 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 6 of 22   |
|||||
| Author | Message | ||||
| naiclub Newbie  Joined: 07/09/2016 Location: ThailandPosts: 6 |
  |
||||
oztules Guru  Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Looks very nice Naiclub... but not sure of the holes on the positive side of the caps, and the neg ones seem to be not connected to the bottom neg layer. You also don't want conductivity on the heat sink holes on some layers... ie bottom layer to heat sink is ok for the pos one, but not the top layer, and opposite for the output heat sinks. Mad... your lack of heat speaks volumes about your unit.... still cant see how you kill it so easily..... with that temp gradient, it sounds like the waveforms for switching are tickety boo...should be bullet proof from there. I am just not understanding the failures  .... one step at a time I guess...........oztules Village idiot...or... just another hack out of his depth |
||||
Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
Dee mak mak Kuhn Niaclub. 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: AustraliaPosts: 2498 |
I had the Inverter running all day at idle with just a fan heater running off it on fan setting. Just put another 4 FETs on the board, so have 12 now and it runs just the same, very clean sine wave at idle, with fan heater on full just that little blip at zero crossing. With this load, the FETs run about 5 degrees above ambient, still with no fan on the heatsink. 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: AustraliaPosts: 1904 |
Good news Madness. A question: where did you buy your EG1810 LQFP-32 chips from? I tried ebay and aliexpress but no items found there. I did order that DIL converter board from aliexpress. Building my own drivers now, sick of the EGS002 boards malfunctioning. Klaus |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
Hi Klaus, Last ones I got where from here Was that typo or is it different on the 002 boards as it is a EG8010. 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: AustraliaPosts: 1686 |
step by step... hopefully the kapton was the problem..... fingers crossed. Tinker, they are EG8010... try that, I bought these last time, a few running from that batch no problems. https://www.aliexpress.com/item/20PCS-LOT-EG8010-100-NEW-EG-Pure-sine-wave-inverter-professional-chip-32-LQFP-EG8010/325 60837636.html?spm=2114.13010608.0.0.Wv9vML I blew a few 002 boards up initially ( stemmed from the power supply part really), but once running and the 120r and 100uf were incorporated, they proved to be very very tough, still have a few left, only went to the 8010 because I wanted some autonomy... but the 002 are very stable and fine, so I am wondering about your open layout. Maybe try a smaller tranny and less fets for a start until you can pinpoint the problem.... I have run out of ideas, as I am not there. Mad now has a normal board, so he has no excuses in that regard.... but looks like he may have cracked it, and maybe he knows why. ...........oztules Village idiot...or... just another hack out of his depth |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
Looks like I beat you to the punch OZ  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: AustraliaPosts: 2498 |
Now have 16 MOSFETs on board, idle was really clean, but since adding the last 4 there is just a slight distortion on idle. I am not sure if I should change the last 4 FETs to if there is any change. I am kind of over hearing them go bang.  And here it is with a load.  The blip is slightly larger than before but the other (slight) distortion seems to have disappeared.  It is a mess on the bench ATM but it is working.  That Toroid is one I knocked together quickly a while ago for testing, turns are not quite right as I just wound 22 turns over the original primary.  MM there is connected to 75MV 100A shunt, so input current is .8A DC voltage is 49.8V. So less than 40 watts with a fan running off of it. 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: AustraliaPosts: 2498 |
I have been thinking about the dead time, does it need to be a little longer to stop that blip? DT1, DT0: Dead time setup “00”: 300ns “10”: 1us “01”: 500ns“11”: 1.5us Currently, they are both grounded so running at 300ns, have you experimented with this Oz? 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: AustraliaPosts: 1904 |
The dead time will be difficult to change on oztules board, I think I will provide for the solder bridge selection like on the 002 board. Ditto for the unipolar/bipolar selection once somebody explains to me what the difference is and how it affects the inverter. This project provides plenty of opportunity to get clued up how these inverters work, I appreciate all the input here. Klaus |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
With a Dremel or some editing before etching would be easy to modify Oztules board. I have held off etching anymore til I see the version with current overload protection when that is released. Just thought of a simple way to do it for experimenting, just flip those two legs up on the adaptor board and connect to them there. I will have a play with it tomorrow night on my little test power board. There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
| naiclub Newbie Joined: 07/09/2016 Location: ThailandPosts: 6 |
