Driving solenoids in a V8 solenoid engine

Thread Starter

Ephex

Joined Jul 4, 2021
83
Hi guys!

I need some help with the power electronics that will control- and drive the solenoids in my solenoid engine:

1664908208022.png

There are 4 switches with bearings on the end of them, which get activated in four different intervals of the axle angle when an excentric part of the axle pushes them in. In short, there are 4 SPST switches that control when each pair of the solenoids should pull on their pistons. The axle is in 'flat-plane'-configuration which means that pairs of solenoids are always in sync (there will be 4 pairs, one switch corresponds to one pair).

The solenoids in the picture are obviously not spun with copper yet, but I have a couple that I've wound to use for testing. They will all be spun with ~78 g of 0.25 mm copper wire (tested once but the number of wounds is unknown) and each have a resistance of ~55 ohms. The maximum voltage (at the moment) that I've tried with them is 37 V when using a small test circuit powering two solenoids in parallell (circuit shown below). The current through each solenoid then becomes ~670 mA (this decreases a bit as the temperature of the resistor increases), but this makes the solenoid become very very hot after a while. This has however been in a test when I only turn the motor manually at very low RPM.

This leads me to
Q1: Do solenoids generate less heat when "pulling" on e.g. a mild steel piston, compared to just sitting and letting current pass through as if it were a resistor? My thoughts go "if there is nothing nearby that can be pulled/pushed, all the power (37 * 0.670 W) goes to heat, but if it's excerting a force on an object, all that power can't go to heat, some of it has to be 'spent' on pulling the object".

Regardless, I have not decided on how to supply the thing with power so for now I will use a PSU that goes up to 37 V, but this might very well be too much for the solenoids.

The testing circuit is a proof of concept for my thoughts on how to activate one pair of solenoids depending on the correct switch, and here's a representation:

1664909372614.png

There is an N-channel MOSFET (Q1, data sheet attached) that controls when the solenoid pair should be activated. This gets activated by the switch (SW1) which connects the output of a voltage regulator (U1). This voltage will only need to be above 4 V (the threshold voltage for the MOSFETS used in this circuit) and is decided on the ratio between R1_1 and R_2. The 'coils' L1_1 and L1_2 represent the two solenoids connected to the circuit, with D1 being the flyback diode.

The test circuit was made on a perf-board and simply applies the voltage VCC over the solenoids when the axle rotates to the position when the piston is "allowed" to pull, and disconnects VCC when the piston no longer is allowed to pull. All working as expected.

However: I will need to be able to control the forces that the solenoids excert, and therefore the current through them (and therefore the voltage them) so that I can implement some sort of throttle. I have some ideas but I'd like to hear your thoughts:

1. PWM the gate voltage and add some big capacitor near the drain of the MOSFET
  1. Sort of how you'd PWM an H-bridge DC-motor controller, but with the 'PWM-switch' being in series with SW1 so that the axle-switch still decides when current is allowed to flow through the solenoids. Perhaps with a 555 timer that changes duty cycle from a potentiometer? This might however cause issues with the driving of the MOSFETS if the frequency gets too high(?). I might not get away with my super-simple way of activating a MOSFET as I did with my testing circuit.
2. Reduce the voltage of VCC with PWM(?)
3. Use a variable step-up circuit that takes e.g 12 V or 24 V and outputs anywhere from 0 V to 37 V (or somewhere in the vicinity) depending on a potentiometer?
  1. This also enables me to use a lower voltage powersupply (although higher current) so that it turns more into a desktop toy. I will also be able to more easily use it with batteries on future RC-cars or similar projects which I hope to be able to do.

#2 and #3 would be nice since the gate of the MOSFET would not be affected by the big changes, enabling me to get away with the simple activation method. What do you think? Any other ideas?

