On Tue, 20 Feb 2024 17:32:51 +0000, liz@poppyrecords.invalid.invalid
(Liz Tuddenham) wrote:

>John Larkin <jl@997PotHill.com> wrote:
>
>> On Tue, 20 Feb 2024 12:27:10 +0000, Cursitor Doom <cd@notformail.com>
>> wrote:
>>
>> >
>> >Can I just get some clarification on one point here. The two spools
>> >are not speed controlled as such and just spin or drag (as the case
>> >may be) at the same speed regardless of the tape speed selected? So
>> >it's only the capstan motor that needs precise control speed? That
>> >seems to be implication of what's been posted here so far and it would
>> >make things much simpler if there was only one motor's speed to worry
>> >about.
>>
>> I don't recall any decks that changed the reel motor drives at
>> different capstain speeds. Certainly none that were aware of the
>> amount of tape on each reel.
>
>I have a vague recollection that there was a machine that boosted the
>takeup motor temporarily at high tape speeds to reduce 'billowing'
>during startup -- but I can't remember which machine it was (EMI BTR2 or
>Marconi-Stille ??). The Ferrograph Series 7 had two torque settings to
>deal with small and large-hub reels - but it didn't seem to make much
>difference.

I did wonder what that knob was for! Thanks, Liz.

>
>The Collaro 'pushmi-pullyu' deck had a constant speed spool motor with
>variable friction drive to the spool hubs controlled by tension arms.
>The spool motor and the capstan motors were identical (apart from
>diection of rotations) and swapped functions when the deck went into
>reverse.
>
>Grundig 'Stenorette' dictating machines had a constant rotational speed
>spool permanently built into the machine, and no capstan. The tape
>cassette had a loose end hanging out with a loop which you hooked around
>a pillar on the drive spool and the tape gradually sped up as it built
>up on the spool hub. As the recordings had also been made on the same
>type of machine, the pitch didn't vary on playback.
>
>Wire recorders almost all used spool drive, but there was one which used
>a capstan with the wire wrapped around it in a single turn with no
>pressure roller.

On Tue, 20 Feb 2024 15:55:42 -0500, DJ Delorie <dj@delorie.com> wrote:

>Robert Roland <fake@ddress.no> writes:
>>>The usual BLDC doesn't speed regulate. Speed depends on the DC supply
>>>voltage and the loading. Basically, it tries as hard as it can.
>>
>> That's correct. But some speed controllers also support a constant
>> speed mode,
>
>And when you get to industrial BLDC motors, you add smarter controllers
>and position feedback, and you can control pretty much anything wrt that
>motor - speed, torque, position, acceleration, etc. Yup, model
>helicopters and big CNC machines use the same type of motors :-)

Small fan-type BLDC motors often have 4 pins: V+, ground, PWM in, and
tach out. The PWM input controls speed, not very accurately, from zero
to max.

On Tue, 20 Feb 2024 14:53:00 -0800, john larkin <jl@650pot.com> wrote:

>On Tue, 20 Feb 2024 15:55:42 -0500, DJ Delorie <dj@delorie.com> wrote:
>
>>Robert Roland <fake@ddress.no> writes:
>>>>The usual BLDC doesn't speed regulate. Speed depends on the DC supply
>>>>voltage and the loading. Basically, it tries as hard as it can.
>>>
>>> That's correct. But some speed controllers also support a constant
>>> speed mode,
>>
>>And when you get to industrial BLDC motors, you add smarter controllers
>>and position feedback, and you can control pretty much anything wrt that
>>motor - speed, torque, position, acceleration, etc. Yup, model
>>helicopters and big CNC machines use the same type of motors :-)
>
>Small fan-type BLDC motors often have 4 pins: V+, ground, PWM in, and
>tach out. The PWM input controls speed, not very accurately, from zero
>to max.

Hence the need for feedback. I wonder if there's a motor that can spin
at a given speed accurately without f/back?

On Tue, 20 Feb 2024 23:09:10 +0000, Cursitor Doom <cd@notformail.com>
wrote:

>On Tue, 20 Feb 2024 14:53:00 -0800, john larkin <jl@650pot.com> wrote:
>
>>On Tue, 20 Feb 2024 15:55:42 -0500, DJ Delorie <dj@delorie.com> wrote:
>>
>>>Robert Roland <fake@ddress.no> writes:
>>>>>The usual BLDC doesn't speed regulate. Speed depends on the DC supply
>>>>>voltage and the loading. Basically, it tries as hard as it can.
>>>>
>>>> That's correct. But some speed controllers also support a constant
>>>> speed mode,
>>>
>>>And when you get to industrial BLDC motors, you add smarter controllers
>>>and position feedback, and you can control pretty much anything wrt that
>>>motor - speed, torque, position, acceleration, etc. Yup, model
>>>helicopters and big CNC machines use the same type of motors :-)
>>
>>Small fan-type BLDC motors often have 4 pins: V+, ground, PWM in, and
>>tach out. The PWM input controls speed, not very accurately, from zero
>>to max.
>
>Hence the need for feedback. I wonder if there's a motor that can spin
>at a given speed accurately without f/back?

AC synchronous. Old clocks use them.

Or a stepper. Or a "torque motor"

In article <m6cati52vmtk96c5lf6q64irgbjpi68jon@4ax.com>,
Cursitor Doom <cd@notformail.com> wrote:

>>Small fan-type BLDC motors often have 4 pins: V+, ground, PWM in, and
>>tach out. The PWM input controls speed, not very accurately, from zero
>>to max.
>
>Hence the need for feedback. I wonder if there's a motor that can spin
>at a given speed accurately without f/back?

