Small break, but testing continues. I returned to first favorite (No3 from post # 50) Sony drum-head motor. This one rotates @750 rpm @50hz, useful.
With above learning now I run it with really low V, just to keep it very hard to stop on shaft, but easy to stop on rotor. Result is approx 0.8V RMS and about 0.75 A per phase. Coils are about 1.1 Ohm.
Not just that is also super silent and vibration free despite it has stator and motor poles dividable by 2 (12 coils and 12 magnets). So small V required to spin it allows for just a current buffers to drive it, Volts can be sufficiently generated by Supaspin, and 5V phone charger (beefy one) can be used to power whole TT.
Good enough that I ordered pulley for it. Given it size, I also ordered brass and big diameter flywheel to increase momentum. Hope it will come in faster than last one.
With above learning now I run it with really low V, just to keep it very hard to stop on shaft, but easy to stop on rotor. Result is approx 0.8V RMS and about 0.75 A per phase. Coils are about 1.1 Ohm.
Not just that is also super silent and vibration free despite it has stator and motor poles dividable by 2 (12 coils and 12 magnets). So small V required to spin it allows for just a current buffers to drive it, Volts can be sufficiently generated by Supaspin, and 5V phone charger (beefy one) can be used to power whole TT.
Good enough that I ordered pulley for it. Given it size, I also ordered brass and big diameter flywheel to increase momentum. Hope it will come in faster than last one.
That's nice to hear what sort of voltage is likely to be required. I have three TDA2030 amplifier boards bought at great expense (about £3 each!) that are fine on a single rail 12V supply. They'll do nicely without having to change their capacitors for a higher voltage rating. Your flywheel will need a lot of inertia to improve on the motor's own flywheel of that outer rotor.
One thing that I do struggle with is measurement of platter speed... I till don't have shaft encoder yet and I use 1kHz signal from test LP.
Yesterday I discovered an stupidity that I was performing. In all "speed dancing " videos I posted there is about 0.8 % dance (8 Hz given base frequency of 1kHz).
Well, I realized that in REW I was using FFT lenght of 16k and sampling frequency of 48k.
Copy paste from REW instructions:
The FFT Length determines the basic frequency resolution of the analyser, which is sample rate divided by FFT length. The shortest FFT is 8,192 (often abbreviated as 8k) which is also the length of the blocks of input data that are fed to the analyser. An 8k FFT has a frequency resolution of approximately 6Hz for data sampled at 48kHz.
So my maximum resolution was 3Hz as that was the size of the bin, sample data block.... No wonder I couldn't get better results grgghhh... Changed FFT lenght to 512k and immediately I get 0,4% speed deviation (+-0.2%).
New problem is that my laptop is too slow to work fast with larger FFT lengths, its pain to wait for screen to change, actually LP track will finish while I wait, its not trusted..... Need to get my wife to lend me her laptop...
In preamp thread Hans Polak thought me quite few things about measurements, appreciate that a lot! But now I need to call for help:
@ Pyramid: How did you get your polar plots? Is there free tutorial or you can teach? As said, only test LP 1kHz available at moment....
Yesterday I discovered an stupidity that I was performing. In all "speed dancing " videos I posted there is about 0.8 % dance (8 Hz given base frequency of 1kHz).
Well, I realized that in REW I was using FFT lenght of 16k and sampling frequency of 48k.
Copy paste from REW instructions:
FFT Length
The FFT Length determines the basic frequency resolution of the analyser, which is sample rate divided by FFT length. The shortest FFT is 8,192 (often abbreviated as 8k) which is also the length of the blocks of input data that are fed to the analyser. An 8k FFT has a frequency resolution of approximately 6Hz for data sampled at 48kHz.
So my maximum resolution was 3Hz as that was the size of the bin, sample data block.... No wonder I couldn't get better results grgghhh... Changed FFT lenght to 512k and immediately I get 0,4% speed deviation (+-0.2%).
New problem is that my laptop is too slow to work fast with larger FFT lengths, its pain to wait for screen to change, actually LP track will finish while I wait, its not trusted..... Need to get my wife to lend me her laptop...
In preamp thread Hans Polak thought me quite few things about measurements, appreciate that a lot! But now I need to call for help:
@ Pyramid: How did you get your polar plots? Is there free tutorial or you can teach? As said, only test LP 1kHz available at moment....
😉 I was too fast and prematurely built 3 channel amp based on TDA2030's, see posts #39 for PCB and in #41 there is picture of complete amp.
I powered it with +18VDC and at first tests it went in exploding smoke 😵
Than figured that these motors are not loudspeakers, they need far less V, but impedance is very low. Sony drumhead that I just mentioned has only about 1 Ohm per coil R.
