Solenoid Drive

Go To Last Post
35 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Greetings all

I made a little circuit to drive a solenoid which grips cable in a machine.

It is a replacement for the manufacturers part, so I can't change the inputs or anything.

The circuit is inspired by the manufacturers circuit, which is very unreliable and complicated (I don't even fully understand it). My version is simplified, but I have used their ideas, such as running on un-filtered rectified DC. (just in case you were wondering why I would do such a thing...)

The circuit works OK, the solenoid grips the wire correctly etc...

BUT...

The mosfet gets quite hot just sitting idle, leading me to think that perhaps it is not getting switched off properly.

The voltage on the Gate is normally 0V, going up to 5V when triggered.

If you are familiar with mosfet circuits, and can spot the problem, please let me know!

Thanks!
Pete

PS: This is a work project, so any suggestions over the weekend can't be tested until Monday, so don't think I'm asking for your help then ignoring you! All expert tips are appreciated!

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

TI makes a solenoid driver chip that tuns on at 100% pwm, then reverts to hold mode with 50% pwm. Clever?

Imagecraft compiler user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

bobgardner wrote:
TI makes a solenoid driver chip that tuns on at 100% pwm, then reverts to hold mode with 50% pwm. Clever?

Yeah, but too clever for this application!

My circuit is not heating up when holding, its heating up when sitting idle.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It might be that the Opto is not in full saturation. It might have a about a volt of Vce. I would think a volt meter would help determine this, or you can keep wondering why.

There are many possible improvements, however in your desire for simplicity you are on your own.

It all starts with a mental vision.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

KitCarlson wrote:
It might be that the Opto is not in full saturation. It might have a about a volt of Vce.

Oh, good point.

It measures 0.08V under resting conditions.
It doesn't seem much, but it's not zero...

Latest:
If I short out the Collector-Emitter of the Opto, then the mosfet stays stone cold.
So it appears that that little voltage is slightly switching the mosfet on, and generating heat.

Thanks for the pointer!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The mosfet should be rated for 600V at least. What mosfet are you using?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:
The mosfet should be rated for 600V at least. What mosfet are you using?

Ummmm... 250V 8A (FQPF9N25C RS Part #671-5329)

I sort of thought that as long as it was rated a bit more than the actual voltage I would be OK.... :oops:

Is this likely to cause the problem I described, or just shorten its lifespan?

Thanks!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

About the 0.08V, I would think there is also a peak component. A scope is often used to view that.

A fast recovery diode on the coil to freewheel the release current my help reliability.

A zener or other means to limit the MOSFET gate to source voltage to specification may also help reliability.

It all starts with a mental vision.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

petenz,

Are you sure you have the FET hooked up correctly? Its source pin should be connected to Ground. There is an internal diode in every HEXFET which normally does not interfere with circuit operation. But if you get the polarity of the source & drain mixed up, it can play havoc with your circuit. It is a good idea to draw this "parasitic" diode as a part of your schematic symbol for the HEXFET - it acts as a reminder to you and will keep you out of trouble with it. (See the schematic symbol on the data sheet's Page 1 for the polarity of the diode.)

Also, recheck the wiring of your bridge rectifer/diodes to make sure it is actually connected as your schematic indicates (which is correct). If you have the bridge hooked up in some odd way, your circuit may appear to work,but in a way you did not intend.

A good maxim to keep in mind when playing in the Wonderful World of Electronics:

"You can be Right for the Wrong Reason many times in a single day."

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

KitCarlson wrote:
About the 0.08V, I would think there is also a peak component. A scope is often used to view that.

A fast recovery diode on the coil to freewheel the release current my help reliability.

A zener or other means to limit the MOSFET gate to source voltage to specification may also help reliability.

Thanks, I will check it with the scope on Monday, and try the mods you suggested.

Cheers
Pete

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The diode on the coil not only may help. It is mandatory and it's absence may explain part of your problem. The coil seems to have a serious inductance if it's powered without capacitive filtering.
The other part of the problem may reside in the poor transfer ratio of the optocoupler or incorrect driving of it's led.

