The decoupling/boost caps look reasonably sized. If you want to supply an external bootstrap voltage, just plug it into Vb, and reconnect the 1uF between Vb and ground so it can decouple the 2-3A current spike that will happen when the mosfets turn on.

If you do decide to just keep the diode/cap bootstrap configuration, just read through what I wrote before about not having the lower mosfet. It may work sufficiently well with out it for your configuration, but you'll need to test it out.

I can't make out the diode ratings, but they're in the right direction. You should also clamp it to ground as well - hypothetical question: what happens when the motor terminal spikes down to -20V with the gate being held at ground?

If you go through the data sheet and app notes, the only other thing might be to put 6x ~100 ohm in series with each of the mosfet gates to prevent some high frequency oscillations. Otherwise, using the gate driver chip is about as easy as it is going to get.

James

 

Ok, sow in fact i could better eliminate the diodes and fit 1 mosfet?

problem is that i need/don´t want overlapping of the mosfets because then the lower one will be burned out quet easy....not?

Will fit 2 to ground and 2 to vdd if they blow i will fit more??

they are huge! will try to add as much of them as i can.

the gate part i don´t untherstand what would you do? thats in fact what i need to know...

Tks

 

Correct, if the circuit tries to brake with the motor spinning at a good clip, it could burn out - but I'm also saying that without a mosfet there, the upper mosfets may not turn on fully when you approach no-load speeds. If the motor is always moving relatively slowly (back-EMF is low), the circuit should be okay as long as the PWM signal is present.

Are you talking about the diodes?
They need to be sized so that they can handle the voltage spikes that happen when the mosfets switch. The average current they need to handle is relatively small. A "1 Amp diode" might have an instantaneous current rating of 3-5 Amps, so read the data sheets.

Ahh, don't worry about it. It's something you do when people complain about radio interference or mysterious part failures - look at the app notes for details. Lots of engineering time goes into writing those things, and they are very useful.

James

 

Ok, i wanted to fit 4 3amp (rated) diodes.. but if i hear you than i should be done with 2 of each in both direction right??

Can i eliminate the resistance of 10R?? i don't have a 10R with significant wattage...

guess its 1/4 or less..

Tks

p.d. the circuit should be able to do a full motor start with load on it.. i don't have a overamp limit built in because of the reason that fitting resistor in the feeding line with a starter as consuming doesn't look wise... also because i will limit the max rpm/amps with the max pwm the controller feeds...

anyways 6fets of my size should give me 600amps continous and 2400ams pulses sow i gues it should be quit bullet proof also i hope that the amd cpu cooler will let the lot stay cool.

is it a trick if i fit a 27volt zener from the motor to the 1uF cap??

it would need to stay open when the cap doesn't has de 27volts on it...

sow in fact i though of charging the cap with the back emf of the motor?????
maybe a diode to prevent to blow the chip? before the pin?

as you can see i try to do my best on this

Thanxs already for your help!

 

The diodes i use are the 1N5401 they have 3amps continous and :
Peak forward surge current 8.3ms single half sine-wave superimposed on rated load (JEDEC METHOD) at Tl=105degrees = 200Amps

have 2 of them in parallel sow i gues we have finished that stage...

wel i'm gonna hook up the caps and tomorrow test it..

a small motor i don't have one, only a bike starter instead of a car starter..
(dunno wich is worse) think i start with a bulb a couple of amps...

wich frequency would be the safest to start with? the lower the better right?? regarding to mosfet in non fully conduct fase..??

Regards,

Tks

 

I'm now realising that the only reason for me to use this chip is that its for purposes of positive feed to the N channel Fet..

Is it possible to fit a bootcap to a normal driver for example the 9amps TI one i earlier posted??

Tks

some pics to!

Attached Images

	6Fetsparralel2400ampspeak!!.JPG (13.5 KB, 22 views)
	AMDCoolerBigCapInputsOutputsDriver.JPG (13.8 KB, 21 views)
	Spores On the Board.JPG (14.8 KB, 21 views)
	BigCapReverse.JPG (13.7 KB, 21 views)

 

James,

i was studing the datasheet and i saw that COM = low side return, the above pic its just ground..(page 1)
watching at VCC its Low side supply is it true that bove are for me ground?? how the hell that cap can then charge it self??

wich sense has the diode?? or is it that the back EMF (negative vols charge that cap??

Tks

 

okay, where to start...

The diodes should be more than sufficient.

Mosfets have capacitve gates - there is only current going through that resistor when the mosfet turns on and off, otherwise there is no current going through them, and I'd expect acual power dissipation in them to be in the milli - microwatt range.

Considering "0000" copper wire (i.e. 1/2 inch or 13mm copper rod) is rated for 400amps, and I don't see any of that, I'd expect the realistic rating is a bit lower...

? the back emf from the motor won't cause the voltage on the capacitor to increase (by much). If the motor is being forced to turn externally, then you'll need to worry about over-voltage.

If you mean: "The only reason to use this chip is to drive a high-side FET", then yes, that would be it (and a good reason as well). You can find less capable chips, and might be able to save ~20% of the cost of the chip or something.