Higher VPS = more heat amp needs to dissipate. 1 ohm load was too much for TDA @18VPS.
What Im thinking now is that motor must be chosen first, than amp (or current buffer) built accordingly.
Even if motor starts with drive in clipping, its fine, as long as it stably runs in sine wave after speed is at target and VRMS supply is reduced. Higher V PS than required by motor simply means much more heat dissipation and burning amp.
Anecdote: Once a mechanic friend told me why car engines are not good for boats; in car engine goes from 1000 to 3000 revs occasionally , on boat engine rews at 5000 with full torque continuously, like going steep uphill with car forever.... Car is more like speaker, boat like TT.
Next problem I found is that chip amps made for cars (12V battery supply) still have min VPS of 8V, and I actually need only 5 for most of those motors.
Now I'm looking for least components , most simple (no V gain) current buffer that can deliver amp or 2 RMS from 5VPS. THD irrelevant, any up to 5% will do...
PS, for testing I'm using beefy AV receiver with 3R 10W resistor in series between amps output and motor coils. Cool thing is that receiver has volume knob so I can adjust V at fly. Still, with guaranteed min 3R load, few hundred watt receiver boils up after some time delivering just few W. It is because its +-60Vdc Ps needs top be dissipated somewhere while delivering only 1 V.
Hello, have you read through this thread?
https://www.diyaudio.com/community/threads/turntable-speed-stabilty.309349/
https://www.diyaudio.com/community/threads/turntable-speed-stabilty.309349/
Whoops! All good points. Of course, one possible solution would be to rewind the motors. Seriously! Unlike a loudspeaker, it doesn't matter greatly how tidily each coil is wound (you only have to look at the original windings to see that). Another solution would be a three-phase transformer, although I presently have no idea how to wind one. Finally, three amplifiers that don't mind the horrible load (probably the easiest solution, but also easily the least elegant).
Another point that should be considered is that there is no requirement whatsoever to operate the motor at 50Hz. I power my Garrard 301 from 60Hz, which makes it easy to differentiate between motor issues and general mains hum. A higher operating frequency might ease the load...
How about the simple current-boosted opamp?
For a 5v minimum supply opamp (+-2.5v) you could get away with around +-3.5v rails (need up to ~1v for bipolar Vbe's). Should be fairly quick and easy to prototype with on-hand parts.
It's fundamentally what I've used to drive step-up transformers for high-voltage motors.
That very neat and should work just fine. Distortion doesn't appear to be the important parameter in turntable motor drive, it's frequency modulation and sidebands of the drive frequency that need to be minimised.
Hi Laszlo, thanks for hint, yes I was reading that, not all, its huge.
But it seems Dr Feickert tools are not available any more....
Hi, I use speed @50hz just as comparison of motor characteristics..... Old Papst was at 22 Hz , capstan at 61Hz, drumhead will be around 33 Hz, all to get 33 rpm.
Go fully differential with preamp than there will be no need to think of mains or motor hum 😎
Hi, thanks, and yes sure, just need to change TL071 for one that takes 3VDC PS (think minimum for TL is 6V).
To run all from one 5V wall wart is very convenient, I will see if it is manageable...
Opened my opamp drawer and reminded myself of motor driver opamp:
L272 has min VPS of just 4VDC and can deliver 0.7 A output current:
Around crossover 0 Volt opamp is delivering current to motor. As soon as that current reaches approx 70mA respective transistor opens and offloads current from opamop...
Specified voltage swing for L272 is 1 V under V PS (max 1.5V) that gives us 3.5 - 4 Vp-p left from opamp - 2x Vbe = min 2 Vp-p, normally it should be 2.7Vp-p, just fine to make 0.85 VRMS as drumhead motor liked. As said, I think clipping is not an issue at motor start, only continuous run needs nice sine wave.
Does anyone sees mistake here?
Otherwise I will built this and try what I get, depending on parts tolerances it will clip or not. Worst thing is that if it doesn't work, I will need 7-8 VPS after all...
L272 has min VPS of just 4VDC and can deliver 0.7 A output current:
Around crossover 0 Volt opamp is delivering current to motor. As soon as that current reaches approx 70mA respective transistor opens and offloads current from opamop...
Specified voltage swing for L272 is 1 V under V PS (max 1.5V) that gives us 3.5 - 4 Vp-p left from opamp - 2x Vbe = min 2 Vp-p, normally it should be 2.7Vp-p, just fine to make 0.85 VRMS as drumhead motor liked. As said, I think clipping is not an issue at motor start, only continuous run needs nice sine wave.
Does anyone sees mistake here?
Otherwise I will built this and try what I get, depending on parts tolerances it will clip or not. Worst thing is that if it doesn't work, I will need 7-8 VPS after all...
