Hi - I'm a newbie to this forum, so forgive me if I don't get the etiquette quite right on this first post.

I searched the forum for related topics, but didn't quite find what I was looking for.

The venerable Bob Pease drew up this little circuit to drive a solenoid at reduced voltage after an initial "spike".

Paul Rako put together the OrCAD drawing and posted it along with the hand drawn sketch below.

If I'm understanding the concept, the idea is that at "switch on" the C1 capacitor has no voltage drop, hence the solenoid receives the initial "spike or hit" at full voltage of the source. Then after maybe 1/4-1/2 second, the capacitor is fully charged, current drops to zero, and the R2 resistor now divides the voltage and all current flows through the resistor. It seems that the transistor Q1, resistor R3 and control voltage V2 are simply usedt to"switch" the solenoid off and on.

I'm a very bare minimum kind of guy, so I just want to understand how to make this circuit as simple as possiblef for my application.

For my situation, I have a Clippard ET-2M-12 solenoid that opens and closes a valve (to control CO2 flow).

The solenoid has a nominal resistance of 220 ohms, so with a 12V source it draws about 0.055 amps, and at that voltage consumes 0.67 watts. In order to extend the life of the solenoid, Clippard recommends a "hit and hold" solenoiddriver - one that will supply the initial full voltage for minimum of 20-25ms and then drop down to 6V.

I have a 14V, .2 amp wall wart that I plan to use for this application. Also, I don't think I have need for the transistor portion of the circuit since I plan to use a simple mechanical timer that will energize the wall wart in the morning, and turn off at night.

In line with the Paul Rako article in Engineering Design (https://www.electronicdesign.com/te...0/whats-all-this-solenoid-driver-stuff-anyhow),
I have come up with the following:
R1, solenoid resistance 220 ohm, L1 not measured, but assume it's around 2000mH.

So if I want to reduce steady state voltage by half (and thus reduce current on the solenoid by 75%), I will use R2 of 220 ohms, C1 of 470uF,. Now since the circuit will not need the transistor, I wonder if I need the flyback diode D1 at all?

It seems that for my very simple CO2 operation, I would only need to use C1 and R2 in parallel and not require the transistor or diode at all. 14V 0.2 amp wart controlled by a clock mechanical timer would be the rest of the circuit. Does that sound right, or am I missing something important that's going to fry my mouse solenoid!?

Kind regards,
Gregg G.



h.

 

Looks like it should work OK. Probably don't need the diode but I would leave it installed

 

The diode will minimize arcing of (and possible damage to) the timer contacts, so I would leave it in as sghioto suggested.

 

You might need to adjust tghe resistor value but certainly the circuit should work. I suggest checking the temperature of the wall wart after an hout to verify that it is OK for constant duty. It might be that you would need another capacitor across the 14 volts from that wall wart, if the voltage tends to drop excessively so that the valve does not move correctly.

 

If the solenoid inductance is 2H you might need to have C1 > 470u to give the solenoid current time to rise sufficiently.

 

However, If you use transistors you can use a much smaller capacitor and better pulse width control.

 

The TS did express a strong desire to have as few components as possible, probably because such an assembly would take quite a bit of effort to put it into a durable package. And if it will be used in any sort of business environment then it will need to be able to survive contact with a lot of unknowing and uncaring individuals. Not all of the participants here are hobbyists.

 

Here's an LTspice simulation showing the solenoid current response using a 470μF capacitor.

 

Here's an LTspice simulation showing the solenoid current response using a 470μF capacitor.

View attachment 221215

OK, that is close. Probably 330 ohms across the cap would drop the current to half, or a bit less. The plot does not show dropping below about 32 mA , but that is still a good drop. So now the TS has the information needed to make a choice. For a longer pulse a higher value cap would be required, so probably an examination of the valve response time is in order.

 

The TS did express a strong desire to have as few components as possible, probably because such an assembly would take quite a bit of effort to put it into a durable package. And if it will be used in any sort of business environment then it will need to be able to survive contact with a lot of unknowing and uncaring individuals. Not all of the participants here are hobbyists.

Not sure which post you are referring and not sure what "hobbyist" has to do with it, but if post #6, I simply wanted to give the TS an idea of what would be given up without the transistors.

"And if it will be used in any sort of business environment then it will need to be able to survive contact with a lot of unknowing and uncaring individuals."

I'm also well aware of users and these types of environments..

 

Not sure which post you are referring and not sure what "hobbyist" has to do with it, but if post #6, I wanted to give the TS an idea of what would be given up without the transistors.

Referencing post #6, with the transistors., and also line 9 in post #1, about minimum parts.

 

Referencing post #6, with the transistors., and also line 9 in post #1, about minimum parts.

"It seems that for my very simple CO2 operation, I would only need to use C1 and R2 in parallel and not require the transistor or diode at all. 14V 0.2 amp wart controlled by a clock mechanical timer would be the rest of the circuit. Does that sound right, or am I missing something important that's going to fry my mouse solenoid!?"

