 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 : Inverter building using Wiseguys Power board and the Nano drive board
  Page 38 of 38 |
|||||
| Author | Message | ||||
| Solar Mike Guru  Joined: 08/02/2015 Location: New ZealandPosts: 1250 |
That looks great, big thermal mass to suck the heat out of your mosfets. You could almost get them drilled for water cooling, then all you need is a PC type water cooling pump and small radiator. Mike |
||||
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 2251 |
Thanks, as a matter of interest, this Inverter was running at 1kW for 15 minutes with a thin 1 mm thick test Heat Bar, used for checking hole alignment and final sizing before getting a couple of 10 mm Bars made, and the thin sheet heat bars were still Cold. I no longer monitor Heat Sink over temps with the NANO, just the Toroids, as the HS never gets hot enough to warrant it, even when running above the Board Rated Power Level of 6kW per Power Board, or in my case, at 14kW for short periods in the Dual. NANO:Inverter V 8.4ks - Linux AvrDude GUI script V4.1 |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 2251 |
I tested a 3mm deep hole in the top of the HTB under the FETS, recessed screw through the bottom, and a nut on the top, the nut is pulled into the hole and held in firmly as the corners of the nut bight into the round hole, perfect. The recess hole is smaller than the blank plastic relief in the FET Heat tab, the Nut pull nicely below the surface of the Bar under the FET, clean, no ridges or high spots. Lay the Power Board over each HTB and it drops into place, put an insulating washer on the split LO side Bar screws and a Fibre or Metal washer over the HI side BAR screws, spin the nuts on and tighten, and there is a ton of clearance below the CAP boards, again off-the-shelf M3 nuts and bolts. When tightening the nuts, we are not so worried about over tightening onto a Silicon insulator, this is only true for what I call, "Live Bar mounting" of the FETs, IE the Drain connection to the FETs are connected to either +53v terminal (Long BAR), or TX1, TX2 terminals, Short (split) LO Side Bars, as no small single FET silicon insulators are used. When the bars are tightened down on to the Silicon insulation MAT on the Main Heat sink, the load is spread across a large area, 10mm thick 40mm wide Bars, and I have found that there is no need to reef down on those, it's not a small critical mechanical interface like the Heat tab of a FET, there is a lot of thermal mass protection in this configuration. So tomorrow I take these to the CNC guy and get the changes made for the final Heat Transfer Bars, I hope to include mounting for temperature sensors as well. Price should be way below the effort to do them manually, and it's so dam quick when everything just fits.  : Edited 2026-08-27 15:32 by KeepIS NANO:Inverter V 8.4ks - Linux AvrDude GUI script V4.1 |
||||
Revlac Guru Joined: 31/12/2016 Location: AustraliaPosts: 1303 |
Looking Very good  , would love to convert the old 3Phase Milling Machine and Lathe to CNC and do things like this, need a lot of power to run them though.  Cheers Aaron Off The Grid |
||||
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1306 |
Are you going into production ? They look just like bought ones - you do a neat job of assembly, and I do agree with making life easier and getting some tasks out-sourced. It especially makes sense when you save time and increase the precision and remove any chance of making a mistake on the last hole. The wider heatsink bars that solved the hidden middle mounting screw on the long bars was a good solution. A lot less screws to be removed leaving the Fet/bar complete. Do you have a Hakko hand held de-soldering tool? Whenever I am soldering larger pins ie HY5608 I use a hand soldering iron to first solder the pin with a slightly generous amount of solder on the solder side (dont go too crazy with solder) if you inspect the pin after soldering it almost never wicks to the component side of the board. If you heat up the Hakko hand desoldering tool to around 400 degrees and place it over the pin just soldered within 1-2 seconds, the solder flow to the other side of the board in a very neat joint. I also lift the tip back a mm or so and re touch the pad to help "pump" the solder through the joint. The pumping part is probably not required but it feels like youre helping the process...No extra flux or solder required it works with most larger pins and really helps with 2oz board soldering and all 16 FET joints are re-done in 1-2 minutes. The pumping I refer to is definitely not the desoldering pump or you'll have to start all over again. Edited 2026-08-28 12:54 by wiseguy If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 2251 |
Thanks, I believe the machine they use for these has a 2.6kW 3P motor with VFD, tiny cutting bits though. NANO:Inverter V 8.4ks - Linux AvrDude GUI script V4.1 |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 2251 |
