Stepper driven spindle

[Kludge]

Has anyone tried driving a lathe or mill spindle using a stepper motor? I know they turn in jerky little steps but I wonder if it would matter if the step size presented to the spindle itself were minimized via reduction (pully or gear).

Just an idle thought while I sort out tools ...

Best regards,

Kludge ... who probably thinks too much.

 

[ksouers]

I don't know of anyone driving a lathe with stepper, but they use steppers to drive the platters on computer hard drives. The fastest I've seen is 20,000 RPM. The need the stepper for super accurate speed control, so you would get the same control on a lathe.

Come to think about it, maybe CNC machines are driven by steppers.

 

[Mcgyver]

sorry guys, this one is headed for the scrap heap instead of the patent office

steppers have maximum power standing still, that is they can withstand the greatest torque load. the weaken as the speed up and almost no torque at higher speeds.

don't know of anyone driving a lathe with stepper, but they use steppers to drive the platters on computer hard drives.

they don't use steppers to drive the platters, they use steppers to control the read arms/heads.

 

[ksouers]

Sorry to disagree, Mcgyver, but yes they do use steppers to drive the platters on hard drives. The heads are located using a voice coil and opposing magnets. I've taken too many apart for salvage. A convenient source for Torlon bearings if you need any. The steppers are a two-phase design rather than the four-phase commonly used in industry.

Much older hard drives used to use steppers to locate the heads, but that hasn't been done for 20 years. Floppies still use steppers to locate the heads.

I've never seen a torque curve for a stepper so I can't argue there, but there are some quite powerful ones available for driving leadscrews on machinery, though I admit that's at very low RPM.

 

[BobWarfield]

It's close to that, Mcgyver, but not quite so catastrophic. What I mean is there is useful torque at greater than 0 rpm, but it falls off quickly. You can get the torque curves for them. Here is one such example:

http://www.kelinginc.net/KL23H276-30-8AT.pdf

A typical stepper is 200 steps per revolution. That chart shows half steps, so we'll say 400 half steps per revolution. Now let's take a 1700 rpm motor (forget 3400!). On that chart we're at:

(1700 rpm / 60 secs/min) = 28.3 revs/sec * 400 steps/rev = 11,333 pulses per second

Referring to our chart we're at circa 0.3/1.5 Nm = 20% of peak holding torque. Bummer!

Of course our regular motors lose torque rapidly (without a vector drive or other aid) when we slow them down so we have the reverse problem.

The solution to the regular motor problem is some form of gear train: step pulleys, gear box or whatever. The same will work with a stepper. Can you achieve acceptible performance in the rpm range you require? Perhaps.

Why would you care? Well, at the risk of continuing to be long winded, it's pretty handy to be able to index or otherwise precisely control your spindle speed electronically. At the low end, this is handy for threading, for example, where you can thread with an electronic gear box. But that's easier done in other ways. See for example the Electronic Leadscrew Yahoo Group (started by a crazy discussion I launched on a lot of boards a long time ago, but you can actually buy/build one and it's cool).

Alternatively, if you are into CNC, as I am, it is pretty cool. If I can index the spindle under CNC control, I can stick my chuck with an air spindle on the cross slide, but a drill bit in it, and drill a bolt circle in the lathe as part of making a part.

Hmmm. But how do I get the best of multiple worlds?

Well, I can use 2 motors. I can use a regular motor for turning, and a second stepper to index the spindle. I may want to use a disc brake to lock the spindle while indexing as well. Similar things can be done to automate rotary tables and indexers too, for example.

How do the big boys do it? It's quite common (almost mandatory in fact) to use a very large servo to drive your spindle. I've mentioned at least one cool thing that CNC lathes do with that (bolt circles) and there are many more. Another is polygonal turning. If you've never seen that, it is mind boggling. If we rotate a cutter in precise synchrony to the spindle on the lathe, we can "turn" polygons. I can make round stock into a hexagon, for example. Sorta like a rotary broach, but different. Pure vodoo magic!