Dor

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Why not have the opto's output in SERIES with the gate rather than parallel? (Pin 4 to R3 as is now, pin 3 to R5)
The EC junction will have several 100 mV across it as you have noted.

John Samperi

Ampertronics Pty. Ltd.

www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

220VAC has 325V peak voltage. You need 400V FET there.

When the FET is supposed to be on, what kind of waveform it has on its gate? What if gate voltage drops to zero at zero crossing, and has to rise again during rising sine wave? I would at least have some kind of diode+capacitor network to provide the 20V supply, so the FET can be turned on even during zero crossings.

And yes the protection diode over the solenoid would seem like a good if not necessary idea.

And what kind of current would be flowing through the solenoid (or FET)? 1A with 0.08V means 0.08W, 10A with 0.08V means 0.8W.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thanks guys, now I have a few things to work on when I get back on Monday. Will let you know the results...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
When the FET is supposed to be on, what kind of waveform it has on its gate? What if gate voltage drops to zero at zero crossing, and has to rise again during rising sine wave? I would at least have some kind of diode+capacitor network to provide the 20V supply, so the FET can be turned on even during zero crossings.

I too think the problem lies there. The FET is modulated by the 100/120Hz of the rectified mains. The current through that solenoid does not stop suddenly so the FET still has switch a current while it slowly goes off. With a purely resistive load the problem would probably go away.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Shouldnt there be a cap somewhere? On the 20V line maybe?

Imagecraft compiler user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It seems all the voltages measure are average voltages, whereas the FET will respond to peak voltages.

Charles Darwin, Lord Kelvin & Murphy are always lurking about!
Lee -.-
Riddle me this...How did the serpent move around before the fall?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The original circuit was probably more complicated for good reason. 'Simplfying' it really hasn't helped the cause has it?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I seem to have been told off twice for using the word simplicity.

I don't mind adding extra parts to make a better circuit, I just meant that I didn't want to copy a circuit that I didn't understand, and that I couldn't really even decipher (2nd generation fax copy...)
Better to start with the basic building blocks, I thought.

Latest things I have tried:

Cap on the 20V Line.
Diode across coil

No improvement in either case.
(but will leave them in as they are there for a good reason!)

In the course of attempting to scope the output of the opto, there was some slight blowing-up of the bridge rectifier. Might see if I can dig out an isolating xfmr...

So I will now order some new bridge rectifiers, and correctly-rated mosfets to carry on with.

Thanks for all the suggestions.
Pete

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
some slight blowing-up of the bridge rectifier

I would've thought it went up like a Kiss concert.

As for a circuit you don't understand, why not post it here - there might be some educational value to be had.

Also, there's off the shelf modules called 'brake rectifiers' that will probably do something similar. They're used to switch the big solenoid on motor brakes.

Is there any reason you can't just use a relay?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:

Is there any reason you can't just use a relay?

Originally I started off with an off-the-shelf solid state mains-switching relay. it triggers brilliantly, but it won't let go, (I think) because the rectified ac doesn't go below zero, to tell it to stop conducting.

I had discounted the DC ones, cos they are really expensive, but if you total up the amount of time I have spent on this so far, they are starting to look like good value...

Will tidy up the circuit and post it, in case anybody wants to see it - give me a few mins...

Thanks
Pete

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Here's the original...

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yikes. I dont understand that one too well either.

Imagecraft compiler user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

bobgardner wrote:
Yikes. I dont understand that one too well either.

Phew, so its not just me! :D

The Boss has given approval to get a DC Solid State Relay, so hopefully that will be a quicker way out!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Huh, why just not connect the AC solid state relay before the rectifier? Solenoid sees rectified AC, solenoid sees AC..

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Note the use of a varistor. There's also what looks to be an active clamp so when the mosfet turns off, V1 turns on channelling the energy.
where does the 110vac come from?

