Go Down

Topic: What killed this MOSFET driver? (Read 3460 times) previous topic - next topic

MarkT

Those circuits are bad circuits, they put 28V (+/- spikes) directly on the gate of a MOSFET with a
+/-30V absolute maximum gate-source voltage rating.

I've explained this already, and why this is not a high-side MOSFET driver.
[ I will NOT respond to personal messages, I WILL delete them, use the forum please ]

saximus

BTW do you have to use high-side switching?
Unfortunately, yes. The lamps are in an assembly which grounds through the body.

Thanks very much for the input, Allan. We didn't use any series resistance because the datasheet shows the load capacitance connected directly to the outputs and thus I assumed the chip would handle the current limiting internally.

Boardburner2

Unfortunately, yes. The lamps are in an assembly which grounds through the body.
.
Is this an automotive application.?

There are robust high side switches designed for this , although i am not sure if they are suitable for PWM.

allanhurst

probably not with a 28vdc supply,,, but it might be a lorry...

Allan

saximus

Is this an automotive application.?

There are robust high side switches designed for this , although i am not sure if they are suitable for PWM.
It's a mock aircraft application (hence the annoying bloody 28V).

The irony of all of this is my original NPN driver worked on the bench with zero hiccups at all and I decided to ruin it by trying to make it better. That old adage just keeps repeating itself in my head. "If it ain't broke..."

Boardburner2

probably not with a 28vdc supply,,, but it might be a lorry...

Allan
:)

I was thinking boat , but they generally do not have body return.

allanhurst

#21
Mar 17, 2017, 02:57 am Last Edit: Mar 17, 2017, 03:01 am by allanhurst
Not many modern small aircraft have a metal fuselage - they're usually composites.... so do you really need a grounded load?

And if this is for Boeing or Airbus, I take it you should know what you're doing, and would be unlikely to be using arduinos..


just curious.

Allan

saximus

#22
Mar 17, 2017, 05:47 am Last Edit: Mar 17, 2017, 05:55 am by saximus
Not many modern small aircraft have a metal fuselage - they're usually composites.... so do you really need a grounded load?

And if this is for Boeing or Airbus, I take it you should know what you're doing, and would be unlikely to be using arduinos..


just curious.

Allan
Haha there are no airworthiness concerns. Hence the "mock aircraft" description.

Without going into too much detail, it is something similar to a simulator which will be using a scrap aircraft which does have a fully metal fuselage. It will use many of the exact same systems as the real thing but they don't need to be work the same way and there is no certification requirement. In instances like this, we have had to "invent" a light dimmer because the real OEM one literally costs in the region of tens of thousands of dollars

tauro0221

Hi,
Why you do not install the mosfet instead supply voltage provide the ground to the light bulb. Then connect a resistor of 1k from the Ardiuno pin to the mosfet gate. Make sure the Arduino pin can do PWM. Also when your are working with  high voltage it is a good idea to protect the micro pin with a zener  diode of 5.1 volts. Since you are connecting the pin thru a 1K resistor if something goes wrong  the zener will protect the pin from high voltage. Attached it is a drawing  how you can do it. I think you do not need to use the TC4431 to drive the mosfet. You can adjust the intensity by controlling the pwm frequency.

allanhurst

#24
Mar 17, 2017, 11:17 am Last Edit: Mar 17, 2017, 11:42 am by allanhurst
Not the PWM frequency - the mark/space ratio.

 analogWrite() does this for you.....

And the arduino limited drive current means the PWM frequency should be fairly low, as the switching time will be significant, which would lead to large transient losses with a high frequency. The standard 1kHz / 500Hz should be OK.

And I'd replace the zener with a pull-down resistor - say 10k - to ensure the mosfet is off when arduino power is removed.

But otherwise OK.  Probably fine for driving filament bulbs.

Allan.

ps a simple 555 circuit would do what you want..   no arduino required.

MarkT

#25
Mar 17, 2017, 02:15 pm Last Edit: Mar 17, 2017, 02:16 pm by MarkT
And I'd replace the zener with a pull-down resistor - say 10k - to ensure the mosfet is off when arduino power is removed.

You could put the 10k in parallel to avoid losing the protection of the zener.  Low-side driving like
this needs a logic-level n-FET of course.  Alternatively my high-side design is outlined in #7
[ I will NOT respond to personal messages, I WILL delete them, use the forum please ]

saximus

Hi,
Why you do not install the mosfet instead supply voltage provide the ground to the light bulb.
Hi Tauro, thanks for the input. I don't quite understand this but from your diagram, it looks like you're recommending an n-MOSFET. As I pointed out earlier, I need it to be a high side switch because the light assemblies earth through the body and thus don't have a dedicated negative wire.

tauro0221

Hi,
Then may you need something like this. Same circuit but using a transistor and a mosfet. See the circuit attached.

saximus

Thanks again for the input but it seems like you haven't read any of the other posts in this thread. That is essentially the exact circuit I had designed originally. I have said a couple of times now that I chose to try a purpose built driver instead as it was my understanding that something designed for the job would be superior to an NPN driver.

allanhurst

#29
Mar 18, 2017, 04:36 am Last Edit: Mar 18, 2017, 04:48 am by allanhurst
Just for fun, the 555 approach.

R2 and R38 are a 10k potentiometer shown near one end....

And r37 should be 100.

Allan

Go Up