Yes. Some brushless DC motors have integral controllers which are
designed with this in mind. Since the controller for a BLDC motor has
to be aware of the rotor position (in order to commutate the phases at
the right times) it has the information it needs to control the speed.

There's one such in my LP turntable, for example... it has a couple of
speed-adjust pots. Once set properly, it keeps the platter rotating
at a stable 33 1/3 RPM, despite variations in the torque required to
overcome drag (from the stylus, record brush, etc.).

The simpler BLDC motor controllers simply hard-switch the supplied DC
voltage to the coils at the proper times - for these motors, the speed
depends on the supplied DC voltage.

The more sophisticated controllers will PWM the supplied DC to the
coils. As I understand it, the timing of the switching between coils
depends on the rotor position, while the PWM duty cycle (and thus the
average voltage applied) is altered to control the speed.

On Tuesday, February 20, 2024 at 6:09:18 PM UTC-5, Cursitor Doom wrote:
> On Tue, 20 Feb 2024 14:53:00 -0800, john larkin <j...@650pot.com> wrote:
>
> >On Tue, 20 Feb 2024 15:55:42 -0500, DJ Delorie <d...@delorie.com> wrote:
> >
> >>Robert Roland <fa...@ddress.no> writes:
> >>>>The usual BLDC doesn't speed regulate. Speed depends on the DC supply
> >>>>voltage and the loading. Basically, it tries as hard as it can.
> >>>
> >>> That's correct. But some speed controllers also support a constant
> >>> speed mode,
> >>
> >>And when you get to industrial BLDC motors, you add smarter controllers
> >>and position feedback, and you can control pretty much anything wrt that
> >>motor - speed, torque, position, acceleration, etc. Yup, model
> >>helicopters and big CNC machines use the same type of motors :-)
> >
> >Small fan-type BLDC motors often have 4 pins: V+, ground, PWM in, and
> >tach out. The PWM input controls speed, not very accurately, from zero
> >to max.
> Hence the need for feedback. I wonder if there's a motor that can spin
> at a given speed accurately without f/back?

The AC synchronous motor is probably the typical example. However there are some situations that make it hard to control, such as starting and exceeding slip spec. (Not the best choice for use in a R-R IMHO

A stepper motor can do this but there are challenges in other areas as well, they can miss steps, and starting speed vs torque control.
Any book on electric machinery will provide a good understanding of motor types, classifications, and performance curves.
In general, If you want good controllability (speed, position) a DC motor driven by power amplifier will do well....dig out your books on classical control theory and modern (state space) control for understanding on how to design various position, speed controllers/regulators. Central to this is understanding static and dynamic loads the system will experience along with motor mechanics (e.g. inertia, etc). IF you really want to redesign the R-R motor system, you should get matlab and simulink and the control theory tool box and have at it. Sure beats the bad old days of hand drawing bode plots and gain phase plots and overlaying the effects of compensation networks....

If you have never implemented discretized PID, filters, and other control approaches, you will have fun discovering things like integral windup, digital limiting, and correctly implementing digital filters and precise periodic sampling. These issues and solutions have been well documented, but if you never have done this before, or know where to look, you are in for a wonderful learning experience.
You just might want to get new belts and clean up the motors for the R-R deck....

On a sunny day (Tue, 20 Feb 2024 08:11:44 -0800) it happened John Larkin
<jl@997PotHill.com> wrote in <mvi9ti90okknq83gn85bma09du2msqtuu3@4ax.com>:

>On Tue, 20 Feb 2024 12:27:10 +0000, Cursitor Doom <cd@notformail.com>
>wrote:
>
>>
>>Can I just get some clarification on one point here. The two spools
>>are not speed controlled as such and just spin or drag (as the case
>>may be) at the same speed regardless of the tape speed selected? So
>>it's only the capstan motor that needs precise control speed? That
>>seems to be implication of what's been posted here so far and it would
>>make things much simpler if there was only one motor's speed to worry
>>about.
>
>I don't recall any decks that changed the reel motor drives at
>different capstain speeds. Certainly none that were aware of the
>amount of tape on each reel.
>
>Big sci-fi movie type computer tape drives used air columns to buffer
>the reels. That reduced the effective inertia of the reels to about
>zero. The capstains could start/stop every record, ballpark an inch of
>tape, as needed.
>
>I designed a tape controller, Ampex 9-track to PDP-11. I'd forgotten
>all about that until this thread.

The video quadruplex AVR1 from Ampex had the air-column buffers
it used photocells to see were the tape was in the air compartment
and a servo on the suppply reel motor to keep just enough tape in that air space
https://www.flickr.com/photos/watvhistory/3365534784/in/photostream/
note the holes in the capstan drive, no rubber wheel! it sucked the tape fixed...
The white thing top left is motor with a wheel with 4 rotating heads....
more pictures in that link, look a the electronics
https://www.flickr.com/photos/watvhistory/3365533536/in/photostream/
we had several of those, often not working when powered on in the morning
finaly it was decided to leave it on all the time,...
Used for video editing, a reel with tape was really heavy...
Had to keep it running, some factory mods were applied over time too IIRC.
vacuum system?
https://www.flickr.com/photos/watvhistory/3365534384/in/photostream/
Expensive machine...

On Tue, 20 Feb 2024 23:09:10 +0000, Cursitor Doom <cd@notformail.com>
wrote:

>Hence the need for feedback. I wonder if there's a motor that can spin
>at a given speed accurately without f/back?

