New DC drive system for TT -RIM drive- starts here!!!

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Stefano,

The flywheel is no problem, yes I believe I can do it.

I would like to try the Igus bushing bearings though. I believe with the proper low friction igus bearing which is like teflon or glass reinforced teflon and a good shaft, we can have very low noise. I am not sure how exactly to solve the support of the shaft, similiar to the ceramic ball or saphire ball that is used for platter spindles, but I don't think it will be difficult to solve.

I think the Flywheel pulley might also be very nice for allowing someone to use a belt, if we just just a different size pulley.

Jamie
 
Arch,

I am curious about what you are saying about feedback coming too late. I was thinking that even with a very heavy platter that eventually the motor and controller must see a "signal" to add torque or whatever, even if this signal is just a difference of angle of rotor to armature, and that maybe the heavy platter just slows down this signal, or delays the signal in time, such that it really doesn't "solve" the problem.

Is this what you are saying, in a way?

I am still very curious about this needle dragging the record and slowing the platter. If this is true Wow and Flutter and even FFT measurement cannot measure this effect since they require test tones. Is the test tone the same drag on the needle?

This is why I am curious about Pyramids BLDC idea. He may be able to see the signal of speed change in the AC Polyphase motor much quicker than any other feedback control. It might be so quick that it would be hard to measure, except as measured in the control itself. If would then be up to the Golden Ears to say whether it has solved the "needle drag" phenomenon.

I do not understand controls well enough. But I think Pyramid is thinking he is working in microseconds in feedback. I'm talking out my **** a bit, but I think the control concept is orders of magnitude meaning 10X, 100X, 1000X faster than what is currently done in AC Synch or DC. The AC synch relies on the phase lag of the rotor to the field and will correct itself, but it has no intelligence to adjust current, voltage, or dampen the control input to avoid hysteresis, overshoot, undershoot of input.

Jamie

Jamie, the drag doesn't come from tones, but rather from more complex grooves such as the one on classical music or other complex passages.
These complex grooves will cause the platter to slow down for a fraction of a second due to the elasticity of the belt and then as a spring effect the speed will increase.
This is all because of the belt. With RIM drive you won't have this problem.
That is also why I am looking at a motor with a little more power/torque which will keep speed up under every conditions and the motors I have linked here...especially the second one are more than up to the task.
Consider that the added mass from the flywheel and the use of heavy platter will also have inertia that together with the RIM drive, will not fear the worse groove modulation.

Constantly correcting using encoder will be more negative to sound than having a 0.01% slow down event for a fraction of a second. :magnify:
 
Hello Everyone
I'm extremely interested in rim drive. Had a few experiments a couple of years ago but gave up as I couldn't get the motor to provide a consistent speed at the very slow speeds the motor needs to run at in this configuration. People may be aware that Teres do a rim drive (Think its known as the Verus?) and Vic of Transfi uses it on his Salvation turntable. By all accounts the results are excellent. Therefore I'm sure that people with a much better idea of design than I can come up with a workable solution- I'm sure it'll be worth it.
Regards
Chris

Chris,

would you mind explaining a bit further the set up you had? i.e. motor type and model and also motor controlled used supply voltage and current capability etc etc...this might be useful for development and I can comment it
 
Stefano,

I mean I am understanding a little bit about the DC solution. I think there are still maybe some issues with DC, but now I understand better how your idea for the controller will be better, or solve some issues that have been included with DC.

I am still not really understanding Torque though. When the platter is up to speed isn't the torque at zero in the motor? It is free running only against friction. When the platter slows down, then the torque must increase to overcome the inertia of the platter to make it go faster, but only until it is at the right speed.

In a simple way, I am thinking that it is not torque that is the problem, it is the delay between the platter going slower and then torque being applied. Some people have mentioned that they like some friction in their platter bearing such that some amount of torque always is applied by the motor. But, this is still just adding inertia to the system, not solving the delay for adding torque.

I understand what you are saying about constantly correcting the encoder may be bad for the sound, but this is only if the encoder is too slow and is not correctly damped. Damping the encoder electronically stops it from overshooting speed. Using faster electronics with much more precise speed measurement may solve the encoder being too slow.

I can understand that possibly encoders used in the past may have oscillated. But, would not assume that someone could not fix these problems, just as you might fix the problems with DC speed control:)

There are a lot of theories about which approach is better, but these are very general assumptions and when we talk about control systems, new technology is being added everyday, and the quality of the control is very dependent on the designer. Speed control in a digital system can use very advanced mathematics, predictive control, etc. all because the digital controller can make 1000's of calculations per second. I think you are assuming that all encoders might use PID control, or other simple exponential smoothing or integrating algorythms which are prone to some amount of overshoot. This does not need to be the case when the controller could possibly use any MatchCad expression to control even learning from it's past mistakes of overshoot, etc.

These kinds of controls are used to temperature control in semiconductor manufacturing, etc. They occur very slowly in time, but the math works the same in very fast time, depending on the speed of the signal processor.

I do not know if any of this technology works for diy, but this is where Pyramid is headed with his ideas of the BLDC controller.

