 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 5 of 22   |
|||||
| Author | Message | ||||
Madness Guru  Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
Cutting the holes was not that hard actually, I used a 1mm cutting disc and cut a few mm inside the hole I needed, as I worked around it cut slightly larger. Then used normal grinding disc to tidy it up. I think 1 KV caps might have been sufficient but the 3 KV where readily available. 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 |
The Inverter is now on the wall, I figured it will be far easier to assemble it from here that way as there is no way I could lift it when finished.  The major parts are now in place, top right is 3 filters from output of Aero-sharps. There is a perspex panel to go on the top left that will have screen and buttons for Arduino based monitoring system I am working on as well. It will control fans as well as data logging temperatures, current and voltages, this will be accessible via ethernet and display current information on a 3.5" TFT LCD with touch screen. Big problem I have though is bad wave-form as shown below, something wrong the timing, have tried 3 different power board layouts, get the same result with all of them.I am going to try ordering components locally for the control board to see if that rectifies this. Images are with no load, adding up to 2.4 makes the wave blurred and lots os of squiggles in it, the toroid and choke make lots of horrible noises when this happens also.   These are from the the 2 sides of the secondary output. 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 |
If all else fails Aliexpress is having a sale and this 7500VA Inverter looks tempting. A trick to save a bit of money on Aliexpress that I discovered also is to add an item to your wishlist or basket and leave it there and they quite often discount it to tempt you to buy. There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
| Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
I did purchase one of these when they were on sale couple of weeks ago for my other inverter and it arrived a few days ago, complete with 12 bolts and big copper terminals for the connections and a 5uf mains cap. It looks quite well designed, note the 4 groups of Tip 41/42 complementary drive buffer transistors on each leg of the H bridge, this is a very good idea as it isolates the drive chips and keeps the high current drive signals connections very short and the same length. Main mosfets are HY4008. The SPWM drive card seems a bit different from the usual ones on sale in that it has a remote phase detection input; but it came with a note not to use it ?. Main supply caps are 5 x 4700uf 63 volt, may need an upgrade here. Will let you know how it performs when I fire it up. Mike |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
Thank you for that report Mike, being able to use it as a UPS was one reason I was considering it. I think I will rule it out and keep persevering. Thanks to Tinkers tip of connecting the CRO across the AC out (did not do that due to my inexperience) I am getting a on a small test board with 4 FETs. Also, I get the same with the full-size power board with just 4 FETs. However, fully populating the board is a failure. Works with no load, although Sine wave has some distortion, running at 2400 watts causes the sine wave to be very distorted. Yet with just 4 FETs it will run the same load happily and the heat sink only rises 2 degrees above ambient. Any suggestions welcome, what I am planning to do next is add FETs 4 at a time and then test with the same load to see when it starts to get pear shaped. Not sure if it could be some of the FETs are faulty, they all test fine at around 2 M ohm with the meter. Also, I wonder if it needs the complementary drive buffer as mentioned in the previous post. There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
| Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
Mad you have made some progress, you have discovered that the waveforms and power dissipation seem ok when just 4 fets are in the H Bridge, then get worse when fully populated. If you are using the IRFB4110 fets then each has a gate cap of 9620pf multiplied by the number of fets in cct, perhaps your driver chips are running out of grunt to switch the mosfets cleanly when more are populated, or there is a higher resistance or bad connection in the ribbon cable between the control and power board. I would put your scope probe between 0v and each of the mosfet gates to check that the gate drive signals are clean sharp steep sided transitions with no oscillations. Perhaps swap out the driver chips, one may be damaged from previous issues. Personally I dont like using ribbon cables for these sort of drive circuits, short duration high current pulses are required switching the gates, ribbon cable and associated crimp connections isnt exactly the ideal here, preferably use twisted pairs of thicker wire for each the gate drives. Mike |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
Yes making progress, 3 steps forward and 2 steps back  but it is progress.I am using IRFP4110 which is TO247 version of the IRFB4110, as far as I am aware it the same silicon in a heavier case and much more substantial legs. Capacitance is identical, also these FETS end up with very short legs with just 4mm between the case and PCB. Wether this lower resistance here is a factor I don't know, power boards using the TO220 version would have at least 10mm of much skinnier legs As OZ has it working without any of these issues I am scratching my head, I will replace the ribbon cable with some heavier wire with ground wire twisted on each gate drive and try it. I have looked at FET gates with the scope and everything looked clean and identical to the output of the of the IR2110's. I will be looking closely at it as I build the new PCB adding 4 FETs at a time and testing at each stage. Also, I am planning to test each FET individually with power on it and check them as per this video and group them. This will also help to eliminate any that may be faulty. There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
Grogster Admin Group  Joined: 31/12/2012 Location: New ZealandPosts: 9308 |
That's an interesting 7kW5 unit - not a bad price at all either! Hmmmmmmmm...... Smoke makes things work. When the smoke gets out, it stops! |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
If you are planning to build a single core Torroid this Inverter is very good value. There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
| Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