The final solution will be two PCBs attached to the front- and back side of the motor. I am obviously not finished but here's a preview at least:

1664912559407.png

The plan is to have each PCB control two pairs of solenoids each, but the above PCB only has one side filled up at the moment.

Q2: What would a good decoupling capacitance between VCC and GND be? Is it needed?

Q3: What would a good value for C1 be in this case? I only added it so that if I needed it on my PCB


I hope that I could sufficiently explain the problem. Grateful for all answers!

// Ephex
 

Attachments

Last edited:

shortbus

Joined Sep 30, 2009
10,045
The axle is in 'flat-line'-configuration which means that pairs of solenoids are always in sync (there will be 4 pairs, one switch corresponds to one pair).
That is hard to understand. Can you please give a drawing of what you mean? The configureration of the crankshaft and it is a crankshaft not an axle.
 

LowQCab

Joined Nov 6, 2012
4,023
The first problem with this setup is the Mechanical-Switches,
they will create "Timing-Scatter" which will vary with RPM, and generally
throw in a bunch of arbitrary confusion.

The ideal "Switches" would be 4 Optical-Interrupters, spaced 90-degrees apart,
and a Disc with a small slot cut in it, which is attached to the Crankshaft,
which activates the Optical-Interrupters in sequence.

The PC-Board, or Perf-Board, that the Interrupters are mounted to must have a means for
adjusting their position,
to create what would be, the "Ignition-Timing" in an actual Internal-Combustion-Engine.

The Coil heating issue that You were concerned about can be handled by a
"Current-Regulator-Circuit", ( rather than by a "Voltage-Switch" ), ( PWM is just a fast Voltage-Switch ).
A Current-Regulator will allow You to run much higher Power-Supply-Voltages, without over-heating.

If You have aspirations of running this Motor on substantially less Voltage,
You will probably have to re-wind the Coils with much heavier Wire.
This will also mean that it will require very High-Current capable Lithium-Ion type Batteries,
Double-AA Alkaline-Batteries will be a tremendous disappointment.

All this really depends on just how much
actual "Useful-Torque" You will be expecting this Motor to produce.
This type of Motor is a "Novelty-Toy", and is highly inefficient, even when perfectly executed.
It is not practical to use it in a Model-Car unless it is substantially "Geared-Down",
and ~5-MPH is all the speed that You might reasonably expect.

However, for a Desktop-Toy, it's pretty-cool,
and with accurate Timing, and careful design of the Electronics,
it could easily be capable of spinning ~5000-rpm or more, ( with no Load ).

Are You prepared to start over with the Electronic-Controls of this project ?
( what You currently have to work with is going to provide less than impressive performance )
.
.
.
 

Sensacell

Joined Jun 19, 2012
3,432
Your metallic coil bobbins will act as "shorted turns" through transformer action - this will limit the speed performance as the effect will be worse the faster you try to switch the coils.

You can use conductive bobbins if you slit them so they do not form a complete circular path.
 
Last edited:

Ian0

Joined Aug 7, 2020
9,667
That is hard to understand. Can you please give a drawing of what you mean? The configureration of the crankshaft and it is a crankshaft not an axle.
Do you mean a flat-plane crankshaft? So that the solenoids on one bank are 90° out of phase with the other?
 

Thread Starter

Ephex

Joined Jul 4, 2021
83
That is hard to understand. Can you please give a drawing of what you mean? The configureration of the crankshaft and it is a crankshaft not an axle.
Oh sorry, I got axle and crankshaft confused.

The crankshaft is flat-plane since all the cranks all lay on one plane, as opposed to e.g. a cross-plane crank shaft. There are some nice videos on youtube that shows animations of the two so I recommend them! However, it's not relevant in this case since I have already solved the timing with the switches. What's relevant for this post is that each switch is activated when "its" pair of solenoids need to activate.

