starter solenoids

[strantor]

A while back I was working on my controller for a big DC permanent magnet motor. That project has fizzled out for the time being, but one thing I was considering at the time was making an H-bridge out of automobile starter solenoids and providing PWM via a single array of high powered low side mosfets instead of 4 arrays. While working on a family member's car I encountered this drawing:
View attachment 64027
According to this (I don't know if it should be taken literally or not) the start signal engages only half of the solenoid coil, after which the solenoid closes and the remainder of the coil is powered through it's contact. This also should result in the solenoid latching closed IMO, which doesn't happen. So I'm a little confused about how the solenoid works, and if this drawing is flubbed or not.

1. If I wanted to use solenoids as I described above, is it possible to bring the negative side out the coil out so that it's not a chassis ground anymore?

2. If I wanted to use solenoids as I described above, when I apply voltage across it, will I be applying voltage across the entire coil or only half of it?

3. Is one side of the coil in fact internally connected to one side of the solenoid contact, and if so, why does it not latch?

4. How much current can a starter solenoid contact sustain for long periods of time? (keep in mind, it would not be switching heavy loads a whole lot, just remaining closed)

5. How much current does a starter solenoid draw?

 

[tcmtech]

I would say that schematic is wrong or at least improperly drawn.

Relating to automotive starter contactors they come in several different configurations. Most factory issue ones are intermittent duty and will burn up their coils if ran continuously for more than a minute or two but many aftermarket ones have continuous duty ratings.

As far as connections go There are some that have just one coil terminal and some that have two coil terminals and others that have one terminal for the coil and the second one is a NO connection internally attached to the big moving contact that connects to the two big high current lugs when the contactor is energized.

As far as power handling most will easily take at least 100 - 200 amps continuous duty without issue with peaks well over 1000 for short term.

Relating to doing PWM control with them unfortunately no. They are basic on/off type mechanical switch's with slow response times and trying to use them in a fast make/break PWM operation would probably weld the internal contacts together in short order!

 

[ronv]

Depends on the solenoid. Give this a read.

https://www.electro-tech-online.com/custompdfs/2012/05/strtctr.pdf

 

[strantor]

Relating to doing PWM control with them unfortunately no. They are basic on/off type mechanical switch's with slow response times and trying to use them in a fast make/break PWM operation would probably weld the internal contacts together in short order!

I wasn't planning to PWM the solenoids. Just energize them in one direction or the other, and leave them there. PWM would be through them (as they remain closed) via a giant mosfet.

Something like this:
View attachment 64033

 

[tcmtech]

That would work just fine but I would suggest using a 600 - 1200 volt IGBT instead of a 100 volt Mosfet for a 96 volt motor.

 

[strantor]

That would work just fine but I would suggest using a 600 - 1200 volt IGBT instead of a 100 volt Mosfet for a 96 volt motor.

haha, yeah that would be sensible. Problem is I already bought the mosfet and now I need to use it. I might just run 72V instead, for a little head room.

 

[strantor]

Depends on the solenoid. Give this a read.

https://www.electro-tech-online.com/custompdfs/2012/05/strtctr.pdf

that's a good PDF. Thanks. Now I know that it would likely take considerable modification of the solenoid to make work this way. Also it will drain a good amount of power, which isn't good.

 

[tcmtech]

If you are running that high of current levels along with mechanical contactors 28 volts wont be anywhere near enough headroom.

Odds are you will blow right through that 100 volt Mosfet and destroy it the first time you use it.

 

[dr pepper]

Some marine diesel starters have 2 windings on the solenoid, there is a pulling in winding and a holding winding, as the name implies both windings are used to pull in the pinion and engage it onto the flywheel ring gear, and then when its full engaged a switch is made that disconnects the pulling in winding, leaving just the holding winding in circuit.

I assume the pulling in winding cannot be on too long as to get the field strength in the size available underated copper winding wire is used.

You could look into solenoids used for fork lift vehicles and smaller mobile plant, older vehicles used some pretty beefy contactors and solenoids, these older kind of vehicles tend to be at the end of their life now so salvage shouldnt be that difficult.

 

[ronv]

Do you already have 4 of the Big FETs? How big is the motor?

 

[strantor]

Do you already have 4 of the Big FETs? How big is the motor?

No I only have 1, and they cost 190$ each.