Once he has layouts for both boards set, I believe he can make as many of them as needed, so once I'm happy with final pieces, I'll post the price here, if anyone is interested I can pass the contact details on to them. Yes, I do have a Hakko 474 Desoldering Station, and, as you previously mentioned, I had little trouble removing one of the small Gate Drive PCBs from the Test Power Board  I've actually had little trouble soldering the Power board, and for some reason the FETS went quickly and smoothly. I wondered if anyone would notice the wider bars with the outside mounting, I had them drill the same hole on the split bars, and make a 3 mil wide cut through the hole, so only one screw for the centre of both full and split bars, obviously a large insulated fibre washer needed that spans the 3 mm gap across the split Bars. That middle hold down was what I wanted to solve with the wider Bars, I've had the end holes offset slightly inbound, to allow even clamping force, as I've done this before, I know that it does not need much force for excellent heat transfer with the surface area of the two Heat Bars and silicon mat. I'm currently trying a simple way to make an isolated mount in the Heat sink itself, this removes the need for precision, and the need for any special mounting hardware to isolate the Heat bar hold down screws for Live Heat Bar mounting, and if it works, it would also make mounting isolated Heat Bars really simple as well. I should know how it works out later today, initial test was really encouraging. EDIT: I don't think I quite explained the term "Live Heat Bars" to new builders, it means that the Long aluminium Heat Bar is connected to +53V, and the Split Heat Bars are connected to TX1 and TX2, the Toroid / Choke terminals, and all via the Drain (Heat Tabs) of the FETs when no individual FET silicon insulators are used. : Edited 2026-08-28 15:19 by KeepIS NANO:Inverter V 8.4ks - Linux AvrDude GUI script V4.1 |
||||
| wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1306 |
Yes I note the tripod screw approach for the heatsinks and the split bar solution. A philosophical comment, I have never advocated the live heatsinks, gut feel and experience wont let me do it. The 4 small bead chokes are now all effectively in parallel along with the FET drains. The ringing of the FET drains was solved by isolating each Drain from the outside world with a high impedance choke, by using the live heatsink approach we effectively have one humongous FET with 1/4 of the initial impedance/inductance on its combined drain. I am not saying it wont work or that it will/wont cause issues but it is entirely new untested territory for me, especially with the HY5608s. Earlier when I was using HY4008s ringing was still evident with the drains all combined on an unisolated heatsink so I abandoned doing it. My contention is that the 4 x FETs run very cool on the heatsink bars even with a silpad washer so I will probably continue to isolate my FET drains from the bars. If you have already exhaustively tested the live heatsinks approach and resulting emi and circuit noise etc that is reassuring, but I felt the need to clarify what it means. Of course as a spin-off the split bar is then not needed for the lower FETs on the negative bus and then you may not need isolation for the 8 FET screws as they may be the 8 earthing points you use from the negative PCB traces to the chassis (not for high current path use!). Edited 2026-08-28 16:56 by wiseguy If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| nickskethisniks Guru Joined: 17/10/2017 Location: BelgiumPosts: 483 |
Allthough they are (a lot) more expensive, I like to use some "aluminium nitrate ceramic pads" under the mosfets instead of the silca pads. With the pads you need to make sure there is no debris that could puncture and ruin the isolation. |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 2251 |
Hi Mike, we have previously discussed the issue with the small beads being effectively shunted on the High Side when the Drains are paralleled directly to the Heat Sink, but good idea to bring that up so others can make an informed decision. Yes, I've tried both methods of FET mounting and I've never really noticed a difference on the Scope during testing, as I have two identical New power boards, it will be easy to do a retest between FET HI side isolation and paralleled mounting, paying specific attention to ringing under various load conditions in the HI side FETs. I doubt I'll notice any change in THD, but I'll still check this as well. BTW, I tested that relatively simple way I found to mount the HTB to a BIG heat sink without the hassle of breaking a thread taping bit in the Heat Sink, possibly screwing the Heat Sink as the broken cutter can be almost impossible to get out of a heavy base plate without drilling around the sheared off cutter, fortunately, if that happens, this method will save the Heat Sink providing the broken cuter was not embedded in a FET mounting thread, which it won't be if using Heat Transfer Bars. Some of these old very heavy Heat Sinks with thick base plates that I've removed from old Grid-Tie Inverters feel almost as tough as steel. : Edited 2026-08-29 08:34 by KeepIS NANO:Inverter V 8.4ks - Linux AvrDude GUI script V4.1 |
||||
| analog8484 Senior Member  Joined: 11/11/2021 Location: United StatesPosts: 215 |
Nice bars ... it's silly but I love the round corners. Reminds me of the aesthetics of the nice DIY amplifier builds I've seen. |
||||
