Go Down

Topic: Problems with a MOSFET relay (Read 11513 times) previous topic - next topic

JChristensen

Try it as a low-side switch instead of high-side?

I haven't worked with MOSFET relays like that, from the datasheet I assume pins 3 and 4 would be interchangeable, but I might also try swapping them around.

KirAsh4


Hi, I think you may have a problem with supply and filtering, you mentioned that the supply to the array with the mosfet relay was down to 3.2v.
These arrays have micrcontrollers built in them, probably running at 3.3V. If the main supply drops to 3.2V then the regulation will not be available to keep the microcontroller stable.
To fix this either use a small mechanical relay, or parallel up the output of three or more of these mosfet relays units to drop the switching  resistance.
To check, measure the voltage across pins 3 and 4, AT THE opto mosfet relay and see if you are loosing the 1v there and not your wiring or too small width tracks to the opto mosfet relay.

Tom


No.  This is a custom made strip.  Just LEDs and drivers.  The drivers are capable of running on 2.5V and it runs fine without the relay.

KirAsh4


Try it as a low-side switch instead of high-side?

I haven't worked with MOSFET relays like that, from the datasheet I assume pins 3 and 4 would be interchangeable, but I might also try swapping them around.


Yes, you can flip pins 3 and 4, just not 1 and 2.  Though the datasheet also shows pin 3 being grounded, which would seem to indicate the MOSFET being on the low side.  So I did try both, exactly like your schematic.  In one case, pin 3 can get grounded, on the high side you don't ground it.  Neither fixed the problem.

So my next step is trying to push the MOSFET with a higher current.  Right now it's getting about 16mA, which is more than it's minimum of 10mA.  But I'm going to push that to 20-25mA and see what happens.  Its max is 30mA so I should be fine with 25 even.

However, that will have to wait till this evening because I left the unit running overnight (without the relay) so I can drain the battery and get a low voltage measurement base point for the voltage indicator being built-in the code.

KirAsh4

Alternatively, can anyone suggest a different setup, with maybe different parts?  I don't *have* to use this specific opto MOSFET.  I just need to be able to switch at least 2.5A from a Li+ battery (4.1V - 3.0V).  I picked this one because I had one handy, it can switch up to 3A and it's small - size is a requirement here.

JChristensen

#19
Aug 01, 2013, 07:01 pm Last Edit: Aug 01, 2013, 07:11 pm by Jack Christensen Reason: 1
How about just a MOSFET? Since you're tying the grounds together on the two sides of the opto device, I assume isolation really isn't necessary. The MOSFET part in the schematic will switch 5-6 amps, has RDS(ON) of 27m? @ VGS = 2.5V, and is in an SOT-23 package.

alnath

maybe it is defective ?
I would try it alone, using only the schematics on the right of the electrical characteristics
have you only one of these ?

KirAsh4


How about just a MOSFET? Since you're tying the grounds together on the two sides of the opto device, I assume isolation really isn't necessary.


Correct.  The final product will be running off of the same battery, so I don't need the isolation.  It just happened that I had one of these suckers handy.

So "just a MOSFET" he says ...  All of the high Amp MOSFETS I find are large packages and I don't have the physical space for four of them.

KirAsh4


How about just a MOSFET? Since you're tying the grounds together on the two sides of the opto device, I assume isolation really isn't necessary. The MOSFET part in the schematic will switch 5-6 amps, has RDS(ON) of 27m? @ VGS = 2.5V, and is in an SOT-23 package.


Ok, now I find myself asking the question, "Just what in the world were you searching for?"  Those are certainly small enough to work with.  :)

JChristensen


So "just a MOSFET" he says ...  All of the high Amp MOSFETS I find are large packages and I don't have the physical space for four of them.


I just modified Reply #19 with a few items from the datasheet. Fits the specs as far as I've heard them.

KirAsh4

Yeah, that'll work ... bit of an overkill on the voltage, but that's fine.  It's only switching 4.2V (max, nominal is 3.7V, minimum will be 3.0V).  Signal will be same voltage (range).

Is there a through hole counter part?  I like doing breadboard testing before I commit an SMD design for manufacturing.

JChristensen


Yeah, that'll work ... bit of an overkill on the voltage, but that's fine.  It's only switching 4.2V (max, nominal is 3.7V, minimum will be 3.0V).  Signal will be same voltage (range).


Not at all, that's just a bit of a safety margin. Nothing says it has to operate at 99% of the spec. It's pretty common to operate significantly below VGS Max. 40V to 60V transistors are very commonly used in 5V circuits.

Quote

Is there a through hole counter part?  I like doing breadboard testing before I commit an SMD design for manufacturing.


Dunno. I'd have to check IR's web site.

KirAsh4

Ok, so assuming pin 1 on the strip is VCC and pin 4 is GND, is this the proper configuration, or do I need to turn things around?

And how did you calculate what the value of R2 should be (you have it as 220 in your schematic)?

JChristensen

#27
Aug 01, 2013, 10:04 pm Last Edit: Aug 01, 2013, 10:11 pm by Jack Christensen Reason: 1

Ok, so assuming pin 1 on the strip is VCC and pin 4 is GND, is this the proper configuration, or do I need to turn things around?


Your schematic is correct.

Quote

And how did you calculate what the value of R2 should be (you have it as 220 in your schematic)?


Mmm, embarrassing. Didn't really calculate it. Well roughly, mental calculation. Its purpose is to limit current from the MCU pin as the MOSFET gate capacitance charges. It is a good safe value, will not overload the MCU pin, and should turn the MOSFET on plenty fast as long as you're not turning it on and off at several kHz or higher rate. FYI, I am using the identical circuit, at 3.3V and much lower current, and it works fine. I usually don't cycle it faster than a few hundred Hz.

EDIT: I also use 220 and 10K in another circuit that switches 4-5A at 12V, different MOSFET, but that works well also. Actually it's a motor speed control driven by PWM from the Arduino, so again 400Hz or so IIRC.

KirAsh4

Not planning to cycle it in that way, at all.  It'll function as a simple ON/OFF to turn the string on or off via a mechanical switch.  The only reason I want to do it this way is because of the high amperage being drawn by the string.  I would've needed a rather large mechanical switch, for which I do not have the physical space for.  So going for a much smaller, lower amp switch means having to do something like this ...

The setup is a single Li-Ion battery powering *everything*, both the MCU as well as the strings.  However, there's an ON-OFF-ON switch that on one side will only turn on the MCU so I can dick around with it without needing the strings also on, and when flipped the other way, both MCU and the strings will be on.  Whether I connect that signal to the MCU, or straight to the battery (via the switch) remains to be seen.  I probably will as I've discovered that the LED drivers being used have a tendency to come on at full duty cycle when first powered ... so you get this blindingly bright white string till it received a signal from the MCU to do something else.  So that's one argument to use an MCU pin and bring it up only after it starts sending data, potentially an 'OFF' instruction for 2 seconds while it turns the string on as well.  Kind of a dirty hack, but oh well.

JChristensen

Guess I'd be surprised if the LED string could receive data without being powered on. Maybe that's not what you're saying.

Go Up