Below is a picture of a flat-plane crankshaft

1664990165164.png
 
Last edited:

Thread Starter

Ephex

Joined Jul 4, 2021
83
...
However, for a Desktop-Toy, it's pretty-cool,
and with accurate Timing, and careful design of the Electronics,
it could easily be capable of spinning ~5000-rpm or more, ( with no Load ).

Are You prepared to start over with the Electronic-Controls of this project ?
( what You currently have to work with is going to provide less than impressive performance )
.
.
.
I only aim to have this as a desktop toy, yes. With that said it would be fun to see if it could be able to be used in a small model car, but as other people mentioned, this will be after a considerable amount of gearing down. The thing won't travel fast (if it even travels at all) but that is fine by me.

I am by no means looking to find the most optimal solenoid engine, atleast not for this version. Your optical-solution for the timing does sound interesting, but for this exact model I am going to stick with the mechanical switches. This is partly because everything regarding the switches is already done, but also the fact that they sound amazing when the crankshaft turns...

The switches all activate when they should and I have them screwed in place in 'slots' which allow me to unscrew them and push them closer to/further away from the timing-part on the crankshaft, changing when they activate and for how long. This is not perfect but it adds some adjustability.

So no, I am not willing on completely scrapping everything regarding the electronics. I would really like to keep the mechanical switches but if I make a new motor and/or new PCB for it, I will definitely use either optical-timing or something like hall-effect sensors :)
 

Thread Starter

Ephex

Joined Jul 4, 2021
83
Firstly, I must say, your model looks really nice.
Thank you!

But here's the thing. The key point about solenoid design is to allow the plunger to complete the magnetic circuit.
[...]
I think this would give you something like an order of magnitude improvement in force per current times turns.
This all sounds pretty complex, and it would probably increase the performance a LOT but I sadly don't own the lathe, which is purely meant for brass and similar alloys so turning lots of new parts (and certainly not steel) is not an option for me. I will however keep this in mind if I decide to optimize the motor!

Your photo shows the plungers sticking out from the end of the coil - in that scenario there will be virtually no force as the solenoid has already done its best to complete the magnetic circuit through air!
When I designed everything I did it in a way that would make the plunger long enough to reach down to its link arm while also making sure that the centre of gravity (CoG) of the plunger would come almost right up to the CoG of the copper windings when it's the furthest away from the crank shaft. What I mean by 'maximum extension' is when the CoG is the furthest inside of the bobbin while the 'minimum extension' is when it's the furthest out of it (right part of the picture).

1664993290118.png

The thing is that the solenoid won't have to pull on the link arm, since the crank shaft is positioned at a point where no torque will be generated.

But still: Do you recommend me e.g. shortening the plungers to shift the CoG in any way?


I do like your idea of controlling voltage with (a single) LM317 although I'd get the opto-interrupters to turn BJTs hard on - simple and less fussy than MOSFETS. Maybe one transistor per coil in case the coils are not perfectly matched. Only showing one coil, maybe something like this:
View attachment 277708
When the rotating disk interrupts the opto interrupter diode turning its transistor off then the transistor connected to the coil is turned on. The pot adjusts the output voltage of the LM317. I favour many turns of finer wire.
The LM317 I use is only to get a lower voltage than VCC that I can route through the switch to the gate of the MOSFET. I did not plan on use it to drive the solenoid it self. If I change VCC, I will have to boost it if anything, since I'd gladly supply the motor with 12V or 24 V which then gets boosted in some way to <= 40 V. I don't quite grasp the functionality of the circuit. The output voltage drives the solenoid while also activating the transistor that closes the circuit for the solenoid? I've never used transistors, only MOSFETs. If this circuit could be adjusted to work with MOSFETs, that would be cool! (check the circuit below). Thanks for the custom drawing :)

I don't think I will do the steel-cylinder modification at this time, but I will perhaps cut a slit in the cylinders to reduce the magnetic resistance that @Sensacell mentioned. I assume this is something like what occurs when you drop a magnet through a copper tube? However, how much will this affect the whole thing? Will there be a considerable amount of magnetic resistance (is this even a term lol)? 10%? 20%? Or is it something like "the maximum RPM will go from 4000 RPM to 5000 RPM if you slit the cylinders?" The cross area of the brass bobbins seem really small to me. It just feels really risky to cut them all in such a way.

