switching using FET IRF540 P40NF10L

[lilshawn]

I've started a project that will replace a convoluted series of switches and solenoid relays with a single circuit. the original circuit implemented in the 80's had a 24v relay close a switch stack that turned on high voltage to a secondary relay contactor coil operating at +40v. (switching the coil ground to activate) due to this design, it was found that the contacts would burn out due to the high load.

it was later replaced with a weird relay with a switch stack consisting of NO and NC switches and operated a large TO-3 PNP? NPN? transistor. (i'm not entirely sure, as someone had got to it before I had... a transistor in anycase) but it was just as unreliable and prone to blowing out.

so my project was to replace all this with a FET and be done with it. i originally attempted to use a IRF540N but when testing the circuit, i could not conduct it properly. i thought the 5V supply i had would be able to switch it on but i guess not.

so I changed to a P40NF10L since it's datasheet seems to say it's Vgs is lower than the IRF540N but still can't seem to get this operating properly when testing. Do I need to find something lower?

attached is the circuit i built. the fet is not quite accurate as the internal diode is not included, but is accurate enough to explain my circuit. any help explaining where i've gone wrong would be appreciated. the FET need only be active ~300ms at a time, faily intermittently. (not constantly) the circled area is representative of the high voltage contactor relay coil and it's associated wiring.

the basic switching circuit is good. it operates when it should, i just cant seem to drive the fet to conduct the HV to ground to activate the contactor

EDIT: the reason for the switch on the activation line is a representation of the driving boards activation of the original relay circuit. the driving board activates this relay also by switching it to ground, but it's capabilities to do so are very limited to a very small load, hence the weird circuit. so it is being represented here by a switch to a 0v ground.

[Kim Christensen]

the basic switching circuit is good. it operates when it should, i just cant seem to drive the fet to conduct the HV to ground to activate the contactor

Strange. Do you have 5Vdc between the gate and source terminals of the P40NF10L when the relay is on?
It is possible that the 7805 might be unstable. Have you tried a 0.1uF capacitor on the output of the 5V regulator? Both bypass capacitors should be as physically close to the regulator as possible.

Oh, and how long is the wire between D1 and the FET?

[lilshawn]

I currently have the circuit on the bench testing it, so technically D1 is infinity away from it? When installed, The diode is attached directly across the contactors terminals and a wire would be run back to the board for switching (about 4 inches)

I have not installed a bypass cap on the output side of the 7805 but I could certainly add one. Can't really hurt to add it...I'd initially thought adding one might cause the relay to delay turning off while it burns off the charge, but now that I've thought a little more about it, that won't be the case at all.

The voltage being output into the gate is 5.003v when the input is activated. There is a 10k resistor across the fet to tie the gate low. There is no decernable voltage between the gate and source when activated or not.

[Kim Christensen]

The voltage being output into the gate is 5.003v when the input is activated. There is a 10k resistor across the fet to tie the gate low. There is no decernable voltage between the gate and source when activated or not.

Well, that doesn't seem right. Something must be miswired. Is the "HV Sol" power ground tied to the 24V ground?

[DavidAlfa]

What's the contactor resistance?
Definitely, sounds like a isolated ground, you should measure something there!
Also that's why relays are still used a lot, they provide a simple method of isolation.

[lilshawn]

Well, that doesn't seem right. Something must be miswired. Is the "HV Sol" power ground tied to the 24V ground?

That's entirely possible. I'll be going over it again to be sure. Yes they share a common ground circuit. The main power supply feeds several voltages used by the whole system.

What's the contactor resistance?
Definitely, sounds like a isolated ground, you should measure something there!
Also that's why relays are still used a lot, they provide a simple method of isolation.

The contactors coil resistance is ~3 ohms. I know it seams really low, but its low resistance is not much of an issue, due to it's extremely low duty cycle...and a fairly high mechanical resistance it has to overcome. It only activates for about 300 milliseconds to overcome the spring tension and close the contact, a mechanical interlock toggles over to latch the output. A secondary relay coil is activated later to unlatch the contactor contact interlock to release the contact and open the circuit.

As a side note, my test setup involves not using the full voltages specified here, as I don't currently have a 40v supply handy, and I'm not currently able to take the machine down for testing.

As an alternative, I'm driving the "low voltage" 24v side currently with a 12v supply and my "40v"  is being substituted with a 18v supply using a couple incandescent bulbs in series as a "load" to indicate the switching. (Since it is a purely resistive load, there is no flyback diode attached to it) I've also tried measuring continuity through the FET with the VOM to check for a path through to ground when activated to see if it reads a "short" but nothing seems right Perhaps my testing methods are flawed... Perhaps my wiring is wrong. I will be double checking it to be sure.

Thanks for the help guys, it's much appreciated 👍🏻

[DavidAlfa]

If you have 5V between the gate and source, the MOSFET should turn on, end of story.

Take a new MOSFET, connect the multimeter in diode between drain and source, toggle the gate between 5V and gnd, the multimeter should show a short when 5V is applied, and become open again when setting it to gnd.


It might sound dumb, but are you sure the connections are correct?

[wraper]

I have not installed a bypass cap on the output side of the 7805 but I could certainly add one. Can't really hurt to add it...I'd initially thought adding one might cause the relay to delay turning off while it burns off the charge, but now that I've thought a little more about it, that won't be the case at all.

7805 must have a cap on its output, otherwise it may became unstable. Unless P40NF10L is fake or faulty, 5V gate voltage is completely enough.

[lilshawn]

7805 must have a cap on its output, otherwise it may became unstable. Unless P40NF10L is fake or faulty, 5V gate voltage is completely enough.

Sorry, to clarify I'd had a thought about doing things differently originally by switching on and off the 24v powering the 7805 with a 24v relay, after thinking about the logistics of it, and why I would even do that... (Sorry I'm scatterbrained AF and my thought processes go all over the place when coming up with ideas and solutions to a problem.) I decided to go ahead with with switching the output... The fact there was no cap on the output side is a leftover fragment of that initial thought process. I have since added one to the circuit when it was mentioned earlier yesterday.

The fets are good parts. I ordered them from mouser a while back. I use them to replace some nla driver FETs on pinball solenoid driver boards.

I have some other things going on today, but I will return to this later this evening to go over how the fet is hooked up. But in all honesty it could be something as stupid as my brain not understanding the pinout of the fet on the parts side up, vs the pinout when it's solder side up. Wouldn't be the first time. Thanks again for the fingers pointing in the right direction.

[DavidAlfa]

It's plain simple (Though that drawing is pretty terrible) 

[lilshawn]

after some diagnosing and testing, i went ahead and swapped the P40NF10L for another one. in my test setup i ended up in a state where was measuring 18v flowing through when nothing was applied to the FET gate.

I did find some splashy solder on pads around where the FET was installed after i'd pulled it out... (i'm using some prototype perfboard that for some god forsaken reason has square solder pads all over it making it fairly easy to bridge connections.) I may have had some stray solder messing things up from when I changed from the IRF540N to the P40... or it ended up there when I desoldered the P40 FET to change it to a new one.

either way, after cleaning up the pads around it and changing the FET out, to seems to be working now.

thank you guys for all your pointers! always something to learn!