Hard drive motor question
Jim
have salvaged so far are from 3 inch drives and all have 4 connections which
go to the motor, one of which is a ground. I haven't been able to tell much
from tracing things out on the PCB except that I now know that the motor
does not run from the 5 volt or 12 volt power supply directly.
That's all I want to do is get the motor turning at something close to it's
original speed.
I guess that there must be some fast switching going on amoung the 3
remaining connections to the motor, and if this is so, is there some way to
duplicate this that isn't too complicated?? I made a wrong assumption that
since a hard drive motor is always turning when the computer power is on,
that it was running directly from the power supply. So now I need some help
in figuring things out.
Thanks
Jim
jb...@cfw.com
Ken Taylor
http://www.ti.com/sc/docs/storage/products/data/silspin.pdf
Ken
"Jim" <jb...@cfw.com> wrote in message news:31pr9.4557$46.3661@fe01...
Chuck Simmons
Actually spin is controlled by a micro or DSP in a drive. However. the
motor needs some explanation.
Typically, the motor is a three phase brushless DC motor but it does not
have commutation sensors. Usually a DC brushless motor has some sort of
commutation sensors or encoder. A 3 phase motor using linear hall effect
sensors has 11 wires, 3 coil, 2 hall bias and 3 differential pair hall
outputs. These motors are found in many CD and DVD players. A 3 phase
motor using hall switches will usually have 8 wires. 3 coil, 2 for hall
switch power and ground and 3 hall switch logic signals. With linear
sensors, the motor driver has three differential comparators to convert
the hall voltages to logic levels. If hall switches are used, the logic
from them goes directly to a part such as the TI/Unitrode UC3625 which
produces the correct signals for the coil drive transistors.
Hard disk drives use a different approach. The motors have only the
three coil wires. To start the motor, the MCU or DSP blindly attempts to
start rotation by operating it as a stepper motor. A back EMF detector
will start working when the motor is spining fast enough. When the BEMF
detector starts to work, the MCU or DSP detects that event and uses the
BEMF detection from one phase to "dead reckon" optimum commutation based
on a timer. The motor is then accelerated to desired speed.
There are methods of optimal starting covered in various patents but
this is needless detail since the blind stepper start does work and has
been used in many drives.
Running a hard drive motor will require adding a commutation sensor or
using an MCU as mentioned above. It is not all that trivial.
Chuck
--
... The times have been,
That, when the brains were out,
the man would die. ... Macbeth
Chuck Simmons chr...@webaccess.net
Uwe Zimmermann
>
> How can I get a motor salvaged from a hard drive to run? All the ones I
> have salvaged so far are from 3 inch drives and all have 4 connections which
> go to the motor, one of which is a ground. I haven't been able to tell much
> from tracing things out on the PCB except that I now know that the motor
> does not run from the 5 volt or 12 volt power supply directly.
>
> That's all I want to do is get the motor turning at something close to it's
> original speed.
>
> I guess that there must be some fast switching going on amoung the 3
> remaining connections to the motor, and if this is so, is there some way to
> duplicate this that isn't too complicated??
These motor contain three coils (if it has 4 wires) which have to be
switched in sequence in order to get the rotor rotating in the
rotating magnetic field created by the coils - much licke a stepper
motor.
For a simple test you could use a three stage decimal counter and a
set of power transistors. The switching frequency will directly
determine the rotation speed, i.e. for 7200rpm you'd have to switch
each single coil 120 times per second.
The problem is that the motor will probably not start moving, unless
you ramp up the switching frequency slowly from 0Hz until you reach
the desired rpm.
Uwe.
>
> Thanks
> Jim
> jb...@cfw.com
--
------------------------------------------------
Uwe Zimmermann uw...@geocities.com
Dipl. Phys. uw...@kth.se
Solid State Electronics
Department of Microelectronics and IT
Royal Institute of Technology
Electrum 229
S-16440 Kista-Stockholm
Sweden
www.ele.kth.se/FTE www.geocities.com/uwezi
------------------------------------------------
Chuck Simmons
>
> Jim wrote:
> >
> > How can I get a motor salvaged from a hard drive to run? All the ones I
> > have salvaged so far are from 3 inch drives and all have 4 connections which
> > go to the motor, one of which is a ground. I haven't been able to tell much
> > from tracing things out on the PCB except that I now know that the motor
> > does not run from the 5 volt or 12 volt power supply directly.
> >
> > That's all I want to do is get the motor turning at something close to it's
> > original speed.
> >
> > I guess that there must be some fast switching going on amoung the 3
> > remaining connections to the motor, and if this is so, is there some way to
> > duplicate this that isn't too complicated??