Jamie
 
It is the same discussion that about the amplifier, the correction arrives each time after the defect has vanished adding a new defect.
That is why some people use amplifiers witout feedback.

Music is always unpredictable, unless you use only one record.
 
Last edited:
It is the same discussion that about the amplifier, the correction arrives each time after the defect has vanished adding a new defect.
That is why some people use amplifiers witout feedback.

Music is always unpredictable, unless you use only one record.

Exactly! You got the point! It works exactly the same. My expertise is on audio design and I usually design gears with little feedback. I do think that too much feedback is bad as well as no feedback at all.
So there is a balance there and I want to bring this concept into the plate here. :snoopy:
 
Stefano,

I mean I am understanding a little bit about the DC solution. I think there are still maybe some issues with DC, but now I understand better how your idea for the controller will be better, or solve some issues that have been included with DC.

I am still not really understanding Torque though. When the platter is up to speed isn't the torque at zero in the motor? It is free running only against friction. When the platter slows down, then the torque must increase to overcome the inertia of the platter to make it go faster, but only until it is at the right speed.

In a simple way, I am thinking that it is not torque that is the problem, it is the delay between the platter going slower and then torque being applied. Some people have mentioned that they like some friction in their platter bearing such that some amount of torque always is applied by the motor. But, this is still just adding inertia to the system, not solving the delay for adding torque.

I understand what you are saying about constantly correcting the encoder may be bad for the sound, but this is only if the encoder is too slow and is not correctly damped. Damping the encoder electronically stops it from overshooting speed. Using faster electronics with much more precise speed measurement may solve the encoder being too slow.

I can understand that possibly encoders used in the past may have oscillated. But, would not assume that someone could not fix these problems, just as you might fix the problems with DC speed control:)

There are a lot of theories about which approach is better, but these are very general assumptions and when we talk about control systems, new technology is being added everyday, and the quality of the control is very dependent on the designer. Speed control in a digital system can use very advanced mathematics, predictive control, etc. all because the digital controller can make 1000's of calculations per second. I think you are assuming that all encoders might use PID control, or other simple exponential smoothing or integrating algorythms which are prone to some amount of overshoot. This does not need to be the case when the controller could possibly use any MatchCad expression to control even learning from it's past mistakes of overshoot, etc.

These kinds of controls are used to temperature control in semiconductor manufacturing, etc. They occur very slowly in time, but the math works the same in very fast time, depending on the speed of the signal processor.

I do not know if any of this technology works for diy, but this is where Pyramid is headed with his ideas of the BLDC controller.

Jamie

Jamie,
BLDC motor are noisier than DC motors to begin with and like you said if encoder is not designed well and specially the feedback loop, the overall system will be unstable which is bad.

Feedback introduces phase delay no matter what even on BLDC, but, if used in the right direction it will be a terrific solution, but if used wrong it will be terrible and bad for the sound. It is just like amplifiers!
Nevertheless, the main difference on our design is that, the project is based on the quality of each element.
This I include quality of the motor, quality of the flywheel/mechanical design and especially quality of the supply.
If all these elements are in the right place with a RIM drive, the platter will really not slow down not even under the most critical situation. A little current feedback will almost be there for show and perhaps apply that little minimum correction needed (the plan is to have the user have the flexibility to engage and disengage current feedback based on their preference!!), while the tachometer will correct for temperature drift which are only long term correction.
When motor heats up the resistance of the coils will go up, therefore the platter will slow down.
The tachometer is there to compensate for this effect ensuring rock solid stability over time.
All these commercial systems, will take 20-30minutes to get to correct speed waiting for the motor to warm up and this is really inconvenient because when you change disc, the motor will cool off a bit and speed will be slightly affected!

I think we should start on the mechanical design right now if you guys agree with that. Pyramid and I will need that to start designing electronic and software for the system.

We might want to start discussing how we can implement the flywheel. Also I am thinking I would like to see the metal enclosure for the external pod milled from the aluminum for minimum vibration. The mass of the pad should be pretty high to allow maximum dampening.

Please let me know your thoughts. I am going to start drawing a more specific concept design shortly.
 
Stefano,

Sounds good.

I am not thinking very much yet about the mechanical design as I was waiting to see what RPM you were thinking for the motor, which then might suggest motor pulley size and flywheel size. I am also curious about the motor dimensions.

I am thinking the "Pod" might need adjustable height. For it to work with the most TT's it should be short as possible with the ability to be adjusted for more height. Or, owners will need to add support for the Pod to raise it to the rim.

Jamie
 
If the drive pulley is a bit larger it allows easier placement to reach the platter over the plinth, but then the motor needs to run slower.

Do you have a "typical" platter TT height, platter distance from plinth edge we should build to? Maybe your VPI is a good reference. It's about typical height, some TT's with thicker platters might be taller, and only a few TT's shorter, which might not be good candidates for Rim drive since they might have thin plinths and too light.
 
If the drive pulley is a bit larger it allows easier placement to reach the platter over the plinth, but then the motor needs to run slower.

