What is a better solution to drive a IRF3205 with an NE555?

[Chriss]

Oh hey, those schematic symbols I drew so many years ago.

Tim

Hi!
If you have practical experience with this schematic can you pls check my scope diagram from my Reply #24 and
what you think? is it good enough for using it or should I maybe tune a bit more the G-D signal?

Thanks.

[T3sl4co1l]

What about it?  There's a bit of turn-off drain ringing, a modest R+C damper would help with that (maybe 22R + 1n).  Turn on is slow and smooth.  Doesn't look like it would be a problem at the switching frequency.  If the transistor is getting too hot for your liking, use a bigger heatsink, or consider improving drive speed (or, consider getting a smaller and faster transistor).

If you expect highly inductive loads, the R+C might be increased, or a clamp diode or TVS added, to clamp excessive voltage (better to divert peak voltage somewhere, even if it's burned in a TVS, than to make the MOSFET handle it).

I don't recommend the complementary emitter follower, as the 555 output does not swing full +V/GND as it is, and the follower drops 0.7V above that.  Discrete alternatives are more complex (a rail-to-rail amplifying stage, then a follower), and integrated alternatives you probably don't have on hand (e.g., TC4420 gate drive IC).

Tim

[Mechatrommer]

The FET is cool like a dead man after 48h. 

at what current are you testing? IRF3205 is a 80A mosfet, 8mOhm Rdson. if you are testing the 25W lightbulb @ 12V, thats 2A current. P = I2R = 4 * 8e-3 = 32mW on the mosfet, thats nothing. real test should be near mosfet rating or half of it to be safe ie 40A, but for your sake its 10A max if i see it correctly. at 10A, the mosfet will dissipate 0.8W fully on, assume half of it with zero loss switching 50% PWM, there's where if you want to sense some heat and decide if your gate drive is efficient or not. the push pull in the OP you posted should be the better and easy solution if you find any issue with 555 directly driving the fet. other beffy IC should be better but i think will be overkill for your 120W application. energy loss can rise up due to overkill driver (fast switching) i think. ymmv.

[Chriss]

T3sl4co1l:
Thanks for your help mate! I really appreciate it.
At the moment I have a 16A schottky dioede parallel with the DC motor,
the ringings are still there but much lower. Like you see on my last scoping.

I will try to put the R+C dumper as you mentioned and see how much will that help.
That R+C dumper I should put on the D and GND or between the +12v and GND?

Mechatrommer:
I'm testing the circuit around 10-12A current.
Everithyng is cool even on the 10, 50, 90% duty too.

Can I say does this circuit is actually ok for a current of 10A?
I assume the max current what could be delivered through this circuit is 16A because of the rating of the schottky diode.
Of course with a heta sink on the diode.
A m I right?

[T3sl4co1l]

Oh, no wonder then -- the diode belongs at the source of the transient: the switch.  Put that thing right down on the board, beside the transistor, and beside a +V to GND capacitor.  All gone!

Tim

[Mechatrommer]

I'm testing the circuit around 10-12A current.
Everithyng is cool even on the 10, 50, 90% duty too.

then the switching loss is not much concern.

Can I say does this circuit is actually ok for a current of 10A?

from your description, it should. you just have to concern about the transient and back EMF because mosfets are sensitive to over voltage, energy doesnt have to be much to toast a mosfet. i would say guard the mosfet as much as you can afford with transient diodes and snubber.

I assume the max current what could be delivered through this circuit is 16A because of the rating of the schottky diode.
Of course with a heta sink on the diode.
A m I right?

i'm not sure which diode but if its for transient or voltage reversal protection and a big diode, 16A usually a continuous DC current when diode operating in forward mode, in transient mode, the diode usually rated in hundreds of ampere of surge current. your circuit using IRF3205 should be able to draw near the spec current of 80A (at 100degC) it just a matter of your cooling system. i would say 5-10W of mosfet dissipation for cheapish (fan) cooling system, thats around 35A @ 8mOhm Rdson. thats in DC, maybe you can get by with 70A in 50% PWM not really sure havent dont that  you have to work out the SOA chart and derating factor as well... ymmv.

and btw another thing is, also check your driver (555) it may not be happy with 1.2A of gate current sourcing and draining, but it may.. ymmv.

