Hi Pyramid,
I'm trying to correlate speed, torque and current with equipment I have at hand , with purpose to understand effects better. Keeping in mind this is belt drive, not DD.
So, between power amps and motor I have installed 3 x 3R resistors. I did it for safety purposes while testing.
One input of soundcard (CH2) I took over resistor to measure current, other input is connected to voltage (CH1), both at the same phase.
Picture 1; voltage supply with motor disconnected, just for reference, clean sine V supply, one phase:
Picture 2: Voltage (blue) and current (yellow) platter spinning :
The sine wave is slightly deformed in V, and little bit more in I. Of course there is expected phase shift due reactive load.... I think V sine would stay pristine if there is no 3R resistor to allow current to modify V.
The picture is same if I play record so I did not attach, instead here is video to show it dynamically; first 10 seconds TT is turning , than for rest I put needle on the record. I can't see any difference...
In above I cant see any except micro current changes.
Now when typing, I got idea that I should have check the same without belt, just light motor... Could be my heavy platter is doing its job too good.... Back to basement...
So here it is: motor spinning alone, no belt on it:
More or less the same.
Interesting is also that current is less with load of platter, than without it.
Then I tried 45rpm spinning platter, here current dances bit more and current is reduced :
Here I can see little bit more current dancing left and right, is this speed - current oscillation you talk about?
Can you take anything out of this? Thanks in any case.
I'm trying to correlate speed, torque and current with equipment I have at hand , with purpose to understand effects better. Keeping in mind this is belt drive, not DD.
So, between power amps and motor I have installed 3 x 3R resistors. I did it for safety purposes while testing.
One input of soundcard (CH2) I took over resistor to measure current, other input is connected to voltage (CH1), both at the same phase.
Picture 1; voltage supply with motor disconnected, just for reference, clean sine V supply, one phase:
Picture 2: Voltage (blue) and current (yellow) platter spinning :
The sine wave is slightly deformed in V, and little bit more in I. Of course there is expected phase shift due reactive load.... I think V sine would stay pristine if there is no 3R resistor to allow current to modify V.
The picture is same if I play record so I did not attach, instead here is video to show it dynamically; first 10 seconds TT is turning , than for rest I put needle on the record. I can't see any difference...
In above I cant see any except micro current changes.
Now when typing, I got idea that I should have check the same without belt, just light motor... Could be my heavy platter is doing its job too good.... Back to basement...
So here it is: motor spinning alone, no belt on it:
More or less the same.
Interesting is also that current is less with load of platter, than without it.
Then I tried 45rpm spinning platter, here current dances bit more and current is reduced :
Here I can see little bit more current dancing left and right, is this speed - current oscillation you talk about?
Can you take anything out of this? Thanks in any case.
I've not seen one like that, but quite a few manufacturers discovered that a cap under the motor to lift the rotor was worthwhile. Surprisingly, few realised that what they were doing was equalising the drives of the two motor coils to the common shaft. From the motors I've seen, they were all derivatives of the Philips motor, which was intended for driving louvres in air handling systems...
@Pyramid. Thanks for the explanation. I can imagine that the shape of the pole pieces and their interaction with the permanent magnet could cause variations in instantaneous velocity as the rotor rotates, but that seems to me to be akin to the rotation having a fundamental frequency plus harmonics. I'm curious about your oscillation statement because that seems to be occurring at a lower frequency? I have previously wondered about belt drive turntables being an LC filter with the mass of the platter being the C and compliance of the belt being L. Such a system always has some resistance that damps (or not) its resonant frequency, usually achieved by the belt or platter bearing friction. The motor is habitually driven from a voltage source, but they can be current driven, and operation between those two extremes allows damping to be adjusted more readily than changing belt. Are you perhaps suggesting a similar effect in the direct drive turntable where low frequency oscillation could occur with interaction between platter mass and motor torque characteristics?
The motor is essentially and electric torsional spring. If the rotor and field are exactly aligned there is no torque being produced and the "spring" is relaxed. If a load is applied, the rotor will fall behind by some number of degrees but stay in synch with the rotating field. The higher the load, the higher the lagging angle and the higher the torque being applied. This remains true until the angle exceeds 90° and which point the transmitted torque starts decreasing and the motor stalls or falls out of synch (you break the spring). The higher the drive voltage and current, the "stronger" the spring and the more torque it can deliver before falling out of synch.
