DIY Open Source EV Charger

[arnolde]

Originally Posted by MazdaMatt

Arnolde, was there a rating in "C" rather than "pF" for the gate? Could you link a datasheet plz?

http://docs-europe.electrocomponents...6b804732b9.pdf
Oh... Am I late? *ggg*

So I take it, that 100mA gate current (NE555 output through a 100R resistor) is plenty fast enough?

[MazdaMatt]

I noticed in this sheet that it mentioned a rise time of 20nS and a fall time of 55nS. Does that mean that we are safe to switch in this time frame, or would we be generating huge voltage spikes on the 150uH inductors? If we could do 66nS switching time, for example, we could increase that current by 30 times, so you'd be pushing 1.5A into them... Not sure how safe that is for spiking, though... that leads to a conversation about GS capacitors, which i believe that you are not using in the controller.

[MazdaMatt]

Arnolde - you said you had an excel sheet for calculations. See if you are happy with 1uS switching power loss compared to say, 0.1uS. Eventually you get diminishing returns by going faster and faster and you get increasing downsides.

[arnolde]

Originally Posted by MazdaMatt

Arnolde - you said you had an excel sheet for calculations. See if you are happy with 1uS switching power loss compared to say, 0.1uS. Eventually you get diminishing returns by going faster and faster and you get increasing downsides.

In other words, a switching frequency of 100kHz instead of 1Mhz?
Of course I'm happy with that :-) I'd even go lower if necessary... What I want is optimum efficiency (95% would be nice ;-) and the only reason I'd go faster (like, over 25kHz) is to reduce the size of Caps and L's.

EDIT: 0.1µS would even be 10Mhz... or am I mixing something up here?

[MazdaMatt]

Oh god, don't do that to those poor mosfets.

You're confusing switching time and switching frequency. You want the minimize the % time that your mosfets are partially switched (with consideration to your other components). Just to ballpark it, try 50kHz and see how that affects your inductors and capacitors.

If you were switching at 10Mhz, you would never be fully on or fully off and you'd be burning insane amounts of power in the mosfet.

[arnolde]

I'm doing the excel sheet myself, so I can only implement what I've already understood. Which doesnt include this issue yet ;-)

Anyway, now at least I understand why in the abovementioned Eq.12 on Page 52 of the magazine, the gate current influences the switching losses.

Now, how am I to determine if I want my switching time to be 0.1µS or 1µS or 10µS? Basically I'm limited to the NE555 output capability, and according to the circuit design, the gate is sunk directly but sourced (pulled up) by a 10k resistor to 12V, so the current is limited anyway to way under 100mA unless I change that. But it seems the thread opener's setup is working fine in that respect.

[NiHaoMike]

In general, lower frequencies result in lower switching losses, but increase the necessary size of components which might offset the savings.

[MazdaMatt]

Mike, please fill me in on what way a lower frequency increases the size of components. Which components?

If you are switching at 100kHz, for example, and you take 1uS to switch, you are at 1% of your time switching. going to 0.5 or 0.1% of your time isn't really a big difference in the grand scheme. Changing from 10% down to 1% is definately a big deal.

Run a few numbers to size your parts according to 5uS, 1uS and 0.1uS and see what you get for costs and heat generation.

[arnolde]

So if I switch at 250kHz, I can get by with really small components :-)
Hey, maybe I'll try to fit it into a wristwatch *ggg*

[NiHaoMike]

At lower frequencies, the inductors and capacitors need to be larger to have the correct time responses.
The capacitors are probably not that much of an issue since it's generally ripple current ratings, not capacitance, that is the limit. But the inductors would need to be larger.

If you're using several power stages in parallel, it is possible to stagger the clocks and make the effective switching frequency higher, but keep the actual switching frequency the same.