Look at this thread: https://www.diyaudio.com/community/threads/polar-plot-python-script.409469/
I'm using an older version of Anaconda. If you use a newer version, there may be compatibility issues. You can search the web for an archive that has all older versions available.
Clumsiness rules 🤣 🤣 🤣 🤣
Here:
Another thing, + input of amp I need to reference at Vcc/2, or at 2.5V.
In Supaspin outputs of opamps are already at that DC. If I only bypass C4. C5 and C10 looks like it is fine and no additional components are needed?
As the PWM from the PI Pico is 3.3v, the output of the opamps will not be exactly centred on 5v/2, so I'd keep the ac-coupling caps and use a 100k/100k potential divider from 5v instead of the pull-downs to bias the outputs (so I extra resistor on each).
Didn't post in a week, but there was some work going on. Unfortunately 2 defeats.
First is my attempt to use Sony drumhead motor, I was seduced by share beauty of this machine and so low voltage it is running at, ignoring 2 important things; it has ball bearings and equal number of windings and magnets..
Pulley arrived much faster this time, large chunk of brass that cost me another 50 Eur. here is picture where Im finishing pulley with nail file.
The hard plastic jar where I put motor is sideways discovery, it is acoustical amplifier, started to call it pain amplifier (ref movie Dune) as is shows position of every ball in bearing. Once I thought motor is quiet, putting it into jar amplifies every possible noise it produces.
Then it was small hammer and screwdriver job, small steps kicking parts until balance (and lowest noise ) was achieved, this is motor assy without rotor:
Such big and heavy pulley adds rotation interia, there is space for further improvement by adding heavy metal skirt under it...
Mounting of this motor seemed more simple than for capstan motor, well not so. I machined (with my hand tools, 2 pcs of MDF and arrange distance between them to be flat with 3 m3 screws, than I applied standard sanitary (bathroom ) silicone patches as mechanical decoupling of the motor:
Finally motor assambled in temporary TT rig so I could run it:
Top is painted black as I dont like too much brass color...
Anyway, motor is noisy, ball bearings simply cannot be quieted to the level where they are not present.
Worst thing is that speed is far less stable than with capstan motor (or premotec, or papst) . If with capstan I get 0.01 RPM changes every few minutes and if at all (as measured by Supaspin, one rotation average) with this I got dancing of about 0.1 RPM every rotation.....
Waste of time and money......
First is my attempt to use Sony drumhead motor, I was seduced by share beauty of this machine and so low voltage it is running at, ignoring 2 important things; it has ball bearings and equal number of windings and magnets..
Pulley arrived much faster this time, large chunk of brass that cost me another 50 Eur. here is picture where Im finishing pulley with nail file.
The hard plastic jar where I put motor is sideways discovery, it is acoustical amplifier, started to call it pain amplifier (ref movie Dune) as is shows position of every ball in bearing. Once I thought motor is quiet, putting it into jar amplifies every possible noise it produces.
Then it was small hammer and screwdriver job, small steps kicking parts until balance (and lowest noise ) was achieved, this is motor assy without rotor:
Such big and heavy pulley adds rotation interia, there is space for further improvement by adding heavy metal skirt under it...
Mounting of this motor seemed more simple than for capstan motor, well not so. I machined (with my hand tools, 2 pcs of MDF and arrange distance between them to be flat with 3 m3 screws, than I applied standard sanitary (bathroom ) silicone patches as mechanical decoupling of the motor:
Finally motor assambled in temporary TT rig so I could run it:
Top is painted black as I dont like too much brass color...
Anyway, motor is noisy, ball bearings simply cannot be quieted to the level where they are not present.
Worst thing is that speed is far less stable than with capstan motor (or premotec, or papst) . If with capstan I get 0.01 RPM changes every few minutes and if at all (as measured by Supaspin, one rotation average) with this I got dancing of about 0.1 RPM every rotation.....
Waste of time and money......
Second defeat was trying to run Python script from unfortunately late Mr Wurcer and Pyramid .
@Pyramid, this is very useful tool and all my thanks to both of you, but I spend hours and my app directory now looks like snake pit 🤣 , no way to run it. In original thread you recommended Pyton v2.0 and older, but that is year 2000 release..... No way it will run today.
Is there a plan to rewrite it , or is there a point to ask on your original thread, might be someone runs it at W11? I'm total code writing dummy, would need step by step instructions to do it...
@Pyramid, this is very useful tool and all my thanks to both of you, but I spend hours and my app directory now looks like snake pit 🤣 , no way to run it. In original thread you recommended Pyton v2.0 and older, but that is year 2000 release..... No way it will run today.
Is there a plan to rewrite it , or is there a point to ask on your original thread, might be someone runs it at W11? I'm total code writing dummy, would need step by step instructions to do it...