Which MOSFET would you prefer to use? IRF1010NS vs IRLR2905

[Chriss]

Hi!
I have to solve a repair task where I have to change the original MOSFET which is a hybrid one without package and no
labeling on it with some of which are available on the market.
The electronic device is a diesel high pressure pump control electronic so called Bosch VP44 pump.
Many sources on the internet are talking about a compatible mosfet what they use is the IRLR2905, but I also know from
technicians who deal with this repairings does the repairing fails again and again after short period of time when they done the repair. The IRLR2905 fails again.

The task the mosfet hase to do is to control a solenoid valve on the pump. Nobody knows does the mosfet fully open and close the desired solenoid or maybe he have to adjust the valve from open to close vice versa.

In an original documentation from Bosch I found an information does the control unit have to handle a current around 20A.
If I compare the info to the real situation here is what I get:
I have two solenoids on the pump what is controlled.
S1 and S2 are let's say the solenoids.
R_s1 = ~14 Ohm
R_s2 = ~ < 1 Ohm

That would mean if I take U_bat=~13V:
I_s1 = ~ 1A at switching ON it could maybe be take 3A-5A for short period of time I assume.
I_s2 = ~ 13A at switching ON it could maybe go over 17A-18A for short period of time.

In average it could be somehow around 15A of current delivering by the electronic control unit under normal condition.
I have also calculate with the ambiental heat ( heat of the engine when running ) which makes more worse the situation.
Somehow this speculation is compatible with the writing in the mentioned document above.

I'm in doubt which mosfet should I better use, the suggested one on the internet the IRLR2905 which are not commented from any person why to use that mosfet and they have troubles with it, or should I use the IRF1010NS which I have chosen most because of the very low R_DS(on)=11mOhm compared to the IRLR2905 where the R_DS(on)=27mOhm,
and
the IRF1010NS Id @ Tc=100C = 60A but the IRLR2905 Id @ Tc=100C = 30A
and the IRF10NS I_DM=290A but the IRLR2905 I_DM=160A

So, in any case for me it is most safer to use the IRF1010NS vs the IRLR2905.

What are you think about my thinking?

Thanks for any suggestion and help.
My best regards.

[Etesla]

If the mosfet is controlling a solenoid and dying, there is a chance that the thing killing the mosfets is not current, but instead voltage spikes produced when the solenoid (an inductor) is switched off. In that case, a mosfet with a higher voltage rating, or maybe a flyback diode or suppressor circuit, could solve your problem. If it really is the current though, either of your choices should be sufficiently overrated assuming you were right about the mosfets actual purpose in the circuit. .

[magic]

Using higher voltage MOSFET will only result in higher amplitude spikes. At 20A the coil likely has some significant energy in it and transferring that energy to maybe 1nF of capacitance in the FET and cabling will consequently produce a spike up to hundreds or thousands of volts unless something breaks down before.
The relevant MOSFET spec is maximum avalanche energy at Tc=100°C or whatever the ambient is. A freewheeling diode would help too if it can be fitted into the circuit and if there isn't one already there, I would check that first before assuming that FETs are dying from breakdown.

IRLR is a logic level FET, which means it works with 5V gate drive and this could be the reason why it is recommended. Check if the driver works at 5V or 10V/12V.
Compare figure 1 in the respective datasheets and you will see that while IRF1010NS is capable of higher current overall, IRLR2905 has higher conductance at 4.5V or even 5V gate voltage.

The reason why IRLR2905 keeps blowing up may be dV/dt. 13A is 13nC/ns, this flowing into 400pF=0.4nC/V produces voltage rising at 32V/ns, well above the 5V/ns maximum allowed. Additional capacitance would slow it down to a safe value and there may be additional capacitance already there.

I typed that pump number in google and found people mentioning IRLR3636, which seems to be better in pretty much every way.

Before installing anything, make sure that the driver and gate resistors/diodes aren't blown.

[Chriss]

Thanks for the nice replays, I'm searching for the flywheel diode but it looks like there is no one on the pcb.

The original mosfet was killed because of a bad cabling insulation which is going from the controller to the solenoid.
Two wire was in short.
Actually shorted out the mosfet D and power wire. Of course the mosfet die.

I tried the IRLR2905 and put a load of 12v/21w bulb to the output of the controller.
I left them shining overnight and everything was fine.

When we put back the controller to the car and connected the solenoid, the car started but in several hour the controller was killed again. The mosfet was killed again, the same way as the original one.
That means, the mosfet was in short between the G and D.

I talked to some of friend who are also involved in such of problems and they also repair that, and they approved the same situation when thy are in trouble. Also the short between G and D.

Could a bad soldering joint when I/we solder make the bridge from the G to the pad on the pcb of the controller?
I ask this because the pad on the PCB is really small and coated with some material what is really hard to solder.
I'm thinking if this happens and there is a resistance between the wire which goes to the G and the pad on the pcb that could maybe prevent the mosfet to fully open and the mosfet overheats and die.
Could this thinking be correct?

Why does the short appear between the G and D all the time?

[OM222O]

as others said, it's due to a large voltage spark (which in turn breaks down the oxide layer between drain and gate and shorts them out). just add a schottky diode to prevent the spikes. also you can try adding an RC snubber network to dissipate its energy in a resistor.

[magic]

The original mosfet was killed because of a bad cabling insulation which is going from the controller to the solenoid.

Obviously make sure there are no more shorts, including intermittent shorts caused by engine vibration.

When we put back the controller to the car and connected the solenoid, the car started but in several hour the controller was killed again. The mosfet was killed again, the same way as the original one.

IMO this transistor is likely to fail here due to dV/dt.

Could a bad soldering joint when I/we solder make the bridge from the G to the pad on the pcb of the controller?

I think it's unlikely, did you test this connection after the FET failed?
I'm repeating myself, but make sure that the gate resistor (if any) is OK.

Why does the short appear between the G and D all the time?

When transistors fail they often turn into one metal and silicon pulp where everything is shorted to everything.

Any picture of that board?

[Chriss]

Here are the pictures:


The mosfet is not on the pcb, it should be put on the place where the D letter is.
That big cooper plate under the D is actually the Drain of the mosfet.

[Zero999]

There appears to be a lot of chip on board components there.

I think a freewheeling diode might be a bad idea. It would reduce the likelihood of killing the MOSFET, but it might render the device non-functional or reduce its performance. There will be a reason for it not having a freewheeling diode: it will drastically slow the turn off time, as current will keep circulating through the diode and the solenoid coil, after the MOSFET has turned off.

Replace the MOSFET for something similar and that can absorb the energy stored in the inductor without being damaged.

[magic]

I would start with a 4.7~10nF capacitor from drain to source to limit turn-off dV/dt to a safe value. I did it once in an HV flyback transformer driver and it gave me no problems whatsoever. I didn't try without the capacitor because the FET was somewhat pricey

It will add a turn-off delay too but only tens of nanoseconds so unlikely to be significant.

It of course does nothing to address the avalanche problem, but if you are lucky it isn't avalanche killing your FETs. By the way, can you measure inductance of that solenoid and its cables to estimate the amount of energy it holds?

[Chriss]

No, I can't measure the inductance, to complicated to hook up a scope and resistor etc. to make the measurement on the car.

I'm also thinking about, does it is possible does in the original version of the electronic there is an IGBT and not a mosfet?

This is gonna through my mind because nobody was dive into the trouble or even didn't checked the circuit how it is built up but they use some mosfet transistors and they change them until it last a couple of months or similar.

A mosfet just to switch a 15A solenoid without a flywheel diode on the pcb etc. that is a bit hard to believe to me.

I'm not sure anymore does it it is in original with a mosfet made...

[SilverSolder]

Maybe the solenoid is actually controlled with PWM, rather than bang-bang?  There are other examples of automotive control systems that operate solenoids this way (e.g. evap canister purge).

That would explain the absence of a flywheel diode...

Might be a good plan to put a scope on the thing and see what it is doing?

[magic]

Wouldn't PWM involve having a freewheeling diode in place so that the whole thing works akin to a buck converter and solenoid current stays mostly constant?

[Chriss]

I know, all this is like walking in the dark without any light source...

I ordered some of these mosfet IRLR3110Z just to try what will happen if I interchange the IRLR2905 with IRLR3110Z.

If the car will start even just for a short time I will try to measure the signal with the scope.

But, I will also hook up the scope when the unit is on the bench and power them on.
Just to see if there is some PWM sent to the solenoid wires.

[teus]

Did it work for you?
I've got my VP44 electronics module sitting here, and I want to know if the current MOSFET is bad. From basic testing, diode and resistance test seem fine.

[teus]

I've removed the leads to the original chip, so I could do a correct measurement. It's bad alright.
I'm not able to solder to the PCB at all. 2 hours of trying with hot air, gas soldering iron, etc. the original solder is long gone on the PCB, now I'm left with a naked pad on a PCB that's probably aluminium. I'm way angry.

[magic]

Alu is unsolderable without special fluxes. And solderable with the right flux.

[exe]

I heard those fluxes ruin tips very fast because they are highly acid.

[teus]

hmm, OK. any pointers on a small quantity of flux or solder, available from Farnell, Mouser, RS, Digi-Key, ..?