Electronics > Projects, Designs, and Technical Stuff

Experimental 48V->200V boost

<< < (3/3)

ocset:
I did a quick LTspice representative sim of one of your Boost stages, and its as attached if you wish. (it mightn’t be too accurate due to your using a non linear inductor). If I were you id use ferrite….will make the slope comp easier to calculate.

170V is quite high for a boost output voltage, and I reckon your reverse recovery losses in the fets and diodes will be too much.
(as you know, diode rev rec will also heat the fet in a booster)
Youll probably have to increase the fet gate series resistor enough so that you reduce the reverse recovery spike a bit.

Reverse recovery is the grim reaper of high power / high vout boost converters in CCM


But in truth, with  deep CCM , and your high vout, you need SiC diodes, and bigger ones than you have. I couldn’t find the datasheet for your E5… diodes.
As you know, with hard switched topologies at fairly high power, and   in deep CCM, you especially need to tune the gate drive series resistor and current sense filter, otherwise the power supply  just wont work… it  will get swamped with noise.
I think you need a decent heatsink directly on the fets and diodes…not just going through the FR4 to a heatsink below (?)
I would use a current sense transformer aswell (like T3sl4co1L said) ..they do  give a less noisy  signal than a sense resistor,and good for high currents.

beware also using fets in parallel.....check you dont have vgs ringing too much...but you seem to have checked this.

The pdf attached tells how to calculate slope comp

Theres a current sense txfmr sim attached too (in ltspice)..its for a forward but its the same type you shoudl use (i think Tim maybe already said this for you)... The CST will have to go in the drain connection of the fet, which is a noisy switching node, but it has to go there.
I am sure you are aware....the CST inductances and turns ratio has to be correct, otherwise your current sense signal may  slope down instead of up.

So i hope now you are up on the absolute curse, the absolute SMPS graveyard, which is "reverse recovery in high voltage Non isolated converters" (In isolated ones, the leakage inductance catches/staunches the leakage spike a bit)
Those little 4by2mm (non SiC) diodes that you originally had, would have wanted to  run off the board screaming in truth.

Navigation

[0] Message Index

[*] Previous page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod