A perfect application for Genetic Algorithms. Let the GA "evolve" a waveform while optimizing the Fitness Function == constant speed. Let it run ten or twenty thousand generations and look at the speed plots. Are they sufficiently better than the motor does all by itself with no waveform finagling? If so then program the optimum waveshape into the ROM of a DDS engine and Bob's your Aunt's husband.
SciPy has a very nice GA built in, all debugged and tested and ready to go. It's called "Differential Evolution".
As long as you've got the test gear set up and running, have the GA discover (and save!) the optimum waveshapes for nine different settings of tracking force. Or three different stylus shapes. Or numerous different tonearms with different arm shapes, effective lengths, total masses, etc.
SciPy has a very nice GA built in, all debugged and tested and ready to go. It's called "Differential Evolution".
As long as you've got the test gear set up and running, have the GA discover (and save!) the optimum waveshapes for nine different settings of tracking force. Or three different stylus shapes. Or numerous different tonearms with different arm shapes, effective lengths, total masses, etc.
Few months back while searching for motor cogging and how to avoid it I came across some website (I have not saved the link and now don't remember) which showed use of algorithm and software which predicted the motor cogging and rectified it. From whatever I remember it gathered data of a particular motor and predicted the varying torque and compensated it with some electronics.
Regards.
Regards.
Not sure how you would use this here. Model the process or actually feed the waveform into the TT and measure the output and feed it back in? We have a Python script for using a test LP but as Bill said you just end up tuning to the mastering lathe and test LP.
Good call,
As a matter of fact, when designing balanced 60/120V systems per US NEC 647.7(4):
"All 125-volt receptacles used for 60/120-volt technical power shall have a unique configuration and be identified for use with this class of system."
We specify L5-15 twist-lock connectors for use with the balanced power outlets to meet code where the balanced power is a permanently installed system. Where we develop balanced power as a separately derived power source using free-standing cordset transformers this is not covered under NEC so a common 5-15R can be used. However all receptacles carrying balanced power including those on a free-standing transformer must be GFCI protected for obvious reasons.
Different codes apply elsewhere.
Cheers!
Howie
Slightly different in tone, sure, but how is your response to Scott so much different? You spend so much of your
BTW, $1300 in 2007 would be a 30 year mortgage payment on a $210,000 house. Median household income was 50k, already implying a debt to income ratio of ~33% before taxes/insurance, which are usually incorporated into that ratio, so closer to 36%. Surely this just the type to run out and get that little auditory bliss as they leverage their credit cards to the max and forgo any form of savings to help their noisy kids college fund. Yes, audio equipment at this price bracket is the stuff of the proletariat. I'll be generous and assume this argument was poorly formed in the heat of agitation. Tournesol, do you want to defend Richard's logic here?
I recommend you consider the motor waveform to be the sum of (a) an ideal perfect sinewave, read out from a ROM; and (b) an "enhancement" waveform. Then have the GA optimize ONLY the enhancement. With this partitioning, the enhancement is low amplitude, low dynamic range, inverse-distortion waveform that is separate from the big ole honker sinewave "common mode". Another advantage is that you can install "stupidity filters" between the GA and the summer (i.e. upstream of the motor), in case the ignorant GA tries something ridiculous that would really, really hurt the motor. Feed the summation into a 28 bit DAC, apply output filtering, and drive the motor.
Sense platter speed using something other than the groove and stylus. Perhaps optical sensing with 32 equally spaced tick marks around the perimeter of the platter. And a digital tachometer running from an 0.5 ppm crystal. I expect you do want to perform the platter speed optimization while a record is being played by a cartridge/stylus, just in case stylus-down has a teeny effect upon the speed. After all we're trying to optimize away deviations that are, themselves, teeny.
GA reads the digital tachometer to calculate the "Fitness" of each member of the population. Fitness = absence of speed variation.
In 2007, Elektor published a mains conditioner project based on two back-to-back transformers topology and filtering:
Interference-free Sound | Elektor Magazine
They also included some measurement results.
Regards,
Braca
Interference-free Sound | Elektor Magazine
They also included some measurement results.
Regards,
Braca
At this level of sophistication, you'd need to have a very refilled 3d model of the motor + coupled system in 3d and validation. Fun exercise to play with but I'm okay with letting John have his fun with motors. The mischief I'm looking at is getting a 1.5-2 how accurate synchronous motor and a vector drive for the table saw. Ridiculous? You bet.
Buddy of mine rebuilt his bandsaw using a DC motor (!) whose speed was controlled by a phase locked loop that he designed. The poweramp he built, which translated the PLL output, into DCmotor input, was impressive.
He demonstrated the whole apparatus by nonchalantly feeding a 10 inch diameter pine log, into the bandsaw. The sounds it made sure did change, but the motor speed (conveniently displayed on a digital readout attached to the PLL board) remained rock steady.
He demonstrated the whole apparatus by nonchalantly feeding a 10 inch diameter pine log, into the bandsaw. The sounds it made sure did change, but the motor speed (conveniently displayed on a digital readout attached to the PLL board) remained rock steady.
Last edited:
Yes, Scott and Waly are always nibbling at my ankles. Guess they think its fun and that such opinions are a big influence on others.
Yes, 1300 is a bit too much for many people. I cant help that. But there were many lower priced models as well that were offered which had good audible results for people. buy one of those.... models of the 5000 series or 6000 series. All models are available at 'affordable' prices on the used market/internet. I see no point to that cost argument of the past. Just another red herring. I have got used to them as have others. Once in awhile I have to give it back to them and they get all indignant and cry when I do it. So that argument is none sense. No excuses now. Plus, I gave you a cheap-o alternative way.
I haven't read the Elector art but I am sure they did a good job for a universal solution to the recognized problem. .
THx-RNMarsh
Last edited:
Love the story. I converted a 1950s sewing machine to pwm drive (freebie off Craigslist, brute can handle upholstery fabric, which I'm doing). Has a cheap universal motor. Directly rectified 120v was hitting full speed under load with 20% duty cycle so I stole an old transformer from my dad's 70s receiver that died. Pretty much perfect step down and the motor runs nicely. Might hook up a Arduino nano and a Hall effect sensor to the motor shaft and have the foot control potentiometer control motor speed instead of duty cycle.
Silly fun stuff if you ask me.
Jon Risch transformer filter
Here is a link to a Jon Risch project that is of the same idea.
Audio Asylum Thread Printer
Here is a link to a Jon Risch project that is of the same idea.
Audio Asylum Thread Printer
year 2001. yes it has been a recognized problem for a long time. Now with more products using smps, the problem is getting worse in the home.
THx-RNMarsh
It is a base capability in the Delta Tau motion control platform. For any motor where the DT does the commutation, it is possible to generate a lookup table that will modify the current of each phase based on position. By doing that, it is trivial to polish the T.
Steppers are seriously coggers, with torque ripple from 5 to 20% of the peak torque, the cogging is four times the electrical cycle in frequency. Three phase brushless is far more refined..
I had to worry about that for the things I worked on last year, as the torque ripple was exciting a resonance such that two not independent stages went into positive feedback. I had to reduce speed until the resonance mode was away from the stimulus. It would have been possible to mod the torque by position, but I didn't see the need to make it so complex. That would require the maintenance techs learn way too much.
jn
JN,
This is the best cite I could find. Seems to be this century measurements on actual AC generators in service with distortions running under 1%.
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.541.91&rep=rep1&type=pdf
ES
This is the best cite I could find. Seems to be this century measurements on actual AC generators in service with distortions running under 1%.
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.541.91&rep=rep1&type=pdf
ES
You probably need at least 40Hz or so BW on the correction. My cheap belt drive has a substantial 9Hz component (300/33.333) and another forum member's Lenco has one at 25Hz. There's also the problem of having a motor where you can vary the speed.
Last edited:
- Status
- Not open for further replies.