Diode size

[93forestpearl]

This is aimed at Vikash, but anyone else with input is welcome.


The issue is inductance from a solenoid, and the necessary diode to prevent reverse current in the circuit.

Derek (whitelegacyguy) is running a turbo out of a 6.0 Powerstroke diesel which is VNT. He was running the solenoid at 800 Hz with his hydra since he can only choose between 800 and 30, which is on a extra injector driver rated at around ten amps. After messing with it for a while, the little diode in the ecu popped and rendered the cirsuit essentially useless.

Now, he is using a high frequency solid-state relay with a much larger diode to run the solenoid on a regular auxiliary output from his hydra.

My question is how big of a diode is big enough for a solenoid that is between 5 and 6 ohms? It draws a decent amount of current, but his output FET should have been beefy enough to handle it. He determined that the inductance of the coil is much larger than other solenoids and that is why that circuit went south.


Ideas?

I plan to use this turbo on my car, possibly as soon as this winter. My output FETs are rated to 25 amps (Link Plus), but it still makes me nervous. I would obviously use a solid state relay, which are rated at about 40 amps at a ton of voltage (compared to 12v).

[log1call]

Not a clean switch on the 300 hertz causing heat? Perhaps a square wave would improve things if you think the FET was big enough.

[93forestpearl]

We don't have a choice as to how the Hydra or Link is going to switch. It is a PWM signal.

[vrg3]

Derek's right that the inductance, not the resistance, is the issue. When you stop driving a solenoid, a flyback voltage is induced with the opposite polarity of the drive voltage. This voltage spike can be very high and destroy components.

If he popped the diode in the ECU the flyback should have toasted the FET that was driving the solenoid. That's what happened?

I don't know much about VNT turbo control systems, but you might want to check with the manufacturer about the frequency you're running that solenoid at. If it's meant to be pulse width modulated at 30-ish hertz it might be a bad idea to run it at 800 hertz.

Then, make sure the solid state relay you're using is designed for inductive loads. And if it is, the manufacturer might be able give you recommendations on appropriate flyback protection (assuming it doesn't have something built in). There are options other than just diodes.

Do you have any datasheets for the solid state relay or the VNT control solenoid?

But, anyway, if you just want a diode that won't pop, just pick a monster that's also fast-recovery. Here's one I picked at random from Digi-Key:

http://search.digikey.com/scripts/DkSea ... S20BGMI-ND

[93forestpearl]

Good stuff Vikash.



The FET in the ECU didn't visibly pop, or cook the leads. What happened is he all of a sudden smelled the infamous smell of burnt electronics, and then his vanes closed all the way, and stayed that way until he shut the car off. After that, if he would energize that circuit, it would go to fully closed, grounding the solenoid and closing the vanes completely. The FET may be messed up, but I haven't done anything with this stuff for seven years wince I switched from EE to ME.


The factory duty cycle on the solenoid is just over 180 Hz. He's running it now at 30 Hz and it seems to work as well as it did at 800 Hz. This is a Hydra. I can set the exact frequency with my Link when I do go about this setup.


I don't know what the manufacturer of the solid state relay says about inductive loads, so I'll have to ask him. He just gothte largest diode he could get his hands on. If he is running it as 30 Hz, would there really be an issue with recovery? Maybe I give the stuff too much credit.



Anyways, the VNT system on the Garrett GT3782 VNT uses a 5.x ohm solenoid mounted on the turbo center section to control the oil pressure to a piston which moves the vanes. At zero DC the spring in the piston defaults the vanes to open. Adding pulse width to the solenoid closes the vanes down, spooling the turbo, which is a very very fine line between making boost and power, and making boost but choking the motor. It's a lot of work, tuning wise, but should be worth it in the end. The videos posted in the turbo forum showed some promising results right before the issue with the Hydra.


Thanks for your feedback Vikash.

[vrg3]

Cool, the VNT control system sounds about how I thought. So it's using engine oil pressure to actuate the vanes?

Yeah, that FET is (figuratively speaking) popped. It's fused.

Hmm, 180Hz is a lot more like 30Hz than 80Hz. But, again, only Garrett can tell you if one is undesirable.

You still want a fast-recovery diode even at low frequency because you don't want the diode to slow down the activation of the solenoid.

So, yeah, something like what I linked to ought to be okay. If they're affordable, I would also keep a spare of the solid-state relay handy just in case.