Converting 6N to direct drive

[nixiebunny]

I was recently asked to figure out why a belt-drive converted 6N turntable had so much wow. It was converted using a low-cost packaged BLDC motor with a simple UC2625 speed controller that has no PID tuning control, so of course it has a lot of wow.
This got me to thinking about how to make the platter itself into a BLDC motor. For reference, my day job has me working on radio telescopes, one of which has a pair of exotic BLDC motors built into its azimuth and elevation mounts, with a long string of rare earth magnets and several lumps of 3-phase drive coil sets.
My idea is to mount a steel ring inside the outer lip of the platter, and then place a ring of alternating N and S facing rectangular magnets inside this ring.
The ID of the ring would be chosen to allow exactly 2N magnets to fit snugly next to each other.
Then some multiple of 3 coils would need to be wound to drive this homemade rotor, mounted to the 6N base plate. That's not difficult.
What do you think?

[markrob]

Hi,

If you have the chops to pull it off, go for it. If you are looking for a fun project, sounds like a great one. I suspect you will learn quite a bit and spend serious time getting it up and running (the devil is in the details) as compared with the traditional belt drive servo approach. It looks like the 2625 can easily be adapted to external feedback control either using the Hall sensor derived TACH output or via an external encoder (probably a better option).

If you end up going that route, please share the details of the project with us. Before he was banished from here, Todd opcode66, demonstrated his home brew direct drive build. Not sure of the final performance specs that he attained, but it was an interesting idea.

Mark

[nixiebunny]

Thanks. I think I can handle it. So many people don't think that you can build a motor because they're full of magnets and wires, but magnets and wires are pretty darn simple, other than their interesting field effects. So I'm planning to set it up so that I can easily change the magnet geometry and drive current profiles to achieve minimum wow and flutter. I could even go so far as to simulate it, but that's not usually my style.
I have already figured out the basic design. A flat waterjet-cut steel ring under the platter holds a few dozen magnets. A laser-cut plywood ring centers this pole piece ring, and another stacked on it positions the magnets in a circle, and has an encoder zebra ring cut into it on a smaller diameter. The magnets and ring will eventually be glued into place.
The forcer is a steel pole piece bolted to the two original motor bracket mounting holes. It has pointy ends to minimize cogging. The three coils are positioned onto it in 120 magnetic degree spacing increments with stacked laser-cut plywood forms. The encoder pickup and Hall sensors are also mounted on this wood thing.
Now I just need to work out a few little details...

[markrob]

Sounds awesome. Hope to see it in action. Are you planning to open source the design? I think there are many here that would like to build this for our DIY lathes.

[RolfNoot]

Long time ago I designed a BLDC motor driver. It consists of an accurate xtal oscillator (.5 ppm), a phase locked loop, a filter and a microcontroller to generate the 3 sinusoidal drive signals, a triple ADC and 3 voltage controlled amplifiers.

The speed is controlled by the RMS voltage of the signals, not by the frequency of the signals like a stepper motor.

I’ll be a fun project

Regards,
Rolf

[nixiebunny]

Sounds awesome. Hope to see it in action. Are you planning to open source the design? I think there are many here that would like to build this for our DIY lathes.

Yes, I certainly plan to describe what I'm doing in detail, and publish the drawings and code. It's too fun to keep bottled up.
I have just found an open-source Arduino library that does just what is needed for the controller. It's called SimpleFOC, which is a field-oriented control implementation.
This is an algorithm that provides extra-smooth motion using current and phase angle feedback and optimizing the motor currents based on these. It should run easily on a small Teensy or ESP32.
I also bought a bunch of 1/2 x 3/4" magnets to start building a motor with. Next, I will make a drawing of prototype pole pieces for a local waterjet fab house to make some steel for this project to get off the ground.
But first I need to test my new spectrometer at the big telescope.

[nixiebunny]

I made the first version of the rotor today. The ring is 1/16" mild steel from Ace Hardware, cut out with a jigsaw. The magnet guide plate is 1/8" acrylic that was lying around.
The encoder notches are placeholders. I will add many more for a high resolution angle readout.
I also plan to add two analog Hall sensors in quadrature. I can test which is more effective.
Next up. build a forcer and spin the motor as a generator to see the waveform. I'm hoping for a pretty clean sine wave.

6Nrotor-guide.jpg

6Nrotor-ring.jpg

[symatic]

awesome stuff. I'd love to see how this turns out. I'm planning on driving my platter with a turntable but a dedicated motor is the dream

[farmersplow]

A great project! It looks like you (mostly) know what you're doing. The first steps have already been taken. I'm very curious how it will continue - especially what will ultimately come out of it.
Stay tuned - I will continue to follow the project with pleasure.
Greetings from Austria
Thomas