making a plotter with solenoids instead of servos

[krayvonk]

Im thinking about putting together an xy plotter from scratch, but I had an idea that if I was using purely magnetic movement over spoolly based or cog based movement the machine could be alot more accurate by a factor of maybe ~100.

However, I heard somewhere that solenoids at about 50cm arent as efficient as smaller ones,  is this true?  and if so why?

[Circlotron]

Solenoids are fairly nonlinear with regard to force vs distance. Also you would definitely need some kind of position feedback. This is not needed with a stepper motor.

[krayvonk]

My plans involve using a nychrome wire neatly laid along the dimension to give me resistance over distance,  and my circuits output is a pure binary output, of a more than less than compute - so non-linearities are taken care of.

[virtualparticles]

If the movement of the solenoid is nonlinear then the control loop with the resistive feedback will be very "interesting". To avoid oscillation you'll need to design for the worst case loop dynamics. You might also consider piece-wise linearizing the solenoid behavior. I've never seen this done but if it can be done with fins on the back-end of a missile then I don't see why it couldn't be done with a solenoid. You could also use an FPGA like a MAX 10 and do a system with adaptive loop dynamics. That would be the most awesome of all.

[krayvonk]

Might be harder than I was thinking,  maybe ill go back to programming,  give hardware a break for the time being.

[Rerouter]

Easiest way for feedback on an X/Y table is like they used for old fuel level sensors, Have a loop of thin wire strung from one end, around the other and back to the start, You could likely even rip apart an old mercedes tubular fuel sender if you just wanted to pinch the entire mechanism so long as you let it soak in a solvent to remove the smell, and run a brasso metal polish rag up and down the wire a few times

The moving head has a double wiper (top and bottom wiper on each wire to ensure good contact) connected across both wires on the loop, you measure the resistance of this loop however you fancy and use that for really basic servo control, probably just using an op amp and you set the position of the servo with PWM or a DAC whatever your preference.

These old things for reference, https://www.ebay.ie/itm/VDO-dip-tube-sender-279mm-long-for-VDO-fuel-level-gauge-Motorsport-Rally/262101667418?hash=item3d067a0e5a:g:QWAAAOSw5VFWJLBb
Can be found at car wreckers for under $5 at most places I have seen, for a reliable linear feedback potentiometer roughly 5-90 ohms. with an end stop contact to boot.

I'm afraid I do not have any internal pictures on hand, But I repair the things for a living (replacing the winding wire and cleaning them out after 20 years of use) so can say they should be fit for purpose. rpeatability on these sensors is quite high,

If you do want to go down the magnetic only path, use a spring to oppose your motion, it helps linearise things quite a lot, you would still probably want to use a length multiplier such as a gear or lever so your solenoid does not have to be bigger than your plotter,

[krayvonk]

Making a good wiper is something I have to work out,   so far all my tests have involved me pressing them on the back as they go.
Im a great programmer, then im less so good at electronics, then mechanics i am hopeless.

So how big do solenoids get in the industry?    Id imagine very big,  but to make something that just makes miniature pictures maybe it being so big doesnt even matter...

[Rerouter]

You can get cheap ones up to about half a meter, beyond that tends to be in the world of threaded rods for longer stuff, or pnuematics in industrial applications, solenoids only really remain when things must happen stupidly quick, the balance between making the moving part a low mass, but not letting it bend like a banana over its length gets hard.

As for things happening quickly

Lets say you have a 6V solenoid, but you want it to move as fast as humanly possible, well you pulse it with about 165V for 15ms, wait the travel delay then give it a -130V pulse for 17ms to slow it down for the stop.

[krayvonk]

So a lone solenoid, versus a stepper motor with a threaded rod,  would offer no more miniaturizing potential, one over the other?

[edavid]

So a lone solenoid, versus a stepper motor with a threaded rod,  would offer no more miniaturizing potential, one over the other?

A solenoid/linear motor would have to be much larger to generate enough force to move the plotter head, since you don't have the gear ratio of the threaded rod.

[Spirit532]

Solenoid linear actuators exist, they require the use of high resolution linear encoders(glass scale or magnetic strip encoders), they cost a lot, they are fast, and very precise. However, they don't output very much force.

[KC0PPH]

Look up Festo Servo Pneumatics

I remember at a trade show they were doing a PNP PCB assembly with them. Never asked how they worked.

[Spirit532]

Never asked how they worked.

Pneumatics, proportional valves, and encoders on the end effector. Quite inefficient and inaccurate, but scalable.

[FrankT]

You could use a linear stepper - https://www.h2wtech.com/category/single-axis-linear-stepper?gclid=EAIaIQobChMIreSb6NKp4wIVwxmPCh0sXw6BEAAYAiAAEgLvTPD_BwE#productInfo1

The rail they run on is basically a stepper motor rolled out flat.  The ones in the link can travel at 2m/s

[Domagoj T]

Hold on a sec.
Printers (and that one ~2m wide plotter I had a chance to see opened up) use optical strip encoders with DC motors and regularly achieve 1 200 DPI. You want 120 000 DPI? What print head could keep up with that.
Why do you need that? What is the intended purpose? If you need higher quality print in smaller format, go to a photo studio, they have high resolution printers. If you need large format (as in jumbo ads) that is visible from far away, you don't need high resolution.