Re: What we need is a measurement...
I puzzled over this, too, but I remembered a comment someone made about the sine wave being generated by the motor. I checked this out (against the schematic), and indeed it is the case. There is a separate winding (F.G.) that gives frequency (speed) feedback, and it is this that is referenced to the digitally-derived clock. Thus there are two closed loop controls: a phase-locked loop for instantaneous speed correction, and the overall speed governor/regulator that measures platter rotational speed. It is the latter control loop, IMO, that is more likely to induce jerky movement (I'm not yet clear on how the error correction works, though).
Ultimately, yes. For now, we have the service manual and the waveforms depicted there, which are sine-esque in shape.
I must say that this thread (the latter part) has made me study the schematic more closely, and one of the design concerns (for a modern circuit) has been "solved" for me: the generation of 3-phase variable-frequency sinewaves. It requires relatively sophisticated electronics to produce extra 120 & 240 degree constant phase shifted signals for a variable frequency signal (short of making a special transformer). But the SP10 motor already does this. I'm feeling quite motivated now...
I puzzled over this, too, but I remembered a comment someone made about the sine wave being generated by the motor. I checked this out (against the schematic), and indeed it is the case. There is a separate winding (F.G.) that gives frequency (speed) feedback, and it is this that is referenced to the digitally-derived clock. Thus there are two closed loop controls: a phase-locked loop for instantaneous speed correction, and the overall speed governor/regulator that measures platter rotational speed. It is the latter control loop, IMO, that is more likely to induce jerky movement (I'm not yet clear on how the error correction works, though).
Ultimately, yes. For now, we have the service manual and the waveforms depicted there, which are sine-esque in shape.
I must say that this thread (the latter part) has made me study the schematic more closely, and one of the design concerns (for a modern circuit) has been "solved" for me: the generation of 3-phase variable-frequency sinewaves. It requires relatively sophisticated electronics to produce extra 120 & 240 degree constant phase shifted signals for a variable frequency signal (short of making a special transformer). But the SP10 motor already does this. I'm feeling quite motivated now...
Re: Re: What we need is a measurement...
No , it's actually very easy. Once you have two phases in quadrature any other phase relationship can be generated by simple vector arithmetic. I had a "twilight zone" moment the other day when I discovered that the ordinary E24 resistor values will allow round number vector sums to very precise degrees - less than 0.1 degree error in most cases. There seems to be no reason why this should be, but it is.
But I digress.
All you need to do is generate two phases in quadrature, which can be done any number of ways - easiest is a PLL chip which will lock at 90 degrees shift. I prefer other methods -one of my drives uses a scheme where the clock is divided in a cycle of eight and there's a second cycle of eight shifted two steps from the first, giving exact quadrature.
Another circuit I'm working on uses one of the AD DDS chips which is a two channel device with programmable phase shifts of (I think) 0.1 degree precision.
Once you have quadrature you simply add the two phases together via a vector sum to get the phase you want. To go from 0 and 90 to 120 for instance, invert the 0 to get 180 then add 0.5 x 180 to 0.8660 x 90 and the result is 120. I leave 240 degrees to the reader as an exercise (Don'tcha hate that in EE texts)
As an example of the E24 thing, 13k/15 k is almost exactly 0.8660. so the sums are very easily achieved.
BTW the motor doesn't generate the sinewave for the drive and of course there are two frequency generators, the comparison between these is how a PLL works.
Shaun said:
No , it's actually very easy. Once you have two phases in quadrature any other phase relationship can be generated by simple vector arithmetic. I had a "twilight zone" moment the other day when I discovered that the ordinary E24 resistor values will allow round number vector sums to very precise degrees - less than 0.1 degree error in most cases. There seems to be no reason why this should be, but it is.
But I digress.
All you need to do is generate two phases in quadrature, which can be done any number of ways - easiest is a PLL chip which will lock at 90 degrees shift. I prefer other methods -one of my drives uses a scheme where the clock is divided in a cycle of eight and there's a second cycle of eight shifted two steps from the first, giving exact quadrature.
Another circuit I'm working on uses one of the AD DDS chips which is a two channel device with programmable phase shifts of (I think) 0.1 degree precision.
Once you have quadrature you simply add the two phases together via a vector sum to get the phase you want. To go from 0 and 90 to 120 for instance, invert the 0 to get 180 then add 0.5 x 180 to 0.8660 x 90 and the result is 120. I leave 240 degrees to the reader as an exercise (Don'tcha hate that in EE texts)
As an example of the E24 thing, 13k/15 k is almost exactly 0.8660. so the sums are very easily achieved.
BTW the motor doesn't generate the sinewave for the drive and of course there are two frequency generators, the comparison between these is how a PLL works.
Re: Re: Re: What we need is a measurement...