That would be called a synchronous motor. A BLDC motor is actually a
synchronous motor. If it gets blindly commutated at a certain speed,
it will rotate at that speed (but it will be inefficient). It is
possible to abuse a BLDC motor as a stepper motor. If you apply
current to one of its windings, the rotor will snap into one position
and hold that position.

In order to optimize efficiency, the controller needs to know when to
commutate. Hobby controllers are available in two types, sensorless
and sensored. Sensorless systems need almost no additional hardware
for the feedback. They simply measure the EMF produced by the rotating
magnets.
--
RoRo

onsdag den 21. februar 2024 kl. 14.52.19 UTC+1 skrev Robert Roland:
> On Tue, 20 Feb 2024 23:09:10 +0000, Cursitor Doom <c...@notformail.com>
> wrote:
> >Hence the need for feedback. I wonder if there's a motor that can spin
> >at a given speed accurately without f/back?
> That would be called a synchronous motor. A BLDC motor is actually a
> synchronous motor. If it gets blindly commutated at a certain speed,
> it will rotate at that speed (but it will be inefficient).

only at very slow speeds, else it'll just vibrate

On Thursday, February 22, 2024 at 1:30:17 AM UTC+11, Lasse Langwadt Christensen wrote:
> onsdag den 21. februar 2024 kl. 14.52.19 UTC+1 skrev Robert Roland:
> > On Tue, 20 Feb 2024 23:09:10 +0000, Cursitor Doom <c...@notformail.com>
> > wrote:
> > >Hence the need for feedback. I wonder if there's a motor that can spin
> > >at a given speed accurately without f/back?
> > That would be called a synchronous motor. A BLDC motor is actually a
> > synchronous motor. If it gets blindly commutated at a certain speed,
> > it will rotate at that speed (but it will be inefficient).
>
> only at very slow speeds, else it'll just vibrate.

You do need to have some kind of sensor for rotation. If you have got that - no matter how crude - you can feed an acceleration sequence in to the motor coils and spin it up gradually enough that you can be pretty confident is will end up rotating at the synchronous speed.

At EMI Central Research we put together a linear scan driven by a pin on a loop of chain, and a stepper motor drove the gear that moved the chain around its loop.

If the magnet at the pin didn't got past the Hall effect sensor when it should, the motor had been stalled, and the state machine read the 16 step acceleration sequence out of PROM to get the system going again. It was quick and reliable - we did four scans a second - and people would lean on the system to stop it to watch it recover.

--
Bill Sloman, Sydney

On Wed, 21 Feb 2024 14:52:19 +0100, Robert Roland <fake@ddress.no>
wrote:

>On Tue, 20 Feb 2024 23:09:10 +0000, Cursitor Doom <cd@notformail.com>
>wrote:
>
>>Hence the need for feedback. I wonder if there's a motor that can spin
>>at a given speed accurately without f/back?
>
>That would be called a synchronous motor. A BLDC motor is actually a
>synchronous motor. If it gets blindly commutated at a certain speed,
>it will rotate at that speed (but it will be inefficient). It is
>possible to abuse a BLDC motor as a stepper motor. If you apply
>current to one of its windings, the rotor will snap into one position
>and hold that position.
>
>In order to optimize efficiency, the controller needs to know when to
>commutate. Hobby controllers are available in two types, sensorless
>and sensored. Sensorless systems need almost no additional hardware
>for the feedback. They simply measure the EMF produced by the rotating
>magnets.

Thanks. I'm just trying to work out which type would be most suited to
the role of a capstan roller motor to use at 3 fixed speeds (after
gearing down if necessary).

On 22/02/2024 11:40 am, Cursitor Doom wrote:
> On Wed, 21 Feb 2024 14:52:19 +0100, Robert Roland <fake@ddress.no>
> wrote:
>
>> On Tue, 20 Feb 2024 23:09:10 +0000, Cursitor Doom <cd@notformail.com>
>> wrote:
>>
>>> Hence the need for feedback. I wonder if there's a motor that can spin
>>> at a given speed accurately without f/back?
>>
>> That would be called a synchronous motor. A BLDC motor is actually a
>> synchronous motor. If it gets blindly commutated at a certain speed,
>> it will rotate at that speed (but it will be inefficient). It is
>> possible to abuse a BLDC motor as a stepper motor. If you apply
>> current to one of its windings, the rotor will snap into one position
>> and hold that position.
>>
>> In order to optimize efficiency, the controller needs to know when to
>> commutate. Hobby controllers are available in two types, sensorless
>> and sensored. Sensorless systems need almost no additional hardware
>> for the feedback. They simply measure the EMF produced by the rotating
>> magnets.
>
> Thanks. I'm just trying to work out which type would be most suited to
> the role of a capstan roller motor to use at 3 fixed speeds (after
> gearing down if necessary).

The obvious answer is a stepper motor (synchronous motor) with a
crystal-controlled frequency drive.

It will work better if the controller can generate acceleration and
deceleration sequences to make slow and smooth changes in rotational
speed - the spools of tape have rotational intertia and you can't change
their speed of rotation all that quickly.

Once you have got it up to speed, the rotational frequency will be as
stable as your crystal clock. There will some phase lag between the
drive waveform and the position of the rotor - it creates the torque
that counteracts the friction losses, but that should be pretty stable.

You do need some kind of stall detector to accelerate the motor up to
speed again after some ham-fisted user has stopped it's rotation.