This can work in reverse as well. If the voltage/current is too high, the platter can accelerate in front of the rotating field and will start to be pulled back. Because of inertia, it can over shoot which is where I think the oscillations were coming from when I ran the VPI HW40 as an AC synch motor open loop. If the current was too high, the speed would become unstable as oscillations built up. If I carefully controlled the voltage I could eventually get the platter to settle down although W&F were still terrible and any change in load would set it off.
I'm not sure where you are measuring current and voltage at the same time and where your ground references are connected. That may account for the large phase shift as well as distortion. Most of these motors have a power factor in the 90's so they appear as mostly resistive.
The lower current at 45 RPM is a result of the increase in back EMF most likely which increases with speed. It effectively raises the impedance of the motor windings. If you keep increasing speed without raising the voltage, the current will eventually drop so low the motor will stall.
Hi, at the same time, 2 channels of scope at same time.
Ground is negative of voltage probe, so yes. Inputs of my card are fully differential so it doesn't matter
Hi, If I'm not mistaken (and recently I often are 🙁) you are talking old fashion synchronous motor with equal stator and rotor poles?
I took first picture of BLDC from internet, stator and rotor poles are not equal (and not multiplied by 2), all rotor poles and the field are never aligned, and rotor is much slower than rotating field (more rotor magnet fields, slower).
If not removing the issue, I think unequal pole numbers in contemporary BLDC will reduce it a lot. In below picture, top and bottom poles are aligned, rest is somewhere inbetween
Thank you; I think that's the key point. Imagine a system where we drive the platter via a belt having zero compliance. We still have a mechanical filter because of platter mass and the motor "spring" but the only damping to the system is electrical (like an amplifier damping a loudspeaker's low frequency meachanical resonance). Returning to a more conventional belt drive, we can adjust electrical source resistance to the motor to set the Q of the system, or we could adjust belt compliance (rubber, thread, magnetic tape). This is where the differences in belts come in. But damping in the belt is invariably temperature-sensitive. If we use magnetic recording tape (near enough zero compliance) we can adjust source impedance to the motor to set Q. We still have an LC filter that might resonate, but it's more stable with temperature, and we can tweak its Q on the fly. And, if you think about, that's what has to be done for direct drive.
I had occasion to replace the motor in an Oracle turntable and finally settled upon an Anaheim Automation 8-pole BLDC motor and a Maxon DEC 50/5 1-Q-EC controller module. The Maxon controller accepts a DC signal that sets the RPM, so the only other components required are a precision DC source and a DC supply to power the Maxon module. The motor exhibits a slight cogging, but with a bit of isolation (Sorbathane worked for me) motor noise can be reduced to inaudibility.
Attachments
Intrigued by discussions here, today I took bit different approach, I was wondering how was Sony motor (No4 in post #50, inside picture post #118, and back EMF in post #130) that Pyramid also recommended as good candidate is doing with original onboard drive. So supaspin sinewave generator on side for this test.
Pre-driver chip datasheet helped to reveal circuit diagram, it is actually exactly as in datasheet.
3 hall sensors are driving analog predriver which is driving 3 transistor bridges. Speed is controlled analogously by DC supplied to transistors and Vref of the chip. As simple as this:
I connected 12 VDC to VCC as chip specifies and adjusted Vs (Servo out) DC to 3VDC to achieve desirable speed of about 400 RPM.
Like this motor shows some vibrations but very little, similar as I tested it earlier with pure sine directly to coils without original electronics. Nothing special.
However the torque is much less like this, with sine wave drive it was much higher.
Then I checked running voltage directly between one coil and gnd, well 1.4something VRMS...... earlier with sinevave I run it at 2.5 - 3 VRMS, here is explanation for the torque.
Finally, wanted to see how this original drive wave form looks like; this is probe connected between gnd and one phase of the motor:
Ouch, not very pretty. However it does resemble its EMF that I shoved in post #130 , at least approximately.
Now when I learned that in first tests I run this motor with 70 to 100% too high V, I connected it back to sinewe generator, this time adjusted for 1.4 VRMS, and whoah, now it is vibration less as preferred Phillips and as quiet as ball bearings allow. However now it has much less torque than Phillips, even that torque is enough to start and run the platter, but no DJ-ing around with it. Still, it can be used for TT, and so far its confirmed that pure sinevawe drive helps reducing vibration for motors that have close to sine back EMF
PS, this motor in this arrangement was responsible for tape speed of post production of A production movies, found place where they sell it new ("little bit" out of date site). To be bit fair to Sony, there is another set of hall sensors on board used to control servo DC via another set of electronics, and speed for digital is probably less important as it will be re-clocked before put on blueray.... But still, would expect more elaborated drive for this:
https://www.spectratech.gr/en/product/14324/Sony_HDW-M2000P-20
Pre-driver chip datasheet helped to reveal circuit diagram, it is actually exactly as in datasheet.