Mark, this is one case where I'd be very happy to be proven wrong. I don't have any experience with signal generation and phase shifting or PLLs, and I'm actually still learning about it. I would like to make a DIY controller to replace the old electronics of the SP10. I am reluctant to use (expensive) special-purpose chips or micros, as I would like to share the controller with other DIYers (got to keep it simple). You are the first person I come across who actually seems to have done some work in this area. Will you share your ideas/give me guidance?
Please help me out here. I see TR28, etc., but I'm not clear on its purpose.... I don't see another source for a sine wave other than the coils connected to the input of the driver circuit....
I'm doing this thing cold, as I was not planning on working this project right now. This thread is making me work...pheew
Mark, this is one case where I'd be very happy to be proven wrong. I don't have any experience with signal generation and phase shifting or PLLs, and I'm actually still learning about it. I would like to make a DIY controller to replace the old electronics of the SP10. I am reluctant to use (expensive) special-purpose chips or micros, as I would like to share the controller with other DIYers (got to keep it simple). You are the first person I come across who actually seems to have done some work in this area. Will you share your ideas/give me guidance?
Please help me out here. I see TR28, etc., but I'm not clear on its purpose.... I don't see another source for a sine wave other than the coils connected to the input of the driver circuit....
I'm doing this thing cold, as I was not planning on working this project right now. This thread is making me work...pheew
Re: Re: Re: Re: What we need is a measurement...
If you intend using a servo loop I'm no use to you, I don't like them and I am no expert in their design. Designing servo loops is a lot like designing feedback systems except there is an extra layer of complexity due to things like jitter. I can tell you quite a bit about what NOT to do from bitter experience.
If you intend using a fixed frequency synthesis technique you need to find a way of dealing with the cogging torque and the torque ripple which will bedevil this implementation. The basic techniques involved have been known for 20 years, see for instance “Minimization of Torque Ripple in Brushless DC Motor Drives”, H. Le Huy, R. Perret, R. Feuillet:IEEE Ind. Appl. Soc. Ann. Meet., Toronto, 1985, pp. 790- 797.
Chris Brady apparently uses digital signal processing to reduce their effects. I am using different techniques, which I am not willing to share as I am pursuing IP protection on them.
I'm not sure about the specifics of the SP10 electronics because, as I said, I'm replacing them with a completely different implementation.
Shaun said:
If you intend using a servo loop I'm no use to you, I don't like them and I am no expert in their design. Designing servo loops is a lot like designing feedback systems except there is an extra layer of complexity due to things like jitter. I can tell you quite a bit about what NOT to do from bitter experience.
If you intend using a fixed frequency synthesis technique you need to find a way of dealing with the cogging torque and the torque ripple which will bedevil this implementation. The basic techniques involved have been known for 20 years, see for instance “Minimization of Torque Ripple in Brushless DC Motor Drives”, H. Le Huy, R. Perret, R. Feuillet:IEEE Ind. Appl. Soc. Ann. Meet., Toronto, 1985, pp. 790- 797.
Chris Brady apparently uses digital signal processing to reduce their effects. I am using different techniques, which I am not willing to share as I am pursuing IP protection on them.
I'm not sure about the specifics of the SP10 electronics because, as I said, I'm replacing them with a completely different implementation.
A partial list of do nots, based very much on personal experience. There may be some mistakes I didn't make but probably not many.
Don't try to design a servo system without a really thorough understanding of feedback theory. Remember, calculus is your friend.
Don't try to design a servo system without being able to translate mechanical and rotational elements into electrical equivalents so that the phase behaviour of the system is calculable.
Don't try to nest two feedback loops inside one another unless you really understand both of them. Once you do understand both of them, you still shouldn't do it.
Don't try to design a servo system which incorporates jitter reduction without taking the effective time constant of the jitter reduction into account. This one bit me on the bum big time, it took me quite a while to find that this was the source of the loop instability. The first prototype locked and everything worked and I thought I had "IT", but I could never get it to work again.
Don't try to design a servo system without a really thorough understanding of feedback theory. Remember, calculus is your friend.
Don't try to design a servo system without being able to translate mechanical and rotational elements into electrical equivalents so that the phase behaviour of the system is calculable.
Don't try to nest two feedback loops inside one another unless you really understand both of them. Once you do understand both of them, you still shouldn't do it.
Don't try to design a servo system which incorporates jitter reduction without taking the effective time constant of the jitter reduction into account. This one bit me on the bum big time, it took me quite a while to find that this was the source of the loop instability. The first prototype locked and everything worked and I thought I had "IT", but I could never get it to work again.
I have had a few SP10 MK11P tts (ex. BBC) The one I have now has had all of the electronics normally fitted to the motor frame removed and fitted within the large void in the PS case. The two units are interconnected by an umbilical. It is by far the best SP10 I have heard, despite being in a solid obsidian plinth. ( I intend making a wooden plinth similar to the Kinetia concept)
I have to hand a 12"x1" disc of extremely fine grade pure carbon. This I intend having machined to (initally) sit on the platter as a platter mat! Should that bring about the benefits which I believe are possible thereby, I will have it fitted with a bronze back-plate and machined to fit directly on the drive shaft.
