Solenoid driver

[earckens]

In addition to your circuit's elegance: accuracy, simplicity, and low cost, there is another advantage: the NMOSFETs intrinsic diode from drain to source will catch negative voltage swings (kickback) from the solenoid coil.

spec

Great, I am impressed by all yours contributions here! Thank you so much!

 

[chemelec]

Wow, great! Thanks a million.

Just what do you suggest to use for the MOSFET? I happen to have IRL520N and IRLB8721 in stock here..

Almost any Mosfet should be OK.
The IRL520N should be OK.
Considering the Short On Time, Doubtful you should need a Heat Sink on it.

 

[earckens]

Not With them, But Instead of those Other Circuits.

Hi chemelec, so no large capacitor to be discharged over the solenoid coil?

And what if this capacitive discharge helps to get the coil core to move? How should/can/may this capacitive unit be integrated in your design?

 

[chemelec]

No Capacitor needed, Just the Circuit as Shown.
Mosfets have a Very Low On Resistance, Supplying High Current, Unlike Transistors that have a 0.6 Volt loss.

(Changing the 47uF cap to a 100uF cap will give about 1 Second On Time.)

And Switches can be added in so you can Select Various Solenoids, if needed.

If you weren't so far away, I could have sent you a circuit board.

 

[earckens]

No Capacitor needed, Just the Circuit as Shown.
Mosfets have a Very Low On Resistance, Supplying High Current, Unlike Transistors that have a 0.6 Volt loss.

(Changing the 47uF cap to a 100uF cap will give about 1 Second On Time.)

And Switches can be added in so you can Select Various Solenoids, if needed.

If you weren't so far away, I could have sent you a circuit board.

So basically the whole concept of capacitor discharge over an inductive load in order to increase the dissipated energy in a coil can be thrown overboard because the inherent advantage of this capacitive discharge was offset all the time by transistors getting in the way of an efficient current flow, and is now being compensated by a clean very low impedance switch?

I can tell you there will be some old and hard habits to die in model railroading electronics

I am learning to use Eagle (thanks to some very nice inspiration from spec) so if you have some Eagle file starter concept for your circuit I would be very obliged. If not I will work on this and forward you the result. But then I will ask lots of patience from you, I am just beginning to use Eagle (and Kicad).
Erik

 

[spec]

Why? Excuse my ignorance, I had my studies finished 30 years ago

Here are the advantages of MOSFETS:

(1) MOSFETs have an infinite power gain: you just place a voltage on the gate, of between 600mV to 12V, depending on the particular MOSFET, and the MOSFET can conduct up to around 50A, in practical terms. No current flows into the gate. Note that this is at DC. Because of the MOSFETs large effective input capacitance, you do need to provide more and more current to turn the MOSFET on and off fast. But once the MOSFET is on or off, no current, practically speaking, is required.
(2) When a MOSFET is turned on the MOSFET forms a a resistance between its drain and gate and the current can flow in either direction. Some MOSFETs can have a drain to source resistance as low as 3 mili Ohms. This relates to a forward drop of only, 3mR * 50A = 150mV. No other, reasonably priced, semiconductor device could match this.
(3) MOSFETs tend to have a better safe operating area (SOA), then bipolar junction transistors (BJTs). SOA indicates a devices ability to handle voltage and current at the same time.
(4) MOSFETs are incredibly fast compared to bipolar transistors, to a first approximation. But the parasitic capacitances of MOSFETS: Cds, Cdg, and Cgs, means that you need to use special circuits to realise the MOSFET's inherent speed. Cds can be especially significant, depending on the circuit configuration.
(5) MOSFETs have a huge parasitic diode from drain to source. The drain/source diode has similar voltage and current ratings to the MOSFET. The drain/source diode can be useful and other times it can be a nuisance. To block current in two direction a reverse blocking configuration is need where two similar MOSFETS are connected in reverse series.
(6) It is generally easier to connect MOSFETs in parallel to increase current handling.
(7) Because MOSFETs are so widely used, especially in switch mode power supplies (SMPs), they have the advantage of production scale, and are thus relatively cheap. This is in complete contrast to the situation up to around 1995.
(8) An unbelievably wide range of MOSFETs, to suit all applications, are available, from a large number of manufacturers.

Here is a link to the International Rectifier IRFPbF3206 which will give you a feel for a nice traditional high power NMOSFET:
https://www.infineon.com/dgdl/irfp3206pbf.pdf?fileId=5546d462533600a401535628d64a1ff0
There are also PMOSFET versions, but PMOSFEts tend to have double the Rdss.
More modern MOSFETs have lower gate drive voltage requirements.

The above covers medium and high power MOSFETs, but there is another general group: junction FETs (JFETs) and small signal MOSFETs.
These handle currents from 1mA to 500mA and are designed for high speed to extremely high speed operation. They have very low parasitic capacitances.
**broken link removed**

spec

 

[chemelec]

A Good Mosfet, but a bit of an Overkill for your needs. (IRFPbF3206)

On the Circuit I Posted, it should also have a Supply, Bypass Capacitor between Pins 4&8 and Ground.
Such as a .1uf .

Probably Not Needed, but you could also add another 100K from Supply to the Junction of the .1 and the Trigger Switch.
(Would be better, If you need to trigger this Multiple Times in a short period of time.)

