Thread Tools
Mar 24, 2012, 11:10 AM
Registered User
Thread OP
Discussion

Solid State Winch Solenoid


The Solid State Winch Solenoid
Since many of you are handy building stuff and are forward looking, early adopters, I thought I would describe a project that might be of interest. It is a non-mechanical solenoid for energizing your winch. It can be built either as a main solenoid, a safety switch or as a combined solenoid and safety switch.
It simply consists of 6 N Channel MOSFETS that do the switching, some stainless hardware a couple of strips of copper and a copper clad PC board. I found a box/enclosure at Radio Shack that fits well or there are lots of other options.
Benefits:
Extremely low activation energy (no current activating a coil)
No contact degradation and wear out
Extremely low resistance, low loss

The first version had 10 FETs and did not have a lot of mechanical strength so I broke some leads due to vibration when hauling the winch wagon over a dirt road. The design has migrated to fewer FETs and a sandwich construction using a PC board for strength. Works great.

The PC board solved two problems. It allowed the two strips of copper to be connected together mechanically so the leads do not fatigue and it allowed soldering the leads to the board. Ever try to solder to a thick piece of copper? It conducts the heat away so fast you need a torch or you can do it on the stove top. You have one shot and it better be right.

I have been running my 6 FET switch for about 6 months now and I fly about every 2 or 3 days on average with about 5 or 6 flights per session. The winch is a FLS on 12 volts launching 3.5 meter TD planes. Electrically it has been flawless. The only design change has been to connect the copper strips together through a PC board to eliminate lead stress.

The down side:
It does require rewiring your winch. But that is not hard.
The design has a minimum track record but so far, no electrical issues.
Additional diode across the motor as insurance (something you should do anyway)

If anyone is interested, let me know and I will describe some of the not so obvious details that I think are important as well as some pictures. Material cost is about 25 to 30 bucks per solenoid, not counting your construction time.

Mike Wilson
Sign up now
to remove ads between posts
Mar 24, 2012, 12:01 PM
that's gonna hurt...
Mike - more info please, and thanks for posting

Bruce
Mar 24, 2012, 01:45 PM
Registered User
Please provide the part numbers of the FET's, the rest I can do over here in the UK.
I have had a thought about this on occasions, but when the solenoid broke the other month I just bought a new one.

Very interested in this.

Foot switch via a resistor to pull up the gate to the battery voltage and hey presto we have a solid state switch.

I even thought about using an active pull up to help over come the gate capacitance.

Sounds good.
Jon
Mar 24, 2012, 03:03 PM
Registered User
Thread OP

FETs


Jon,
The electrical design is dead simple. I use a passive pull down of about 470 ohms or so and just apply full battery voltage when turning it on via the foot switch. Best to be fully on or off.
The FET's are made by NXP and are PSMN1R8-30PL in a TO220 package. You can look up the specs but they have about 1.8 milliohms Ron. Id of 100 Amps. Good price too.

The trick is the physical configuration. Here are some pics of what I came up with. The basic design is a single switch that can be operated by a foot pedal or an internal 12 volt battery/toggle switch if you just want it to be a safety switch.

The "killer configuration" is to stack two and have a combo safety switch and main winch switch all in one small package. A lot of copper.

I also use FFH60UP40S3 Diode across the motor for extra insurance.

Mike
Mar 26, 2012, 09:10 PM
I do this for fun!
Very interesting!

If it fails, is it more likely to fail open circuit or closed circuit?

If the answer is open cicuit always, that would be a great safety feature. I've had a solenoid fail closed and that wasn't any fun at all.

Ken.
Mar 26, 2012, 11:03 PM
Registered User
Thread OP

No Failures


Ken,
No failures means I don't have and data. But....

The design is done in such a way that really focuses on not stressing the parts. Stress is the cause of failures. The N Channel FETs are just like the ones in your computer. They are reliable because the current density and local heating is low. You computer has either a fan or thermal pipe tower or both to get the heat out because failures are a bad mark on the manufacturer and heat is the main cause. So lowering stress through "over design" and keep temps low is key. I have done hard launches and tried to feel the temperature rise. It is nearly nonexistent. Each of the 6 FET's can handle 1,000 amps for a very short time. They can each handle 100 amps continuously as long as you keep the temperature down. The heat is caused by the on resistance of the FET so selecting devices that have very low on resistance keeps the temperature down... and reliability up.
The N Channel FETs are not a single devices but hundreds of thousands of devices all in parallel. So the current density per device is low.
One possible stress mode is energy from the motor magnetics needing a place to go when the FETs are turned off. The voltage will spike and something breaks down. There is a parasitic diode in the FETs that will break down and it is designed to do just that. You just need to keep the energy under control. I have reduced this occurrence by installing a high current diode across the motor so it cannot produce voltages large enough to case avalanche breakdown of the FETs.
The whole reason for doing this FET project is to get around the chronic wear out and welding of contacts of mechanical solenoids. But..... as a belt and suspenders approach, I have put two switches in series so that I can easily turn things off in the event that there is a failure where the FETs short. The second unit is completely isolated so there are no common elements that will cause both to fail. But I believe the design philosophy should be to fix the design if failures occur and not band aid things with more parts.
Mike
Apr 03, 2012, 11:37 AM
Registered User

Schematic?


Thanks for posting this what a great job and neat project. Can you post a schematic that shows how it all goes together please? Have you thought about adding a 555 timer as a speed controller for launching smaller gliders? Speed control would be easy with this setup. Also why the separate battery to power the gates? Why not drive it from the motor battery? Also the 470 Ohm pull down resistor will be dissipating about 0.3 watts are you using a 1/2 watt resistor or with the short duration it is used are you finding a 1/4 watt works? Sorry for so many questions your post got my mind whirring.
Last edited by ebourlet; Apr 03, 2012 at 12:45 PM.
Apr 03, 2012, 07:09 PM
Registered User
Thread OP

A few answers


Ebourlet,
I will send you a schematic if you like as well as a bill of material. Send me your email address if you want it. If you want to build one, that will get you started but I am sure there will be a few questions.

Great idea about launching smaller gliders, that version is almost ready to test. I am using a 556 with one section as a 140 Hz oscillator and the other as a PWM that goes from about 30% to 100% duty cycle with a pot adjustment. The thought is that you can dial in the torque you want and then just use the pedal. A local guy has a very macho winch but wants to tone it down so I will probably be testing it in the next week or so. It would also be good for bringing in the line. A little safer. Bench testing looks good but need to try it on the winch.

The separate battery is only for convenience so you can stick it anywhere in the winch loop. I have mine running without a battery and it is easy to do but you do need to put the solid state switch in the negative leg just before the battery negative terminal and run a dedicated wire from the positive side of the battery to the foot switch and then to the solid state switch gates.
The safety switch version is done with a battery to make installation simple.
You can stick in anywhere in the loop. It can also be used without a battery if you are willing to put it in the negative leg and add some wiring

I also did the combo version (solenoid and safety switch) this way to make the safety switch COMPLETELY isolated from any possible faults that would make it fail. It is more of a sample of how it could be done if your are really worried about failures. I would rather think about how to make it more reliable rather than adding more stuff in case it does fail.

I am not even sure I used a 470 ohm or 1k ohm. It was something I had that was convenient. It is on the winch. Guess I should check to see what it really is. Either will be fine if sized properly for power. You do make a good point about power. I actually think that a larger resistor may help with limiting the kick back voltage from the motor although it will slow the turnoff and case some heating. I have not tried to stick a scope on the system when launching to see.... and fly my glider at the same time.

Mike
May 22, 2012, 09:28 AM
Registered User
If one of the FET's fails short circuit, it wont stay failed for long
Will probably ignite and turn to carbon very quickly.

I'm liking this thread.
I just wish that I had time to build my own
May 22, 2012, 03:16 PM
Registered User
Well this is the design that I have been looking into.
Its six ST FET's with low RDS on's.
470ohms to ground and apply the full battery voltage to the gates to activate the FET's.

The nice thing is that if one of the FET's fails short cct, it won't stay short for very long.
The smoke will escape and it will get hot before going open circuit.

The current spread between the six FET's will reduce the power consumption of each device (which you can't odd's).

Although the FET's won't all switch at the same time.
The charge slope of the gates will be different for each FET.
However they will be very similar and we are talking less the a few micro-seconds between them.

Any comments welcome, and of course I'm ready to be shot to pieces.