What's priority #1 right now is to be able to control the current through the solenoids. Do you think something like this would work?

1664996083643.png

The voltage that drives the solenoid is now controlled by U2, similar to @Jerry-Hat-Trick 's circuit. It's rated at 1.5 A, and if I use two solenoids in parallell they will draw 0.67*2 A = 1.34 A at 37 V. If I only drive one solenoid with this entire circuit I will have to have all these components x8 instead of x4. What is the problem with having two solenoids in parallell as in the circuit above, if they are not identical to eachother? Is it a major issue? It would really simplify things, and enable me to get away with the small space I have on the front- and back side of the motor!

The other LM317, U1, still only "supplies the switch" with a smaller voltage sufficient to activate the MOSFET.

Thanks for the help so far everyone!
 

Attachments

shortbus

Joined Sep 30, 2009
10,045
The crankshaft is flat-plane since all the cranks all lay on one plane, as opposed to e.g. a cross-plane crank shaft.
Just my thoughts. A flat plane crank won't work very well for this type of "V" motor. While in an internal combustion engine a flat plane crank has some extra horsepower out put, it only works there because the cylinders fire individually, not in pairs like your doing.

For the way you want this to work you would get a better result if you change to a "boxer" style engine like the old VW or Subaru or Porsche use.
https://en.wikipedia.org/wiki/Flat_engine


When I designed everything I did it in a way that would make the plunger long enough to reach down to its link arm while also making sure that the centre of gravity (CoG) of the plunger would come almost right up to the CoG of the copper windings when it's the furthest away from the crank shaft. What I mean by 'maximum extension' is when the CoG is the furthest inside of the bobbin while the 'minimum extension' is when it's the furthest out of it (right part of the picture).
Again not trying to make you feel bad but to help you. Center of Gravity has nothing at all to do with how solenoids work. The way your drawing shows the solenoid plunger, also called an armature, all the way in the coil at all times. Doing that it won't move at all.

The plunger needs to be mostly out of the coil at rest, when the coil is not activated. Then when it gets turned on it will pull into the cylinder. Also the coil is only active, turned on, to pull the crank through toward the coil. At other times it needs to be turned off, if they are all on all the time the motor will be in a locked up condition.
 

LowQCab

Joined Nov 6, 2012
4,023
C-1 could be useful, but it could also be a liability, depending upon it's value.
U-2 is not necessary with accurate, reliable, and repeatable, timing that is only practical
with Electronic-Switching, where the Start, and the End, of the "On-Period" is carefully controlled.
The major problem with your "Electronics" are the Switches.
They are guaranteed to cause erratic-operation, excessive Current/Heat, and will generally
be a perpetual "pain-in the-butt" to attempt to get working, and to continue to work, reasonably well.

As Shortbus said,
Center-of-Gravity has little to no significance in this situation.
A Solenoid needs to have a "Magnet-Circuit", otherwise, you will just be
unnecessarily generating Heat with a coiled-up piece of Wire.
Once the Magnetic Circuit is "completed" the "Pulling-Force" will turn into a "Holding-Force".
.
.
.
Solenoid .png
 

Thread Starter

Ephex

Joined Jul 4, 2021
83
For the way you want this to work you would get a better result if you change to a "boxer" style engine like the old VW or Subaru or Porsche use.
https://en.wikipedia.org/wiki/Flat_engine
Alright, but I am currently not going to redo the model. It might not be perfect but I never intended on it being that way.

Again not trying to make you feel bad but to help you.
No worries!