Here is an animation of polygon turning: http://video.google.com/videoplay?docid=-7831550688320827327

As you can see, synchronization is essential!

Mills do cool things with it too. Thread milling is one, where the spinning cutter follows a spiral path just right to cut thread, just like on a lathe. On a simpler vein, most CNC tapers have "ears" on the taper. When the tool changer engages, the spindle has to index the ears or the toolholder won't go in the spindle. Having the spindle itself as a servo that you can index to any desired position is a happy thing.

So, being able to control more about a spindle motor than just it's speed is a good thing. In other words, precisely controling its position at all times. Steppers can do this, but they don't produce much power at very high rpm. But, we can also do this with servos for a little more money. And yes, it would be valuable to understand that rotational position on a hard drive too.

So there you have it. Much more than you cared to hear on the topic, I am sure.

Cheers,

BW

 

[Mcgyver]

well if you're going to tell me you're the chief engineer at Seagate i guess i have to believe you, but doesn't make sense based on my knowledge about steppers

i know the newer drives don't have them but thought maybe that was where the notion of the stepper came from - because they were in the older drives. The spindle motors on the ones I've had apart don't act like any stepper I've worked with - they have the definitive steps as you rotate it, with the hard drives I've had apart, there's no increment, it moves smooth as silk. seem like well balance dc motor to me likely with a control loop - circuit reads patter and tells motor to speed up or slow down. If it was some newfangled stepper that doesn't have steps, you're talking a 4MHz sampling rate if it was a 200 step unit, seems like a lot to keep up with and I don't know that a stepper that can go 20k (of course that doesn't mean there isn't such a thing) Steppers, as I've known them, also don't provide smooth uniform motion, there can be a lot of noise generated that you'd think would exacerbated at those speeds. What about them tells you or makes you thing they are steppers - I don't have enough expertise in this area to dig in my heels but it seems inconsistent with what i do know and my experience working with steppers

The torque thing is true regardless of the motors size, spin up a 1200 once stepper and torque drops off dramatically at max speed, which isn't very high for these steppers. another advantage of servos which is what the commercial and higher end stuff seems to use

Bob, strikes me there would limited applications for indexing the spindle of a machine tool with a stepper. indexing a workhead is entirely different than a machine tool spindle. Thread milling needs three axis not an indexed spindle and as i said a stepper is a poor choice because of the lack of torque at higher rpms.....obviously it has torque above zero. However it drops of at high rpms and is considered by most as effective at low rpms only, not a characteristic you want for a spindle. Regardless, top speed for a stepper is not all the high to begin with and so bloody noisy it would want to shake the machine apart. you don't have to index the spindle as in 'go exactly here' for most of the ops you mention, you just have to know where it is and it has to be to much much finer resolution than a stepper, hence encoders. Subtle difference granted, but for example when threading on a lathe you would actually have to move the spindle any precise degree of rotation; just spin it and know its exact angular position. btw a servo basically is just a DC motor and a rotary encoder.

In instances where you have to know where the spindle is, like threading in a lathe, or tapping, you will need 10x or more your 200 steps per rev for decent results. gearing down doesnt work as slow speed is already a problem to the notion of driving a spindle with a stepper. So we do this via encoders.... as a diy situation that has always been the challenge, coping with the sample rate. I don't know what cool means in the context of indexing a spindle via stepper; it has function or not and i don't get where there's a compelling function but there are lots of negatives around speed and torque

 

[Kludge]

Oi, vey! This isn't quite what I had in mind!

I know the torque dies off at higher speeds. I also know that steppers are cheap in a range more than adequate to power some of my smaller equipment at their upper speed ranges. The reason I asked is two fold: to find out if the fact that they take little jumps vs smooth rotation going to be a problem, and to see if the fact that they are steppers can add anything to their functionality - indexing being one application I had in mind. The second part has been answered but the first part not so much.