>
> These motor contain three coils (if it has 4 wires) which have to be
> switched in sequence in order to get the rotor rotating in the
> rotating magnetic field created by the coils - much licke a stepper
> motor.
> For a simple test you could use a three stage decimal counter and a
> set of power transistors. The switching frequency will directly
> determine the rotation speed, i.e. for 7200rpm you'd have to switch
> each single coil 120 times per second.
> The problem is that the motor will probably not start moving, unless
> you ramp up the switching frequency slowly from 0Hz until you reach
> the desired rpm.
>
> Uwe.
Ken Taylor provided a link to a TI note on reducing motor noise which
shows driving wave forms for essentially half step drive fro a 3 phase
DC brushless motor. An easy way to get all of the required switch states
is to use a 3 stage Johnson counter which is 3 D-flops with common clock
wired Q to D exept the third which is wired !Q to D on the first D-flop.
All states can be decoded with two input gates.
The motors are never used this way except when starting because there is
a resonance of the rotor in the magnetic field. The resonance is well
known to designers of stepper drives. With a large inertial load (such
as disks), it may be impossible to pass through the speed that excites
this resonance. A hard drive motor and load have this resonance at only
a few Hz.
Another reason that the motors are not normally run as steppers is that
they are very inefficient that way. For optimum efficiency, the phase
switching needs to take place where the superimposed BEMF curves of the
3 windings intersect. This maximizes torque of the motor thus reducing
the current drive required to maintain constant speed. (See my post that
explains commutation sensors and how the motors are normally run.)
Jim
guess I should re-do my question.
There does seem to be a lot of "brushless DC motor drive circuit" IC's
(mostly Phillips) out there that can drive this type of motor. Some seem to
require hall sensors, some not.
Can a TDA5140 or TDA5141 be used as a simple single speed driver for these
motors (with correspondingly few other components)? Seems these chips have
a lot of other functions........is there an equivelent but simpler IC that
simply drives the motor?
Jim
Chuck Simmons
>
> After reviewing the responses and a bunch of data sheets off the web, I
> guess I should re-do my question.
>
> There does seem to be a lot of "brushless DC motor drive circuit" IC's
> (mostly Phillips) out there that can drive this type of motor. Some seem to
> require hall sensors, some not.
>
> Can a TDA5140 or TDA5141 be used as a simple single speed driver for these
> motors (with correspondingly few other components)? Seems these chips have
> a lot of other functions........is there an equivelent but simpler IC that
> simply drives the motor?
>
> Jim
I didn't find a TDA5141 data sheet. Does it have hall switch or linear
hall sensor inputs? Actually most standard products do. The drivers for
sensorless running of brushless DC motors are designed to accept a
current command and commutation commands from a microcontroller. I have
never seen a stand alone brushless DC motor for a sensorlsss motor but
there are many for motors with sensors.
A driver that requires hall switch inputs can be used to drive a
sensorless motor such as used in HDDs by using 3 D-flops as I mentioned
previously in a Johnson counter to generate fake hall states. This can
be used at very low speed and low torque.
Jim
sensors. Instead they use the "back-EMF sensing technique to sense the
rotor position" An example: Philips TDA5143T data sheet, which just
happens to be the chip on one of the drives I salvaged. The schematic in
this data sheet is typical of what I am finding for the "sensorless" drive
circuit ic's....three coil winding connections plus the ground are the 4
motor connections and there are usually 6 or so capacitors rquired. Unless
I'm just missing something in reading and understanding these data sheets, I
can't tell if this is ALL that is necessary to run the motor or not. I may
have to unsolder this IC and find a way to breadboard it (ug...soic, not
looking forward to that).
If this is so, this might be a SIMPLE method to control these motors (there
are some versions of these controller chips that are available in DIP
packages.
Of course this begs the next question....how is the RPM of the motor
controlled using this method? Maybe the chip is controlling the motor
speed to ONLY one rpm....not user variable....not that important to me, but
might be something I'd like to keep in the back of my mind.
"Chuck Simmons" <chr...@webaccess.net> wrote in message
news:3DAF56FE...@webaccess.net...
Chuck Simmons
>
> I good number of the data sheets I have been looking at do not require
> sensors. Instead they use the "back-EMF sensing technique to sense the
> rotor position" An example: Philips TDA5143T data sheet, which just
> happens to be the chip on one of the drives I salvaged. The schematic in
> this data sheet is typical of what I am finding for the "sensorless" drive
> circuit ic's....three coil winding connections plus the ground are the 4
> motor connections and there are usually 6 or so capacitors rquired. Unless
> I'm just missing something in reading and understanding these data sheets, I
> can't tell if this is ALL that is necessary to run the motor or not. I may
> have to unsolder this IC and find a way to breadboard it (ug...soic, not
> looking forward to that).