Center of Gravity has nothing at all to do with how solenoids work. The way your drawing shows the solenoid plunger, also called an armature, all the way in the coil at all times. Doing that it won't move at all.
It does move, actually. After winding two bobbins with two different wire gauges (0.20mm and 0.15mm) and hooking them up with with 37V, they where able to pull the armatures all the way in with them always pulling on them. During testing, I was also able to hook up two solenoids in parallell with my testing circuit. This was when the solenoids where mounted to the engine, and they where able to pull one third of a rotation of the crank shaft. They might absolutely be able to pull more, but the concept works.

The plunger needs to be mostly out of the coil at rest, when the coil is not activated. Then when it gets turned on it will pull into the cylinder. Also the coil is only active, turned on, to pull the crank through toward the coil. At other times it needs to be turned off, if they are all on all the time the motor will be in a locked up condition.
Do you recommend me to shorten the armatures in a way that changes how far they go inside the solenoid?
At the moment, they function with the all-be-it non-perfect switches and current setup but I guess this is to improve the pulling force?
 

Thread Starter

Ephex

Joined Jul 4, 2021
83
C-1 could be useful, but it could also be a liability, depending upon it's value.
U-2 is not necessary with accurate, reliable, and repeatable, timing that is only practical
with Electronic-Switching, where the Start, and the End, of the "On-Period" is carefully controlled.
The major problem with your "Electronics" are the Switches.
They are guaranteed to cause erratic-operation, excessive Current/Heat, and will generally
be a perpetual "pain-in the-butt" to attempt to get working, and to continue to work, reasonably well.

As Shortbus said,
Center-of-Gravity has little to no significance in this situation.
A Solenoid needs to have a "Magnet-Circuit", otherwise, you will just be
unnecessarily generating Heat with a coiled-up piece of Wire.
Once the Magnetic Circuit is "completed" the "Pulling-Force" will turn into a "Holding-Force".
.
.
.
View attachment 277760
I have learned that mechanical switches won't be perfect, and I thank you guys for making me understand that. If I make another model, I won't use them. But, the main reason for this post was to figure out how to control the current through the solenoids with the concept at the moment. If the switches break down after not much testing or if they make the engine stutter or if they restrict the maximum rpm from going to 5000+ RPM, I am totally fine with that. I just want this concept to run. It is very much a desktop-toy. The main reason as to why I chose the mechanical switches was the fact that the engine would sound cooler as they all *click* when activated lol.

Regarding the plungers, should I shorten them in a way that would make the magnetic circuit close when it's the furthest in? Shorten them so that they don't poke out at all? I will also think about the iron/steel shell around them.

So, do you guys think this circuit would work given the mechanical switches?
1665034671780.png

I want to stress the fact that I'm grateful for all the help I'm getting, I'm learning so much. However, I am not after an optimized solenoid engine. I just want my model to run :)
 

LowQCab

Joined Nov 6, 2012
4,023
Run it with a 5-Volt Power-Supply and connect the Switches Directly to the paralleled-Coils.
If the Switch Timing is right, it will run, no electronics required.
But it will run much better with properly designed Electronics,
but the Electronics can't fix a mechanical design problem.
You're trying to throw Higher Voltage and Current at it
to overcome problems with the design of the Magnetic-Parts.

The Crank, Rods, and Bearings, are first, and it looks like you've got that part handled pretty well.
The design of the Solenoids is next, and what You have needs a lot of refinement.
Then we can start talking about designing the Electronics

Making an Electronic-Switch with a FET and Voltage-Regulators is
attacking the project from the wrong end.
The Circuitry would have to be designed to withstand "any" condition that is likely to go wrong,
so everybody here is looking for "what can go wrong with this project",
and how to work around, or eliminate, those potential problems before designing any Circuitry.
What You already have should work .......... maybe.