Okay, let me back up a bit. I've been thinking about building a mini-CNC lathe/horizontal mill. Right now, there are several projects going on in making CNC versions of the Sieg C0 and I think there's a CNC Unimat DB/SL or seven floating around. I've never seen a CNC watchmaker's lathe but that doesn't mean none exist. My next major purchase will be a Seig which will eventually be converted to CNC. (Accessories are available to make it useful as a horizontal mill which makes life so much more easier.)

But (huge word), the Clisby has certain ... ummm ... "merits" that making something of its size ideal for experimentation. It doesn't have to be fancy (although reasonable repeatability would be nice) since all it would be is a test bed for some of my less sane ideas. BW noted elsewhere the fact that "Kludge" is also a verb and this is one place it certainly would apply. It would be built of whatever I have laying about and what looks good on ePay - bed/milling column from 8020 Inc Garage Sale, bearings from some other source, etc. I may even have a spare cross slide from a watchmaker's lathe sitting in one of the thus far unchecked cigar boxes not doing a lot.

So far as I can tell, software doesn't care what size the steppers it's driving are aside from step size so turning small ones works just as well as turning large ones. If I can find a package that also knows about controlling the spindle, this is even more so gooderer since I can use software rather than gears for threading and other synchronized operations. The assistance of a willing (or maybe not so willing, dependent on circumstances - I have no objection to using slave labor on occasion ;D) software 'spurt will likely be needed to create the kludges to make it work but it'll be one of those fun and exciting one-off packages that software weenies love. (I was a hardware weenie when I was still playing with computers and coded only out of self-defense or when I'd run out of other options. You can't kick software when it's not working right. On the other hand, that's one reason I loved the National INS8073 SC/MP chip so much; it had NIBL (National Industrial Basic Language) built in on ROM which was designed for process control applications.)

Ksouers: - I thought the spindle motors were 2 phase brushless motors. Oh, well. I have now proven I can be wrong.

KB & McGyver: This is just for playing with on a small scale. For a larger machine, I'd certainly go with a servo motor instead. As it is, I'm thinking about a 3:1 or so step up via pulleys or gears for higher spindle speeds which will kill torque anyway but keep the steppers in their happy range.

Everyone: Keep on posting. I'm learning a lot of things I never knew that I never knew!

Best regards,

kludge

 

[Mcgyver]

If I can find a package that also knows about controlling the spindle, this is even more so gooderer since I can use software rather than gears for threading and other synchronized operations

kludge, is the spindle you'd put a stepper on or a workhead, like a 4th axis? I'm thinking doing so to the spindle will do little good as it wouldn't give enough resoultion for what you described, gear and threads and such. I think you need several thousands of divs per rev to make a workable thread and you have the inheriant problem of not much speed and falling torque. When you start to calculate the signa frequency thats sent back to computer, it becomes problematic - iirc the common software running on a pc starts to poop out at 25kHz??. Stepping an rotary axis opens doors, many years ago there was some great diy stuff on this in MEW and John Stevenson who posts here has electronically couple an encoder off a spindle to a workhead such that he can hob gears without a mechanical connection between hob and blank.

I don;t think the motion is smooth, that is based on the noise you get (you can play a jingle by tunning frequencies) but I suppose to be scientific one would have to pull data from a linear encoder to see. it certainly feels and sounds more like incremental action than the smooth action you get with pwm on a dc motor.

hey, I hope you prove me wrong and open new doors but in addressing the Q of using a stepper on a machine tool spindle those are the reasons i don't think it do much good

 

[Kludge]

kludge, is the spindle you'd put a stepper on or a workhead, like a 4th axis?

Workhead for sure but the spindle just to see what happens. This is a pretty small machine - Clisby sized (2-1/2"x4-1/2" give or take a skosh) - so the reduced torque wouldn't be a headache. I'm not locked into it obviously and can certainly cobble up a PWM controller for a DC motor. I just happen to have some bipolar steppers put away (and a few unipolar, I think) but I also have some DC motors of a variety of flavors in yet another cigar box or two (I really do need to mark what's in each one some day! ) so my options are fairly open. Come to think of it, I think there may be a fairly torquey brushless motor or two in there as well.