--
Bill Sloman, Sydney

On Thu, 22 Feb 2024 15:03:24 +1100, Bill Sloman <bill.sloman@ieee.org>
wrote:

>On 22/02/2024 11:40 am, Cursitor Doom wrote:
>> On Wed, 21 Feb 2024 14:52:19 +0100, Robert Roland <fake@ddress.no>
>> wrote:
>>
>>> On Tue, 20 Feb 2024 23:09:10 +0000, Cursitor Doom <cd@notformail.com>
>>> wrote:
>>>
>>>> Hence the need for feedback. I wonder if there's a motor that can spin
>>>> at a given speed accurately without f/back?
>>>
>>> That would be called a synchronous motor. A BLDC motor is actually a
>>> synchronous motor. If it gets blindly commutated at a certain speed,
>>> it will rotate at that speed (but it will be inefficient). It is
>>> possible to abuse a BLDC motor as a stepper motor. If you apply
>>> current to one of its windings, the rotor will snap into one position
>>> and hold that position.
>>>
>>> In order to optimize efficiency, the controller needs to know when to
>>> commutate. Hobby controllers are available in two types, sensorless
>>> and sensored. Sensorless systems need almost no additional hardware
>>> for the feedback. They simply measure the EMF produced by the rotating
>>> magnets.
>>
>> Thanks. I'm just trying to work out which type would be most suited to
>> the role of a capstan roller motor to use at 3 fixed speeds (after
>> gearing down if necessary).
>
>The obvious answer is a stepper motor (synchronous motor) with a
>crystal-controlled frequency drive.
>
>It will work better if the controller can generate acceleration and
>deceleration sequences to make slow and smooth changes in rotational
>speed - the spools of tape have rotational intertia and you can't change
>their speed of rotation all that quickly.

That *is* something I'm concerned could spoil the party with the
simpler solutions proposed here. Not sure if it'll make much
difference in practice, but we'll find out empirically I guess.

>
>Once you have got it up to speed, the rotational frequency will be as
>stable as your crystal clock. There will some phase lag between the
>drive waveform and the position of the rotor - it creates the torque
>that counteracts the friction losses, but that should be pretty stable.
>
>You do need some kind of stall detector to accelerate the motor up to
>speed again after some ham-fisted user has stopped it's rotation.

On Thu, 22 Feb 2024 00:40:01 +0000, Cursitor Doom <cd@notformail.com>
wrote:

>On Wed, 21 Feb 2024 14:52:19 +0100, Robert Roland <fake@ddress.no>
>wrote:

>>Hobby controllers are available in two types, sensorless
>>and sensored.

>Thanks. I'm just trying to work out which type would be most suited to
>the role of a capstan roller motor to use at 3 fixed speeds (after
>gearing down if necessary).

The most important difference between the two systems, is starting
torque.

Since the sensorless systems use the moving magnets to determine
commutation timing, the motor must be spinning in order to commutate.
Of course, it needs commutation in order to spin, so you essentially
have a catch-22 situation.

There are different strategies to overcome the startup problem. The
simplest one is to simply commutate "blindly" at low current and see
if any timing signals show up. There are more sophisticated methods,
but common to them all is that they provide very low torque at zero
speed. For propellers or helicopter rotors, this is not a problem, so
sensorless systems are used. For cars, however, starting torque is
important, so sensored systems are used.

The sensors are simply a few Hall effect sensors. There is no need for
any shaft encoders. In hobby products, the sensors are built in to the
motor at the factory, so the end user simply sees a few extra wires
that need to be connected to the controller.
--
RoRo

On 26/02/2024 12:34 am, Robert Roland wrote:
> On Thu, 22 Feb 2024 00:40:01 +0000, Cursitor Doom <cd@notformail.com>
> wrote:
>
>> On Wed, 21 Feb 2024 14:52:19 +0100, Robert Roland <fake@ddress.no>
>> wrote:
>
>>> Hobby controllers are available in two types, sensorless
>>> and sensored.
>
>> Thanks. I'm just trying to work out which type would be most suited to
>> the role of a capstan roller motor to use at 3 fixed speeds (after
>> gearing down if necessary).
>
> The most important difference between the two systems, is starting
> torque.
>
> Since the sensorless systems use the moving magnets to determine
> commutation timing, the motor must be spinning in order to commutate.
> Of course, it needs commutation in order to spin, so you essentially
> have a catch-22 situation.
>
> There are different strategies to overcome the startup problem. The
> simplest one is to simply commutate "blindly" at low current and see
> if any timing signals show up. There are more sophisticated methods,
> but common to them all is that they provide very low torque at zero
> speed. For propellers or helicopter rotors, this is not a problem, so
> sensorless systems are used. For cars, however, starting torque is
> important, so sensored systems are used.

Stepper motors always provide the same torque when they step slowly at
any speed - as long as the current through coil can get up to the
tolerable peak, you will get the same torque.

If the magnetic field lines up with position of the rotor, you won't get
any torque, so the strategy is to start by stepping the magnetic field
slowly enough that rotor can follow the rotating magnetic field, which
gets rid of any initial stiction. At low step rates the rotor can
oscillate around the zero torque position, and you have to avoid steps
rates that match that oscillation frequency. Once you have got the rotor
moving slowly, you know where it is and you can start your acceleration
sequence.

> The sensors are simply a few Hall effect sensors. There is no need for
> any shaft encoders. In hobby products, the sensors are built in to the
> motor at the factory, so the end user simply sees a few extra wires
> that need to be connected to the controller.

The Hall sensors are shaft encoders - the rotor is bonded to the shaft,
and magnets in the rotor are what you are detecting.

The end user may see them as a few extra wires, but sophisticated users
will see them for what they are.

