What did you do Today 2018

The Stepper addition to my EMCO rotary table went well as seen in my various post, but as usual when pushing the envelope with steppers, high rpm and harmonic vibrations are a problem. Not wishing to try find a DSP type stepper driver with active harmonic control ( such as Gecko, etc) I went mechanical..A Rattle harmonic damper, and was most pleasantly surprised. Since the Rotary table's intended use was as a synchronised gear blank drive on a gear hobber, the table's speed is related to spindle ( hob) speed. In my case, the stepper drives the table with a 40:1 ratio.

So, for example, if the Hob spins at 370RPM ( a midrange speed on my mill spindle), and we are cutting a 14 tooth gear with a pitch of 2.5mm, the stepper must spin at approx 1060 rpm - not easy to achieve with the table inertia, friction, etc.

I got close to this by playing with stepper driver voltage, up to 60 volts, and current, but the 370rpm setting was simply not possible. I made up a simple damper, as in the photos, and it works so well that I can even reduce the supply voltage down to 40v with no stall or loss of steps.

I made a rather miserable video, but it gets the idea across..

The wheel is 60mm OD, 20mm thick aluminium. The slugs are 10mm OD, 18mm long, mild steel.

The holes are 11mm ID.

I just covered the slugs with masking tape for the tests..

This shows the damper wheel mounting position.

The three speeds at which I tested the damper - 120rpm, 200rpm, 370rpm.

The video...Sorry, the start seems a bit mucked up..

I did three tests, first with the damper at the three speeds, then with the wheel but without the slugs, then with the wheel completely removed and then the 120 and 200 rpm speeds still work ( 200 does stall once in a while..) but the 370 does not work at all. At 40v supply even the 200 RPM speed fails without the wheel/damper

Edited By Joseph Noci 1 on 07/11/2018 16:58:31

Walked past the 8F under rebuild in the shed yesterday, and saw this staring out at me:-

... and I'd thought Hallowe'en was over...

How does that work Joseph?

Definitely a hint of Cyberman there.

Mike

Be even better when it gets the nose job done.

Edited By V8Eng on 07/11/2018 19:55:59

Not quite today; the last couple of days, actually. "Playing" - as my wife calls it - with Alibre Atom3D. Unwilling to wait for the next instalment, I've continued with the engine design that I've been mulling over for a while. I've decided to re-do the crank with scalloped webs instead of circular ones, but later on. In the meantime I've tried my hand at the cylinder block. It will be a wet-liner, 4-cylinder job with a swept volume of 123cc. I don't know if I'm going about it the right way, but I've built the block up (down, really ) from the cylinder head face in slices, which are then individually sculpted with cut-outsetc, before the next layer is added. Many mistakes and re-tries later, this is what I've go to show for it:

I don't know where the red lines are from, maybe they are boundary lines. ( I'm too wrapped up to read the manual at the moment! ) David Jupp has given me a link to some videos, so I'll have a look at them soon.

It's not the prettiest engine I'd have to agree, but, hey, even Cosworth had to start somewhere!

John

Edited to remove smiley face.

Edited By John Hinkley on 07/11/2018 19:58:23

Hi there Mark,

In essence the device 'breaks' the resonance that the stepper rotor suffers from, and brings it back into the control domain.

A stepper is a spring-mass device - It does not rotate smoothly but jumps from step to step, with a mechanical oscillation at start and end. The applied voltage to each winding, in the stepping sequence, results in a torque which turns the rotor. Under normal operating domains, the rotor position lags the rotating torque vector by anywhere from 60 to 100 degrees. As the rotor speeds up, however, this lag starts to increase, for various reasons - one being that the current does not have enough time to build up in the winding, due to the shorter period that winding is 'on' and the winding inductance. When the lag increases, and approaches 180 degrees, the torque vector is more out of phase with the rotor position, and the rotor stalls, looses steps, or can even go in the opposite direction!

The addition of a flywheel adds inertia to the oscillation at start and stop, adding a peturbation to the oscillation, and retarding the phase lag, helping reduce the lag and the loss of torque. The rollers inside the holes add a second order perturbation to the lag, reducing it drastically. In effect, the rollers attempt to retard the step-rotation of the wheel at the start of the step - they are standing still while the flywheel moves, and then they are moving while the flywheel comes to a near halt at the end of the step, so they continue moving and 'smack' into the side of the holes in the wheel, damping the oscillation at that point.