(Be Yamamura made an enormous tt (full air float as opposed to air lift) some dozen years ago when living in Italy.....It was absolutely stunning in every way.)
When this has all 'happened' I will report back!
I have to hand a 12"x1" disc of extremely fine grade pure carbon. This I intend having machined to (initally) sit on the platter as a platter mat! Should that bring about the benefits which I believe are possible thereby, I will have it fitted with a bronze back-plate and machined to fit directly on the drive shaft.
(Be Yamamura made an enormous tt (full air float as opposed to air lift) some dozen years ago when living in Italy.....It was absolutely stunning in every way.)
When this has all 'happened' I will report back!
Re: power supply
No such thing.
The two available drive methods are to replicate the servo loop speed control or to use a synchronous drive with some method of compensation for the torque cogging and ripple torque of the motor.
The second turns out to be even more complex than the first so neither fits your criteria.
brianco said:
No such thing.
The two available drive methods are to replicate the servo loop speed control or to use a synchronous drive with some method of compensation for the torque cogging and ripple torque of the motor.
The second turns out to be even more complex than the first so neither fits your criteria.
Hello Everybody..... all this talk about motor control....actually Technics sp10 MKII motor control.....I myself would think that the technics engineers knew about torque ripple cogging etc... and would have engineered those problems out of the deck....
Btw my sp10MKII didn't sound all that great until i rebuilt the power supply... i have changed my mind of selling it....that made such a difference i cannot believe ..........that and its sitting on a 450lb stand...
it also has a blue fulton mat and i am using my basis record clamp ....BTW i think the clamp acts as a vibrational sink its damped with dynamat extream ...these made a great difference....
Lawrence
Btw my sp10MKII didn't sound all that great until i rebuilt the power supply... i have changed my mind of selling it....that made such a difference i cannot believe ..........that and its sitting on a 450lb stand...
it also has a blue fulton mat and i am using my basis record clamp ....BTW i think the clamp acts as a vibrational sink its damped with dynamat extream ...these made a great difference....Lawrence
The Technics engineers used a servo loop which does indeed guarantee almost perfect speed stability in the face of cogging torque and torque ripple.
On the other hand the counter torque from the minute drive variations which the servo uses to compensate for these effects is necessarily transmitted straight to the chassis, making the choice of mounting for this beast very important.
Which brings us neatly back to the original topic of plinths for the SP10.
Has anyone heard Jonathan Weiss's slate plinth?
On the other hand the counter torque from the minute drive variations which the servo uses to compensate for these effects is necessarily transmitted straight to the chassis, making the choice of mounting for this beast very important.
Which brings us neatly back to the original topic of plinths for the SP10.
Has anyone heard Jonathan Weiss's slate plinth?
Re: One more once
The first pic in this url is not a Kinyata plinth as implied by the adjoining blurb. The Kinyata 'treatment' includes removing the TT top plate and mounting just the DD motor in the fruitwood plinth. Also the motor drive electronics are discarded in favour of Kinyata's own electronics which are mounted in a seperate chassis/enclosure.
This is not to imply that the plinth shown is not good.
The Obsidian plinth does have problems: I have had both versions. My experience is that the rubber feet used a) perish quite alarmingly and, b) that they are best removed. I tried many conical solid feet - all were better than the originals. The best though were 4 solid graphite (fine grain) feet. This treatment certainly improved the tightness, imagery and noise floor to such an extent that I would say that the obsidian plinth is much better than many believe. I use mine on a sand box on a wall mount shelf stand.
jlsem said:
The first pic in this url is not a Kinyata plinth as implied by the adjoining blurb. The Kinyata 'treatment' includes removing the TT top plate and mounting just the DD motor in the fruitwood plinth. Also the motor drive electronics are discarded in favour of Kinyata's own electronics which are mounted in a seperate chassis/enclosure.
This is not to imply that the plinth shown is not good.
The Obsidian plinth does have problems: I have had both versions. My experience is that the rubber feet used a) perish quite alarmingly and, b) that they are best removed. I tried many conical solid feet - all were better than the originals. The best though were 4 solid graphite (fine grain) feet. This treatment certainly improved the tightness, imagery and noise floor to such an extent that I would say that the obsidian plinth is much better than many believe. I use mine on a sand box on a wall mount shelf stand.
Hi John, sorry to drag up this old thread, but I am cleaning up for a move into the unit (condo) I just bought and I found an SP10 motor unit in a box in my store. I thought I'd sold this one, and I have definitely sold the PSU (easy enough to make another) and through a bit of surfing I found your build for the SP10 base here and then this thread. I plan to copy your design, if that's OK, and would like to know which material should I use for the block; the grey cast iron or magnesium? I can get both just as easily so I'd rather just build once and do it as right as possible first time.jlsem said:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.