Cursitor Doom isn't a sophisticated user, but if he is posting here we
need to treat him as if he could acquire some sophistication.

--
Bill Sloman, Sydney

On Mon, 26 Feb 2024 01:57:01 +1100, Bill Sloman <bill.sloman@ieee.org>
wrote:

>On 26/02/2024 12:34 am, Robert Roland wrote:
>> On Thu, 22 Feb 2024 00:40:01 +0000, Cursitor Doom <cd@notformail.com>
>> wrote:
>>
>>> On Wed, 21 Feb 2024 14:52:19 +0100, Robert Roland <fake@ddress.no>
>>> wrote:
>>
>>>> Hobby controllers are available in two types, sensorless
>>>> and sensored.
>>
>>> Thanks. I'm just trying to work out which type would be most suited to
>>> the role of a capstan roller motor to use at 3 fixed speeds (after
>>> gearing down if necessary).
>>
>> The most important difference between the two systems, is starting
>> torque.
>>
>> Since the sensorless systems use the moving magnets to determine
>> commutation timing, the motor must be spinning in order to commutate.
>> Of course, it needs commutation in order to spin, so you essentially
>> have a catch-22 situation.
>>
>> There are different strategies to overcome the startup problem. The
>> simplest one is to simply commutate "blindly" at low current and see
>> if any timing signals show up. There are more sophisticated methods,
>> but common to them all is that they provide very low torque at zero
>> speed. For propellers or helicopter rotors, this is not a problem, so
>> sensorless systems are used. For cars, however, starting torque is
>> important, so sensored systems are used.
>
>Stepper motors always provide the same torque when they step slowly at
>any speed - as long as the current through coil can get up to the
>tolerable peak, you will get the same torque.
>
>If the magnetic field lines up with position of the rotor, you won't get
>any torque, so the strategy is to start by stepping the magnetic field
>slowly enough that rotor can follow the rotating magnetic field, which
>gets rid of any initial stiction. At low step rates the rotor can
>oscillate around the zero torque position, and you have to avoid steps
>rates that match that oscillation frequency. Once you have got the rotor
>moving slowly, you know where it is and you can start your acceleration
>sequence.
>
>> The sensors are simply a few Hall effect sensors. There is no need for
>> any shaft encoders. In hobby products, the sensors are built in to the
>> motor at the factory, so the end user simply sees a few extra wires
>> that need to be connected to the controller.
>
>The Hall sensors are shaft encoders - the rotor is bonded to the shaft,
>and magnets in the rotor are what you are detecting.
>
>The end user may see them as a few extra wires, but sophisticated users
>will see them for what they are.
>
>Cursitor Doom isn't a sophisticated user, but if he is posting here we
>need to treat him as if he could acquire some sophistication.

Up until that last paragraph I was just about to commend you on being
more like the old Bill Sloman who posted helpful advice here back in
the day. You just can't resist throwing barbs, can you? Sigh...

On Sun, 25 Feb 2024 17:46:41 +0000, Cursitor Doom <cd@notformail.com>
wrote:

>On Mon, 26 Feb 2024 01:57:01 +1100, Bill Sloman <bill.sloman@ieee.org>
>wrote:
>
>>On 26/02/2024 12:34 am, Robert Roland wrote:
>>> On Thu, 22 Feb 2024 00:40:01 +0000, Cursitor Doom <cd@notformail.com>
>>> wrote:
>>>
>>>> On Wed, 21 Feb 2024 14:52:19 +0100, Robert Roland <fake@ddress.no>
>>>> wrote:
>>>
>>>>> Hobby controllers are available in two types, sensorless
>>>>> and sensored.
>>>
>>>> Thanks. I'm just trying to work out which type would be most suited to
>>>> the role of a capstan roller motor to use at 3 fixed speeds (after
>>>> gearing down if necessary).
>>>
>>> The most important difference between the two systems, is starting
>>> torque.
>>>
>>> Since the sensorless systems use the moving magnets to determine
>>> commutation timing, the motor must be spinning in order to commutate.
>>> Of course, it needs commutation in order to spin, so you essentially
>>> have a catch-22 situation.
>>>
>>> There are different strategies to overcome the startup problem. The
>>> simplest one is to simply commutate "blindly" at low current and see
>>> if any timing signals show up. There are more sophisticated methods,
>>> but common to them all is that they provide very low torque at zero
>>> speed. For propellers or helicopter rotors, this is not a problem, so
>>> sensorless systems are used. For cars, however, starting torque is
>>> important, so sensored systems are used.
>>
>>Stepper motors always provide the same torque when they step slowly at
>>any speed - as long as the current through coil can get up to the
>>tolerable peak, you will get the same torque.
>>
>>If the magnetic field lines up with position of the rotor, you won't get
>>any torque, so the strategy is to start by stepping the magnetic field
>>slowly enough that rotor can follow the rotating magnetic field, which
>>gets rid of any initial stiction. At low step rates the rotor can
>>oscillate around the zero torque position, and you have to avoid steps
>>rates that match that oscillation frequency. Once you have got the rotor
>>moving slowly, you know where it is and you can start your acceleration
>>sequence.
>>
>>> The sensors are simply a few Hall effect sensors. There is no need for
>>> any shaft encoders. In hobby products, the sensors are built in to the
>>> motor at the factory, so the end user simply sees a few extra wires
>>> that need to be connected to the controller.
>>
>>The Hall sensors are shaft encoders - the rotor is bonded to the shaft,
>>and magnets in the rotor are what you are detecting.
>>
>>The end user may see them as a few extra wires, but sophisticated users
>>will see them for what they are.
>>
>>Cursitor Doom isn't a sophisticated user, but if he is posting here we
>>need to treat him as if he could acquire some sophistication.
>
>Up until that last paragraph I was just about to commend you on being
>more like the old Bill Sloman who posted helpful advice here back in
>the day. You just can't resist throwing barbs, can you? Sigh...