3 hall sensors are driving analog predriver which is driving 3 transistor bridges. Speed is controlled analogously by DC supplied to transistors and Vref of the chip. As simple as this:
I connected 12 VDC to VCC as chip specifies and adjusted Vs (Servo out) DC to 3VDC to achieve desirable speed of about 400 RPM.
Like this motor shows some vibrations but very little, similar as I tested it earlier with pure sine directly to coils without original electronics. Nothing special.
However the torque is much less like this, with sine wave drive it was much higher.
Then I checked running voltage directly between one coil and gnd, well 1.4something VRMS...... earlier with sinevave I run it at 2.5 - 3 VRMS, here is explanation for the torque.
Finally, wanted to see how this original drive wave form looks like; this is probe connected between gnd and one phase of the motor:
Ouch, not very pretty. However it does resemble its EMF that I shoved in post #130 , at least approximately.
Now when I learned that in first tests I run this motor with 70 to 100% too high V, I connected it back to sinewe generator, this time adjusted for 1.4 VRMS, and whoah, now it is vibration less as preferred Phillips and as quiet as ball bearings allow. However now it has much less torque than Phillips, even that torque is enough to start and run the platter, but no DJ-ing around with it. Still, it can be used for TT, and so far its confirmed that pure sinevawe drive helps reducing vibration for motors that have close to sine back EMF
PS, this motor in this arrangement was responsible for tape speed of post production of A production movies, found place where they sell it new ("little bit" out of date site). To be bit fair to Sony, there is another set of hall sensors on board used to control servo DC via another set of electronics, and speed for digital is probably less important as it will be re-clocked before put on blueray.... But still, would expect more elaborated drive for this:
https://www.spectratech.gr/en/product/14324/Sony_HDW-M2000P-20
with learning from above test, I was playing with reduced Vrms on (so far) preferred Phillips motor, reduced sinewaves to 1.8 V (instead of previous 2.5) . It is silent as an owl now, torque is reduced and motor is stoppable by finger pressure, but given transmission ratio TT platter still has more than good torque, suitable for DJ-ing, Cheap Phillips VCR motor still rules in this set of tests...
The change in vibration is most likely reduced cogging because of lower magnetic attraction between rotor & stator and less rotor overshoot because the drive current is closer to what is needed for constant speed and the electrical "spring" is weaker.
If you look at the total DC current being drawn at 33 RPM, when you change to 45 RPM, increase the drive voltage until the total DC current is the same as 33RPM and you will maintain the same torque at 45 that you have at 33. You need to increase the voltage because of the increase in BEMF at the higher speed. If you are using Rich's Supraspin, I believe he has added the capability for this. It also has higher voltage at start up to get the platter moving but reduces the current (torque) once the platter is up to speed.
If you look at the total DC current being drawn at 33 RPM, when you change to 45 RPM, increase the drive voltage until the total DC current is the same as 33RPM and you will maintain the same torque at 45 that you have at 33. You need to increase the voltage because of the increase in BEMF at the higher speed. If you are using Rich's Supraspin, I believe he has added the capability for this. It also has higher voltage at start up to get the platter moving but reduces the current (torque) once the platter is up to speed.
Yep, this feature is in Supaspin, thanks Richard again for preparing this project throughly! For the moment I use manual adjustments as there is so much to test, but final tt will use this for sure
Hi Pyramid, this analogy of yours with motor as a spring is very useful and applicable (and appreciated).
One thing I take out of it is that use of powerful high torque motors is becoming redundant (belt drive speaking). Many people including myself (until recently) advocated that motor should be much more powerful than whats needed to spin the platter (dog wages the tail). What I learned from this chat is that we reduce field power = reduce torque to level where it is just enough to keep sync speed, meaning regardless how big is motor, optimal running will make it flee power anyway.
So as preliminary ( I reserve to make any final conclusions yet) is to hell with high torque motors.... small one will just do, even if at start up it is overloaded, who cares as its only few seconds.
Another thing to what helps is uneven number of rotor poles comparing to stator poles in most of BLDC motors (my post #150), all poles never align at same time = spring effect reduced.
And yet another thing learned, post #153, is that usual BLDC drive with hall sensors and analogue chip actually has really good opportunity to make perfect drive. Hall sensors are picking up motor back EMF and through analogue driver chip transferring it in copy of EMF wave shape to the current drive of the motor...simple and effective. Stil so far I got better results with pure sine drive to motors that have sine back EMF... to be further examined..
