 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 : TO-247 MOSFETs for a 48V system
| Author | Message | ||||
Dougie Analog Newbie  Joined: 21/02/2025 Location: AustraliaPosts: 14 |
Let's say it's a given that TO-247 is a best fit for their heat dissipation ability compared to TO-220, then is this a fair list of MOSFETs in order of preference for a 48V system? HY4008W 80V, 200A, 2.9mΩ, 195nC, $2.00. Good: Amps - RDSon - Price. Fair: Voltage - Total Gate Charge. Unknown: uniformity compared to the big brands. IRF100P219 100V, 203A, 1.7mΩ, 168nC. $5.63. Good: Voltage - Amps - RDSon. Fair: Total Gate Charge. Not so good: Price. IRFP4110 100V, 180A, 3.7mΩ, 150nC, $4.40. Good: Voltage, Amps, Total Gate Charge. Fair: RDSon. Not so good: Price. HY5608W 80V, 360A, 1.5mΩ, 365nC, $3.90. Good: Amps - RDSon. Fair: Voltage. Not so good: Total Gate Charge - Price. IRFP4468 100V, 195A, 2mΩ, 360nC, $6.23. Good: Voltage - Amps - RDSon. Not so good: Total Gate Charge - Price. IRFP4368 75V, 350A, 1.46mΩ, 380nC, $8.30. Good: Amps - RDSon. Not so good: Voltage - Total Gate Charge - Price. IRFP3077 75V, 200A, 2.8mΩ, 160nC, $6.53. Good: Amps - RDSon - Total Gate Charge. Not so good: Voltage - Price. Other MOSFETs that I know nothing about but the specs and prices look good. SP010N02BGHTF 100V, 235A, 2mΩ, 158nC, $2.30 SP010N03BGHTF 100V, 210A, 3.2mΩ, 100nC, $2.00 GT023N10Q 100V, 226A, 2.1mΩ, 121nC GT016N10Q 100V, 288A, 1.65mΩ, 165nC, GT035N10Q 100V, 190A, 2.5mΩ, 68nC GT045N10Q 100V, 208A, 3.8mΩ, 78nC Are there any others that are contenders for the list? |
||||
poida Guru  Joined: 02/02/2017 Location: AustraliaPosts: 1431 |
the larger gate charge might not be such a problem if the gate drive is strong. e.g. using the not very strong 1.4 Amp IR2184 then gate charge needs to be looked at but if you use some PHAT drive design that can do 4 or 6 or more Amps and switch fast then Gate C is now no problem. I have shown to myself that HY4008 with the large gate charge, driven by TIP41/42 totem pole transistors is good enough. The last inverter build I did is still running after about 5 years. The TIP41/42 are not very good. low current output from them... Let's say the inverter is on for 3/4 of that time. 20KHz switching. 5 x 356 x 24 x 3600 x 3/4 seconds x 20KHz switches applied to the HY4008 Gates. This amount should show a weakness, you think? this is 2.4E12 switch events. My inverter takes the output of ONE TIP41 or TIP42 to drive 4 HY4008 gates it's still running fine. Wiseguy's inverter design is superb. It uses a very stout totem pole gate drive design that employs some very hard hitting FETs to drive about 6 Amps into the power FETs. Some suppliers of HY4008 give fake or poorly performing devices. LCSC is a good source. wronger than a phone book full of wrong phone numbers |
||||
Murphy's friend Guru Joined: 04/10/2019 Location: AustraliaPosts: 671 |
I used to run HY4008's in my inverters, have managed to blow some up. Have changed them to HY5608's, no more blow ups so far. Since one requires fewer HY5608's when swapping them for HY4008's, their higher cost is not so bad. I used 16 HY4008's in a 6KW inverter, now use 12 HY5608 in the same inverter. The much lower RDS on of these allow use of a single heat sink with isolated Mosfets on it, making a much simpler design and external heat sink fins possible. I would recommend Wiseguy's driver, it's miles ahead of the early TIP41/42 drivers forumites here used some years ago. |
||||
| flyingfishfinger Senior Member  Joined: 12/09/2020 Location: United StatesPosts: 117 |
I'm using the higher-voltage FDH055N15A, 150V, 167A, 5.9mOhm, 92nC, at a price. This is still for a 48V system, but my array outputs up to 130V. I've blown up one (in the diode-configuration, no clue why) but otherwise I've barely loaded them and they seem to work fine. For the voltage and current, there aren't that many options for lower RDs. R |
||||
| Dougie Analog Newbie Joined: 21/02/2025 Location: AustraliaPosts: 14 |
Yeah flyingfishfinger, higher voltage for charge controllers is a challenge I've yet to tackle. I can see lots of inspiration for that here. In the back of my head, I'm thinking that I'd use IGBTs but I haven't gone down that rabbit hole yet. Going up in voltage (VDS) I can see the other specs will mostly be less good. Question is, how close to the wire do you go, how much voltage head room is reasonable? The enormous specs of the HY5608 are hard to ignore and maybe it needs to be at the top of the list. I was concerned about the 365nC and the implications for the driver, dead time, system complexity etc. Clearly Murphy's friend, you and others have made them work well. I'll get some and play with them and hopefully learn something. poida, I liked what you said about about TIP41/42 and yet they just work... Cheers da |