Thank you for your advice make it correct |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Mad.... there is noise or something in there causing the sine wobble with 4 fets... not sure from where. Perhaps see if a shorter ribbon helps, or dream up other changes. You have it right where you want it... ie you can effect change now to the shape ( adding fets), and when you can effect change, you can learn.... it is the only time you can really... My waves are very good, so am loathe to change anything at this stage... it alll just works Unipolar and bipolar are two ways of doing the spwm. I am using the unipolar. This is where we run one side of the H bridge with a fundamental square wave of 50hz ( a full 1/100th second each time top and bottom) and modulate the oposite side of the bridge at the same time with 20khz spwm wave.... ie we leave say the bottom of the bridge turned on for the full 100th of a second, and fiddle the top half of the bridge with the sine wave info. Bi polar is where we use the spwm information on both the top of the bridge and the bottom of the bridge at the same time... ie we are switching the top half with 20khz spwm, and the bottom half of the bridge with 20khz spwm at the exact same time to drive current into the primary. I expected the bipolar to be much quieter than the unipolar, as the switching is all above hearing range, but in this 8010, it was noisier. In the PJ it is quieter as a rule. Switching losses should be higher with bipolar. It is also why one set of heat sinks run hotter than the other two. In the current design, both the 002 and my 8010 switch spwm to both high and low side alternatively. I was going to do all the spwm on the two low sides, but the temp of each says that 20khz is easy for the high side to do, as it is usually the same temp as the low side spwm bridge. Both high and low side spwm are hotter than the high and low side fundamental wave side of the bridges. Have not changed the dead time, as works well anyway... however, easy enough to change on the 8010 chip board itself.. then plug it in. Mad, had a chance to test the current scr o/load yesterday but only at 4kw... ran out of time. Seems to work, but with the pot values and CT of 2000t, it is too difficult to set easily, so need to change a few values there before I can test definitively. May even be easier to monitor the battery transformer primary current with a few turns around the inductor e core instead.... this will find any saturation problems from over long DC switching signals as well perhaps. ..........oztules Village idiot...or... just another hack out of his depth |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
Hi Oz, I have a short ribbon cable that I have tried previously with a different board also I have made up a cable with 4 ground wires twisted with the 4 drive wires. As yet I have not tried it as I was getting good idle results until the last lot of FETs went in. Have you tried different dead time? I have been getting that blip under load from the outset, I see you had it also in the early days of testing your boards, did you figure out any cause of it? The Inverter has been running all night at idle, almost 24 hours now, that is a very significant KMstone for me. I will probe around the FETs with the CRO to see if I can narrow down the source of that wobble. 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: AustraliaPosts: 1686 |
No deadtime experiments here... always the same. I never really found out why it has changed ( blips on cross over). Suspect board layout and transformer and inductor changes. Time to bite the bullet and drive it hard???? ......oztules Village idiot...or... just another hack out of his depth |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
Hi Oz, ATM I am limited by 60 amp breaker and test toroid. Could easily change that if I wanted. But I would like to see it running nice and clean before hammering it. Just reheated solder joints on the last FETs and scrubbed the board, no change and nothing visible on CRO. Will try changing ribbon cable this afternoon and reconfigure it. Might even put it in the case with big toroid to see how it behaves then. I still reach around the corner to turn it on, have not been hit by shrapnel yet but wary of it. 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: AustraliaPosts: 1686 |
Mad... encouragement pics. Was testing the current limiting abilities of the board with the scr jury rigged into it. Must have restarted it 100 times under loads etc finding out how it functioned. It functions very smoothly, and is now settable quite nicely from about 8 amps to over 60 amps on the 240v line.... these pics are with 5hp lathe motor starts. The meter is not super fast, but after about 20 starts and stops, it came in every time around the 57-59 amp mark... which is surprisingly consistent.. anyway it starts it fine. Here is the meter on the 240v output 40 amp wire:  And here is the rats nest doing the 14-15kw of start ups.  I was concerned that stopping it mid stride with the scr may lead to spikes... but no.....it cuts it off perfectly smoothly. No sign of distress. ........oztules Village idiot...or... just another hack out of his depth |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Oztules, are you using the SCR in a similar way I posted in the latest schematic in 'building an inverter from scratch'? And are you using the SD line to switch off too? While my little test toroid was working I found that the current to switch off was easily adjustable with my 2K/10 turn trimpot. I really liked that feature. If you have a scrapped Latronics inverter, they use a neat current transformer with a split winding. Just 1/2 winding gives enough signal to work the SCR. That part is easy to salvage. Part # is 57P1805A from Vitec Electronics Corporation. Thank you also for the unipolar/bipolar lesson, I will try to digest the info. But as you say unipolar is quieter and more efficient with the IR2110 I will set it to that. Klaus |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
Hi Oz, I look forward to the new control board with overload protection, sorry if it sounded like I was prodding you. When you have time is okay with me. There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
| Page 6 of 22 |
|||||
 Print this page |
