ChocoHolic said:
i have test most of d-fets from *R
not very good at my project, running at 1mhz,
perhaps my speed is toooo fast,too high dv/dt
d-fet ,very easy to be break
but if u use the driver from ir, perhaps will work well,
because the driver from *r is slowly.
i'v never used it
rg
fumac
The NXP (Philips) devices listed in their 'LV MOSFET Applications' app-note (for the UCD basic design) all appear to be good devices. Although the parts shown are TO220 and D2-PAK parts, if you look on the NXP site there are also D-PAK (TO252) versions of some of them. I'm using D-PAK devices on a fully SM class-D design of mine, (using thermal vias to conduct the heat to the other side of the pcb, where I have a heatsink attached), and the low inductance leads of this package result in a very low-noise design.
poobah said:You guys are scaring me...
I just designed those Direct Fets into a prototype!
Don't worry there are many ways to kill MosFets.
Fumac's trouble might not necessarily appear in your proto. Without detailed screenshots and analysis what went wrong at fumac's place, it is impossible to translate his findings to your project.
I would say 'go on' and let us participate on your findings.
ChocoHolic said:
yes, keeping research, this is the only way to make ur project running .
but , my test told me , that the d-fets just can be use in a low power project,
because the heat-sink is hard to asm on it , include d-park, not easy to use at high power
i use TO-220 at all of my high power projects.
rg
fumac
IMO, to bypass body diode is much easier, than to find proper 200V mosfet.
PS: I've tried directfets 2 years ago in the caraudio 12V H-bridge SMPS, 2mOhm 30nC/13nC (though Trr isn't matter for 12VDC apps), well, i think it's really nice stuff, but very annoying was directfet replacing procedure, due to huge trafo pins, that was extremely close to the directfets. So i guess, i would be damned by repairers, and finally i throws it out. BTW, directfets heatsinking is not so convenient vs to220 ones.
PS2: "IRFZ48V 100V 36A .021ohm 42nC/41nC" hmm, rather irf540z By the way, 540z could be successfully used in the class D too, of course it isn't the best choice though. This is THD/Power sweep of amp 400kHz ir_driver+irf540z +/-36VDC, loop gain are ~36db, 4oHm loaded:
PS: I've tried directfets 2 years ago in the caraudio 12V H-bridge SMPS, 2mOhm 30nC/13nC (though Trr isn't matter for 12VDC apps), well, i think it's really nice stuff, but very annoying was directfet replacing procedure, due to huge trafo pins, that was extremely close to the directfets. So i guess, i would be damned by repairers, and finally i throws it out. BTW, directfets heatsinking is not so convenient vs to220 ones.
PS2: "IRFZ48V 100V 36A .021ohm 42nC/41nC" hmm, rather irf540z
By the way, 540z could be successfully used in the class D too, of course it isn't the best choice though. This is THD/Power sweep of amp 400kHz ir_driver+irf540z +/-36VDC, loop gain are ~36db, 4oHm loaded:Attachments
Qrr is very high.
For example:
http://www.farnell.com/datasheets/1648212.pdf?_ga=1.118420553.119672679.1472661918
500 nC at only 17A, compared to IRFB4227's 430 nC at 46A.
For example:
http://www.farnell.com/datasheets/1648212.pdf?_ga=1.118420553.119672679.1472661918
500 nC at only 17A, compared to IRFB4227's 430 nC at 46A.
I put a little solder on one pad first.
Then use tweezers to hold the part and solder it to the soldered pad.
I then solder the other pad.
500 vs 430 is called 'very' high??
That 14% more.
Qrr is increasing with forward current. Should I explain it further?
Yes, I am aware of that. But 14% is not something that keeps you awake all night? With a proper design it's not something to be bothered about to much.
As in, there are MOSFETs which are better yes, but also which are worse.
D
Deleted member 148505
Comparison is unfair, IPP320N20N3 is very easy to drive.
IRFB4227 upgrade is IPP120N series
http://www.infineon.com/dgdl/Infine...n.pdf?fileId=db3a3043440adf7501440c77d8950149
My Trabant consumes 6 litres of fuel on 50 km, so it is only 20% worse than a Suzuky that drinks 5 litres on 100 km. Right? NO!
Did you get it now?
I repeat: 500 nC at only 17A, compared to IRFB4227's 430 nC at 46A.
If you compare them right, you must take 2 of IPP320N20N3 paralleled against 1 IRFB4227. This way you can see that in the same application (same output power) IPP320 has about 3 times more recovery loss. So the difference is not 14%, but 200%.
Unfair? It's a different contest! Gate charge is one thing, recovery charge is an other. You can't disqualify a runner because he is too heavy.
And driving is not a real issue IMHO. A $0.07 pnp transistor can drive the IRFB4227 easily.
Yes, this one is a good MOSFET, and they explicitely stated "FastDiode(FD)with reduced Qrr" on top of the cover page, unlike IPP320N20N3 that was mentioned. Do you know what is unfair; changing competitor during the contest.
BTW talking about whole series containing significantly different parts is quite confusing. Be specific! IPP120N20NFD. Unfortunately it's 2-3 times more expensive than IRFB4227 (compared cheapest source for both part).
Yes, I can read.
But I was not the one who compared the two.
I was just saying it's a reasonable alternative (for whatever design).
Therefore I would like to go back on topic, completing the list with usable MOSFETs
D
Deleted member 148505
Yes driving is not an issue but the gate drive current requirement will double or triple, issue in class-d if there's no external supply, another $$.Unfair? It's a different contest! Gate charge is one thing, recovery charge is an other. You can't disqualify a runner because he is too heavy.
And driving is not a real issue IMHO. A $0.07 pnp transistor can drive the IRFB4227 easily.
Yes, this one is a good MOSFET, and they explicitely stated "FastDiode(FD)with reduced Qrr" on top of the cover page, unlike IPP320N20N3 that was mentioned. Do you know what is unfair; changing competitor during the contest.
BTW talking about whole series containing significantly different parts is quite confusing. Be specific! IPP120N20NFD. Unfortunately it's 2-3 times more expensive than IRFB4227 (compared cheapest source for both part).
Yup im wrong, there is no IPP120N series, only Optimos 3 FD technology mosfets. If they made them cheaper no one will buy the IRFB4227 - not discontinued yet.
I IPP320N20N3 is an upgrade of IRFB4620/5620
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.