It is good value, I would be carefull though, notice how the gates of the H Bridge mosfets are all joined together for each quadrant, with a single drive resistor\\diode etc powering them, this goes against recommended best practice, just a little bit dubious on that board design. Have a read of this documment it explains best design technique in this area and why each paralled mosfet should have its own gate resistor. AN-978 Application Note may also help with fault finding. This other note is worth reading also AN-941 Mike |
||||
Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Madness, I had a good look at the PCB at the above link and noticed they have paralleled the gates of the 6 Mosfets for each bridge leg and are using just a single resistor//diode network for each. I'm using one for *each* mosfet and wonder why they are doing it different. Also, close inspection of the EGS002 board they have plugged in shows there is a black wire coming from the IFB input to the EG1810 (I traced it) and goes to the motherboard. I wonder what that wire does. Any ideas anybody? Klaus |
||||
oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
I bought a few of those to test out. They are very good for the price, and perform well. The common gates surprised me, but they work fine, and pushed one way past 5kw.... way past. BUT... I pulled the mosfets off the heat sinks, removed the sinks, and cut the output heat sink in half with a gap, and tapped the heat sinks for input and output wires....and replaced. The system they have is too puny..... it now looks like solar mikes. Well worth a play. I have not used one in a long full house testing, as they were a curiosity really, but I feel/know they would handle it just fine. I did open up the current settings to run a big mig welder, and it was fine until I went too far ( 230 amps settings) ...... but that was very much over running it too. ..........oztules Village idiot...or... just another hack out of his depth |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
Thanks for those PDF's Mike, good information there, quite a lot to take in. If you look at the feedback from buyers of that inverter board there is one there from Oztules talking abut modifying the EGS002. Perhaps he will chime in regarding that black wire. 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 |
That video has terrible camera works, the guy never heard about tripods. I gave up watching it and looked at the next one on the list which was much better. I tested each of the new batch Mosfets recently, using 5A drain/source current and measuring the voltage drop across the drain/source pins. I used the same gate drive parts as in my inverter driver board and 12V to turn it fully on. I would have liked to use 10A to test them but I managed to blow my old big power supply with a dickie series resistor. They were mostly 15m Ohm, about twice the spec so I guess I got what I paid for. However, when measuring some Mosfets I scrounged from a junked Latronics inverter their Mosfets were in the 20m Ohm+ range. Klaus |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Will look tomorrow Mad. .........oztules Edit... yep goes to the ifb on the 002 from the lm358 currents sensing circuit on the board... ie goes to pin 1 of lm358... output pin of current sensing op amp Village idiot...or... just another hack out of his depth |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
This document from IR regarding paralleling MOSFETs may be helpful to someone also. 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 |
Madness, perhaps I found a clue to your sine wave distortion problem. Did you try to run the inverter without that big ferrite choke in the primary circuit? I saw that my small test inverter had a badly distorted 240V wave form when I used a ferrite choke I had just quickly wound for my testing. No distortion with no choke at all nor with my big 60mm sq wire one. Klaus |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
Hi Klaus, Thanks for the input, I actually have a very nice clean Sine Wave now thanks to your tip regarding connecting the CRO. However still getting exploding FETs, had 4 FETs on the PCB today, ran it up to 2.2 KW for about 5 minutes running a fan heater, there was just a very tiny distortion at Zero crossing, similar to OZtules early results. Then I turned the heat off and just left the fan running, after about 45 minutes BANG. I need a blast shield for this #%^!@& thing. It ran perfectly for that 45 minutes, just a little buzz in the transformer, which I get in all 3 different transformers I have been using. Postmortem I found 1 Fet with a large crater, 1 with a big crack and other 2 shorted to the gate. Chinese IF2110 that makes the SPWM was also dead. I have replaced the 2 IF2110's with genuine ones I had on hand and replaced the FETs. I have had it running for about 30 minutes with 1200w load on 4 FETs since repairing it. Now going to try with 8 FETs. 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 |
Interesting to note Klouse.. will look at this tomorrow if I get time. You two are having a hard time of this, and I don't know how to advise you, as I cant begin to guess what is different here  I guess you can always fall back on the power jack driver board... but there is something wrong in there somewhere that is iffy, that I have not experienced. ..........oztules Village idiot...or... just another hack out of his depth |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
Just come back in from another session of working on the Inverter. I now have 8 FETs running perfect sine wave at idle, just the tiniest little blip at zero crossing running a fan heater at 2200 watts(would much rather it was an AC with this stinking hot weather) FETs are rising about 2 degrees above the heatsink after running about 30 minutes and heatsink barely above starting temperature. This is with no fans just got it standing on it's end so it can draw air through naturally. I found what I think caused the last explosion, the edge of the Kapton tape I put on the heatsink bottom was pinched under the bottom of one of the MOSFETs. It was marked by it, so was definitely keeping it from making good contact. Tomorrow I will add another 4 and see if it keeps behaving itself. If anyone wants it I have attached to the PDF's to etch the power board, I used 300 X 200 MM 2oz blank PCB from here. 2016-11-13_130157_300X200_PCB_Top.pdf 2016-11-13_130217_300X200_PCB_Bottom.pdf 2016-11-13_130231_300X200_PCB_Silk.pdf If you plan to use these you will need to use A3 to print them as it is 3 MM longer than A4. BTW The extra 4 holes near the safety caps are there to fit MOV's for lightning protection. There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
| Page 5 of 22 |
|||||
 Print this page |