 

[spec]

A Good Mosfet, but a bit of an Overkill for your needs. (IRFPbF3206)

I wasn't suggesting an IRFPbF3206 for this application: you could probably swing full-size Canadian Pacific railway points with that.

spec

 

[earckens]

A Good Mosfet, but a bit of an Overkill for your needs. (IRFPbF3206)

On the Circuit I Posted, it should also have a Supply, Bypass Capacitor between Pins 4&8 and Ground.
Such as a .1uf .

Probably Not Needed, but you could also add another 100K from Supply to the Junction of the .1 and the Trigger Switch.
(Would be better, If you need to trigger this Multiple Times in a short period of time.)

Duly noted. I will add to the drawing. Thanks,
Erik

 

[earckens]

And Switches can be added in so you can Select Various Solenoids, if needed.

How would you select various solenoids using this circuit? I cannot think about anything but a rotary switch but that is not feasible from an operator standpoint. Each turnout is commanded by a SPDT switch, the pushbutton in your schematic would command a pulse to a solenoid. Now how can this pulse be distributed to more then one solenoid coil without a rotary switch?

 

[spec]

I am not answering for chemelec, but this is how I saw the control of the various points in a model railway set up:

(1) You have a bank of leavers, like the levers in a signal box. Each leaver controls just one of the points.

(2) The leavers operate exactly the same as the leavers in a real signal box.

(3) Only a single chemelec circuit would be needed and the routing to the various points would be handled by switches.

(4) You could even get sophisticated and introduce some logic so that there were safety interlocks, once again, just like a real railway. An Arduino could make it really felible.

(5) A simpler set up could just comprise an array of toggle switches where the position of the switch leaver defined the position of the points.

spec

 

[spec]

As an aside to this thread but related:

At one time, the company that I worked for had quite a few contracts with British Rail and the Highways agency: LED display systems (LEDs were the latest thing), processing, communications, and I got the job of designing and building a huge model railway and road layout to advertise/demonstrate our equipment to the customer and at trade shows.

It was an enjoyable couple of months as no expense was spared and I had a completely free hand, including access to the services of the machine shop, carpentry shop, plating shop, paint shop etc etc.

The model had rolling plains, grazing livestock, mountains, roads, tunnels, bridges and, of course, railway track. And the rolling stock and trains were the best (can't remember the type and make). Even, the automobiles on the roads were miniature scale models of Jags, Aston Matins, etc, none of your plastic models out of xmas crackers.

The factory had 1200 head at the time and we got donations from quite a few members of staff who were experts in various fields: sewing, model making, weaving, pottery, dolls houses, wood carving and turning.

The model was a beautiful thing when it was in operation and, naturally, the electronics, including lighting were pretty advanced. I never found out what happened to the model after the demonstrations were over, but one of the directors was a keen railway modeller.

spec

 

[earckens]

(3) Only a single chemelec circuit would be needed and the routing to the various points would be handled by switches.

(4) You could even get sophisticated and introduce some logic so that there were safety interlocks, once again, just like a real railway. An Arduino could make it really felible.

View attachment 101556​

On (3): if SPDT switches are used, and 1 "chemelec board" for the whole, each turnout would be activated (1 of both solenoids per turnout) each time the pushbutton is triggered. How can you avoid this and only trigger the solenoid required (with just hardware)?

On (4): probably a backup concept for (3)?

Erik

 

[cowboybob]

... The model was a beautiful thing when it was in operation...

Man, spec. I am so jealous.

Got a Lionel as a Christmas gift in the mid-fifties. Coolest gift ever, even if only a single, circular track. Puffed "smoke"! My Dad built a mountain range with wooden trestles for it in our basement. Although, do be honest, the locomotive unit spent more time at the repair shop then on the track.

You need to find that sucker... .

 

[chemelec]

The Push Button on my circuit triggers the 555 to give a Pulse out.
The Solenoids should be connected to my circuit with Switches, so you can pick which track is being switched.

earchens, In the Picture you Show, Are These the Switches to Control the different Tracks?

 

[earckens]

Hi chemelec, which picture do you refer to that I show?
The solenoids should be connected to your circuit: ah but there it goes wrong. Each turnout is connected to a control panel where a SPDT switch indicates which leg of the turnout is chosen (left or richt); so there is permanent electrical connection to the solenoid corresponding to the leg (left or right) where traffic is to be directed. If now all these SPDT's are wired with a common to the MOSFET circuit then at every activation of the pushbutton every turnout will have one of both solenoids activated, whether required or not.

 

[chemelec]

Sorry But without seeing an actual wiring diagram of your system, it is a bit difficult for me to understand your setup and how to solve this problem.

 

[earckens]

Sure, sorry for the handdrawn sketch, I am not proficient enough in Eagle yet. Here is a concept drawing.

 

[chemelec]

Looking at your drawing, No matter how you supply power to this, it appears you will activate All the solenoids to the side they are connected to.

Those Switches should have a Center Off, so you can select each solenoid individually.

 

[earckens]

But "center-off" switches do not allow visual feedback of turnout position: you cannot visually see from the switch in which position the turnout currently sits. Is there a hardware solution with double throw switches?