View Full Version : Neat cheap power cross feed for mill "pics"

05-18-2003, 01:45 PM

It is a auto wiper motor with speed controls mounted on a lead screw. Neat ideal, I had to share my finding the website.

You know with autozone giving lifetime warranties on auto wiper motors, someone could really enjoy this ideal.

[This message has been edited by ibewgypsie (edited 05-18-2003).]

Alistair Hosie
05-18-2003, 03:00 PM
I don't think people understand how powerful a windscreen wiper motor is i e how much torque. It is also built to last this is a good idea I think these motors could be used for a host of stuff good idea Ibew Alistair

John Stevenson
05-18-2003, 03:12 PM
Many years ago I built a crane for my nephew using three 12v truck wiper motors. One for slew, jib and rope.
I used 36" long 1/8" welding rods welded up into a braced jib, just like the real thing.
A big truck battery set in the base supplied power and ballast.
This could lift a 56# propane bottle.
The kids had endless fun with this until it eventually broke and was neglected.

John S

[This message has been edited by John Stevenson (edited 05-18-2003).]

05-18-2003, 11:04 PM
Make sure you get one that was origally for a car or truck with a large windshield and makes full rotation. I have one that drives my bead roller. It came from a Subaru and even with additional gearing, it bogs down. (It draws about 1.5amps and 3amps on stall, so get something beefier than that)
I picked up a 120v ac motor from a bed at the salvage yard. It's a 80rpm gear motor and as soon as I find the time, I'm going to try it. It's reversable if I can figure out the wire connections. (Black, Red, Green, White) The tag on the motor says to use a capacitor but no diagram. Any one ever wire one of these up?

[This message has been edited by CCWKen (edited 05-18-2003).]

05-19-2003, 12:05 AM
The best wiper motor assemblies I have seen come from rear window wipers. Because they are compact, they must be powerful for size, which they are. I have seen one offered as a winch motor, which is a story in itself, a ripoff story if I had bit, which I didn't. We were talking 1500 lbs pull, he was saying 'see if you can stop the shaft with your hand, there's lots of torque there'. Larger, like the one in the pic, but better, is one from a ford truck or van, with a much more robust output gear. Anyway, the other way to go might be old satelite positioning motor/gearboxes. Usually, the motors are ok, the parts that usually fail wouldn't be needed to adapt to a leadscrew. Any such motor would, of course, need a feedback system if used in a cnc setup. I think today, it's a lot easier and cheaper to go with stepper motors, but then I haven't checked the cost of apropriate feedback systems, and suitable electronics, for a cnc system. My own project, when i get to it, will be to adapt some satelite positioner motor/gearboxes to the mill, and control by push buttons and adjustable speed controls, for a 'powered manual' system.

[This message has been edited by darryl (edited 05-19-2003).]

05-19-2003, 12:17 AM
This is fine if you use a feedback system like servos use, otherwise how do you accurately control them? Aren't steppers the motors of choice? I realize that one miss step on a stepper could foul up a piece and without a feedback system you wouldn't know that the load caused a miss step. Do you all prefer the approach of servo's or of steppers?

05-19-2003, 06:32 AM
Since we are talking about using motors for what they aren't intended for I'll throw this out. Allelectronics.com sells a wiper motor for about $20.00. I've bought a couple of them for projects, and they work very well. They are rated at about 1 1/2 amps, and are suposedly from a saturn car.

On a recent trip to the scrap yard I aquired a wheel chair with what looks like two very nice motors, and no controller. I've thought of doing a couple of things with them. 1 - power feeds for a mill and 2- a robot. I received a schematic for a motor controller from a friend, but he hasn't built it so I don't know for sure if it works. Any of you out their used these motors, and built your own controller? I would like to be able to use PWM in conjunction with either a basic stamp or RC remote, if I do the robot. Thanks

05-19-2003, 06:59 AM
I have used steppers since the late seventies.. Actually I think servos with feedback are more easier on the screws and nuts. My bridgeport has steppers with no feedback other than pulses out.
It has lost no steps since the power has been stepped up to 7 amps. (that I am aware)
Which is better? oranges and grapes which do you like? steppers are easier to program I think.

05-19-2003, 09:12 AM
There was an article in PIM several years ago that showed a Ford wiper motor used as a feed for a mill drill. It included a schematic for the electronics. I made one for my RF-30 mill/drill and was very happy with it.

The article was in August,'94; author was John F. Kraemer


05-19-2003, 03:12 PM
I bought a 4-pack of car window motors $69 total for a fishing downrigger project (only used 2 for the project which left 2 for the shop). These are 12vdc, reversible, right angle drive and plenty beefy. It does have a funny 9-tooth ouput gear for the window rack. They are a little slow at 70rpm but gearing would take care of that.
search for WINMTR-248-4

On steppers; one thing I found out from the home brew cnc guys is that the way to prevent missed steps at higher rpm's is to up the voltage from 5 to 25 times rated but keep careful control on the current supply. Makes the steppers very stiff but responsive. Add a microstepping drive (like a gecko) and it gets very smooth less ratchety.

