied to the primary against the secondary wrap.



Everything is possible, just give me time.

3 HughP's 3.7m Wind T's (14 years). 5kW PV on 3 Trackers, (10 yrs). 21kW PV AC coupled SH GTI's. OzInverter created Grid. 1300ah 48v.
 
Madness

Guru

Joined: 08/10/2011
Location: Australia
Posts: 2498
Posted: 10:41pm 29 Nov 2016
Copy link to clipboard 
Print this post

My Toroid is virtually the same as your photo with 70mm square welding cable (90 would fit too), also 140 mm fans above and below, air can flow naturally I have had it running above 3 KW for extended periods and the temperatures did not get over 40 without fans. I don't think 60 degrees is anywhere near where there would be a problem. It was 34 degrees ambient when I took those photo's.

PCB is designed for TO247 MOSFETs which don't have flexible legs, the case sits only 3 mm from the PCB.



Here is the case before being fitted out.




There are only 10 types of people in the world: those who understand binary, and those who don't.
 
oztules

Guru

Joined: 26/07/2007
Location: Australia
Posts: 1686
Posted: 11:36pm 29 Nov 2016
Copy link to clipboard 
Print this post

Guarded optimism is creeping in..... over 8kw, and 6kw for extended periods....

You may be onto something now.

I fully expect it to just run now... without problems. Looking like I had a dream run with layouts.... very lucky maybe.


This does not bode well for Tinker I suspect. His different approach may bring him undone like yours did.


...........oztules
Village idiot...or... just another hack out of his depth
 
Madness

Guru

Joined: 08/10/2011
Location: Australia
Posts: 2498
Posted: 01:03am 30 Nov 2016
Copy link to clipboard 
Print this post

Yes Oz you must be Irish or something, although more likely that you are lot smarter than you like to let on. I certainly would have given up a long time ago without knowing your success. All seems so straight forward now I know how to do it, a bit like learning to ride a bike, start off with lots of crashes and then less wobbles until it all goes smoothly.

Now I will have to think about putting a front cover on it.

By the way here is what the old Trace SW4845e puts out that you asked to see.




There are only 10 types of people in the world: those who understand binary, and those who don't.
 
Tinker

Guru

Joined: 07/11/2007
Location: Australia
Posts: 1904
Posted: 02:35am 30 Nov 2016
Copy link to clipboard 
Print this post

  oztules said  
This does not bode well for Tinker I suspect. His different approach may bring him undone like yours did.


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


Well, I have not given up yet oztules .

Am re designing the PCB too (getting the gate resistors/diode closer to the gate pin), they will now be within 2mm. But am still persevering with my different layout, I like to do things differently.

Anybody can copy an existing PCB design and tart it up for better performance......
The challenge is to start from scratch .

Klaus
 
Madness

Guru

Joined: 08/10/2011
Location: Australia
Posts: 2498
Posted: 11:31am 30 Nov 2016
Copy link to clipboard 
Print this post

  Tinker said  
Anybody can copy an existing PCB design and tart it up for better performance......


Guilty! but I am happy now. 22.5 hours connected to the house and it is doing its job without any drama. Edited by Madness 2016-12-01
There are only 10 types of people in the world: those who understand binary, and those who don't.
 
     Page 8 of 22    
Print this page
© JAQ Software 2024