diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Analogue Source (http://www.diyaudio.com/forums/analogue-source/)
-   -   Linn Lingo vs. Dr. Fuß or Square-Wave vs. Sine Wave Oscillator for Motor Control (http://www.diyaudio.com/forums/analogue-source/212688-linn-lingo-vs-dr-fuss-square-wave-vs-sine-wave-oscillator-motor-control.html)

tiefbassuebertr 15th May 2012 11:13 AM

Linn Lingo vs. Dr. Fuß or Square-Wave vs. Sine Wave Oscillator for Motor Control
 
3 Attachment(s)
There are several motor control units (in Germany sometimes called wrongly "Power Supply") for turntables/record players with synchronous motor.
Most of them uses therefore a quarz oscillator (square wave), divider and low pass filter so as one or two power amplifier sections (mostly totem pole topology) for the motor coils.

But now I have heard about an other approach for higher spectral purity of the sine wave signal (i. e. lower distortion):
The Dr. Fuss motor control. Check therefore this URLs (control device with Wien-Robinson-Bridge sine wave oscillator section):
Erfahrungen mit Netzteilen von Dr. Fuss - fairaudio - leserbericht (1)
http://www.frank-landmesser.de/hoer_...obinson-Bridge

The circuit describtions for such an oscillator circuit you will find there:

http://www.ti.com/lit/an/sloa060/sloa060.pdf (page 17/18)
Types of Quadrature Oscillator | eHow.co.uk
http://www.mics.org/getDocum.pdf?docid=1332&docnum=1
Quadrature Oscillator (Part 1) | Liivatera: Diary from the workbench
Liivatera: Diary from the workbench
http://www.ti.com/lit/ds/symlink/lm4766.pdf

Dr. Fuss uses either a Wien-Robinson-Bridge or a Bubba Oscillator/Quadrature Oscillator ( according TI "Wein bridge Oscillator" resp. "Wien bridge Oscillator") and an audio chip amp, model LM4766TF from National Semiconductor.

Linn uses fore the LINGO two quarz oszillators and two power amplifiers for 33 and 45 rpm and for the Valhalla PCB one oscillator and one power amplifier so as an additional capacitor for the second coil (phase shift) - go to the attached files.

What are the pros and cons of both ??

Thank you for your comments.

Mark Kelly 15th May 2012 12:07 PM

Your question is not so much about square wave vs sine wave as it is analogue vs digital.

Analogue techniques give good adjustability and can give excellent waveform purity but this comes at the expense of accuracy and stability. With more sophisticated techniques however it is possible to approach digital levels of stability. 50ppm is the best I've managed with ana analogue generated motor controller, using parts with matched tempcos in a single oven. To achieve better than this would require a double oven.

BTW the Fraser oscillator is a much better choice than the Wien bridge for quadrature circuits.

Digital techniques offer good precision but it takes some sophistication to achieve good frequency variability. It takes another layer of sophistication to achieve precise / adjustable phase difference.

The square wave generators use brute force LPF to get something approaching a sine wave out the other end. There are digital techniques which don't suffer from this, analogous to a DAC with a variable clock. The best purity I've managed with a digitally clocked motor drive is about -90 dB which only served to prove that anything beyond about 60 db is wasted.

valgamaa 15th May 2012 08:17 PM

Hi tiefbassuebertr, I'm not familiar with motor speed control but am very familiar with quadrature oscillators. If you don't need accurate quadrature signals then the phase-shift circuits like Bubba are a good choice, but there is likely to be a variation in harmonic content at each output as there is increasing filtering as you get further from the sustaining amplifier. For variable frequency, the op-amp and resistor can be replaced with a gm-cell (like the LM13700).
For accurate quadrature an alternative is to use an integrator/comparator-type triangle oscillator, with the zero-crossing of one oscillator switching the integrator of another oscillator. For both cases i agree with Mark that 50ppm stability is probably a practical limit if the oscillator has much tuning range.
With a digital divider clocking two flip-flops on opposite edges to generate a quadrature signal the duty-cycle is important, and 2-3 degrees error is about the best you can expect. To improve this a polyphase filter can be used in cascade, and this can reduce error by perhaps a factor of ten. It will also reduce the harmonic content to give a reasonable sine wave.
At this point I have to declare an interest as the author of the liivatera website that you referenced in your post, and although I need to complete the write-up the VCO described has harmonic content >80dB below the fundamental, and fractions of a degree phase error. This is my preferred analogue solution to your problem, and I am currently writing a paper on this for an IEEE conference later in the year. For the digital option, I would go with a quadrature divider and polyphase filter without hesitation.
How critical is the quadrature accuracy and harmonic content in the motor drive signals, as this will ultimately dictate what options you have,

Phil.

Mark Kelly 15th May 2012 11:59 PM

Quote:

Originally Posted by valgamaa (Post 3024782)

How critical is the quadrature accuracy and harmonic content in the motor drive signals, as this will ultimately dictate what options you have,

Phil.

In my experience, both absolute harmonic level and harmonic spectrum are important. You want 2HD below -60dB. 3HD is actually helpful with coggy motors like the Hurst or Premotec, if you can tune it (I designed an oscillator with tuneable 3HD for this).