Sloman's real enemy is Sloman.

On Sun, 25 Feb 2024 10:13:45 -0800, John Larkin <jl@997PotHill.com>
wrote:

>On Sun, 25 Feb 2024 17:46:41 +0000, Cursitor Doom <cd@notformail.com>
>wrote:
>
>>On Mon, 26 Feb 2024 01:57:01 +1100, Bill Sloman <bill.sloman@ieee.org>
>>wrote:
>>
>>>On 26/02/2024 12:34 am, Robert Roland wrote:
>>>> On Thu, 22 Feb 2024 00:40:01 +0000, Cursitor Doom <cd@notformail.com>
>>>> wrote:
>>>>
>>>>> On Wed, 21 Feb 2024 14:52:19 +0100, Robert Roland <fake@ddress.no>
>>>>> wrote:
>>>>
>>>>>> Hobby controllers are available in two types, sensorless
>>>>>> and sensored.
>>>>
>>>>> Thanks. I'm just trying to work out which type would be most suited to
>>>>> the role of a capstan roller motor to use at 3 fixed speeds (after
>>>>> gearing down if necessary).
>>>>
>>>> The most important difference between the two systems, is starting
>>>> torque.
>>>>
>>>> Since the sensorless systems use the moving magnets to determine
>>>> commutation timing, the motor must be spinning in order to commutate.
>>>> Of course, it needs commutation in order to spin, so you essentially
>>>> have a catch-22 situation.
>>>>
>>>> There are different strategies to overcome the startup problem. The
>>>> simplest one is to simply commutate "blindly" at low current and see
>>>> if any timing signals show up. There are more sophisticated methods,
>>>> but common to them all is that they provide very low torque at zero
>>>> speed. For propellers or helicopter rotors, this is not a problem, so
>>>> sensorless systems are used. For cars, however, starting torque is
>>>> important, so sensored systems are used.
>>>
>>>Stepper motors always provide the same torque when they step slowly at
>>>any speed - as long as the current through coil can get up to the
>>>tolerable peak, you will get the same torque.
>>>
>>>If the magnetic field lines up with position of the rotor, you won't get
>>>any torque, so the strategy is to start by stepping the magnetic field
>>>slowly enough that rotor can follow the rotating magnetic field, which
>>>gets rid of any initial stiction. At low step rates the rotor can
>>>oscillate around the zero torque position, and you have to avoid steps
>>>rates that match that oscillation frequency. Once you have got the rotor
>>>moving slowly, you know where it is and you can start your acceleration
>>>sequence.
>>>
>>>> The sensors are simply a few Hall effect sensors. There is no need for
>>>> any shaft encoders. In hobby products, the sensors are built in to the
>>>> motor at the factory, so the end user simply sees a few extra wires
>>>> that need to be connected to the controller.
>>>
>>>The Hall sensors are shaft encoders - the rotor is bonded to the shaft,
>>>and magnets in the rotor are what you are detecting.
>>>
>>>The end user may see them as a few extra wires, but sophisticated users
>>>will see them for what they are.
>>>
>>>Cursitor Doom isn't a sophisticated user, but if he is posting here we
>>>need to treat him as if he could acquire some sophistication.
>>
>>Up until that last paragraph I was just about to commend you on being
>>more like the old Bill Sloman who posted helpful advice here back in
>>the day. You just can't resist throwing barbs, can you? Sigh...
>
>Sloman's real enemy is Sloman.

His only supporter here was 3rdWit, who it transpired was just a
sock-puppet. I'm afraid Bill's become something of a sad and rather
tragic figure here in recent years.

On Sun, 25 Feb 2024 18:52:41 +0000, Cursitor Doom <cd@notformail.com>
wrote:

