Driving solenoids in a V8 solenoid engine

shortbus

Joined Sep 30, 2009
10,045
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?
Did you see the experiment? I didn't either. I base what I say about solenoids from many years of both using them on machines back when working and from every book I have about magnetics.

Like I said earlier there are so many working manufactured solenoids in industry that have NO steel frame around them that making statements about how they are necessary for a solenoid to work is just absurd. The frame around a solenoid is only one way of mounting them, has nothing what so ever with the strength or working of them.

If doing this was actually important to you, you would spend some time reading about solenoids and not watching stuff about them on Youtube. Sorry if that sounds harsh or mean, it's not meant to be but it is also the truth.
 

shortbus

Joined Sep 30, 2009
10,045
Solenoids don't necessarily have to follow all the rules to be "adequate" for
the particular application,
That is such a self serving statement, to keep you from looking wrong in your previous statement and think if you look at how solenoids really work you will know that. Adequate has nothing to do with it, it's pure known physics of how they work. Electromagnetism and solenoids are a well known and accepted science, and adequate never is thought of in that context.
 

LowQCab

Joined Nov 6, 2012
4,029
Oh my goodness ....... Shortbus "proved" me wrong once again !!!

This type of petty commenting does absolutely nothing to assist anyone,
please keep it to yourself.
.
.
.
 

Thread Starter

Ephex

Joined Jul 4, 2021
83
Did you see the experiment? I didn't either. I base what I say about solenoids from many years of both using them on machines back when working and from every book I have about magnetics.
I personally think the probability is quite small that @Jerry-Hat-Trick was straight up lying about the result of the tug of war experiment, but that's just me. If Jerry could provide a video of the experiment or something similar, that would be nice though :)

Like I said earlier there are so many working manufactured solenoids in industry that have NO steel frame around them that making statements about how they are necessary for a solenoid to work is just absurd. The frame around a solenoid is only one way of mounting them, has nothing what so ever with the strength or working of them.
Do you honestly think that the people in this thread claim that solenoids are unusable without a steel/iron frame? I think they claim that it improves the performance.

Regardless, it's a simple modification if I can get my hands on some steel/iron sheet metal, and if I make them C-shaped as recommended earlier, I can always attach/deattach them even when the current version is done. If it happens that there is no difference / no noticable difference then so be it.
 

Thread Starter

Ephex

Joined Jul 4, 2021
83
What's next:

1. The first version of the PCB is (pretty much) ready to order:

1665186672398.png


The two "halves" of the PCB control two solenoids each, but share the VCOM voltage that's used by the switches.

1665186661982.png

Two identical PCBs will be used. There will be one on each side but one will have 4 mm connectors (rated at 5A) to supply power through VCC and GND.

I could have used a single JST connector for each pair of solenoids instead of one connector for each solenoid, but it feels nice to have the ability to replace one at a time if I need to.

During this, I realized something that I'm quite unsure about.

The variable resistors R1_VAR1 and R2_VAR1 (as well as the other two on the PCB on the other side) which constitute the "throttle"will need to be synced. They all control the voltage that is applied on the solenoids, but each variable resistor only controls two solenoids.

I can't use one LM317 to power all 8 solenoids due to its current limit, but what would happen if all four LM317s share the same variable resistor? Aka e.g. R1_VAR1 is connected to on one of its pads, and the other pad is connected to the adjustment pins of all four LM317, as well as all other RX_3 (R1_3 and R2_3 on both PCBs).

What do you think?

2. The shortening of the armatures/plungers/pistons. (Seems like a lot of performance for just cutting off some steel)
I took a closer look at the pistons, and they poke out 12-13 mm at most from the bobbins, so I first figured cutting off 13 mm would do the trick? That way the pistons would pretty much come up flush to the side of the solenoid and then back 16 mm again.

A friend helped me with doing some testing with the solenoids to get an estimate of the force-curve depending on where the piston is situated. We set up a solenoid with a piston as well as an empty plastic bobbin underneath, applied 20 V, reset the weight to 0 g when it was just holding the piston and then pulled it up incrementally by 5 mm and wrote down the (negative) weight in grams at each distance. This is probably not the best way to measure how many newtons it actually can pull, but the only thing that is interesting to me right now is the difference.