A particular problem I see on a regular basis when helping people with their Hot-Rod-Cars is
that they always blame the part that they understand the least, which is usually an old-school Carburetor,
but the problem is quite often something entirely different, and simple, that they overlooked somehow.
I think that same thing could possibly be happening with this project,
it should work, but it doesn't, so the most misunderstood part gets blamed.

If You will redesign your Solenoids to operate based on the picture I provided
regarding creating a proper "Magnetic-Circuit",
this project could be made to run on 2 Alkaline-D-Cells with NO Electronics at all, just Switches.

( A heavy Flywheel will help )

The Coil-Bobbins need to be made out of Plastic, or some other non-metallic-material,
and be completely surrounded by a 2 or 3-piece, Iron or Steel, cover,
thicker, heavier, material is always better.
You might be able to use a piece of Iron-Water-Pipe for the outside-sleeve,
it's relatively soft metal, and so it's easy to Machine on a Lathe.
If it must be thinner material, 1/2" or 3/4" Electrical-Conduit-Pipe might work,
but it is so thin that it will limit the efficiency of the Motor.
The Bobbins must fit snugly inside the Ferrous Iron or Steel Cover.

Once the Magnetics are properly handled,
then we can discuss the type of performance that You would like to achieve,
and if that happens to contain some lofty expectations,
an Optical-Interrupter setup will probably be required,
and if serious high-performance is your goal it might get a little more complex than that,
but we're not to that stage just yet.
.
.
.
 

shortbus

Joined Sep 30, 2009
10,045
This tread is like so many here. People spend time and money and end up with something not working. But because they already have a serious expense of time/money in a non working thing, they are hesitant to start from scratch again to get things that will work.

Trying to help when the project is wrong from the start only make both the helpers and the thread starter upset!
 

shortbus

Joined Sep 30, 2009
10,045
The Coil-Bobbins need to be made out of Plastic, or some other non-metallic-material,
and be completely surrounded by a 2 or 3-piece, Iron or Steel, cover,
thicker, heavier, material is always better.
Not so. There are so many industrial solenoids made and working with a brass center bobbing that it is just not true that they need to be plastic. Need proof? Look at the solenoids used for Red Hat type solenoid valves. The only consideration that needs to be meet in a solenoid bobbing is it can't be magnetic. Those Red Hat solenoids are made from non magnetic stainless steel, as are the pintle the solenoids fit to. The solenoid valve has an iron/steel armature inside the pintle that moves the diaphragm of the valve.

Also not true is that the solenoid needs to be surrounded by a magnetic sleeve or even a U shaped bracket. While there are "field lines" outside the solenoid coil, the real useful magnetic force is within the center of the coil. This is just simple physics that were discovered way before our time.
http://spiff.rit.edu/classes/phys313/lectures/sol/sol_f01_long.html
 

LowQCab

Joined Nov 6, 2012
4,023
Solenoids don't necessarily have to follow all the rules to be "adequate" for
the particular application, and are usually built to a "price-point" as well.
If it works reliably, and fits the maximum number of Valve-Body part-numbers, it's all good.
Maximum efficiency is not required to move a Needle-Valve ~1/8".

"The-Devil-is-in-the-Details"
.
.
.
 

shortbus

Joined Sep 30, 2009
10,045
I confess, I haven't read and followed the proof in your link but, and I'm going to get in trouble here, I'm deeply suspicious of so much that I read on the internet, there is a mountain of mis-information and, forgive me, I'm doubly skeptical of any link which includes "edu" and/or "lectures".
How sad that is. I never went to college and am not an EE but did learn back in high school about how electromagnets and solenoids work. This is simple everyday physics not some hidden agenda involved.
 

Thread Starter

Ephex

Joined Jul 4, 2021
83
How sad that is. I never went to college and am not an EE but did learn back in high school about how electromagnets and solenoids work. This is simple everyday physics not some hidden agenda involved.
He did perform an experiment and compared no-magnetic-circuit with the normal one, and found a difference? Do you simply not believe the results of his experiment?
 
Top