I can think of a few ways to get feed back for spindle speed (spindle speed != motor speed with any speed changes via gearing or belts) and a 0 reference point for position so they're non-issues.

When you start to calculate the signa frequency thats sent back to computer, it becomes problematic - iirc the common software running on a pc starts to poop out at 25kHz??.

Actually, I was thinking about the idea of an external box - a PIC, for example - to keep them synchronized. This complicates things, especially since I have no experience whatsoever with any of the microcontrollers, but it should work. I think. Maybe. In any event, it would kick the PC out of the loop so all the PC has to do is tell it what it wants then lets the controller figure out how to do it on its own. The PC is still monitoring the overall results but not directly controlling how it happens unless something goes oops.

Keeping in mind I have absolutely no experience with PICs, I have noticed them appearing as cheap solutions to life's little everyday problems and having multiples of them on a USB or other buss seems to be a popular way to do things of late. One thing I need to find is a kit to help me learn how to program and use them - something using one form or another of BASIC would be good for now.

hey, I hope you prove me wrong and open new doors but in addressing the Q of using a stepper on a machine tool spindle those are the reasons i don't think it do much good

No worries. I'm not married to it. Heck, we're not even to the heavy petting stage! I was just thinking about it while I was sorting out tools.

Best regards,

Kludge

 

[Mcgyver]

Workhead for sure but the spindle just to see what happens. This is a pretty small machine - Clisby sized (2-1/2"x4-1/2" give or take a skosh) - Actually, I was thinking about the idea of an external box - a PIC, for example - to keep them synchronized. This complicates things, especially since I have no experience whatsoever with any of the microcontrollers, but it should work.

as i recall it JS's coupling spindle and workaxis doesn't use a computer, he's counting pulses on the spindle and synch the rotating axis with the gear blank. I don't know if he's using a microcontroller, hopefully he shows up on this

Keeping in mind I have absolutely no experience with PICs

what a great excuse to delve into a whole new hobby ;D

 

[Kludge]

as i recall it JS's coupling spindle and workaxis doesn't use a computer, he's counting pulses on the spindle and synch the rotating axis with the gear blank.

Hmmm ... Gotta think upon that one a bit. When you mentioned a 4th axis, was it rotational?

I don't know if he's using a microcontroller, hopefully he shows up on this

Well, he'd better or I shall be forced to pout. ;D

what a great excuse to delve into a whole new hobby ;D

Oh, yeah. Not that I need another one but it seems I'm kind of getting nudged in that direction. Guess I've got to find something slightly smaller than my solder pot and copper to build things. ;D

I was thinking about something. (Oh, NO! Not again! ) If I'm getting this all right, a servo is basically a reasonably well behaved motor with feedback to control its rotation. This should be fairly easy with DC - monitor the spindle speed and adjust the input to a PWM controller as needed. It might even be able to handle speed variations due to load within practical limits. Is this a fair assessment? (The only info I've found on servoes on the web relates to R/C servoes and that's an entirely different ball of well used wrapping paper.)

Defining "well behaved" will take a bit since I'm not sure I understand all the demands/requirements of such a motor. I would think that decent responsiveness to speed changes, the ability to hold a constant speed once it's set and useable torque through the speed range would be a good start. After that, I'm losted. (But at least it's familiar territory. )

So basically, a stepper driven spindle's pretty much a loser but maybe I should consider a servo loop instead. Is that closer?

Best regards,

kludge the Konfuzzled

 

[Mcgyver]

Hmmm ... Gotta think upon that one a bit. When you mentioned a 4th axis, was it rotational?

Yes, it was a hobbing op, mill spindle driving an encoder then the work was mounted on a four axis that sync'd itself to the spindle axis. The encoder was ratio'd up via a belt to give high(er) resolution.....because its cutting gears, the 4axis speed is a fraction of the spindle's, ie for a 30 tooth its 30:1 so you can get away with a lower speed on the rotational axis. This is just what i supposed from looking at his set up, he can of course give more exacting info...steppers can work great for low speed applications like linear or rotational axis, its just powering a machine tool's spindle itself that didn't strike me as having much promise.