The distance was taken from the "lip" of the part closest to the crank shaft. The best would be to measure the distance 'y' in the picture below, but it would have been much more difficult:

asdasdas.png


Here's the data:

1665188824612.png

There's a max pulling force at ~23 mm, but with my current length of the pistons, they travel between these positions:

1665189364641.png

This means that the piston currently experiences its max pulling force at the beginning of the pulling stroke, but at that point there is barely any leverage and therefore barely any torque generated. As the crank shaft rotates after that, the pulling force decreases drastically. Just go from x=21 to x=5 in the plot lmao. So yeah, a drastic design-flaw by me. But now I know at least!

1665190736355.png

The goal for now is to shift this entire interval so that when it's half-way through a pulling stroke, the maximum force is applied on the piston. I suppose one would want to maximize the integral of the above plot (maximize the energy after one stroke?) by shifting the interval (of length 16 mm).

Now, is this as simple as cutting the end off, or is it more complex? If I cut the end off by 'L' millimeters, will the force diagram shift 'L' millimeters as well when doing it with a new piston? In other words; when designing the piston to get more force, is the position of the "end" of the piston, namely the distance 'y' a couple pictures above, the most important distance? In that case, 'x' in the picture above should move in the interval:

[ x_max - stroke_length / 2 , x_max + stroke_length / 2 ] = [ 23 - 16/2, 23 + 16/2 ] = [ 15, 31 ]

1665190262956.png

If it's as easy as cutting it off to "change where the piston ends" in a sence, then I would have to cut this off by 15 - 5 = 10 mm, which isn't very far from how much the pistons poke out of the bobbins (12-13 mm).

I only have 8 of these pistons with no extras so there's no going back after cutting them off.

What's more important is the PCB. I will order it on sunday through a makerspace so this version has to be finished until then. For now, I've left the variable resistors as THT-footprints so that I can solder on whatever I want afterwards depending on if it works or not to have one variable resistor for all LM317.

We're getting there guys!
 

LowQCab

Joined Nov 6, 2012
4,029
A proper "Magnetic-Circuit" will produce around ~3-times more Force.
This would mean that You would have 3 times the power to work with, or,
You could reduce the power requirements that You have right now by ~2/3.

This would mean that the "Pistons" would be the exact same length as
the outer "Magnetic-Shell",
which constitutes the remainder of the "Magnetic-Circuit",
and would be exactly even with the top of the "Magnetic-Structure" when
the Crank-Throw for that "Cylinder" is at "Top-Dead-Center"

The Force will be so high that You MUST turn the Power off before the "Piston"
reaches "Top-Dead-Center" or the Engine will instantly lock-up.
This is why I said that this Engine could run on 2 D-Cell-Batteries.
With weak Magnetics, it will require substantially more Current just to run at all.

If You want to continue with what You have ...........

The "Pistons" are going to try to "Center" themselves,
where there is exactly an equal amount of "Piston" sticking-out of the top and bottom of the Coil,
any time Power is applied to the Coil.

If the Pistons are made the exact same length as the Coil,
the Piston will try to center it's self, once again, just like before,
only this time,
both ends of the Piston will be flush with the ends of the Coil.
This is a "closer to ideal" situation.
Now, with the Piston sized correctly,
if You add an external "Magnetic-Circuit" around the exterior of the Coil,
You will have substantially more Magnetic-POWER available,
which will require much less Electrical-Power for the same amount of work produced.
Otherwise,
all that Electrical-Power ends-up being converted into a whole lot of HEAT,
and not very much Engine-Torque.
.
.
.
 

Thread Starter

Ephex

Joined Jul 4, 2021
83
If anyone's confused; I treat "piston"/"plunger"/"armature" as the same thing, since I'm unsure about what the correct term actually is in this case lol

This would mean that the "Pistons" would be the exact same length as
the outer "Magnetic-Shell",
which constitutes the remainder of the "Magnetic-Circuit",
and would be exactly even with the top of the "Magnetic-Structure" when
the Crank-Throw for that "Cylinder" is at "Top-Dead-Center"
By "top-dead-center", do you mean this? And "bottom-dead-center" as in 180° from top-dead-center?

1665229232152.png

The Force will be so high that You MUST turn the Power off before the "Piston"
reaches "Top-Dead-Center" or the Engine will instantly lock-up.
Yes, the solenoids won't be activated precisely when the piston reaches bottom-dead-center (BDC) all the way (180°) right up to top-dead-center (TDC). This depends on how close I slide the switches to the crank shaft, but I plan to have them activate- and deactivate with some what of a safety margin since there won't be much torque generated near BDC nor TDC anyways.

If the Pistons are made the exact same length as the Coil,
the Piston will try to center it's self, once again, just like before,
only this time,
both ends of the Piston will be flush with the ends of the Coil.
This is a "closer to ideal" situation.
Alright!


if You add an external "Magnetic-Circuit" around the exterior of the Coil,
You will have substantially more Magnetic-POWER available
Since this did not require me turning something out of mild steel/iron, and instead make it out of sheet metal it will definitely be something I'll try.

1665238866169.png

Here's what I have at the moment. The two extra holes at each side of the biggest hole are simply for letting the M4 screws pass through to attach the thing to the engine. The thickness is currently 2 mm, and it's 16 mm wide but this is by no means locked in. I have a hard time trying to figure out how much mass there needs to be in the frame. What do you reckon? Thicker? Wider? I will also have to take the thickness into consideration when shortening the pistons, since it will move the solenoid "one thickness" away.

Even if they were useless, it would make the whole engine look much cooler:

1665240049468.png

However, I think I will skip the process of slitting the bobbins before winding them with copper, since I spent a lot of time turning them all, I will not risk ruining any of them. I would probably find a way to increase the friction of the pistons if I do it. Perfect bobbins are for MkII instead!
 

shortbus

Joined Sep 30, 2009
10,045
A proper "Magnetic-Circuit" will produce around ~3-times more Force.
Again your misleading. The "Magnetic Circuit" means nothing in a solenoid. ALL of the magnetic force in a solenoid is concentrated in the center of the coil. The out side of the coil does nothing to the pulling force of the solenoid.
You have made the statement in other threads about you know what you don't know,I may be paraphrasing here but it is close to what you say. One thing you surely don't know is solenoid magnetics. Easy enough to rectify but you won't do it, you would rather make nonsensical statements.

This type of petty commenting does absolutely nothing to assist anyone,
please keep it to yourself.
So you mean I should just let you mislead people with your answers? Good luck with that, when I know a correct answer.

;
;
;
;
 

shortbus

Joined Sep 30, 2009
10,045
@Ephex To prove to you the iron isn't needed outside of the coils have a look at this -

Also look at how he makes his pistons. He reduces most of the diameter of the rod that goes through the coil. Doing this keeps it from getting caught in the magnetic force. The link is to a V4 but he has more of these engines on Youtube.
 

Thread Starter

Ephex

Joined Jul 4, 2021
83
@Ephex To prove to you the iron isn't needed outside of the coils have a look at this -
I have seen that video along with several others, and yes they don't have any steel/iron frame but still work. This is not surprising, I think we all in this thread can agree that solenoids work without a frame. I just have to repeat; the question is not whether a frame is necessary, only if it improves the pulling force of the solenoid. I for one has become convinced that there is a difference, but we shall see when I will be able to compare them.

I don't think so, but I've only just glanced at this. How about using an LM338 in place of the LM317s? They can handle 5A and I think they have the same pinout. With other improvements I reckon you'll have enough juice to drive all the solenoids.
That's a good idea actually. 5 A will be enough considering the fact that there will be no point when all 8 solenoids are pulling current.
Here's the update:
1665253747944.png

Vdrive is now the driving voltage which gets delivered by U1_2 (LM338).
 

Thread Starter

Ephex

Joined Jul 4, 2021
83
I meant to mention, with regards to the question about the width of the “C” bracket, it needs to be wide enough so that width times thickness is at least as much as the cross sectional area of the armature so it doesn’t saturate first. And it would be nice if the area around the armature hole could be wider as that is where you lose area due to the hole.

