Air Bearing Spindle DD Turntable

[ebrewste]

Shaun,

I don't use BLDC controller chips. The price difference between chips and pre packaged drives are small and not worth the headache, in my opinon. I love the Trust drives, I mentioned in the previous post. Good product and good people. Linear drives are really just glorified opamps, anyway -- you don't really need a chipset. Any linear drive will do the job, I bet.

The big deal about pwm / linear for this amp is electrical noise, followed closely by audible noise from the motor. PWM drives are really a neat trick, but they aren't perfect. They are just like DACs in that they really suck unless an infinite amount of work is done to perfect them. DACs have benefitted from this, while industrial PWM drives have not. It's just a matter of market size, not good / bad intentions.

Erik

[ebrewste]

Christopher,

On the high mass / low mass thing, high mass is easier. You can of course make things much better with a good control loop, but there are two major problems:
1. You have to make a good control loop. This is harder than making a heavy platter.
2. In order for a feedback control system to work, there must be an error. As the control loop gets better and better, you minimize the error, but you never eliminate it completely. This is one of those first principle things.

A heavy platter is another type of control loop. It's not really that different than a closed loop feedback system, but it's a lot simpler (heavy = good).

Both systems do the same thing. Double the gain, all else equal, in a feedback control system, and you increase disturbance rejection (good thing) 2X. If you double the inertia of the platter, you increase disturbance rejection 2X.

If it was me, I would make a really light platter and do all sorts of fancy controls stuff. But I'm a professional controls nerd. I would always suggest a heavy platter to the normal person -- it's easier and possibly more effective.

By the way, with a direct drive system like you have, a heavy platter and a high bandwidth control loop are not mutually exclusive, so you could do either at your leisure.

Sorry for the rambling responses, you got me interested

Erik

[Shaun]

Originally posted by ebrewste
I am a controls specialist for systems like those.

ebrewste
As a matter of interest, are you involved at system integration level, or component level?

Thanks for all the invaluable info! Forgive my enthusiasm (as it is not my project!).

Edit: above question already answered..

[ebrewste]

I do both integrations for OEMs, mainly focused around controls theory and general programming and controller R&D.

Erik

[Shaun]

I priced some linear drives, and they do not seem to be in the ballpark of what a cheapskate DIYer like me had imagined.

TA305-A15 $635ea(usd)
TA310-A22 $770ea(usd)

Am I looking at the right product? From what I can glean, I'd need to add a controller and/or PID device...?

[ebrewste]

Those are the right ones. Linear amps are expensive, no matter how you look at it. If you wanted to, you could either go pwm, like the Elmo product or go largely DIY.

A linear motor amplifier is more or less a stereo amplifier. Lots of people (myself included) have made LM3886 based chip amps, which would likely get you going. You just hook up the two amps to two phases of the motor and Kirchhoff will make sure the third phase is right. I think the chip amps have plenty of power for what you will need. I suspect you will get closer to the TA305 price this way than you are comfortable with. You will, however get almost an infinite improvement in your understanding compared to the TA305 route.

You will need a controller, which will have PID or similar. The microcontroller offering can do this. However, you will need to see if they do commutation. There needs to be some solution for this.

I would look at the resources you are willing to apply to this. If this is a learning / infinite time project, make your own amps and use a microcontroller based controller. If you want to do this faster and more simply, check out the Elmo stuff.

[Shaun]

Thanks for the encouragement. I have done some microcontroller programming before (not professionally, but during my studies). I know from experience that I can build it for significantly less money. But you correctly alluded to this being a time-consuming task. Well, time is money, so I guess from that perspective you are right about the cost being similar between the two options.

Chip amps, absolutely; though I had been also considering power op amps or dedicated driver chips.

If one were developing a commercial product, I guess one could start with a building blocks approach such as linear amp, PID, etc., as a proof of concept. Then, with lessons learnt and parameters defined, build a dedicated controller.

[monsieurphot]

Hi everyone-

I apologize for not being more active on the board, which is after all, for my project. I've decided to apply to a watchmaking school located in Seattle, so I'm in the middle of planning a trip up there (I'm currently in Los Angeles) to take the entrance exam.

I had thought of using a development kit from Luminary Micro or Atmel as a controller- there are a few options out there, and I figure at least it would give me a foundation to build on and allow for proof of concept, as well as an introduction to programming. They also seem reasonably priced compared to commercial units. The electronics end (and programming) are somewhat new to me- I have some experience building kits and doing repairs (I worked as a video poker tech for some time), but the theory and thought of fabricating my own PCB's is a little intimidating.

If it turns out the motor cogs too much, I would still like to use the air bearing as the spindle assembly. In that scenario, drive would either be via belt or idler wheel, with the possibility of machining some sort of adapter that would let the spindle be driven directly if a suitable motor could be found.

I'll start working on the mechanical design tonight, and try to post some drawings.

Christopher

[Shaun]

Originally posted by monsieurphot I apologize for not being more active on the board, which is after all, for my project. I've decided to apply to a watchmaking school located in Seattle, so I'm in the middle of planning a trip up there (I'm currently in Los Angeles) to take the entrance exam.

Hey Christopher

No problem, as long as you don't mind me poking my nose in. I think the development kit idea is good, though the programming bit can be daunting.

Good luck with admission to the watchmaking school. It is an unusual occupation in this day and age. Perhaps you can then teach us about which jewel bearings to use in our tonearms.