Minimizing Solenoid Drop-out delay

[Kevman]

A friend of mine is in to drag racing. Big time.

Launching off of the line is accomplished by releasing a large solenoid on the transmission (the last of a series of steps). He's concerned that the solenoid is not releasing quickly and that is causing him E.T.

I know the solenoid will "hesitate" after the power flowing to it is removed as it takes time for the energy stored in the coil to dissipate. I thought that a diode would fix it by essentially shorting it out (not sure if it has one already), but I've read that flyback diodes actually increase the delay. Wikipedia says that it allows the current to "circulate," which I don't think I understand...

What would be the best method for getting rid of this energy? I thought about changing the relay feeding the solenoid to a double pole and having it short the positive lead to ground, but it seems that would be the same thing as a flyback diode?

I guess some sort of delay that decreases the voltage to the solenoid a second or so after power is applied, reducing current after its drawn in...? Is there something simpler?

Googling about this subject is pretty fruitless, most people want to introduce a delay, not reduce it.

[dom0]

Not clamping it at all is the quickest way the M-field will collapse. This is usually not practical due to driver limitations. So clamping to the highest voltage (with a Z-diode) that the driver can withstand will allow the quickest decay of the M-field.

See TE application note 13C3311 for a in-depth discussion of this topic.

[Ian.M]

The main issue that ultimately limits turn-off is what voltage the insulation can withstand.  In practical applications, its usually the voltage limit of the switching device, as without any clamping, if there is enough energy it is certain to break down due to the back EMF.  As you are using a relay, if it doesn't have an existing anti-parallel diode or snubber, I suspect its arcing across the contacts as soon as it starts to open and all the energy is dumped into that arc.   If its got a diode, wired directly across the coil, adding a resistor of up to four times the DC coil resistance in series with the diode may be the best option (assuming a 12V supply)  as that would allow the back EMF to rise to about 50V for faster turnoff than the very slow turnoff the directly connected diode gives

Reducing the holding current will certainly help.

[SeanB]

Dropping the current in the solenoid to the minimum holding current ( a lot lower than what is required to pull it in) and then using a clamp transorb that is just there to protect the switch transistor will drop the release time down a lot. You need a series resistor with the solenoid to reduce the current, and choose a transorb diode ( single or double direction) that is just below the Vceo rating of the drive transistor will reduce this time quite considerably. Resistor will just have to be able to handle the power dissipation, and you use either a large value capacitor across the resistor, or a timed contact closure, to enable the main solenoid to pull in reliably on power on.

[langwadt]

A friend of mine is in to drag racing. Big time.

Launching off of the line is accomplished by releasing a large solenoid on the transmission (the last of a series of steps). He's concerned that the solenoid is not releasing quickly and that is causing him E.T.

I know the solenoid will "hesitate" after the power flowing to it is removed as it takes time for the energy stored in the coil to dissipate. I thought that a diode would fix it by essentially shorting it out (not sure if it has one already), but I've read that flyback diodes actually increase the delay. Wikipedia says that it allows the current to "circulate," which I don't think I understand...

What would be the best method for getting rid of this energy? I thought about changing the relay feeding the solenoid to a double pole and having it short the positive lead to ground, but it seems that would be the same thing as a flyback diode?

I guess some sort of delay that decreases the voltage to the solenoid a second or so after power is applied, reducing current after its drawn in...? Is there something simpler?

Googling about this subject is pretty fruitless, most people want to introduce a delay, not reduce it.

look at it this way, you are turning  the solenoid "on" with the supply voltage, say 12V, with a fly back diode you are turning it "off" with 0.7V

to get the rise and fall time similar you need to let the voltage flyback to ~2x the supply voltage

lots of mosfets are rated for the energy they can handle in avalanche, i.e. clamping at their max rated voltage 

[Kevman]

Dropping the current in the solenoid to the minimum holding current ( a lot lower than what is required to pull it in) and then using a clamp transorb that is just there to protect the switch transistor will drop the release time down a lot. You need a series resistor with the solenoid to reduce the current, and choose a transorb diode ( single or double direction) that is just below the Vceo rating of the drive transistor will reduce this time quite considerably. Resistor will just have to be able to handle the power dissipation, and you use either a large value capacitor across the resistor, or a timed contact closure, to enable the main solenoid to pull in reliably on power on.

I have some 18V TVS diodes from car-powered project I did (tractor pulling distance measurer I posted about before). That would work, right? Would it be that much better than a standard rectifier diode?

I fear the coil may already have a diode embedded into it, in which case I don't think a TVS wouldn't work, because the other diode would clamp "beneath" it, right?  I'd have to  go the current limiting resistor with capacitor route.

[Kleinstein]

The voltage limit by the TVS should likly be higher than just 18 V. Depending on the driver / switch this could be more like 100-200 V - just not more than the switch or coil can sustain.
If there is a diode allready inside the coil, this will limit the voltage - so the external Zener / TVS would not work.

Lowering the current is not an alternative, but a different way to speed it up - you can use both together.

[max_torque]

Unless you are using some spectacularly massive solenoid, chances are the few milliseconds it takes to release are:

1) many times smaller than a humans reaction time
2) Swamped by other mechanical factors (such as the hydraulic response of the transmission system to which it is attached)


If you really want to minimise reaction time, cheat with a machine vision watching the Tree and direct control of the transmission............

[Tomorokoshi]

If you really want to minimise reaction time, cheat with a machine vision watching the Tree and direct control of the transmission............

Not cheating, engineering!

That would be a relatively basic machine vision / sorting problem. Train on the red, yellow, and green lights (or whatever is used). Have an lock-out that the operator enables at yellow. Then the machine is armed to trigger on green.

[Kevman]

The voltage limit by the TVS should likly be higher than just 18 V. Depending on the driver / switch this could be more like 100-200 V - just not more than the switch or coil can sustain.
If there is a diode allready inside the coil, this will limit the voltage - so the external Zener / TVS would not work.

Lowering the current is not an alternative, but a different way to speed it up - you can use both together.

My dad is going down there and he'll look for diodes and try to detect any in the coil. If there's one in the coil I'm not sure what he can do beyond limiting the current. 18V is all I had, I'm sure it'll be better than .7v on a rectifier.

Unless you are using some spectacularly massive solenoid, chances are the few milliseconds it takes to release are:

1) many times smaller than a humans reaction time
2) Swamped by other mechanical factors (such as the hydraulic response of the transmission system to which it is attached)

You'd be surprised. I've seen not-that-big coils take a long time to release (long enough to hear). Also consider that the margin between wins and loss is frequently down to 5ms or less. I've not the seen the coil in question here, but the one the transmission brake on my dad's Vega is bigger than a can of soda.

Also I don't think the transmission is hydraulic- its a Lenco, I think. But regardless the delay in the transmission, starting that process earlier will reduce the response time!

The reason why the response time of this coil is important was explained to me, and it made sense, but I don't remember, sorry. Note that this is all part of a system of events that happen during launch, so helping everything activate at the same time is a help. I *think* he detected the latency from the coil with his vehicle's data logging.

This is a very serious drag car- its a "Pro Mod." 3500+ Horsepower.