I hope you will post a video of the working engine. I signed up to Vimeo for free and I’m pleased that it works, in case you don’t have another way
Alright, so bigger is better? Then I'll just make it as big as I can fit around the cylinder.

Yes ofcourse I'll show a video! It will take at least a couple of weeks (given that the first version of the PCB actually works).
 

Thread Starter

Ephex

Joined Jul 4, 2021
83
Very interesting dynamics of the LM338 ;)
1665317956701.png

Is it R1/R2 or R2/R1? I suppose it's R2/R1 as with the case for LM317
 

Thread Starter

Ephex

Joined Jul 4, 2021
83
Their formula is wrong.
My sarcasm failed.. I was just joking haha

Ignoring Iadj which is small, that voltage is Vout x R2/(R1 + R2) = Vref = 1.25V - a classic potential divider. So Vout = Vref x (R1 + R2)/R2 = Vref x (1 + R1/R2) = 1.25 x (1 + R1/R2)
Isn't it Vout = Vref x (R1 + R2)/R1 = Vref x (1 + R2/R1) = 1.25 x (1 + R2/R1)?

Judging from this: (R1 is constant, R2 is the variable resistance)

1665322179902.png

R2 = 0 would give 1.25 = V_ref = V_out? Imagine "removing" R2 (modified the circuit in paint.net):

1665322482672.png

Maybe we are talking about the same resistor but just with different names
 

Attachments

Thread Starter

Ephex

Joined Jul 4, 2021
83
Strange.
By increasing R1 (or reducing R2) the value of Vout has to increase so that the voltage across R2 is 1.25V.
Isn't the voltage across R1 supposed to be 1.25V (V_ref)?

This is from LM317:
1665329023250.png

It's not the same component but they are doing the same thing, so I suspect that the same formula is supposed to be similar?
 

Thread Starter

Ephex

Joined Jul 4, 2021
83
One more thought. From this rather neat experiment, you should be able to work out the torque on the crankshaft. How about attaching a lever arm to the end of the crankshaft (I always pocket my wooden stirrer at coffee shops, they are so useful) and hang a known weight at distances along it to work out the torque to overcome stiction? Friction will be less, so once it's going it should keep going. This will allow you to work out whether the motor should self start and what voltage you are likely to need?
Definitely! I will however wait with this until I can perform a new test with the shorter pistons and steel frames.
 

Thread Starter

Ephex

Joined Jul 4, 2021
83
The day has come.


v8_1.png

1666819725691.png

The motor actually works! Hopefully these links work: Video 1 | Video 2

After two long weeks of waiting for the PCB to arrive I could finally do the last things that needed to be done for it to work. I managed to scrape together some time to get the motor fixed, but it's very much slapped together with a few problems. I only got to test it a little bit before having to leave for a trip but soon I'll analyze it more (maybe with an oscilloscope).

The PCB
The PCBs worked with no problem, but I was very close to making a big mistake with the two diodes. I followed the footprint's indication of how the diode should be oriented (the footprint says cathode up, anode down) but it seemed kind of strange. I followed the traces and low and behold, it was the wrong direction.

At that point I believed it to be a stupid mistake in the circuit, but it was another issue. The problem was that the symbol for a diode that I used did not have the same pin numbers as the footprint I assigned to it. One of them wanted #1 to be anode, while the other wanted #2 to be anode. Easy fix though, just solder the diodes the opposite direction lol. Other than that, the PCBs were working as expected.

The solenoids
Would you look at that; I had the energy to make 8 brackets out of mild steel. They look easy to make but it took quite a while. It was a pain to make them align with all the holes. But boy do they make a difference.

Testing with and without brackets: for all tests, we have:
The same solenoid, same piston/armature, same voltage (10 V), same current (181 mA). The only difference is the mild steel bracket.

A solenoid test with no bracket, reaching max 22 g of pulling force.

And with a mild steel bracket? I mean, see for yourselves. From what I can see it peaked at around 72 g.

The increased performance is really good, I did not expect it to be so much better. Glad you pointed it out! There were however other problems with the motor.