The TS has asked if missing something important, it may not involve frying the solenoid, but post #6 might still be of interest to the TS. If not, it can be ignored.

 

Thanks for the replies! I'm glad to hear that it will work in concept. The components should be here on
Wednesday, so after I get it put together I'll post on how it works out.

I think I will put together the "simplest" circuit first, and use a 330 ohm resistor. I guess I'm surprised with the
simulation results since I expected 1/2 drop across the resistor and 1/2 drop across the solenoid (which is
measured and spec'd at 220 ohm). I will study more to understand what's going on.

eetech00 (post #6) your design looks like the Cadillac! If it's not too much trouble, would you post a few words
with some details of the theory? If I'm reading it correctly, the transistor and all the other resistors are controlling
the initial pulse, and then the "hold" has all the current going through R6. Is that correct? I wonder why in your
design, a 218 ohm R6 gives just about 1/2 current at the "hold", as compared to my simplistic design?
When the C1 capacitor is "fully" charged, the Q1 transistor switches off, yes? What is the function of the
M1 mosfet? When I have better soldering skills and a little more practice, I'll try to build this circuit as well.
I'm a retired chemical engineer trying to learn some new things to use in my hobbies.

MisterBill2 (post #9) - Clippard tech support for the solenoid valve suggests a 20-25ms pulse, so this looks just about
right as far as the capacitor size goes.

Thanks again!

 

Now I see why the sim from MrBill2 only drops down to 33 mA - because I'm using a 14V wall wart.
If I use a 12V supply, I think I could use a 220 ohm resistor. With a 14V supply I'll have to use a 300-330 ohm
resistor as MrBill2 says.

I am trying to set up ngspice on my Debian box so I can catch up with you sim guys!

 

Here's my sim with a 300 ohm resistor, which gives 6V across the inductor (yellow trace) with an inductor current (green trace) of about 27mA after the initial pulse:

 

eetech00 (post #6) your design looks like the Cadillac! If it's not too much trouble, would you post a few words
with some details of the theory? If I'm reading it correctly, the transistor and all the other resistors are controlling
the initial pulse, and then the "hold" has all the current going through R6. Is that correct? I wonder why in your
design, a 218 ohm R6 gives just about 1/2 current at the "hold", as compared to my simplistic design?
When the C1 capacitor is "fully" charged, the Q1 transistor switches off, yes? What is the function of the
M1 mosfet? When I have better soldering skills and a little more practice, I'll try to build this circuit as well.
I'm a retired chemical engineer trying to learn some new things to use in my hobbies.

Not really a cadillac. That would be a circuit with a monostable timer chip.

Ok...I'll explain.
When power is applied thru the timer contact, C1 begins to charge thru R1. Q1, an NPN with 12v at its base, turns on and completes a ground connection thru R4 to the gate of PMOS M1. M1 turns on (bypassing R6) and provides a direct current path to the solenoid. The solenoid then gets the full 12 volts across its coil and immediately energizes.
As C1 continues to charge up, the voltage at the base of Q1 gradually decreases at a rate determined by [R1|R2R3]C1. As C1 continues to charge, the voltage at the base of Q1 eventually reaches its 0.7v threshold and Q1 turns off. When Q1 turns off, M1 turn off and removes the bypass of R6. Current now flows from the 12v supply thru R6 and the solenoid. The solenoid DCR and R6 form a voltage divider, so the voltage across the coil is about half that at full voltage.
When the timer contact is opened, the circuit is reset and ready for another cycle.

Sorry if it takes a couple of transistors, but I wanted to show that a smaller cap could be used and better control of the solenoid "on" pulse can be achieved if the transistors were retained. Basically, R1C1 and Q1 form a simple comparator.

 

crutschow - Thanks for the sim with 300 ohm! I almost have ngspice/eeshcema working. I would like to be able to reproduce
your results (for learning purposes). If the capacitor is resized to 750 uF, that should raise the voltage a bit and maybe widen the pulse, yes? I will try to simulate those conditions.

eetech00 - Thanks so much for the detailed description of your circuit! I see you get a "square wave" voltage pulse of just the right width. This is very interesting and I will plan to also reproduce your sim with ngspice. I'll put together the simple
circuit to get things in operation, and then build your circuit as my assembly skills improve. I'm in Washington state, near Olympia.
The rainy weather is coming and I'll have lots of "inside time" very soon.

 

If the capacitor is resized to 750 uF, that should raise the voltage a bit and maybe widen the pulse, yes?

It will widen the pulse but not change the peak significantly.
Sim below:

 

I don't know if I can find the circuit, but years ago I made a solenoid driver for a refrigerant recovery unit that used a 2 input NAND Schmitt trigger gate that on enabling, applied full power to the coil, then after a short time, went to PWM for holding current.
It worked very well.

 

that on enabling, applied full power to the coil, then after a short time, went to PWM for holding current.

Below is the simulation of such a circuit, but it's not really needed here since it would only save about a quarter watt of power over using a resistor to limit the current.