They add very little audible noise in this specific application, as heard from the video - they are also very effective at very low speeds, especially at the natural resonant frequency of the motor, but at that rpm ( between 20 and 150, for typical motors) they are very noisy! For those application there is a viscous damper - a flywheel attached to the shaft, surround with thick silicon oil, and the whole thing encased in a circular enclosure that surrounds the flywheel completely - the enclosure rides on bearings on the shaft to which the flywheel is attached, The enclosure is also quite heavy. The flywheel then tries to accelerate the enclosure through the friction of the silicone oil, but the enclosure lags, and when the flywheel decelerates , the enclosure continues spinning - again a second order damper, but completely quiet. But not so easy to DIY!

Joe

Well done John!

Neil

The rattle damper is intriguing, I cannot say I fully understand its theory but seeing your test results I started thinking about whether reducing the mass of the slug carrier would improve it. Adding a rigid 'flywheel' to the spindle of a motor that theoretically at least, is moving in a series of start and stops seems a counter productive procedure.

Without the damper fitted how much effect on the maximum stepper RPM does the tightness of the belt have? If the belt was slightly loose then the mass of the driven pulley, worm, and the pinion load, would be decoupled from the stepper shaft thus allowing it to more easily keep up with what the motor windings are telling it to do. I know this is at variance with your description of the rotor oscillations but adding some torsional compliance (as long as it does not introduce too much backlash) might be a way of maximising stepper speed.

Ian

Ian,

To everyone, please forgive the verbose post - essentially why I did not add it to the first post on these tests.

If there is better place in another topic heading, would a moderator please move it so..I thought with many folk playing with steppers ( and steppered rotary tables) the info may be useful, but it all gets lost so quickly down the pages.

This post is spilt in two as demanded by the forum word-counter...

I did not give the long description of tests I did with the stepper damper, but the addition of just the wheel without the slugs, that is, just a flywheel, showed a great improvement in achievable upper speed limit, and good improvement with regard to lack of stall at very low speeds ( 20 to 60rpm range - I am talking of stepper rpm here, not hob spindle rpm).

The test method was - apply 40v to the stepper drive, do the run at the three stepper RPM's, and see what happens. Increase the voltage to 50v and repeat, then to 60v and repeat. I also tried at 75v ( my driver can manage 80v) but the stepper motor was not happy there, even with no load.

Additionally a low speed test was done - from 0rpm to 100 rpm over 3 seconds at the three voltages.

At each of the voltages the test run was first with no flywheel attached, then flywheel only, then slugs fitted in the flywheel.

The motor was always connected to the load, with the drive belt taut - maybe 3mm deflection at centre span with a good finger push..

The three stepper speeds were approx 340rpm (spindle = 120rpm), 570rpm(spindle = 200rpm) and 1060rpm(spindle = 370rpm).

The mill spindle reached terminal rpm within 230ms(120rpm), 270ms(200rpm) and 305ms(370rpm) - takes longer to spin to higher spindle RPM due to the spindle inertia. This was measured using an oscilloscope timing the rotary encoder pulses, and seeing when they reached terminal width. The stepper has to accelerate and achieve terminal rpm in the same time.

I did not do any tests with the belt loose - I am trying to have as little backlash in the system as possible. For interest I may try with a loose belt, although not sure to what purpose..I suspect from the tests with no flywheel that is will be worse - the rotor is free to oscillate as it wishes, with no damping at all.

Incidentally, removing the flywheel, and using one of the rotary table locking handles to partially 'lock' the table, ie, add friction to the tables rotation, also improved matters. Anything that dampens the rotor oscillation has a positive effect, and this is also used commercially, with friction brakes, or maybe clutches, on the rotor shaft that have centrifugally activated pressure plates. Not efficient, but this was used in the 'old' days..

It is common in commercial applications to add just a flywheel to the stepper shaft to help the motor through resonance spots - many photo-copiers, (old) photoplotters, big printers, flat bed plotters, etc have this applied.

Joe

Edited By Joseph Noci 1 on 08/11/2018 06:30:38

The second part of the post on the voltage/rpm tests on the stepper and damper..

With no flywheel.

40V supply
340rpm no problem. 570rpm fails occasionaly. 1060rpm fails instantly. Low speed test - occasional stall around 20 to 30rpm

50V supply.
340rpm no problem. 570rpm achieved but very audible note getting there. 1060rpm fails instantly. Low speed test - occasional stall around 20 to 30rpm, but less often.