Stupid is what stupid does
I made it very clear that I was not aiming for anything perfect, which meant that I was definitely not going to end up with something perfect. However, I am extremely happy and satisfied to see it just run. It makes the project all worth it for me. With that said, here are the things that are bad with the motor:

1. Friction
It is in desperate need of lubrication. Even if it's quite easy to rotate by hand, whenever the solenoids pull, they sometimes get pulled to the sides a bit and grind on the surfaces (friction). What kind of oil do you think suits?

2. Shorting of copper wire in the solenoids
This is quite serious. I got a shock when checking the heat of the motor, and began measuring the voltage of the steel brackets and some of them actually sometimes have close to the driving voltage. This makes it very unsafe to use if you have people around who might want to poke around, so I will from now on not allow anyone to use it but me until I fix this issue.

It must have been some part of the copper wire that lost its insulating coating that then gets in contact with the brass bobbin which touches the steel bracket. I will make sure to disassemble enough so that I can put down some electrical tape between the different pieces to hopefully insulate it more. I might have to cover the entire solenoids in some sort of insulant..

3. The stuttering of the motor
I noticed that it was quite stuttery with a very low RPM even at 37 V. It was actually not running very well (could be mostly due to the friction problem). The videos you see of it running is VCC = 37 V with the throttle maxed. There is definitely potential here, it should be able to reach much higher RPMs!

One theory I have is that the flywheel is not adding enough moment of inertia. I believe that the motor is having trouble continuing between each stroke, so maybe having a bigger flywheel could help it ramp up it's speed and keep it for the entire rotation? Once the voltage is applied, it reaches max RPM instantly which could be a sign that the moment of inertia can be increased, no? On the other hand, this does not fix the issue of it having trouble beginning to rotate from complete stop, even at 37 V.

There are probably lots of more things that I haven't discovered yet, but these are the three I found during the little time of testing I had.


Regardless of how far I am willing/able/determined take this project from here, I still see it as a win. IT RUNS! It was my goal with this project; to make a V8 solenoid motor run.

Thanks so much guys!
 
Last edited:

LowQCab

Joined Nov 6, 2012
4,029
Excellent Sound, and looks great !!!

Switching-Timing is probably the biggest problem.

This Motor should spin at least ~2000 RPM without even trying.

Oil is not a viable Solution, and really-REALLY-messy.
Maybe some thin Teflon-Sheet rolled into a tubular sleeve would help.
Coating the Pistons with Teflon, or similar coatings,
would be best, but could get quite expensive.
.
.
.
 

Thread Starter

Ephex

Joined Jul 4, 2021
83
Switching-Timing is probably the biggest problem.
As suggested by LQC, maybe adjusting timing would help.
That sounds reasonable. Giving the pistons more time to turn off could be the solution.

Regarding the oil; I've seen people use very-low viscosity oil to lubricate their solenoid engine:

1666878518903.png

However, he has the solenoids in a horizontal configuration. I get that the oil will run down the piston to the link arm etc in my engine due to them being angled 45° up, but is there no lubricant viscous enough that I can apply? Honestly I wouldn't mind getting the thing messy as long as I don't have to refill the thing with oil to keep it lubricated.
 

shortbus

Joined Sep 30, 2009
10,045
Like the test done by JHT your not testing solenoid pull but are testing an electromagnet. I'm sorry that saying this upsets you guys but it's true. My ban from posting in this thread that you guys cryed to the mods about was up on 10/19/22

A coil with a steel outside and the steel armature in it makes an electromagnet. That is how junk yard lifting magnets are made and the magnetic work holding tables in surface grinders. When the is both a steel core and the surrounding steel it makes an electromagnet that uses both poles of magnetism. A solenoid only uses the concentrated magnetism in the open core, not the stray magnetism out side of the core hole.

As to why your motor doesn't turn very steady it comes down to the same thing the coils are electromagnets and even when turned off they are holding on to the "pistons" until the next one to be turned on gets enough magnetism to over come the residual magnetism.

The no steel surrounds on the coils is why the other ones of these motors shown on Youtube work so much better. They are using solenoids and not electromagnets.

Look here to see why I'm right - https://en.wikipedia.org/wiki/Electromagnet

;
;
;
;
 
Last edited:
Top