||||
| nickskethisniks Guru Joined: 17/10/2017 Location: BelgiumPosts: 462 |
You also have the hy5110 to add to your list if you want to have more headroom. |
||||
| Dougie Analog Newbie Joined: 21/02/2025 Location: AustraliaPosts: 14 |
I've been wading through treads starting at 2017, taking notes and refining my ideas and designs. Because the forum works in the way that they do, old threads get bought to the top of the list by the addition of a new post. As a result, there is a lot of back and forth chronologically speaking, while wading through to the more recent threads. Super interesting learning about the evolution of ideas and which ideas get traction and the various rabbit holes and dead ends. So my process of methodically retracing threads from back to front has meant that I'm only now getting to the latest thinking about how the world should look from an off grid perspective. My reading has now taken me to March last year, which brings me to wiseguy's nano R6 thread. I'm at page 6 and suddenly I feel like I've been living under a rock. Brilliant work! I'm looking forward to reading KeepIS's thread soon (25 pages!!!!). Anyhow, the point to this long preamble is that the questions that I raise may well have been covered ad nauseam in threads that I'm yet to read. I'm going to ask them anyway ;) Looks like the venerable HY5608W is the reining king at least for 48V inverters. It also looks like the high gate charge is a non issue with the Warp inspired wiseguy's buffered inverse opto drive. Murphy's friend is using 3 x HY5608W to replace 4 x HY4008W. The HY5608W story is compelling. However at the risk of flogging a dead horse, I'd like to tease out a couple of issues. Correct me if I don't have the right handle on this but the early FET blow-up issues could be mostly laid at the feet of EGS002 pre mod issues, poor choke choice and the lack of a BJT driver buffer and to a lesser extent the lack of or poorly matched secondary winding capacitor. There was some conjecture about the poorly spec'd TIP41/42 buffer. Even though they are old parts, they may by virtue of their lack of performance be masking the worst effects of shoot through and still functioning well enough to power 4 gates. If the above accounts for 98% of the blow-up issues then in keeping with the KISS principle, bare bones, simple to make and maintain inverter, why is the modified EGS002, non isolated, TIP41/42, sendust and HY4008W combo a less than good choice for the DIY off-gridder? The other issues are: Using more HY4008 FETs means more drain and source pins to carry the current and more surface area to dump heat into the heat sink, although this might be a moot point because it has a higher RDSon. Will the non isolated, TIP41/42, 0V to 12V buffer cut it to drive 3 x HY5608W FETs with their rather large gate charge? Has anybody done the modified EGS002, non isolated, TIP41/42, sendust and HY5608W combo? I'd like to try the above with 2 x half bridges of 100 x 100mm PCBs, 6 FETs on each board. If this works well enough then stocking a collection of spare boards could see a failure restored in minutes by changing out the offending bolt on board. Spare plug in or bolt on boards would be: FET / FET driver generic half bridge, modified EGS002, AC feedback and capacitor boards. I can see the utility of the simple approach for people who truly are off grid, may have rudimentary skills and need access to power to repair and maintain any damaged boards and the system. Recapping tonight's stories... Does the HY5608W rise to the top of the list? Can the HY5608W work well enough in a bare bones EGS002, non isolated, buffered system? that's all she wrote... cheers da |
||||
KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1816 |
