DIY direct drive mount

[EarthLife]

Hello

 

Has anyone out there tried using a couple of these for DIY direct drive at all ? .. https://www.aliexpress.com/item/32959689032.html

 

It wouldn't be for a large telescope, just a smallish (6") SCT. The control board is easy enough to design (a decent cortex M4) and build so that's not a problem, plus a high resolution outer tape system for position fixing (as used on high precision printer/cnc positioning and cheaply available). Relative (as opposed to absolute) positioning is fine once the scope has locked onto a couple or so known stars/objects for calibration.

 

Hi from the centre of England

 

[malc-c]

Looking forward to following your project

[markse68]

Those motors are designed for larger camera stabilised gimbals- if you got your scope perfectly balanced they could probably be made to move it precisely with high enough resolution encoders but they’d not have much torque for dealing with wind etc. Good luck though if you try it and do keep us updated on your progress

Mark

[malc-c]

No expert, but looking at the specs of a NEMA 17, the holding torque is given as 0.5nM, where as the spec of those direct drive motors state its 1.6Nm... 3x greater !

I think the main issue will be with precision, as most drive systems are geared in some way to take advantage of the 1/8th microstepping of the NENA motors.

[markse68]

6 minutes ago, malc-c said:

No expert, but looking at the specs of a NEMA 17, the holding torque is given as 0.5nM, where as the spec of those direct drive motors state its 1.6Nm... 3x greater !

I think the main issue will be with precision, as most drive systems are geared in some way to take advantage of the 1/8th microstepping of the NENA motors.

You don’t normally mount a telescope directly to the shaft of a Nema17 motor though? Usually it drives through a worm gear with ratio 160+:1. And the worm gear being non-backdriveable isolates the motor from the load with vastly higher torque capability for resisting peak loading from wind etc

Mark

Edited October 22, 2021 by markse68

[malc-c]

34 minutes ago, markse68 said:

You don’t normally mount a telescope directly to the shaft of a Nema17 motor though? Usually it drives through a worm gear with ratio 160+:1. And the worm gear being non-backdriveable isolates the motor from the load with vastly higher torque capability for resisting peak loading from wind etc

Mark

Exactly... but with more torque you could reduce the gearing.... but agree, you couldn't use it as a direct coupled direct drive 

[markse68]

But op was asking about direct drive

Mark

[DaveS]

I think DD mounts use a three phase motor with each phase being independently controlled. High resolution Renishaw encoders aren't cheap either.

The fact that only a couple of high-end manufacturers make them should be a clue, as if they were simple I reckon more companies would make them.

I have two such mounts, and yes they are brilliant but cost an arm and both legs!

[EarthLife]

Thank you for the replies, all very welcome !

 

Yes we making a direct drive mount, ie no gears, belts etc will ever be harmed in the making.

 

Decided the position encoder is also going to be DIY using the capacitive vernier type method. That way we can get very high stepless precision very cheaply using simple PCB tracking,with it's own dedicated cortex M4 running the required position processing algorithms etc - The cheap digital vernier callipers you can buy these days use the same method of position sensing, although we will have somewhat higher resolution than a £3 digital calliper.

The aim of the project is to create a cheap DIY'able direct drive + cheap high precision DIY'able position sensor mount.

 

Looks like the motor itself is also going to be DIY, the magnets have been ordered, the copper wire for the windings is about to be (3D printer to be used for making the winding formers etc). The most expensive part we think is going to be the bearings and the steel/iron rings for the magnet mounting, the coils will be iron'less so that there won't be any cogging.

 

The fun is in the experimentation, the learning and the DIY of it all. 

Edited October 30, 2021 by EarthLife

[Horwig]

calpoly18sas2008paper.pdfDirectDriveWorkInProgress.pdf

Don't know if these papers might be of interest

 

Huw

[newbie alert]

I know of a mount, very similar to the mesu that are using nema 17 on a friction drive ...100lb class...  whether it's going to work or not time will tell

[markse68]

21 hours ago, EarthLife said:

Looks like the motor itself is also going to be DIY,

awesome! Good luck with this and really looking forward to hear your progress. Have you done any calculations on what sort of diameter you’ll need?

[EarthLife]

1 hour ago, markse68 said:

awesome! Good luck with this and really looking forward to hear your progress. Have you done any calculations on what sort of diameter you’ll need?

Well the current 3D delta printer here can do up to around 200mm circular I think, so a 190mm diameter motor would be easy to do, which is actually quite big really.

Have 50 neodymium rectangular bar magnets on the way for just £10 a pack, also about to order a pre-made steel 3D printer round bed plate for the magnet backing (helps a lot with the magnetic flux).  Due to the huge uptake of 3D printer tech it's made available some very nice usable frame work items (solid bed plates, fine bearings, hardened chromium steel guide rods, etc etc etc) which helps a lot.