This chip has two *different* sections. One side is built to drive a low-side mosfet, the other side can drive a high-side mosfet. The low-side mosfet is easy to drive and doesn't need the bootstrap capacitor or diode. The high-side on the other hand -

The idea is that when the high-side mosfet is off, the motor terminal will be at ground. At this time, the capacitor will have 11.4V across it (because of the bootstrap diode). When the high-side mosfet turns on, the motor terminal will be at 12V, the capacitor will *still* have 11.4V across it, and the upper capacitor terminal will be at 12+11.4=23.4V, which is enough to turn the mosfet fully on.

James

 

Don't untherstand, the gate resistor can be anything?? where are the high amps needed for the fast charging???

also i now use no lower side mosfet...its just the upper one and the diodes...

Problem to use one lower is that the pwm signal should then be splitted or tapped and i will need an inventor to make from my low pulse a high pulse and feed that signal into the chip, problem is that its scaries me because overlapping is then serious a problem altough it would be in the nanosecond...

anyway i guess it should work atleast it does if the gate charge is wat you say!

 

am i correct in saying that the Bootstrap cap has to be the impedance charge of all the fets + a margin for sure + a load to be at the same load as them sow in fact 6x charge per fet * 2 ??

just a logical thinking really...

Have soldered a bit more Pb on the strips to the fets... to make sure they are better balanced... my old soldermachine couldn't cope with the heat (think because of the heatsink)

but i'm sute uts quite balanced now..

is the lower one cruisucial??

Tks

 

Each mosfet has a gate charge of ~100nC. Current is defined as I = dq/dt, so if you want to turn on a mosfet in 100nsec, you need 1A of current. The peak power on a 10 ohm resistor is going to be I^2*R = ~10W, but this is over a very, very short period. Since I don't want to remember how to integrate this properly - average power per transition is <~1 microwatt. So if you have a 1KHz PWM signal, dissipated power would be < 1mW.

I made a bit of a mistake though - I didn't realize this chip was only capable of driving a small amount of gate current - 200mA sourcing, 400mA sinking. The amount of gate current is important because you want to drive the mosfet on and off as fast as you can in order to reduce the amount of power wasted in the mosfet.

As I've said before, only using a high-side mosfet should mostly work, but you'll need to try it out and see if it has any issues - if the high-side mosfets get too hot or something else happens, then we'll know it wasn't good enough.

 

James, is it a option of cutting out a couple of fets..

for example try it out with 4 of them...(400amps continous in theory) then the capacitance will go lower and the mili amps per gate higher right? I have in the post mail some 7/9 amps drivers but then i need the bootstrap solution externally??

isn´t their a way of using one mosfet just for turning on the rest of them???

Then we have 200mili amps for 1 mosfet and the rest of them are switched on by that mosfet???? For example: the source of 1 fet is connected to the gates of the rest of the 5 exept his own gate. Sow in fact the first 12volts go very fast then the last boost to 24volts is done by the cap.

also i tought of a 27volt zener (have them in stock) to protect the gates from overcharging i know that a silly bit over the limit ruins them... maybe a good mod to-be-sure if something happens we then know that the gate voltage wasn´t to high..

TKs

p.d. anyways i will lower the Frequency to 1Khz the less switches per second the cooler it will run..

 

I have 5 of this chips each:

http://ww1.microchip.com/downloads/e...Doc/21420d.pdf

Item: TC4421EPA inverted
Item: TC4422EPA non inverted


http://ww1.microchip.com/downloads/e...Doc/21418C.pdf

Item: TC4404EPA
Item: TC4405EPA

If i take just a normal driver chip would i be able with a zener or something to make a bootstrap sow that the gate will be charged up??

also in the last link we see a sheme where the gate is just hooked up to VDD
page 7 fig 3-4 in a full bridige config...

 

mhh studing the fact for the zener i just noticed that there is a absolute max rating of 20volts gate source voltage!

sow i source is 12volts then gate can be max 32volts right??

but when device is of i have to ensure the cap is empty not?

Tks

decided that i first try with this chip, else after this chip i will fit some 9A drivers.. maybe 2 of them in parrallel will be enough!

perhaps with an extra bootstrap??!

in fact to protect it right i need a 19volts zener and put it to source right?

 

Don't worry about changing the number of mosfets. When you test it out, you can figure out if you need to improve the switching speed. Although I would suggest that you just use 1 fet while you are initially testing just so you don't blow out an entire bank of mosfets at the same time.

If you want to protect the mosfet gates, put a 15V - 19V zener between the gate and source connections, after the 10R resistor. This is usually only needed in extreme circumstances - If you look at the capacitor, it is relative to the mosfet source. The capacitor voltage will never be greater than 11.4V + V(mosfet source) and there is no way of desroying the mosfet with this capacitor unless something gets shorted/disconnected.

You could use any of these chips to generate a high voltage if you want. The idea behind the bootstrap capacitor is that on every PWM pulse, the capacitor will charge up a little bit, and it is always safe. If you use an external high voltage supply, you will need to add a zener to protect the mosfets. (With the bootstrap approach, the gate zener is *not* required).

James