If I'm getting this all right, a servo is basically a reasonably well behaved motor with feedback to control its rotation. This should be fairly easy with DC - monitor the spindle speed and adjust the input to a PWM controller as needed. It might even be able to handle speed variations due to load within practical limits. Is this a fair assessment?

I can't speak for all of mankind on the nomenclature on this, but servos, at least in the context of cnc mean a bit more. What you described is eminently doable and done all the time, feedback on the motor's rpm controls the pwm so that the rpm stays regardless of load. To me that is not a servo as the word is used in cnc discussions. A servo is the same idea but with a encoder attached on the spindle - that tells a lot more than what the speed is. You can tell the speed simply by counting revolutions but an encoder tells where the spindle in parts of a revolution - it outputs so many pulses per rev. clear as mud?

well behaved. you want the spindle to stay a constant speed under varying loads and a feedback loop will do this, but it is my understand that incredible accuracy here is not required; that it is not the ability to perfectly control spindle speed that is the basis of say being able to thread accurately. Instead its the ability to very accurately encode the spindle, like several thousand increments per rev, and then move the carriage to suit. i suppose the inertia of the carriage is less and its easier to provide direct force than changing a rotating speed of a high speed heavy spindle.

 

[Kludge]

Yes, it was a hobbing op, mill spindle driving an encoder then the work was mounted on a four axis that sync'd itself to the spindle axis.

Okay, I'm having a problem visualizing it this way. I can see a rotary table on a milling adapter with its own drive so each tooth gets cut individually but the way you're describing it got me crosseyed. I think I'd need to see a picture or video of it to understand.

RE: Servo loop description ...

I can't speak for all of mankind on the nomenclature on this, but servos, at least in the context of cnc mean a bit more. What you described is eminently doable and done all the time, feedback on the motor's rpm controls the pwm so that the rpm stays regardless of load.

Okay, good. I've gotten that far in my lessons.

A servo is the same idea but with a encoder attached on the spindle - that tells a lot more than what the speed is. You can tell the speed simply by counting revolutions but an encoder tells where the spindle in parts of a revolution - it outputs so many pulses per rev.

I've seen them in other applications - basic pulse counters for each revolution. Actually, a mouse ball sensor did that and a track ball I think works the same way. But I would think soft ware that could count pulses like that would also be able to use the speed sensor and do the math to get time/revolution then be able to chop that into convenient slices for whatever the application is. On another appendage to be identified at another time, software takes time and hardware doesn't which gives a definite advantage to hardware solutions.

clear as mud?

Why, yes. The mud is perfectly clear.

you want the spindle to stay a constant speed under varying loads and a feedback loop will do this, but it is my understand that incredible accuracy here is not required; that it is not the ability to perfectly control spindle speed that is the basis of say being able to thread accurately.

I don't think perfect speed can be maintained outside some high falutin' shops that do government work and can afford all the fancy toys.

Instead its the ability to very accurately encode the spindle, like several thousand increments per rev, and then move the carriage to suit.

How about 256/revolution and interpolate between them? Getting to the electronic side for a moment, it's not that hard to synthsize however many divisions you want using a smaller number of real ones. Actually, on reflection, this could be done with a tach input as well. You can also take that 256 pulses/revolution and create 360 or 720 or 1440 or whatever pulses/revolution; it doesn't have to be an even multiple.

i suppose the inertia of the carriage is less and its easier to provide direct force than changing a rotating speed of a high speed heavy spindle.

The carriage doesn't have to move anywhere near as fast. With the newer 1.8 and .9 degree steppers complete with half step (and smaller?) capability, the stepper speed (in steps/second) has to increase to match the movement of older 3.6, 7.5 and 15 degree steppers which loses torque in the process. It's an interesting tradeoff - precise positioning vs torque.

I think we've reached a danger point, that is I'm dangerously close to Learning New Things.

