Very low power miniature motor control

[ssmith579]

I have a problem I'm trying to solve and hopefully someone may have some suggestions here. I am designing a control circuit for a 6mm 1.2V DC motor. I'm using a pic 10F222 to control turning on and off the motor. I want to turn on the motor when I detect the shaft rotating and turn off the motor after a period of non-rotation. I can do this with a magnetic reed switch and a small magnet mounted on the shaft but the lifetime of the reed switch is limited. Also I want to power the whole thing from a 357 battery and it needs a 6 month shelf life. I have a 37uA power budget when idle. I don't have an on/off switch so it's always powered. I am looking at a Holtek HT7727 switcher to power the pic. The problem is figuring out how to detect when the shaft is turning. I need to generate an interrupt to the pic to wake it up and I also need to figure out if the shaft is stopped, moving slowly or free running.

Thanks,
Steve

 

[jraef]

How much rotation must it detect? 1 radian, 1 revolution?

 

[IRstuff]

encoder
resolver
hall sensor
back emf


TTFN

FAQ731-376

 

[ssmith579]

I need to detect when the shaft has rotated 5 to 10 times. I would love to have a tach output on the motor but can't seem to find one.

I looked at hall effect sensors but they are too power hungry. The back emf is only about 20mV so amplifying it to a usable level would kill my power budget. A dry contact closure once per revolution would be ideal.

 

[logbook]

I would love to know what this application is, if possible. You are driving the motor when somebody else is already driving it!

You mention the current drawn when off, but not when on. If the life of a reed switch is inadequate then any mechanical switch is also out. That would include a rotary shaft encoder with brushes. An optical encoder would be pretty power hungry.

The motor back emf can be increased in voltage without an amplifier. Use a transformer!

 

[MacGyverS2000]

I can think of one similar application... CD drive trays. If you press in on them (instead of hitting the button), they close. I have always assumed they looked for a voltage on the motor leads, and if one exists the motor is being driven by outside forces (i.e., you're pushing on the tray). If the motor fails to move the object in question (i.e., the motor has stalled), current draw increases beyond a specified range. this again should be fairly easy to detect.


Dan - Owner

http://www.Hi-TecDesigns.com

 

[ScottyUK]

Hi McGyver,

CD players (audio) used to have a tiny switch often made from two sprung copper leaves which detected when the tray came off its fully extended position. This initiated the tray closure. Don't know if things have changed since those days - that was a quite a while ago when fixing the things was cheaper than replacing them!


----------------------------------

Sometimes I only open my mouth to swap feet...

 

[MacGyverS2000]

Good to know... it would certainly be easier/cheaper to implement.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[Skogsgurra]

I can not understand why life of a reed switch is too short.

A reed switch has something like 10^8 cycles life. And yor device does not seem to need to be working more than every tenth second or so.

That translates into a little more than 30 years. That's a lot longer than most gadgets live these days.

BTW, I have used PIC processors in remote measuring devices that run from batteries and my mean current consumption was 13 uA for a fairly complex circuit that measures temperature, supervises 230 V AC, a few switches and also "takes the line" when you phone it. So, I think that you should be able to shave a lot of current consumption off that application.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[ssmith579]

The application is for a very small spool of thread that fits into a pen. I want to pull on the thread to start the motor and then have it shut off if the thread breaks or it doesn't move for a period of time. The life of the reed switch may not be a big issue. The motor spins at about 500 RPM if the thread breaks. It's more of a problem with assembly. I'm looking for something that is very cheap and easy to assembly.

I'm not sure if the PIC input needs a pull-up though. The datasheet says a weak pull-up is 250uA. That would eat up the battery pretty quick.

Thanks,
Steve

 

[Skogsgurra]

Most (all?) PICs have internal pull-ups when pin configured as input. A good strategy is to tri-state when inputs are not being used. And design your circuitry so that it consumes "no" power in the normal state. Also, if you use as low an operating frequency as possible (I have used 32 kHz watch xtls for the clock) you will have very low power consumption. With 32 kHz, it is, for instance, still possible to detect and discriminate ring signals, generate some rough tones and also do measurements with reasonable accuracy without using any power consuming A/D.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[bogeyman]

Reed switches must not have any apreciable capacitance across them, the discharge currents destroy the contact materials.
Look here

http://www.cotorelay.com

 

[Skogsgurra]

Which can easily be avoided using a small series resistor. Why would there be any capacitance in an application like this? Cable length 0 feet.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[DaveScott]

A different idea might be to paint half the shaft black, and then use a photodiode. But the pen would need ambient light to reflect off the shaft.

 

[watchjohn]

I can´t remember the exact name
but their seems to be a technique where the power
drawn by an electrical motor is observed..it seems
that pulses can be observed on the power line
those pulses represent the motor-speed and are in turn
used to control the engine speed
...so it´s some kind of feedback without additional sensors
No idea if this works for all kinds of electromotors

I have seen even special ic´s for it..just don´t remember
their names...maybe looking for "fan-speed-control-ic"
will give some result


WatchJohn

 

[itsmoked]

I can easily imagine commutation noise that could be monitored for it's frequency then the number of poles taken into account.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[pluckr]

Maybe you could use a rotating magnet and a sense coil - it would not consume any power in operation or in standby.

 

[don01]

how about looking at itsmoked & pluckers sensing ideas?. - something like a coil or pick up on the side of the motor. It should be as noisy as hell when moving and quiet when not.
Amp or switch the "signal" back to the PIC ?

but:-
enjoy trying
Don