In terms of a relay, I mean an electromechanical one. Somehow i think the electronics are necessary to control the back emf of the solenoid. Ifyou redraw the schematic to highlight the different sections it should make it easier to understand. Broken down into its building blocks, its reasonably simple.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It is hard to read the schematic, have to guess the connection dots and pass overs. Looks like the D2 and R10 (MOV) protect the MOSFET. Most of the rest is a power supply for the gate trigger circuit. A linear, direct off the line supply may be troublesome even for this low current application.

I would think a SS DC relay, would be less expensive than a redesign and even the individual parts to do it right.

It all starts with a mental vision.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I think your redesigned & simplified HEXFET circuit will work OK. But I think you may not have the level of expertise to debug it. What is level of experience with HEXFET circuits like this one?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

petenz123 wrote:
Here's the original...

When the input is on (5 volts), V3 and V2 are turned off, V4 is on and keeps base and emitter of V1 at ~50V pulsed DC and this keeps a steady 24Vdc on C1 (that is the power supply of the gate driver). There is no current in coil and this was the simple part.

Now the hard part.
When the input goes off (0 volts - is that sure? there are no short pulses?), V3 opens with the base current supplied by R5-C2( RC time ~0.15s), V2 goes on, current though coil is established, voltage on V2 drain goes to ~zero and V1 doesn't supply anymore the driver power line which goes lower than 24Vdc. When the voltage on C2 has raised at the same point where the driver supply voltage went lower (after ~0.15 seconds in the first instance), V3, V4 and V2 goes off and V1's base is now at 196Vdc feeding faster the driver's power line and opening again V3, V4, V2 and so on, but doing for now a high speed PWM.
As conclusion, IMHO it seems that when the input goes off, the coil is fully energized for 0.1-0.2 seconds and after this time the driver do some PWM to keep the average coil current and temperature lower. If this is the case, a Solid State Relay (does it work with high inductive loads?) will overheat the coil compromising safety maybe.

Dor

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Puzzled by the term

Quote:
slight blowing up

Seems similar to

Quote:
slightly pregnant

My mind says that either it has blown up or it has not. Blown up means "exploded" and no longer working (more or less, some times less than more, some times more than less).

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Dor, excellent circuit interpretation and conclusion.

It all starts with a mental vision.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Right, here we have a simple, cheap, working solution!

An electromechanical relay with a few supporting components.

So, why didn't I do this to start with?
Because I wanted to make a plug-in replacement for the existing PCB.

You will notice that this requires a 24V line, so I have to run a couple of wires from the logic board.

The 5V trigger signal is not enough to trip a relay, so I need a low-voltage supply for the driver transistor. It can't come from this circuit, as it needs to be isolated.

I'm quite happy to abandon my design spec of plug-in replacement, and change it to "plug in replacement and run a 24v wire!"

So, a massive thanks to everyone who added suggestions and comments.

As you can probably tell, I'm not very familiar with FETS (although I'm starting to use them in my low voltage stuff, they certainly have advantages), and I hate anything to do with the mains!

Thanks all !
Pete

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

petenz,

One performance parameter I don't think you mentioned so far is the duty cycle of this solenoid. How often is it powered and for what period of time? That is, frequency and period, off-time/on-time, activations per minute, etc, however you might express it for this application.

I am a big proponent of electromechanical relays for reasons of simplicity, but they do have their limitations and deficiencies.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Chuck-Rowst wrote:
petenz,

One performance parameter I don't think you mentioned so far is the duty cycle of this solenoid.
I am a big proponent of electromechanical relays for reasons of simplicity, but they do have their limitations and deficiencies.

It would vary depending on the customer's use, but probably ON for about 10 seconds, then OFF for 10 seconds - but it would be unlikely that this duty cycle would be kept up indefinitely.

I think the easiest thing to do is send the machine out for testing with the relay-based driver in it. If it blows up they will be sure to tell me... :-)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The relay should be OK. Just make sure you put a flyback diode across the solenoid coil, or the relay contacts will burn up.

I like your idea! Outsource your QC department to your customer - just like the Chinese.