Just for fun and my own personal education I thought I'd build my own LT4320 based rectifier for my test bench. Looking at the LT4320 manual it looks pretty straight forward. Just make sure you choose MOSFETs with a low RDS on and you're good. The only thing that has puzzled me a bit is the calculation of C Load. The manual seems to talk about C Load in thousands of uF when I've seen much lower values in other designs here. As a non-EE what am I missing?
Regards,
Dan
Regards,
Dan
Last edited:
Avery good piece of info for me because I was going to put my scope on them 'unloaded' to see what the output looked like.
Regards,
Dan
Thanks. Very interesting point. I'm going to run through the design criteria to supply an F5.
Regards,
Dan
Not quite
Like jean-paul mentioned it's more than just selecting mosfets on rds.
Here's a very good post from Tivol with a mosfet selection guide.
Here's my selection.
It also helps not to think in "will use it for device X" but look at it from a technical point of view so a specific rectifier (with way lower Vf than diodes have) that feeds a filter cap bank and a load. What voltage/current/power is needed for the load? What is the maximum voltage the load endures? Is the specified transformer not too high in voltage considering the now higher filtered voltage? What does one allow for maximum ripple voltage or, what does the load allow to be maximum ripple voltage?
With certain higher loads it may be necessary to use either CRC or CLC filtering. Maybe the builder wants even lower ripple voltage (than being accepted as OK or good enough) as a device is as good as its PSU... In general the allowable power on surge current with ideal rectifiers is higher than with the specified diodes so larger value filter caps and higher rated transformers are now possible.
OK now power on and off effects may occur so maybe "device X" that has very low ripple voltage now has some quirks as heavy plops and woofers going everywhere and even "farting" behavior when powered off. Then maybe speaker protection or just simple and cheap delayed speaker switching with fast switching at power off (often omitted in better audio) may be be a necessity.
Just some remarks for better understanding impact from a different angle...
With certain higher loads it may be necessary to use either CRC or CLC filtering. Maybe the builder wants even lower ripple voltage (than being accepted as OK or good enough) as a device is as good as its PSU... In general the allowable power on surge current with ideal rectifiers is higher than with the specified diodes so larger value filter caps and higher rated transformers are now possible.
OK now power on and off effects may occur so maybe "device X" that has very low ripple voltage now has some quirks as heavy plops and woofers going everywhere and even "farting" behavior when powered off. Then maybe speaker protection or just simple and cheap delayed speaker switching with fast switching at power off (often omitted in better audio) may be be a necessity.
Just some remarks for better understanding impact from a different angle...
Last edited:
So far this is the lowest cost MOSFET that I found that meets all the specs.
IRF60B217 Infineon / IR | Mouser Canada
Regards,
Dan
IRF60B217 Infineon / IR | Mouser Canada
Regards,
Dan
So far this is the lowest cost MOSFET that I found that meets all the specs.
IRF60B217 Infineon / IR | Mouser Canada
Regards,
Dan
This one is good choice as well. Double the voltage and current, lower Rds On. Works happily in my PS. Even survived several accidental short circuits on PS output, that burned 0.6 mm solid core copper wire.
TK56E12N1,S1X Toshiba | Mouser Europe
Gate charge 69 nC and gate input capacitance 4200 pF from the specs, I would definitely have skipped this one... Avoid oversizing the MOSFET the LT4320 datasheet says. May work OK for a while and in some cases but it certainly does not meet the criteria as described and maybe it is wise to keep to datasheet wisdom for guaranteed operation.
This one meets a few criteria and it is isolated: FDPF085N10A
This one meets a few criteria and it is isolated: FDPF085N10A
Last edited:
This is a ballpark calculation:
Using capacitor i-v relation, with 5 nF gate capacitance and 1 mA LT4320 output current capacity, MOSFET gate will change voltage by 3V in 15 uS. That’s fast enough.
The targets I'm aiming for are less than 10mOhm RDSon, a VGS th between 2-3V, and a total gate charge as low as possible.
That's why I posted the IRF60B217. RDS 7.3mOhms, VGS 2.1V, and gate charge of 44nC. On top of that it was 50% cheaper than most of the other suitable candidates.
The nice part is that there doesn't seem to be any shortage of suitable MOSFETs.
Regards,
Dan
That's why I posted the IRF60B217. RDS 7.3mOhms, VGS 2.1V, and gate charge of 44nC. On top of that it was 50% cheaper than most of the other suitable candidates.
The nice part is that there doesn't seem to be any shortage of suitable MOSFETs.
Regards,
Dan
That MOSFET is a good choice. It is both cheap and adequate. I just asked as you maybe have a preference for isolated versions (TO-220F). The ideal rectifier is not so ideal with regards to buildings costs
Yes, the electric car, photovoltaic stuff etc. have stimulated MOSFET development.
Yes, the electric car, photovoltaic stuff etc. have stimulated MOSFET development.
Last edited:
That MOSFET is a good choice. It is both cheap and adequate. I just asked as you maybe have a preference for isolated versions (TO-220F). The ideal rectifier is not so ideal with regards to buildings costs
Yes, the electric car, photovoltaic stuff etc. have stimulated MOSFET development.
I'm just going to do a small amount of tinkering for the fun of it and to see if I can find an ideal combo for my Firstwatt projects.
Regards,
Dan