Everything is going direct drive these days, even your home washing machine is now there, all the simple camera mounts on todays quad copters are very basic £2 direct drive motors (produces amazingly super smooth/stabilized video), yet the one application that has always screamed out for a cheap direct drive system is the amateur telescope, yet most are still using back-lashy gears and/or stretchy/bouncy belts and pully's. Their is no reason at all why a simple direct drive system has to cost the earth these days (apart from the obvious one of charging what they like - because they can). The electronics, firmware and software is the easiest part of it all so that's not a problem (control theory, PID's etc), high current full H-bridge drivers, MCU's etc have all come on leaps and bounds in the past 10 years. The home user can now get 10 decent sized multi-layer PCB's made for less than £20 now. Simple techniques for very fine position sensing has and is coming on too over the years, it is a challenge to get 0.1 arcsec feedback resolution without spending £100's, but there are ways and means.

A good many people now have 3D printers at home (if not, then someone they know will have one), home CNC beds are also becoming more common. Time to make good use of them

[DaveS]

5 hours ago, newbie alert said:

I know of a mount, very similar to the mesu that are using nema 17 on a friction drive ...100lb class...  whether it's going to work or not time will tell

That's not a Direct Drive mount.

[newbie alert]

19 minutes ago, DaveS said:

That's not a Direct Drive mount.

As in the mesu?  

I'm sure it's not got a worm , so if it's not a direct drive what's the correct  terminology?

I know you're going to say friction drive but I'm sure it's driven from the motor directly

Edited October 31, 2021 by newbie alert
Added info

[dmki]

direct drive means that the mount/telescope are driven directly by the motor, with no gearing or pulleys between them.
in principle very simple, but you need to have a closed feedback loop for the drive electronics, thus you need encoders with a very high level of precision... which is a specialist item.
an interesting question would could you replace that encoder with synthetic encoder created from a combination of plate solving and star tracking?

 

to be honest i would be more tempted to build something out of, perhaps used, harmonic gears... high torque and low backlash

 

But best of luck with this, i am very interested in how it goes

Edited October 31, 2021 by dmki

[DaveS]

53 minutes ago, newbie alert said:

As in the mesu?  

I'm sure it's not got a worm , so if it's not a direct drive what's the correct  terminology?

I know you're going to say friction drive but I'm sure it's driven from the motor directly

 

38 minutes ago, dmki said:

direct drive means that the mount/telescope are driven directly by the motor, with no gearing or pulleys between them.
in principle very simple, but you need to have a closed feedback loop for the drive electronics, thus you need encoders with a very high level of precision... which is a specialist item.
an interesting question would could you replace that encoder with synthetic encoder created from a combination of plate solving and star tracking?

 

to be honest i would be more tempted to build something out of, perhaps used, harmonic gears... high torque and low backlash

 

But best of luck with this, i am very interested in how it goes

Just so. In Direct Drive Mounts the motor *is* the axis, no reduction needed. And no backlash or periodic error. In ASA mounts, even the "low end" DDM 60 or 85 the encoders tick at 0.02"/tick.

[EarthLife]

43 minutes ago, dmki said:

an interesting question would could you replace that encoder with synthetic encoder created from a combination of plate solving and star tracking?

You can indeed, simple star tracking (with sub-pixel precision) is an easy first step to keeping it exactly on target (without a position feedback sensor).

 

6 hours ago, newbie alert said:

I know of a mount, very similar to the mesu that are using nema 17 on a friction drive ...100lb class...  whether it's going to work or not time will tell

Friction drive is just another form of geared drive, just that it uses smooth surface contact rather than toothed gear contact.

[newbie alert]

43 minutes ago, DaveS said:

 

Just so. In Direct Drive Mounts the motor *is* the axis, no reduction needed. And no backlash or periodic error. In ASA mounts, even the "low end" DDM 60 or 85 the encoders tick at 0.02"/tick.

So the same as a mesu then, no PE, no backlash  and with the Sitech encoders they are also highly accurate

Only thing that's different is the reduction in the friction drive 

[DaveS]

The Mesu still has periodic error. Any drive involving reduction will, even harmonic drive where the reduction is built into the motor. In DDMs there is *no* reduction at all, and the motor *is* the axis.

[markse68]

1 hour ago, DaveS said:

The Mesu still has periodic error. Any drive involving reduction will, even harmonic drive where the reduction is built into the motor. In DDMs there is *no* reduction at all, and the motor *is* the axis.

which has to be weighed against lower stiffness and more susceptibility to external forces 

[DaveS]

The ASA DDMs have encoders with 0.02" sec resolution and 100 Hz feedback. Susceptibility to external forces is very low.

[markse68]

9 minutes ago, DaveS said:

The ASA DDMs have encoders with 0.02" sec resolution and 100 Hz feedback. Susceptibility to external forces is very low.

yes you need that kind of ultra high resolution to provide error signal big enough for the controller yet small enough for ap. Are there any specs for the torque of their motors and current draw?

[DaveS]

Probably, but I don't have them to hand. Think current draw of the DDM 60 and 85 peak around the 5A mark at 12V.

[markse68]

20Nm continuous for the dam 85 apparently but it can take up to 45kg load. Peak will be much higher