Kapanadze Cousin - DALLY FREE ENERGY

[Dog-One]

The only issue is the need for creation of an isolated dc pwr supply (B2) for driving mosfet's gates. Personally i prefer an isolated dc voltage for tl494 driving.

Almost all my custom boards have DC/DC converter modules on them.  Such as these attached.  You can draw your circuit up just as though you have little batteries and place them wherever you need them.  Completely isolated, no mess, no fuss.

[magpwr]

Well, for the push pull i think it is a good choice for substituting ir2110 with two tlp250.
Some comparison:

TLP250: http://web.itu.edu.tr/yildiri1/mylibrary/data/tlp250.pdf
Io=1,5A
Propagation delay time=0,15uS typical (0,5us max)
output rise/fall time=6us

IR2110: http://www.irf.com/product-info/datasheets/data/ir2110.pdf
Io=2A
Turn-on propagation delay=0.12us typical (0,15us max)
output rise/fall time= around 25ns

There is a big difference between rising and falling times of the output pulse, but up to 25KHz that we need it for is fine. And of course what ever it happens at mosfet's side, our driving circuitry is protected up to 2500V
rms (min.)

The only issue is the need for creation of an isolated dc pwr supply (B2) for driving mosfet's gates. Personally i prefer an isolated dc voltage for tl494 driving.
The link below provides details on how to implement this.

http://tahmidmc.blogspot.gr/2013/05/using-tlp250-for-isolated-mosfet-gate.html

hi Jeg,

If you want super efficiency and lowest switching lost in uJ go go for Cree Silicon carbide mosfet.It's only around 16 usd each at 1.2kv .The gate capacitance is mere 950pf.You can figure out which one in digikey.com

My experiment conducted few years ago related to isolated power supply using optically isolated mosfet driver 2Amps ended in failure all 7 pieces.Replacement after Replacement.

My advise stay clear of those shitty item.Good ones are Adum1200(dual) or Adum1100(Single) is powered via 5volts and output is connected to mosfet driver of your choice.
In the end you will still need a isolated power supply module eg:Delta or etc.

I do own nearly 20 of isolated power supply module.I aim for 24volts output <400mA for SIC Mosfet.

If you take a look at Akula the (-) ground is common from supply to output to hv.In other words there is no supply isolation at all.


If you are using 12volts battery instead of variable power supply please do consider using self resettable circuit breaker eg:5Amp(12 or 24volts depending on voltage) rated as example.

Please do think about it.It's your call from here.

[Jeg]

@D1
Nice idea! I have never mess with dc/dc conv. I took a look of DJ06S2412A which outputs 12V. It provides only 500mA at its output. Do you think is enough for fully opening and closing the gate at 17Khz?

@Mgpwr. Thanks for the advice. This cree mosfet looks interesting even expensive for experimental use.
Do you remember what went wrong and you blew all of your seven pieces?
self resettable circuit breakers. I use some fast glass breakers, but sometimes are not as fast as it needs to. I will searchfor what you are suggesting. Thanks

[magpwr]

@D1
Nice idea! I have never mess with dc/dc conv. I took a look of DJ06S2412A which outputs 12V. It provides only 500mA at its output. Do you think is enough for fully opening and closing the gate at 17Khz?

@Mgpwr. Thanks for the advice. This cree mosfet looks interesting even expensive for experimental use.
Do you remember what went wrong and you blew all of your seven pieces?
self resettable circuit breakers. I use some fast glass breakers, but sometimes are not as fast as it needs to. I will searchfor what you are suggesting. Thanks

hi Jeg,

The parts mentioned can be found in digikey or mouser.Circuit breaker are more like medium to slow blow fuse.Only difference it is resettable.

The optical driver just fail merely driving mosfet with no HV or load connected.
It fail during burn in test for 1hr.

There some articles somewhere in internet which mentioned about these failures from opto mosfet driver.

The Adum1200 iCoupler is a good choice because it can do 10Mhz.But it depend on the suffix character.Just take a look at it's datasheet.

Think about the IX series mosfet driver can only do around 2Mhz max base on experiment which is good enough .

[Dog-One]

@D1
Nice idea! I have never mess with dc/dc conv. I took a look of DJ06S2412A which outputs 12V. It provides only 500mA at its output. Do you think is enough for fully opening and closing the gate at 17Khz?

Easily.  Connect a smoothing/storage cap to the output--do not exceed the size listed in the spec sheet and you will be good to go.  Remember, the gate driver isn't delivering full amperage to the MOSFET gate for the whole cycle of the pulse.  It's a typical capacitor ramp-up curve; once it overcomes the initial charge disparity, voltage takes over and amperage drops off to nothing.

Now if you really want to overcome the ringing associated with a MOSFET connected to an inductor, you can use a bipolar arrangement of these DC2DC convertors to force the MOSFET off.  How?    Just put two in series, this is your drive swing.  Say you use two 12 volt modules, in series this is 24 volts delivered to the gate driver.  Now level set the MOSFET by attaching the Source in between to two DC2DC convertors.  Now you will have -12 volts on the Gate when the MOSFET is off and +12 volts on the Gate when the MOSFET is on.  Why would you want to do this...?

When you drive an inductor, what happens when you turn off the switch?  Recall an inductor fights a change in current.  It fights by way of creating a massive voltage potential.  And where does this potential go?   You have one side of the inductor fixed to the positive rail and the other side is essentially floating since you turned the MOSFET off.  So lets say the inductor and MOSFET Drain now shows a potential voltage of negative 40 volts.  Say what?  Yeah, that's inductive kick, CEMF, Lenz Law, whatever name you want to apply to it.  Look at where the Source of the MOSFET is connected--ground.  That should be zero volts rights?  And the positive rail should be say 24 volts right?  From the perspective of the inductor though, ground is now been pushed to minus 16 volts.  For a brief moment, what do suppose the voltage is at the Gate?  That's right, positive 16 volts referenced to the inductor.  Sixteen volts for a nanosecond is just enough for the MOSFET to begin to trigger again, even though the MOSFET driver tells it to be off.  The issue here is charge and its propagation speed.  It can get to the ground rail faster than it can get through the gate driver and back to the MOSFET Gate, so the MOSFET begins to switch back on.  You'll see this ringing on the scope when the MOSFET turns off with normal/simple drive circuits.

So pulling the Gate negative reference to Source is one way to help with ringing; the other way is to ensure you don't turn off the current to the inductor so fast that it kicks back.  Do both and you will have a nicely optimized drive circuit.  Attached is another "snubber" circuit that helps with kickback.  You will notice those resistors are 10 watts dissipation and they do heat up.  You can probe around with the scope all day trying to figure out why.