I was thinking of using a basic stamp or PIC to control the stepper. Add an encoder and just monitor it for change. If you send a pulse in and the stepper doesn't change position then it might be stuck. Send a few more and keep track on how many are missed. Once you have too many (your choice) then stop the system and throw an error signal, light, buzzer.
If you wanted to have an auto reverse then you need to get fancier. Figure out how far each step takes you, backlash comp and such so you can tell it how far to go so it knows how far to come back.
Or use a couple of travel switches like on a surface grinder. That might not be too bad for simple cuts and surfacing.

05-19-2003, 05:36 PM
Problem is getting a stepper with the same pulses as the encoder. I usually overdrove the encoder. I got a nice BS2 encoder program somewhere. (found it) Cut and paste to your free parralax editor. I got extra BS2's too.

'** Actual subroutine developed for Clarostat 128 ppr encoder **
' David Cofer, Service tech
' Po box 837
' Tunnel Hill, Ga 30755
'Note: Quadrature encoder works via 2 phases 90 degrees out of synch.
'_[^]_[^]_[^]_] phase A of encoder var anew read as 1 or 0
']_[^]_[^]_[^]_ phase b of encoder var bnew read as 1 or 0
' Picks word bits set by previous movements and present
'quadrature ttl logic to decode movement into forward or backward motion
'Aold or bold=previous logic state stored for next condition
'anew=Phase A of encoder logic input, bnew=Phase B of encoder logic

'CCw logic here
'(1)if anew+NOT_bnew+Not_aold+Not_bold then ccw+1 (0000)
'(2)if anew+bnew+Aold+Not_bold then ccw+1 (1110)
'(3)if Not_anew +bnew+aold+bold=ccw+1 (0101)
'(4)if Not_anew +Not_bnew +Not_aold+bold =ccw+1 (0001)

'Cw logic here
'(5) if Not_anew+bnew+Not_aold+Not_bold=CW+1 (0100)
'(6) if anew+bnew+Not_aold+bold=cw+1 (1101)
'(7) if anew+Not_bnew+aold+bold=cw+1 (1011)
'(8) if Not_anew+Not_bnew+aold+Not_bold=cw+1 (0010)
'^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ ^^^^^^^^^
inq var in5
pha var in6 ' phase a of encoder on input 6
phb var in7 ' ' ' b input 7
enc var byte
anew var enc.bit0 ' new input bit for phase a
bnew var enc.bit1 ' new input bit for phase b
aold var enc.bit2 ' old input bit for phase a
bold var enc.bit3 ' old input bit for phase b
loc var word ' real world digital Location
begin:'******************************************* ******
if inq=1 then resp
if pha=aold and phb=bold then begin
gosub increment
goto begin
'****************** Conditional statements ***************
if enc=8 then upcnt '0001
if enc=14 then upcnt '0111
if enc=7 then upcnt '1110
if enc=1 then upcnt '1000

if enc=4 then downcnt '0010
if enc=13 then downcnt '1011
if enc=11 then downcnt '1101
if enc=2 then downcnt '0100
'************** subroutines to increment cnt ******
loc = loc + 1 'increment +1
'debug dec loc,cr
aold = anew
bold = bnew
debug ishex loc,cr
goto enco
loc=loc - 1 'increment -1
'debug dec loc,cr
aold = anew
bold = bnew
debug ishex loc,cr
goto enco
'****************** end of subroutine Encoder *************
goto begin

05-19-2003, 08:10 PM
Actually you don't need much just to detect movement. Unless you're moving to a sudo cnc type of solution you don't need the up/down counts or quad decoding. If you're using a 200 ppr stepper (the most common) then you only need a disk with 100 slits (200 transitions). That could be printed on any laser printer. The slit should change state on each pulse. If it doesn't then the motor didn't turn.

Or use any encoder that is higher than the stepper's ppr and as long as the count is different after the step pulse as from before, regardless of how different, you're ok.

You only need the quad decoding of you are implementing a true pid error correction and positioning algorythym and the stamp doesn't have enough juice to run that at speed. BS's are only giving 5,000 to 10,000 ips(instructions per second) where as the slowest PIC (4mhz) is doing 1,000,000ips. A 20Mhz PIC does 5,000,000ips (divide clock by 4). The basic interpreter has a lot of overhead.

Of course I'm just talking about detecting a stuck motor on a plain, single axis semi-manual power feed. Nothing more complex then that.
So on my mini-mill I have 20tpi leads.
To move at 6ipm I need 120rpm (6/0.05).
A 200ppr stepper uses 24,000ppm to do that.
Which gives 400 pulses to send, detect and deal with every second. A BS could do that. But generating a PWM waveform uses a lot of cycles too.

What would be real nice is to be able to tell the controller what you lead is, stepper rate and such and then to just tell it what ipm to move at instead of twisting a knob to something that just seems right.

Tie that back in to a tach/speed controller that you can tell - I'm cutting 12L14 steel with a 2" fly mill at 500rpm and I want a finish cut of 200 cut's per inch and let it go. A dos box starts looking good now. It can handle all that with no problems. And programing is easy. Add a cheap surplus LCD touch screen and that's a cool toy.