Quadrature accuracy: again if this is tuneable you will find an optimum for any motor. In my exerience this is usually a little off exactly 90 degrees and changes with each motor.

tiefbassuebertr 19th May 2012 11:04 AM

Thank you for your comments.
After looking additional motor control units by several users, I discover an additional possibility for the kind of driving the motor coils - in fact two low voltage power amp and two transformers, approximately 1:10.

Thus there rises up an additional question: What are the pros and cons between a totem pole high voltage amp like Linn Lingo and a normal push-pull power amp with two additional adaptive transformers for the 110VAC motor ?

To evaluate the context of the different kind of oscillators and the different kind of driving stages for the synchronous motors from the audible view, one must spend more time in this matter.

Necessary are both kind of motor driver amps and both kind of oscillators to investigate the differences - once for comparison of the spectrum analysis (measuring record) and once for comparison by listening test.

Perhaps one of the member has already done this.

sq225917 20th May 2012 04:51 PM

Funny you should mention this..

I ended up recently using a 110v/500 va transformer and a couple of 0.1uf caps to power my Kuzma. This was after trying and measuring, and modding and remeasuring An original Valhalla, A Lingo and a Hercules. The Hercules was the best of the Ac regenerators but it still wasn't as good as the trusty UK mains and a big hunk of copper.

sq_kuzma_polar | Flickr - Photo Sharing!
Shows the standard hercules vs the basic Armageddon (transformer PSU) clone.

sq_geddon_recap_polar | Flickr - Photo Sharing!
Shows the modded Hercules, (Q-speed diodes, Vishay caps, phase trim caps set exactly to motor values) vs the Armageddon clone using the same phasing capacitor values.

I ended up using the transformer psu and modding the bearing via increased viscous drag to load the motor and smooth out cogging. I assume the extra drag increases magnetic flux in the motor which somehow smooths it out. My deck is unsuspended so 3rd order effects can be pretty much excluded.

tiefbassuebertr 21st May 2012 03:30 PM

2 Attachment(s)
Quote:

Originally Posted by sq225917 (Post 3030550)
Funny you should mention this..

I ended up recently using a 110v/500 va transformer and a couple of 0.1uf caps to power my Kuzma. This was after trying and measuring, and modding and remeasuring An original Valhalla, A Lingo and a Hercules. The Hercules was the best of the Ac regenerators but it still wasn't as good as the trusty UK mains and a big hunk of copper.

sq_kuzma_polar | Flickr - Photo Sharing!
Shows the standard hercules vs the basic Armageddon (transformer PSU) clone.

sq_geddon_recap_polar | Flickr - Photo Sharing!
Shows the modded Hercules, (Q-speed diodes, Vishay caps, phase trim caps set exactly to motor values) vs the Armageddon clone using the same phasing capacitor values.

I ended up using the transformer psu and modding the bearing via increased viscous drag to load the motor and smooth out cogging. I assume the extra drag increases magnetic flux in the motor which somehow smooths it out. My deck is unsuspended so 3rd order effects can be pretty much excluded.



thank you very much for this explanations. Unfortunately my knowledge and my english isn't enough to understand all this exactly, particularly the marked aeras.
Follow questions rises up:
1) what means "geddon" ? Perhaps the Armageddon from Naim Audio as to find about
Naim Audio Forums: Armageddon - Powered by Hoop.la
2) What I can observe by your diagrams, attached here as PDF?

sq225917 22nd May 2012 10:38 PM

Geddon= Naim Armageddon PSU= a big hunk of copper (transformer).

My Kuzma deck has been re-engineered to take an LP12 main bearing. This modded bearing runs a very high viscosity fluid instead of oil. This provides some resistive drag against which the motor must work, and appears, and measures to help smooth out the motor and reduce the effects of cogging.

The graphs display a frequency demodulation of a 3150khz tone recorded from a test record. The closer to a perfect circle they are, the better the speed stability. Each image compares two different samples, red and green. The 'Geddon recap' sample shows a speed deviation of between 0.01-0.02% In fact most of the error shown appears to come from the off center pressing of the test record.

Each rotation on the graph = 1 rotation of the record, 5 are used for each test sample.

Si

tiefbassuebertr 23rd May 2012 07:18 AM

Quote:

Originally Posted by sq225917 (Post 3033195)
Geddon= Naim Armageddon PSU= a big hunk of copper (transformer).

My Kuzma deck has been re-engineered to take an LP12 main bearing. This modded bearing runs a very high viscosity fluid instead of oil. This provides some resistive drag against which the motor must work, and appears, and measures to help smooth out the motor and reduce the effects of cogging.

The graphs display a frequency demodulation of a 3150khz tone recorded from a test record. The closer to a perfect circle they are, the better the speed stability. Each image compares two different samples, red and green. The 'Geddon recap' sample shows a speed deviation of between 0.01-0.02% In fact most of the error shown appears to come from the off center pressing of the test record.

Each rotation on the graph = 1 rotation of the record, 5 are used for each test sample.

Si

Now I can comprehend all this - thank you very much

tiefbassuebertr 25th June 2012 12:33 PM

A DIY project in German:
Roehren-und-Hoeren :: Thema anzeigen - Luxusnetzteil für Synchronmotoren mal anders


All times are GMT. The time now is 11:08 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2