I can't find these data sheets with a google search so I'm in the dark
about what is in these chips. For example, do they have the starting
strategy built in or does that have to be external? Do these chips have
incremental inductance measurement to discover initial rotor position
for optimal start? The last disk drive company I worked for used SGS
Thompson (now ST) chips that did very little. These were custom designed
for the company I worked for. I got out of hard drives in 1996 so I
haven't seen the new parts (optical disks use completely different
drivers for various reasons).
> If this is so, this might be a SIMPLE method to control these motors (there
> are some versions of these controller chips that are available in DIP
> packages.
>
> Of course this begs the next question....how is the RPM of the motor
> controlled using this method? Maybe the chip is controlling the motor
> speed to ONLY one rpm....not user variable....not that important to me, but
> might be something I'd like to keep in the back of my mind.
I don't know how the wind is blowing in the disk industry today but when
I was in it, the preferred method of speed control was to measure the
time for a revolution by reading servo sectors (there is no hard index
of any kind). The difference between desired time and actual time is
used to calculate a current command to the driver. I suspect that this
is still the preferred method because the speed must be controlled to
within about 0.1% or 0.2% based on the system clock from which write
clock is synthesized. This minimizes the wasted write splice area in
each data sector which is required for speed error and write to read
recovery.
Got a URL for these motor drivers? I could have a look and see what they
do.
Note that I am familiar with spin in HDDs but I only did spin a couple
of times from 1980 to 1996. I mainly did head actuator hardware and
software. Now that I am doing optical, I do all of the motor control
software and hardware for up to 7 motors (simple decks have 5 motors but
more complex ones have 7).
Jim
Here's the location of the data sheet:
http://www-us.semiconductors.philips.com/pip/TDA5143T
This one is pretty representative of Philips' other similar chips. The
small schematic in the data sheet definately gives the impression that aside
from a few cap's, not much else is needed to run these motors. No sensors
are needed here......the ST data sheets and other similar ones DO indicate
that Hall effect sensors are needed.....but I would prefer to stay away from
these methods which seem to over-complicate the situation.
I'm not sure though, how run speed (rpm) is determined. Not that it is a
big deal to me, but I'd be interested to know if speed can be adjusted using
this type of driver.
Thanks
Jim
"Chuck Simmons" <chr...@webaccess.net> wrote in message
news:3DB0AAE2...@webaccess.net...
Chuck Simmons
>
> Here's the location of the data sheet:
> http://www-us.semiconductors.philips.com/pip/TDA5143T
>
> This one is pretty representative of Philips' other similar chips. The
> small schematic in the data sheet definately gives the impression that aside
> from a few cap's, not much else is needed to run these motors. No sensors
> are needed here......the ST data sheets and other similar ones DO indicate
> that Hall effect sensors are needed.....but I would prefer to stay away from
> these methods which seem to over-complicate the situation.
This motor controller will do the trick. The capacitor used in the
starting circuit will have to be decided based on the inertial load on
your motor. The commutation delay capacitors are a bit less critical.
The driver assumes that the center of a "Y" wound motor is available to
connect. If the center of the "Y" is not available (typical of small
motors), a "Y" of equal value resistors will give you the center voltage
so the driver should still work. The full data sheet has an example of
using the the uncommitted transconductance amplifier for motor current
control. In normal operation, this amplifier would be driven by a DAC on
an MCU.
> I'm not sure though, how run speed (rpm) is determined. Not that it is a
> big deal to me, but I'd be interested to know if speed can be adjusted using
> this type of driver.
The FG output is defined to be a square wave at half the comutation
frequency. In a disk drive, FG would be timed by a timer in an MCU and
the MCU would compute a current drive for the transconductance amplifier
connected as in the scanner example. A disk drive will typically switch
to timing some other internal event such as servo sectors for speed
control after servo header acquisition.
To use FG for speed control, a one shot such as a 555 can be connected
to produce a fixed pulse width from FG. The pulses can be filtered to
produce a voltage proportional to speed. An opamp can be used to compare
the speed voltage with a reference you provide and the output of the
opamp can drive the transconductance amplifier connected as in the
scanner example. This configuration is stable unless the gain is too
high or the filter time constant to obtain the speed voltage too long.
Read the full data sheet and try the minimum circuit configuration that
is also given in the data sheet.
If you do not know which coil was connected to which output of the
driver for your motor, you have to experiment. Two of the six possible
connections will work. You can use either. The two configurations that
work rotate the motor in opposite directions.