[Chriss]

Here I'm again 

I played with the RC snubber but without success.
But when I put the schottky diode beside the FET and not on beside the motor here is the result:
The type of schottky diode is all the time MBR1645DT.
Pic1 is the scoping when the diode was on the motor and
Pic 2 is the scoping when I put it near as I can to the FET.

What do you think, is this switching off ringing acceptable when the pwm signal is on 50% duty and
the load is 10A DC fan motor?

Everything is still cool, no matter if the duty is 10% or 90%.
Tested over 30min continuously.

[David Hess]

Current through the diode is initially close to the drain current so it needs to be sized for that; the leakage inductance is trying to maintain the current.

I would add a big decoupling capacitor across the series connected switch and load to minimize the loop area of the high frequency AC.  That may be contributing to the flyback you are seeing.

The output current from the 555 is limiting switching speed which is likely an advantage.

Some DC brushless fans do not like being PWMed like this leading to early failure.  Old motherboards which do this have simple LC output filters and a diode and operate like an open loop buck regulator.

[T3sl4co1l]

Yes, much better.

Tim

[Chriss]

Hi David!

I would add a big decoupling capacitor across the series connected switch and load to minimize the loop area of the high frequency AC.  That may be contributing to the flyback you are seeing.

How big should this cap be? any suggestion?
Thanks.

[T3sl4co1l]

Offhand, a 1000uF electrolytic capacitor rated for at least supply voltage, should do.

Tim

[David Hess]

This is one of those cases where a film or ceramic capacitor is worse than a solid tantalum or aluminum electrolytic because of dampening considerations.  (1) The rule of thumb is usually something like 47 to 100 microfarads of aluminum electrolytic capacitance per amp which is consistent with T3sl4co1l's 1000 microfarad suggestion.  (2) A solid tantalum part would be expensive even though it could have lower capacitance and failure due to surge current becomes a reliability issue.  A 100 microfarad or larger polymer aluminum electrolytic capacitor might be suitable but I would want to test that with an oscilloscope.  The test I use is to increase the capacitance with a worse case load until diminishing returns of reducing power supply bounce are reached and then doubling that value but maybe someone can suggest a more scientific approach; I'm used to dealing with poorly defined loads so it is just simpler this way.

(1) This does not mean that an additional large ceramic or film high frequency decoupling capacitor should not be used because it usually should.  But the large aluminum electrolytic capacitor is still required for dampening.

(2) This is the same rule for simple regulators and why small bulk decoupling capacitors are located where power enters a board.  High performance power supplies may use a small ceramic or film output capacitor with a low value series resistor and careful frequency compensation in order to improve performance but none of this is required here and the brute force approach of a relatively large aluminum electrolytic capacitor is better.

[Chriss]

David:
Can you pls post some sort of schematic with this cap in circuit?

Just to be sure I'm understand correct everything, because the language differences pls.

Thanks.

[David Hess]

It is not that complicated.  It needs to go between the source of the MOSFET close to the MOSFET and the Vcc connection to the load close to the load like at the cathode of the diode.

[Chriss]

I put the cap as mentioned but nothing significant was happen.
Also tried with other values of cap but nothing. 

The ringing peek voltage is 14.5v at 50% duty on a motor load of 10A.

Is this peek actually possible to reduce or complete eliminate?
The datasheet say V(BR)DSS 55v min.

Is my controller actually good enough with this setup and this scope diagram to handle my fan or whatever
with a current of 10A?

Should I in real reduce this ringing?

Thanks.

[T3sl4co1l]

The ringing is small, it won't damage anything.

It could still cause interference, who knows.

If you have any ferrite beads handy, you can try one on the drain pin, or in series with the diode.

Tim

[Chriss]

Ok.
Just for the science and fun, I will find a ferrite bead.
Actually I saw somewhere I think in PC PSU. I will hunt down and give a try. 

[Mechatrommer]

Increased ringing i suspect...