>On Sun, 25 Feb 2024 10:13:45 -0800, John Larkin <jl@997PotHill.com>
>wrote:
>
>>On Sun, 25 Feb 2024 17:46:41 +0000, Cursitor Doom <cd@notformail.com>
>>wrote:
>>
>>>On Mon, 26 Feb 2024 01:57:01 +1100, Bill Sloman <bill.sloman@ieee.org>
>>>wrote:
>>>
>>>>On 26/02/2024 12:34 am, Robert Roland wrote:
>>>>> On Thu, 22 Feb 2024 00:40:01 +0000, Cursitor Doom <cd@notformail.com>
>>>>> wrote:
>>>>>
>>>>>> On Wed, 21 Feb 2024 14:52:19 +0100, Robert Roland <fake@ddress.no>
>>>>>> wrote:
>>>>>
>>>>>>> Hobby controllers are available in two types, sensorless
>>>>>>> and sensored.
>>>>>
>>>>>> Thanks. I'm just trying to work out which type would be most suited to
>>>>>> the role of a capstan roller motor to use at 3 fixed speeds (after
>>>>>> gearing down if necessary).
>>>>>
>>>>> The most important difference between the two systems, is starting
>>>>> torque.
>>>>>
>>>>> Since the sensorless systems use the moving magnets to determine
>>>>> commutation timing, the motor must be spinning in order to commutate.
>>>>> Of course, it needs commutation in order to spin, so you essentially
>>>>> have a catch-22 situation.
>>>>>
>>>>> There are different strategies to overcome the startup problem. The
>>>>> simplest one is to simply commutate "blindly" at low current and see
>>>>> if any timing signals show up. There are more sophisticated methods,
>>>>> but common to them all is that they provide very low torque at zero
>>>>> speed. For propellers or helicopter rotors, this is not a problem, so
>>>>> sensorless systems are used. For cars, however, starting torque is
>>>>> important, so sensored systems are used.
>>>>
>>>>Stepper motors always provide the same torque when they step slowly at
>>>>any speed - as long as the current through coil can get up to the
>>>>tolerable peak, you will get the same torque.
>>>>
>>>>If the magnetic field lines up with position of the rotor, you won't get
>>>>any torque, so the strategy is to start by stepping the magnetic field
>>>>slowly enough that rotor can follow the rotating magnetic field, which
>>>>gets rid of any initial stiction. At low step rates the rotor can
>>>>oscillate around the zero torque position, and you have to avoid steps
>>>>rates that match that oscillation frequency. Once you have got the rotor
>>>>moving slowly, you know where it is and you can start your acceleration
>>>>sequence.
>>>>
>>>>> The sensors are simply a few Hall effect sensors. There is no need for
>>>>> any shaft encoders. In hobby products, the sensors are built in to the
>>>>> motor at the factory, so the end user simply sees a few extra wires
>>>>> that need to be connected to the controller.
>>>>
>>>>The Hall sensors are shaft encoders - the rotor is bonded to the shaft,
>>>>and magnets in the rotor are what you are detecting.
>>>>
>>>>The end user may see them as a few extra wires, but sophisticated users
>>>>will see them for what they are.
>>>>
>>>>Cursitor Doom isn't a sophisticated user, but if he is posting here we
>>>>need to treat him as if he could acquire some sophistication.
>>>
>>>Up until that last paragraph I was just about to commend you on being
>>>more like the old Bill Sloman who posted helpful advice here back in
>>>the day. You just can't resist throwing barbs, can you? Sigh...
>>
>>Sloman's real enemy is Sloman.
>
>His only supporter here was 3rdWit, who it transpired was just a
>sock-puppet. I'm afraid Bill's become something of a sad and rather
>tragic figure here in recent years.

Neither supporters nor enemies make sense in an electronic design
forum. This ain't social media.

On 26/02/2024 4:46 am, Cursitor Doom wrote:
> On Mon, 26 Feb 2024 01:57:01 +1100, Bill Sloman <bill.sloman@ieee.org> wrote:
>> On 26/02/2024 12:34 am, Robert Roland wrote:
>>> On Thu, 22 Feb 2024 00:40:01 +0000, Cursitor Doom <cd@notformail.com> wrote:
>>>> On Wed, 21 Feb 2024 14:52:19 +0100, Robert Roland <fake@ddress.no> wrote:

<snip>

>>> The sensors are simply a few Hall effect sensors. There is no need for
>>> any shaft encoders. In hobby products, the sensors are built in to the
>>> motor at the factory, so the end user simply sees a few extra wires
>>> that need to be connected to the controller.
>>
>> The Hall sensors are shaft encoders - the rotor is bonded to the shaft,
>> and magnets in the rotor are what you are detecting.
>>
>> The end user may see them as a few extra wires, but sophisticated users
>> will see them for what they are.
>>
>> Cursitor Doom isn't a sophisticated user, but if he is posting here we
>> need to treat him as if he could acquire some sophistication.
>
> Up until that last paragraph I was just about to commend you on being
> more like the old Bill Sloman who posted helpful advice here back in
> the day. You just can't resist throwing barbs, can you? Sigh...

It's a moral obligation, in your case. Your undiscriminating enthusiasm
for pro-Putin, pro-Trump and pro-climate change denial propaganda means
that I do need to remind people that you shouldn't be taken seriously.

I do try to educate you, but you don't seem to want to learn.

--
Bill Sloman, Sydney

On 26/02/2024 5:13 am, John Larkin wrote:
> On Sun, 25 Feb 2024 17:46:41 +0000, Cursitor Doom <cd@notformail.com> wrote:
>> On Mon, 26 Feb 2024 01:57:01 +1100, Bill Sloman <bill.sloman@ieee.org> wrote:
>>> On 26/02/2024 12:34 am, Robert Roland wrote:
>>>> On Thu, 22 Feb 2024 00:40:01 +0000, Cursitor Doom <cd@notformail.com> wrote:
>>>>> On Wed, 21 Feb 2024 14:52:19 +0100, Robert Roland <fake@ddress.no> wrote:

<snip>

>>>> The sensors are simply a few Hall effect sensors. There is no need for
>>>> any shaft encoders. In hobby products, the sensors are built in to the
>>>> motor at the factory, so the end user simply sees a few extra wires
>>>> that need to be connected to the controller.
>>>
>>> The Hall sensors are shaft encoders - the rotor is bonded to the shaft,
>>> and magnets in the rotor are what you are detecting.
>>>
>>> The end user may see them as a few extra wires, but sophisticated users
>>> will see them for what they are.
>>>
>>> Cursitor Doom isn't a sophisticated user, but if he is posting here we
>>> need to treat him as if he could acquire some sophistication.
>>
>> Up until that last paragraph I was just about to commend you on being
>> more like the old Bill Sloman who posted helpful advice here back in
>> the day. You just can't resist throwing barbs, can you? Sigh...
>
> Sloman's real enemy is Sloman.