Do you have a "typical" platter TT height, platter distance from plinth edge we should build to? Maybe your VPI is a good reference. It's about typical height, some TT's with thicker platters might be taller, and only a few TT's shorter, which might not be good candidates for Rim drive since they might have thin plinths and too light.

Yes I was thinking the same. VPI is a standard design also probably one of the most diffused table around the world. Tonight I will take all the measurements and try to create a quoted design and maybe we can tweak it if there can be made even more universal .
Remember I am not ME so my drawing won't be the best you ever seen in your life eheheh!!!

Anyway, here is my detailed idea.
The motor will be mounted on a flange which will have the possibility to be adjusted longitudinally on the main enclosure to set for proper tension between pulley and flywheel.
I choose a square flat belt to have more surface area therefore more grip but if somebody thinks otherwise please let me know.
Both pulley and flywheel will have grooves to perfectly fit the rubber belt; this will allow for a more precise/vibration free installation and the pulley can be easily set up to the right height and distance.
The rubber belt should be properly sanded/milled for max smoothness on a lathe so that minimum vibration will be transmitted to the platter due to the irregularities of the ring.


Please feel free to comment on this.

The main key points here are that the motor position can be adjusted and secured on the metal pod and sits on some sort rubber or whatever material thus isolating it further from the pod. The rubber square rubber belt belt will increase grip on both pulley and platter and be less prone to slippage. Flywheel will also absorb along with the mass of the flywheel any residual vibration coming from the motor therefore any vibration should be ideally drained away from the platter resulting in a quieter reproduction.

If this idea is ok with folks here, we can proceed to create a mechanical model work how spindle and flywheel can be created as well as metal pod (very important to kill vibrations and it has to be reasonably heavy). While mechanical design is being created Pyramid and I, will start doing the preliminary design work.
 

Attachments

  • design concept_1.JPG
    design concept_1.JPG
    168.1 KB · Views: 153
I am planning on having the motor spin at about 400-500 RPM for 33 1/3 RPM which will give us some headroom for supply and torque control. since the flywheel will need to be a little wide to accommodate different tables and also to have a decent mass, we can have the motor spin a little faster.
I will do some scaling work tonight and try to dimension the pod.

See if I can allocate some time to do that tonight so you guys can get started on the mechanical design.
I might have an evening of downtime from my project since I am awaiting for parts and prototype which should arrive on Friday, so I will take advantage of that.
 
I'm a complete 'noob' regarding TT but this thread looks interesting. I don't understand why you have a flywheel in this design, the turntable will be a flywheel so why add another wheel ? - looks to me that it is more a way to provide some mechanical isolation between the motor and the TT to reduce noises ?
 
thanks for your interest. I am glad this project is turning out to be of interested to so many people, it is really exciting.
The flywheel is there to have more mass, although not much compared to the platter, it will still add some extra flywheel effect, but mostly, it will help in lowering vibration going to the stylus which is the main problem.
Considering the average platter will be around 15-20lbs adding a 5lbs for example is almost adding more than 25% of mass to the entire system.
Bul like I said the most use will be to lower vibration. On RIM drive it is very critical. The lower the noise the better the overall results.
The flywheel is not powered up directly by the motor but it will spin on a low noise/friction reversed (?!?) spindle and this is what goes in contact with the platter.
There is a big difference from having the platter powered directly by the pulley or by the passive flywheel. :magnify:
:cool:
 
Stefano & others:

I want to relay the information that was given to me over the years in talking about turntable design/drive systems etc.

When I told Thomas ( TW Acoustics ) that I wanted to build my own turntable we talked for hours as I had many questions. Thomas being a former high school teacher and as a friend explained several key points of all the high end turntables that were out giving me both the strong points and weak points of each design. He explained that a rim driver system will never equal a good belt system sound wise for the following reasons.

1. Thomas made several variations of a rim drive system and none were as quiet as a belt drive.
2. Too many moving parts and the parts have
to be ultra precise and balanced to be acceptable.
3. The "O" rings or mini belts has to be precision ground and ultra precise.
4. When not in use, the drive system would have to be totally disengaged or the " O " ring would get a small flat spot or dimple and can be heard while listening to music.
5. He ( Thomas ) has personally owned highly modified ( by him ) and stock Lenco, Gerrard 301 & 401 turntables and found them inferior in overall performance.
6. If the rim drive system sounded better he would be using it.
7. On a properly designed drive system a flywheel is unnecessary and adds complication and noise which is unavoidable.

This information is given not to bad mouth rim drive turntables as I have no personal experience with them. It is knowledge gained by Thomas over many years of research and practical experience and cost lots of money.

Rim drives were built to survive being played at broadcasting stations day after day much like the Technics SP 10 series and not the ultimate for audiophile use.

Thomas uses a large pair of Cessaro horn speakers and top Tron electronics as his reference system.

No drive system is perfect, however the belt drive is the least offensive ( as of now ) So pick your poison.

Hope this helps.
 
Hi,

One needs a little humility to attempt expensive state of the art
design. Here we just have self-opinionated musings on what
does and how things work, which are not well informed.

Good luck with it, but I'm not going to engage in another
slanging match about how much you really know about
what you are attempting to do as in another thread.

rgds, sreten.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.