Just a few random thoughts I had in the little time I have ATM. IMHO, it's a bit more complex, one of the major factors is how you build the Inverter, especially Cable sizing selection, inverter layout and resulting cable lengths, connection integrity, transformer size and winding ratios, types of loads, and AC cable lengths and loss between the inverter and bigger loads. Yes early Chokes with less than ideal Inverter construction would make early inverters susceptible to FET death, IE very little margin for error. The Secondary CAP were mostly in the ballpark, but a lot of emphasis was placed on Toroidal resonance, I basically shot that down when I went down a rabbit hole on that subject. Another realization was the magnitude of DC input current peaks with seemingly tiny loads, one or two 12v plug packs can randomly generate Peak input currents well over 300A at power on, with a small inverter and low margin for error (build, components, etc) it is not surprising to see them go bang for seemingly little loads. You are right about Wiseguy and his Inverter design evolution, he and Poida have carried out some brilliant time consuming and thought inspiring testing and waveform / data generation posts. This is how my Dual inverter build was constructed, Power board swap measured in minutes, same for the controller should it ever fail. Spares already built. Building this way with room to play, makes it a simple task to modify or upgrade if needed, and a joy (seriously) to work on or fault find. NANO Inverter: Full download - Only Hex Ver 8.1Ks |
||||
| KeepIS Guru Joined: 13/10/2014 Location: AustraliaPosts: 1816 |
I build one around HY4008w devices with TIP41/42 etc, it was very good and took a real hammering, even using one of the toriods I'm using in the Dual unit, however, I would never go back to an EGS002 SPWM driver / controller. IMHO, the Nano controller is an absolute Game changer in every way, I simply cannot fault it . Edited 2025-03-20 16:17 by KeepIS NANO Inverter: Full download - Only Hex Ver 8.1Ks |
||||
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1193 |
I think you would be better off to put a couple of my designs together, essentially what you are after, but with (potentially) proven performance. The only issue is that both boards are untested in their current form as they are later revisions with refinements of previous working designs, but I would expect them to work first time. It needs 2 x half bridge 6 FET PCBs, both 100mm x 100mm and can pair with an EG8010 controller PCB also 100mm x 100mm and removable 2 x capacitor PCBs that screw onto the half Bridge PCBs. The controller uses the proven inverse opto drive system used for the Nano driven Power PCBs, which eliminates any chance of cross conduction or other glitches from the EG8010. It also has 2 channels of Fan control from 10K NTC thermistors (targets set by trimpots). The second half bridge PCB can daisy chain from the first Half bridge PCB or the controller PCB. There are jumper links on the half bridge PCBs which must be set opposite on 1 PCB to drive as a full bridge. The controller with isolated drive and isolated power to the FET PCBs is the best topology with least chance of wiring induced issues IMHO. The EG8010 controller could later be replaced with a Nano controller (plug and play) and both/either boards can also use the same front panel PCB. I have attached a PDF of screen shots of the EG8010 controller and Half bridge FET PCBs. I can send you a controller PCB, 2 Half bridge PCBs and 2 Capacitor PCBs for $20 + postage if you want to have a play. But the warranty is void when I put a stamp on the envelope  I will also supply BOMs schematics etc.The EG8010 controller also has circuitry to provide a current economiser function for 24 or 48V solenoids (without internal/integral economisers) both from the 48V supply or to just drive an economiser type 48V solenoid. (slight mods needed to the PCB) PCB's: EG8010 & 3FET HalfBridge.pdf and Other Sides.pdf Schematics: Controller EG8010R4.pdf and 6 FET 4KW MaxiPWR R0.pdf Edited 2025-03-20 18:29 by wiseguy If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| Dougie Analog Newbie Joined: 21/02/2025 Location: AustraliaPosts: 14 |
I'm IN!!! I'd gladly take that generous offer wiseguy. Could I pay extra and get the Nano version as well to play with? In the mean time I'm still wading through your thread and digesting it all. I'm still looking forward to reading your thread KeepIS. One step at a time or chunk it down... I haven't been this excited since Fonzie gave me this electronics kit.  Why won't he let my sister use it? I'll PM you wiseguy with my contact info. Cheers brother da |
||||
 Print this page |
 |
The Back Shed's forum code is written, and hosted, in Australia. | © JAQ Software 2025 |