John Larkin divides the world into people who praise him as he feels he
deserves to be praised, and the rest, who are his enemies.

He'd be a lot more successful if he had a more realistic idea of the
limits of his capabilities, so we aren't actually his enemies, even if
he likes to think we are.

An enemy is somebody who tries to damage you. Vanity is a vice, and
feeding somebody's vanity is what an enemy would do.

--
Bill Sloman, Sydney

On 26/02/2024 5:52 am, Cursitor Doom wrote:
> On Sun, 25 Feb 2024 10:13:45 -0800, John Larkin <jl@997PotHill.com> wrote:
>> On Sun, 25 Feb 2024 17:46:41 +0000, Cursitor Doom <cd@notformail.com> wrote:
>>> On Mon, 26 Feb 2024 01:57:01 +1100, Bill Sloman <bill.sloman@ieee.org> wrote:
>>>> On 26/02/2024 12:34 am, Robert Roland wrote:
>>>>> On Thu, 22 Feb 2024 00:40:01 +0000, Cursitor Doom <cd@notformail.com> wrote:
>>>>>> On Wed, 21 Feb 2024 14:52:19 +0100, Robert Roland <fake@ddress.no> wrote:

>>>>
>>>> Cursitor Doom isn't a sophisticated user, but if he is posting here we
>>>> need to treat him as if he could acquire some sophistication.
>>>
>>> Up until that last paragraph I was just about to commend you on being
>>> more like the old Bill Sloman who posted helpful advice here back in
>>> the day. You just can't resist throwing barbs, can you? Sigh...
>>
>> Sloman's real enemy is Sloman.
>
> His only supporter here was 3rdWit, who it transpired was just a
> sock-puppet.

Only in Cursitor Doom's imaginary universe.

> I'm afraid Bill's become something of a sad and rather tragic figure here in recent years.

Only in Cursitor Doom's imaginary universe. Cursitor Doom is fond of his
fatuous and implausible delusions. John Larkin's delusion that he
actually designs his electronic circuits is the same kind of
self-serving error.

--
Bill Sloman, Sydney

On 26/02/2024 6:09 am, John Larkin wrote:
> On Sun, 25 Feb 2024 18:52:41 +0000, Cursitor Doom <cd@notformail.com> wrote:
>> On Sun, 25 Feb 2024 10:13:45 -0800, John Larkin <jl@997PotHill.com wrote
>>> On Sun, 25 Feb 2024 17:46:41 +0000, Cursitor Doom <cd@notformail.com> wrote:
>>>> On Mon, 26 Feb 2024 01:57:01 +1100, Bill Sloman <bill.sloman@ieee.org> wrote:
>>>>> On 26/02/2024 12:34 am, Robert Roland wrote:
>>>>>> On Thu, 22 Feb 2024 00:40:01 +0000, Cursitor Doom <cd@notformail.com> wrote:
>>>>>>> On Wed, 21 Feb 2024 14:52:19 +0100, Robert Roland <fake@ddress.no> wrote:

<snip>

>>>>> Cursitor Doom isn't a sophisticated user, but if he is posting here we
>>>>> need to treat him as if he could acquire some sophistication.
>>>>
>>>> Up until that last paragraph I was just about to commend you on being
>>>> more like the old Bill Sloman who posted helpful advice here back in
>>>> the day. You just can't resist throwing barbs, can you? Sigh...
>>>
>>> Sloman's real enemy is Sloman.
>>
>> His only supporter here was 3rdWit, who it transpired was just a
>> sock-puppet.

Only in Cursitor Doom's imaginary universe.

>> I'm afraid Bill's become something of a sad and rather
>> tragic figure here in recent years.
>
> Neither supporters nor enemies make sense in an electronic design
> forum. This ain't social media.

It's individuals interacting, which makes it a social medium.

When the individual are as flawed as Cursitor Doom and John Larkin, some
of the interactions can get downright ugly, and they resent being called
to account.

--
Bill Sloman, Sydney (not perfect, but closer to it than some I could name).

On 2024-02-18, Cursitor Doom <cd@notformail.com> wrote:
> Gentlemen,
>
> Can motor speed control ever approach the effectiveness of the old
> style drive belts and pullys approach?

you mean like a centrifugal governor?

> Would simple PWM be enough or would there be some additional trickery
> needed?

PWM could. if you sample the back EMF during the off time of the PWM and feed
that back to the regulator... (or read the motor speed some other way,
you could have an interruptor typse sensor and control speed using a
PLL)

Oldschool when they weren't using centrifugal governors they would put
a compensating negative resistance in series with the motor and feed
the combination from a fixed DC voltage or fake that result.

--
Jasen.
🇺🇦 Слава Україні

On Mon, 4 Mar 2024 07:45:43 -0000 (UTC), Jasen Betts
<usenet@revmaps.no-ip.org> wrote:

>On 2024-02-18, Cursitor Doom <cd@notformail.com> wrote:
>> Gentlemen,
>>
>> Can motor speed control ever approach the effectiveness of the old
>> style drive belts and pullys approach?
>
>you mean like a centrifugal governor?
>
>> Would simple PWM be enough or would there be some additional trickery
>> needed?
>
>PWM could. if you sample the back EMF during the off time of the PWM and feed
>that back to the regulator... (or read the motor speed some other way,
>you could have an interruptor typse sensor and control speed using a
>PLL)
>
>Oldschool when they weren't using centrifugal governors they would put
>a compensating negative resistance in series with the motor and feed
>the combination from a fixed DC voltage or fake that result.

Thanks for your input, but this has been thrashed out by now. This
thread is stale.