| analog8484 Senior Member Joined: 11/11/2021 Location: United StatesPosts: 215 |
They look impressive on paper. Definitely much more expensive but still cheaper/simpler than liquid cooling? Do you also use soft silicone pads/grease with them (since they are hard)? What kind of thermal improvement have you seen? I wonder if the KeepIS 6kW inverter build can become a 9kW+ inverter using them. |
||||
| analog8484 Senior Member Joined: 11/11/2021 Location: United StatesPosts: 215 |
Hmm ... so, the beads might not be necessary? |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 2251 |
These are still prototype, initially I was going to do more than just rounded corners, but due to the Thickness and the time, the cost would increase for no physical gain, but yes, rounded corners etc will be on the final units. You mean the Dual, which is consertive @ 12kW, especially as it has run up to 15kW. The previous 6kW build was a Single Power board with FETS mounted directly to the Heat Bars. The Power Board limitation was not the FETs or FET heat, it was the 2oz double copper sided PCB, and how much heat you are willing to accept in the tracks and PCB power connections. If you had forced air cooling across the PCB and CAPS then yes, it would easily do 8kW for some time, which I have tested. Not for the survival of the Inverter as I ran without any HI side Beads on the first two builds, but that does not mean that it ran without a little unwanted ringing and more radiated noise, so the Ferrite Beads would likely improve that. I have the two power boards almost ready to compare, but even then, I may not hit the conditions where the beads actually help, wiseguy found they did and I do not doubt that at all. At the very least, I proved that the Inverter could run way past it's design parameters and not fail without them. : Edited 2026-08-30 10:32 by KeepIS NANO:Inverter V 8.4ks - Linux AvrDude GUI script V4.1 |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 2251 |
Saved a heat sink with holes in the wrong place, right where I wanted to mount the rails, and a damaged mounting hole that had to be drilled out. I found this is so quick to do, literally takes a few seconds to drill 11mm holes with a small Drill Press, drill press preferred so that holes are 90° to the surface. The plugs are a form of Polycarbonate rod, had them laying around for years. They were driven in with a hammer and hardwood block  you could also use a big vise and press them in, finish with them slightly below the surface.For initial testing I didn't glue them, I tried to pull one out by tapping an M4 thread and a fitting to act as a stud puller, M4 screw tightened to where the screw driver was stripping the Head of the screw, but it would not budge, left overnight - no movement. Now it's so easy to tap with those cheep Chinese combined Drill and Tap bits meant for an Impact driver, takes about 6 seconds for an M4 tap. If you somehow manage to break a bit, it only takes seconds to drive the insert out from the the FIN side and pop another in. So I rescued a couple of old heat sinks (getting scarce now), and as a plus, if you decide to use FETs mounted directly to the Heat Bars, now there is no need to use insulated hold down screws as the mounting screws are already insulated by these inserts. Which is where the idea initially started, while staring at a damaged Heat Sink, I was about to bin it for another project but decided to try this instead. BTW the Power boards must be mounted in exactly this position on the Heat Sinks to fit into the existing Dual Inverter Housing.     NANO:Inverter V 8.4ks - Linux AvrDude GUI script V4.1 |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 2251 |
I've been testing two Power boards, one board with the HI Side FET beads effectively paralleled by the FETs by being mounted directly onto the Heat Bar, the other board with all FETs separated with individual FET Insulating pads. Mike (WG) is correct, the results are definitely in favour of keeping the Ferrite beads in operation. Under some load conditions, there can be a few cycles of ringing in the waveforms on the trailing transitions, and there was a very real reduction in that ringing under various load conditions with the Ferrite beads operating correctly. When I first tried ferrite beads a few years back, they did not reveal the results that I've noticed today, yes, there was a slight ringing reduction but that appeared to be overshadowed by the inductive kick of the beads, however, these beads are ones that I have not used before, they are slightly smaller that the originals I used, and I was looking at the same waveforms locations back then as well, this could have had an influence on the results. In setting this up, I found that I really like having fixed FET retaining screws in the Heat Transfer Bars, it makes it so easy to drop all the FET insulators on in a few seconds without any of them moving when the board is placed over the screws and gently lowered onto the silicon pads, this keeps them perfectly aligned even without a smear of heat transfer compound. : NANO:Inverter V 8.4ks - Linux AvrDude GUI script V4.1 |
||||
| Page 38 of 38 |
|||||
 Print this page |
 |
The Back Shed's forum code is written, and hosted, in Australia. | © JAQ Software 2026 |