Sorry if my posts are kind of brain stream, I publish as I test and those tests are repeatable. at the end Ill make summary. My believe is that small prepared tests is way to get bigger picture in place.....
About home brew shaft encoder:
Permanently attaching optical shaft encoder (with several thousand precision PPR) would be absolutely right way to measure and proof exact platter speed oscillations in real time (I did not use wow and flutter as these historical measurements are eagghhh; historical). Doing this (except of obvious cost consequences) is mechanically not trivial at all. Another option is custom made laser engraved optical disc, but still cost + mechanical assy + maintenance in this case as its not enclosed any more (dust).
Really fine precision optical shaft encoders are available from servo and robot industrial technology, or as motor encoders or as 0-shaft encoders responsible to synchronize whole machine. Since this tech is mature, many are available from scrap yard, I actually know something about it, so ask if doubting.
Test LP helps here only to certain extend due to LP tolerances. This is many zeroes less accurate than industrial shaft encoder.
In posts #95 to #101 we (Tuberadio and EC8010) discussed optical reflective sensor to be used as kind of amateur shaft encoder. In particular I use CNY70. I said that I don't know how small can "dark" pach be for sensor to pick it up repeatably, now i tested it.
So, with CNY70 attached as close as possible to platter rim, I managed to get smallest possible "black patch" of about 1,5 mm , and speed reading still works robustly. At this level its sensitive to ambient light, result is with shadow above sensor:
Since we also need a "white spot" , whole patch is 3mm wide.
My idea was to engrave (or print + glue) stroboscope paper looking ring at the bottom of the platter, with CNY70 looking at it, precisely glued inside of the plinth. Distance between sensor and strobe lines would be less than 500um and since its below platter, its in shadow from ambient light,
Given max platter D of about 300mm, if this is installed beneath platter at say 280 mm diameter, we can get max 360 PPR (one degree resolution) which @ 33,33333 RPM gives 200Hz signal. To be safe probably 180 lines would be better, giving typical 50hz strobe pattern and 100Hz output.
I think this can be printed and glued.... Still need to calculate best printer DPI to see how will it integrate. Printer dpi are xy oriented, and this is angle oriented precision needed...
Frankly Im not sure if 100 (or 200) Hz is good enough to measure "speed dance"
About home brew shaft encoder:
Permanently attaching optical shaft encoder (with several thousand precision PPR) would be absolutely right way to measure and proof exact platter speed oscillations in real time (I did not use wow and flutter as these historical measurements are eagghhh; historical). Doing this (except of obvious cost consequences) is mechanically not trivial at all. Another option is custom made laser engraved optical disc, but still cost + mechanical assy + maintenance in this case as its not enclosed any more (dust).
Really fine precision optical shaft encoders are available from servo and robot industrial technology, or as motor encoders or as 0-shaft encoders responsible to synchronize whole machine. Since this tech is mature, many are available from scrap yard, I actually know something about it, so ask if doubting.
Test LP helps here only to certain extend due to LP tolerances. This is many zeroes less accurate than industrial shaft encoder.
In posts #95 to #101 we (Tuberadio and EC8010) discussed optical reflective sensor to be used as kind of amateur shaft encoder. In particular I use CNY70. I said that I don't know how small can "dark" pach be for sensor to pick it up repeatably, now i tested it.
So, with CNY70 attached as close as possible to platter rim, I managed to get smallest possible "black patch" of about 1,5 mm , and speed reading still works robustly. At this level its sensitive to ambient light, result is with shadow above sensor:
Since we also need a "white spot" , whole patch is 3mm wide.
My idea was to engrave (or print + glue) stroboscope paper looking ring at the bottom of the platter, with CNY70 looking at it, precisely glued inside of the plinth. Distance between sensor and strobe lines would be less than 500um and since its below platter, its in shadow from ambient light,
Given max platter D of about 300mm, if this is installed beneath platter at say 280 mm diameter, we can get max 360 PPR (one degree resolution) which @ 33,33333 RPM gives 200Hz signal. To be safe probably 180 lines would be better, giving typical 50hz strobe pattern and 100Hz output.
I think this can be printed and glued.... Still need to calculate best printer DPI to see how will it integrate. Printer dpi are xy oriented, and this is angle oriented precision needed...
Frankly Im not sure if 100 (or 200) Hz is good enough to measure "speed dance"
Last edited: