My DIY Friction Drive Mount #maxesp3 #esp32 #cnc #diy #mount


Hi everyone, this is my 1st post here...and first I want to THANK all on-step group's members for their GREAT job. THANK YOU!!!

My name's Luca, I'm 37 yrs old and I'm from Italy. I'm an amateur astrophotographer and since I started this hobby I've always tried to find "DIY" solutions ...This includes a DIY EQ platform for my Dobson, several motorized focusers, automated dust caps, adaptors, fittings and so on. I dont' have a real workshop, but I own a CNC router I built about 8yrs ago and a mini-lathe (an old HobbyMat MD65 that I restored few months ago) and...well... The router is in my attic and the lathe is outside, on the terrace (and yes, my wife is a pretty understanding person....) .
I own an HEQ5 I retrofitted with SKF bearings and some other improvements, and I'm pretty happy with it, but I wanted something "different" why don't make my own mount?!?!  I opted for a friction drive mount, as most parts can be machined by myself, except for the discs... so...this is what I drew in Solidworks ("Sadly" I use it everyday...):

During this months I've machined almost every single part I needed , except for the discs: they were machined, heat-treated and grinded by a professional mechanical workshop , based on my drawing , with stricted machining tolerances.

Some technical infos:

- Disc diameter: 180mm
- Pignon diameter: 8mm
- Motor: "High precision" 200steps  NEMA 17 reducer 1:10 (with minimum backlash)
- Total reduction: 1:675
- Theorical resolution: 0.15 arcsecs/microstep with 1/64 steps 
- ESP32 module + 2130 stepper drivers ; Same settings as MaxESP3 - GoTo rate set to 5.0°/sec

Now it's almost 95% done, I've not any issues with the friction system, everything run smooth and without backlash or slipping if well balanced with minimum preload... I can't wait to try it outside, hoping everything's good.

Here's two short clip I made this morning:


Moving around...

And few images:


Khalid Baheyeldin

This is an amazing piece of work.
Shows talent and determination ...

But I have a question: you used a 10:1 integrated gearbox in the motor.
It is understandable that you want more physical gear reduction for more resolution and more torque.

But with a friction drive, the attractive part is no periodic error.
Now with the gearbox there will be non-periodic error.

I don't know what the solution to this is, beyond a larger wheel, which is not easy.
Maybe pulleys instead of a gearbox?

Or perhaps this is all "theoretical concern", and when you test it it works as intended.
Specially as you have a guide scope/camera anyway ...

andrea tasselli

This is unbelievably good piece of work you have done there. I cannot wait for you to report how well does it work for real. In the meanwhile I'm curious to know what kind of CNC router you have been using. I mean, short of industrial grade machinery, all I have seen isn't supposed to carry out that much pocketing you have done, never mind the other bits!


Thank you Khalid,

I forgot to mention a belt and pulley reduction (1:3) between the gearbox and the friction system (so that overall reduction is 1:675). You're right, maybe there will be a non-periodic error with those gearboxes, but I know other manufactorers use them (Open loop,  With no encoder) ... I'm pretty confident it won't be an issue with autoguiding. I tried last month to autoguide with RA only and a "slice" of 30º I machined from aluminum (it was TERRIBLE...) and it didn't go beyond 0.7" RMS ... so...FINGER CROSSED!!! 🙂

@andrea tasselli I built my router on my own , first with MDF panels and with that one I machined all the other parts in Aluminum. Here's few pics and a video of another cool project (a carbon fibrr/aluminum wristwatch) I managed to achieve with it:

Jay Murphy

Beautiful, bravo, wonderful craftsmanship. 


Beautiful...I'm green with envy.

On Sat, May 22, 2021 at 1:32 PM, Jay Murphy
<stercust@...> wrote:
Beautiful, bravo, wonderful craftsmanship. 

Howard Dutton

Amazing levels of skill in several disciplines at work here.

Nicely done, looking forward to reports as to the performance.


Thank you all :-) 

I want to share some "tests" I did before the mount was complete. As I said before, I tried to set-up only the RA for auto-guiding and the first thing that surprised me was the calibration datas I got from PHd2:

It wasn't showing any evidence of I'm pretty happy with it. Then I tried to evaluate the Backlash with no-load, stopping the tracking and moving the RA axis at 1X sidereal rate...and again the "inversion" occurred without any lag:

Backlash Test

An this final clip shows why I decided to have the discs machined by a professional workshop (and also why I'm so proud of the results I achieved with my little old lathe)



That's far more than just awesome Luca.. exceptional work and your wristwatch matches your skill here
Can I ask how much it cost to make, more than your eq5?.. I've just joined this group and thats certainly a wow moment, keep the posts coming as interested in how it performs in the future. What do you think it's payload is?


Beautiful work on both the mount and the CNC! 
What mechanism do you use to apply normal force between the friction drive shaft and the disk to achieve sufficient friction force?


Thank you 😊...

Here' a clip about preload system (far more easy to explain it ...):


I calculated that a preload of 250N can stand a 15Nm unbalancing torque, that's really a lot, before slipping. 

I can't tell how much the payload is, but I can reasonably say it must be under 10Kg for astrophotography use... 
Total cost is way below an used HEQ5 (but it required me about 4 months of work in my spare time...). Anyway...The most expensive parts are the discs - and you'd understand why if I show you the drawing...

Khalid Baheyeldin


I added your mount to the Showcase page, under custom equatorial.

Please update this thread when you test it under the stars, with the PHD2 graph, and a sample image.


Hi there!

Well, last night It has the FIRST LIGHT (yuppeeeeeee!!!) , despite the weather conditions were not the best was cloudy and little windy, but I decided to give it a try.

My aim was to check polar align mechanism first, then test the plate solving and pointing accuracy and in the end collect some datas from phd2 guiding...out of the box! Here's a small rep:


ES 102 apo carbon + ASI533 MC Pro + DIY Autofocuser
Zenithstar Z61 apo as guiding scope + ASI 120mm mini
N.I.N.A. as imaging Software
phd2 with multi-star guiding 
Sharpcap for polar alignment

- Polar alignment with sharpcap was easy as usual and in a few seconds it was done issues except from some kind of flexure I had due to 3D prinded bushes that I plan to make in Brass "one day"...Once achieved an "Excellent" result I switched to N.I.N.A.

- I looked for a pair of common objects (M81 , M3) and the mount didn't had any issues finding and pointing them ...after two tries both the objects were straight in the center (yep, not at the same time obviously...😅). Please note that I didn't aligned the mount using alignment stars, I've not been doing it for a long time , after discovering the magic of plate solving and syncronization 🙂

- I switched on phd2 and calibrate .... I didn't get a straight 90º angle between the axis but never mind, it was good anyway...Choosed a bunch of stars between moving clouds and start guiding...and this is the first graph I got:

I was really happy with this results, considering such a bad weather conditions... usually I can't go below 0.9" with my Tuned HEQ5 pro !! Then I tried to point to another region, and I got this guiding graph:

...and again.... I was really surprised (by the amount of dirty I have on my ASI120...)

Last but not least I tried to get some infos about dec backlash, using the guiding assistant ....even if the wind was growing:

Really really happy with this overall results OUT OF THE BOX!!!!

In the next days I'll dig more deeply in the tweaking and tuning (especially on the sidereal rate values..) and I hope to try my mount without auto-guiding too. Then I'll collect more data to be analyzed with phdLogviewer...

Mike Ahner

Incredible & Beautiful! Congratulations!

Khalid Baheyeldin

I can't read the RMS numbers in the guiding graph screenshots.
Can you please send them as a reply, in text form? Everyone can read text.


Oh, I'm sorry Khalid.

In the first image  was

0.71" RMS, the 2nd one 0.66" RMS 🙂

Howard Dutton

Not too much difference between RA (0.53") and Dec (0.47").  Most noise happened in bursts, perhaps wind.
Looks promising to be capable of better.

Gildo Bosi

Fantastic project, congratulations, what city are you from?

Gildo (Bagnacavallo, Ravenna)

John Petterson


This is phenomenal work.  Thanks for sharing what is possible. 

With a friction drive, you might want to consider adding a set of encoders to the mount that would allow you to maintain the OnStep's alignment if the mount gets bumped, or if you manually swing it to another target.  I know there are steps you can take through the menus to sync it to the new location as well.  But the machine work on this is very nice.


Thank you John for yor reply and suggestions. Yes, an encoder may be worth it , and I'll take it in consideration in the next "release" if there will be one....but It will be an High resolution one for sure. 

@Gildo I'm from Brescia 🙂