60V supply.
340rpm no problem. 570rpm achieved , still very audible note getting there. 1060rpm fails instantly. Low speed test - no stall around 20 to 30rpm, but very noisy around those rpm's.

Flywheel only.

40V supply.
340rpm, 570rpm, no failures. 1060rpm fails almost immediately. Low speed test - No stalls,  - subjectively the motor seemed smoother at low rpm.

50V supply.
340rpm, 570rpm, no failures. 1060rpm fails after 1 -2 seconds running. Low speed test - No stalls, - subjectively the motor seemed MUCH smoother at low rpm.

60V supply.
340rpm, 570rpm, no failures. 1060rpm would fail once in 2 or 3 tries. Low speed test - No stalls, - OBJECTIVELY the motor was MUCH smoother at low rpm.

Flywheel and slugs.

40V supply.
340rpm, 570rpm, no failures. 1060rpm would fail once in 2 or 3 tries. Low speed test - No stalls, slugs noisy between 15rpm and 60rpm - motor much smoother at low rpm.

50V supply.
340rpm, 570rpm,1060rpm, no failures. Low speed test - No stalls, slugs LESS noisy between 15rpm and 60rpm - motor much smoother

60V supply.
340rpm, 570rpm, 1060rpm, no failures. Low speed test - No stalls, slugs quieter between 15rpm and 60rpm - OBJECTIVELY the motor was MUCH smoother at low rpm.

some observations -

Higher voltage IS needed at high rpm's - Di/Dt issues, and needed to overcome the back-emf at higher speeds.

A simple flywheel does wonders

The added slugs make a small miracle..

Joe

edit - usual typo's..

Edited By Joseph Noci 1 on 08/11/2018 06:38:42

John, the red lines are each of the sketches, there should be an option to have them visible or hidden but it is slightly different in Atom so I won't confuse you by showing how I can do it. You can also hide the green Axis & plane if desired.

Engine looks good

I'll answer that for you, as Nigel has'nt yet.

All Triumph twin alloy c/heads with a stub, ie push over exhausts, were threaded into the head. Stopped in 71 with the advent of the retrograde step of push in exhausts.

Nice little earner back in the day for me, converting push in heads to push over heads. All done on a Colchester Student with the tool post removed, a jig bolted to the top slide and drilled and tapped from the headstock.

Must of done nearly a hundred heads....

Thanks for the tip, Jason. I'll investigate later after this morning's Tesco raid. As for the engine, I fear it will remain a virtual one - I can't see the block being cast easily and I would imagine the cost of an aluminium billet that size would be unaffordable. I suppose, if I kept all the swarf, I could get half my money back! Maybe I should be looking into getting a 3D printer and scale it down a bit?

John

An update on the thread **LINK**

Have now cleaned up the two WW1 150mm German Howitzer cartridges and mounted them to make a couple of gongs. The 'hammer heads' were made from a semi hard ball from a toy shop covered in wet chamois leather. I now have to decide which auction site would be best for selling them.

In Atom3D under viewing prefrences you can 'toggle sketches' and 'toggle references' as well as a number of other options.

I've missed a lot of this but it looks wonderful. To hob a 14 tooth gear with a 40:1 reduction worm requires virtual gearing 14:40

Assuming 400 half steps/rev on the motor it works out beautifully 400 x 14 / 40 requires 140 steps/rev.

I'm guessing you are 3d printing quadrature encoders to go on the spindle.

The easiest way to read a quadrature encoder is to combine new data with previous data, turning 2 bits in to 4 and then using a look up table to get step and direction.

I know I am a nosey git, but am I right?

I rebored a pair of motorcycle cylinders earlier. A tricky setup, they were standard size but worn, the pistons I want to use are 0.25mm over size so not much to come out and be sure the bores would clean up. the suns shining so I'm going to jump on the Suzuki and take them the be honed.

John

Posted by Neil Wyatt:

"In Atom3D under viewing prefrences you can 'toggle sketches' and 'toggle references' as well as a number of other options."

Thanks for the pointer, Neil, that did the trick. I really should read the manual!.

John

How do you display a proper quote?

Click 'quote' under a posting You can delete any bits you don't want

If you 'lose' the blank line at a the bottom you can get a new one by hovering top or bottom with the mouse and you get a red new line icon that inserts a line that isn't part of the quote.