I have simulated this on Symetrics and on paper it works really well.

Some notes are on the attached PDF.

Cheers
Jon

edit, attachment removed as I have updated the circuit with an "Active" gate drive !
Last edited by ceres-syd; May 29, 2012 at 09:23 AM.
May 24, 2012, 09:09 PM
Detail Freak
target's Avatar
I'm still interested in this Mike, but also sorry for sitting on your sample for so long.
I'm getting the B winch I have back together slowly, but I really think that a macho TD winch would be a better test (More Amps and More tapping!)

G/L.
R,
Target
May 27, 2012, 09:04 AM
Registered User
Having some experience with such circuits, I advise caution!

The biggest risk is that the battery voltage realized at the circuit board after the wiring losses can drop below what is necessary keep the FETs fully turned on during pulsing of the winch motor. Pulsing of a winch motor represents repeated application of stall current to the FETs which can exceed 300 amps for a Ford long shaft motor with appropriate wiring and battery. If the instantaneous battery voltage realized at the gates of the FETs should fall below about 5 volts for even a few microseconds, the FETs will overheat and melt. The failure mode in this situation is shorting of the FETs which will result in destruction of the model on the line and possibly a fire.
The solution to this problem is to use a diode and large capacitor to store the voltage of the unloaded battery for driving the FET gates and incorporating a low voltage cutoff circuit to prevent operating the circuit with insufficient battery voltage.

There is a problem with parallel FETs in this configuration that can result in their oscillating due to something called Miller Capacitance feedback. The solution to this problem is to insert a small resistor of about 47 ohms in series with the gate of each individual FET. Again, the failure mode associated with this problem is shorting of the FETs in the on state.

Finally, the pull down resistance on the combined gates of the FETs appears too large to insure instantaneous shutoff. The gates of the FETs appear as large capacitors, the spec sheets state the gate charge in coulombs that must be drained from the FETs to turn them off. Failure to turn the FETs off quickly enough during pulsing can cause overheating and shorting of the FETs. The best solution to this is to use an active device, such as another FET to short the gates of the parallel FETs to ground through their isolation resistors when they need to be turned off.

The solid state solution to controlling a winch looks attractive at first blush, but nothing is really simple in this world! It is a law of nature.
May 27, 2012, 11:58 AM
It's a Boolean thing.
Miami Mike's Avatar
Good information there!

The power dissipated as heat by a transistor is the product of the voltage across the device and the current passing through the device. When the device is fully off the current is zero, and when the device is fully on the voltage is zero. In both of those cases the power is zero. But when the transistor is in a state where both the current and voltage are greater than zero then it dissipates power. In a circuit like this one, that state needs to be avoided as much as possible.
Latest blog entry: The Shepard Tone Vario
May 28, 2012, 11:20 AM
Registered User
I do admit that the gate drive solution of my switch design is very limited.
I am working on version 2 with an active gate drive using a push pull pair of transistors and a logic level FET to switch them.
The component count goes up a wee bit though, but its all do-able.
J
May 29, 2012, 03:05 AM
Registered User
I have made an update to the schematic diagram following some feed back.
The main change is to the FET Gate Drive.
The dischagre rate of the parallel switching FET's would be fairly tardy.

I have changed this to an active driver using a couple of transistors to drive the gate high and low and force the voltage change over a smaller magnitude of time.
This will eliminate the potential oscillations that may occur during switch on or off, and the potential undesired heating effect of the switch FET's (Where all the smoke and fire escapes from the 1/4" square area of silicon.).

Anyway please pass constructive comments and I will amend the schematic accordingly.

Hope this helps, and I must admit that I havent yet got past the simulation stage of this design, so please use at your own risk.

Cheers

Jon

P.S I have now deleted the second version of schematic and ameded it so the foot swicth works correctly.
Last edited by ceres-syd; May 30, 2012 at 02:21 AM.


Quick Reply
Message:

Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Sailplane Winch Solenoid Safety Buzzer Miami Mike DIY Electronics 7 Aug 16, 2019 01:14 PM
Mini-HowTo Solid state winch electronics circuit Jurgen Sailplane Talk 2 Jul 17, 2009 06:50 AM
Dual solenoid light check for winch ferradas Thermal 2 Oct 27, 2003 05:57 PM