**broken link removed**.

Designed for 24-72 VDC battery operated equipment. Capable of 300 amps peak and 125 amps continuous. Totally open, fan cooled motor.
Power: 8Hp continuous & 18 Peak Hp

Those are the specs provided by the manufacturer, but I have found several instances where this motor has been run @96V, up to 400A peak and 200A continuous, especially when forced air cooling is employed.

 

[dr pepper]

I dont know if its any interest or relavance, however older electric vehicles controlled speed and directionby rotating the brush holder plate, easier and cheaper, but not necassarily more efficient.

I have a IGBT module right here, its a double transistor, and is rated at 200amps per transistor, they are 100 quid each.

SKM200 by semikron.

 

[ChrisP58]

You can accomplish the same thing as the marine solenoid mentioned in post #9 by reducing the voltage feeding the coil a few mS after the contacts are closed. Dropping the voltage by 75% will reduce the power in the coil by almost 90%

I have often used this technique with relays by putting an RC circuit in the positive end. The resistor is about 3 times the coil resistance, and the capacitor is enough to hold most of the applied voltage a little longer than the transit time of the relay.

 

[strantor]

You can accomplish the same thing as the marine solenoid mentioned in post #9 by reducing the voltage feeding the coil a few mS after the contacts are closed. Dropping the voltage by 75% will reduce the power in the coil by almost 90%

I have often used this technique with relays by putting an RC circuit in the positive end. The resistor is about 3 times the coil resistance, and the capacitor is enough to hold most of the applied voltage a little longer than the transit time of the relay.

Like this?
View attachment 64056

 

[ChrisP58]

Almost. Move the mosfet to the other side of the coil, with the source pin to gnd, and the drain pin to the coil. That way you can easily drive the gate from 5V logic.

You will also need a diode across the coil to catch the inductive spike from when you turn off the coil current from killing the mosfet. The cathode of the diode goes to the positive end of the coil.

 

[strantor]

oops crap I put a n-channel. meant to put a p-channel, because I'm still not sure if the coil end will be connected to chassis internally. If it is, then I won't be able to switch low side.

 

[tcmtech]

Just do the whole thing in IGBT's and be done with it.
This is how I get my big power switching devices cheap.

1200 volt 600 amp $49.95.
https://www.ebay.com/itm/PRX-IGBT-C...052?pt=LH_DefaultDomain_0&hash=item2a11b2409c

If you are willing to watch the auctions you can get even better deals. A while back I picked up 6 of these brand new still in the factory packaging for about $30 each.

 

[dr pepper]

Thats a proper tranny that is.

IGBT's have a slightly diffo switching characteristic than mosfets, but they seem popular for serious amps.

I look after a couple of machines with 200a dual versions as mentioned, and they are reliable, last time one failed it did funny things, it'd mis trigger and give a low output but then when warmed up worked ok.
Usually when they go theres just the metal backplate left.

 

[strantor]

When I first started looking into this last year, I spent countless hours researching material on the topic. I've forgotten most of it already. I think the reason why I went with the mosfet module over an IGBT is because initially I was planning to parallel several smaller (TO-247) devices to achieve the Amp capacity that I need. IGBTs have a positive temp coefficient and I read that trying to parallel more than 6 of them in discrete packages is an exercise in futility. So, I was all about MOSFETS. I quit reading about IGBTs and spent weeks reading about mosfets. Then I ended up buying a semiconductor module, which negates the original reason for choosing MOSFETS in the first place. The only thing I have going for me now, is that 96 (or 72V) is a little lower than the generally recommended (according to white papers) 250V cutoff voltage for IGBTs.

The IGBT technology is certainly the device of choice for breakdown
voltages above 1000V, while the MOSFET is certainly the device of choice for device breakdown voltages below 250V.
Between 250 to 1000V, there are many technical papers available from manufacturers of these devices, some preferring
MOSFETs, some IGBTs.

I read else where that the losses due to slower switching speeds and, forward transconductance losses being > than Rds(on) losses, the MOSFET looked like a better option.
All that being said, right now I'm wishing I had some big IGBTs and just be done with it. The IGBT you linked to has more capacity than the solenoids I've been looking at. Maybe I'll get some....decisions, decisions...

 

[shortbus=]

Just do the whole thing in IGBT's and be done with it.
This is how I get my big power switching devices cheap.

1200 volt 600 amp $49.95.
http://www.ebay.com/itm/PRX-IGBT-CM...ifferent things from them in last year or so.