Best regards,

Kludge

 

[ksouers]

... What about them tells you or makes you thing they are steppers - I don't have enough expertise in this area to dig in my heels but it seems inconsistent with what i do know and my experience working with steppers

Mcgyver,
The motors that I've pulled out of HDs have multiple leads going to them, usually 4 though I've seen more. When energizing any two leads the rotor indexes in discrete steps, it does not continue to rotate. By energizing the leads in turn you can step the rotor around in a full circle. I can also feel the steps when rotating by hand, granted it's very weak. The mass of the platters can easily overcome the steps, but then the motor is only designed to spin the platters, not move machinery.

Several years ago there were a few garage/hobbyist based businesses selling driver boards for these motors, don't know if any of them are still around. Lately the motors are being built integral to the drive frame, not a separate removable unit as they used to be.

I got interested in them as some experimenters were using them as micro alternators powered by windmills to keep batteries topped up on their sailboats. I thought I'd give it a try.

In all cases they were referred to as "steppers". It seemed logical and accurate. With the faster drives coming onto market the makers had to have better speed control than just a feedback loop. A stepper would seem to fit the bill.

I only mentioned it for the speed control. I thought that aspect might be useful.

No, I'm not an engineer. And I admit my knowledge of motors, and electronics in general, is limited. But the motors act like steppers and others have referred to them as such. I have no reason to believe otherwise.

Ksouers: - I thought the spindle motors were 2 phase brushless motors. Oh, well. I have now proven I can be wrong.

Don't know. Can they act like steppers? I could be wrong as well. Won't be the first time.



If it looks like a duck, quacks like a duck, it must be a duck. Or maybe it isn't...

 

[Kludge]

Quote from: Kludge
Ksouers: - I thought the spindle motors were 2 phase brushless motors. Oh, well. I have now proven I can be wrong.

Don't know. Can they act like steppers? I could be wrong as well. Won't be the first time.

Not to the best of my knowledge. A fair example of them is the motors used to turn computer fans - not a hairly lot of power but they have the same number of windings.

By the way, two windings kind of sounds like a bipolar rather than the "got lots of leads" unipolar steppers more often found in CNC apps. Bipolar is easier to drive but I can't remember what the step sizes usually found are. For an application like that, I can't see having an ultrafine step size would be important since there is a fairly decent flywheel attached. Hmmm ... a point to ponder upon ...

If it looks like a duck, quacks like a duck, it must be a duck. Or maybe it isn't...

Depends ... was is specified by the DoD? :big:

Best regards,

Kludge

 

[Mcgyver]

Okay, I'm having a problem visualizing it this way. I can see a rotary table on a milling adapter with its own drive so each tooth gets cut individually but the way you're describing it got me crosseyed. I think I'd need to see a picture or video of it to understand.

let me see if i can get the head spinning at the same time

this'll be tough if you're new to gears and how they are generated....I'm not sure it can explain it well in a paragraph or two......Study a picture of a hob. think about the involute shape of a gear tooth and how is formed by an imaginary string unwinding form its surface. Now note the shape of the teeth on a rack - they have completely straight sides (just a gear of infinite dia, hence the line unwinding is infinitely long, ie straight). Well a gear hob is just taking that rack tooth form and wrapping it around a cylinder in a helix to make a cutting tool (gashing and relief etc is add, but I'm keeping it simple(ish).

Now mount the hob on a universal mill (a universal mill is a horizontal mill where the table can be angled) and set the table at the same angle as the helix. For the gear blank, whose axis is in line with the table and angled to the hob, when it looks at the hob see the straight sides like rack, not the helix. The spindle rotates and the gear blank rotates but its coordinated - for each rev of the spindle with the hob, the blank rotates an amount equal to 1/# of teeth. advance them into one another and a spur gear is generated. The motion is very similar to a worm in worm wheel, except instead of being stuck in one position, the worm is a cutting tool and moves across the blank creating a spur gear. Helical gears are the same idea but offset of the table is not at the same as helix's

i bet that is really really hard to follow without some pics, tried to find some good imagery, not luck yet,

Kevin, interesting. I don't know enough off hand as to accurately define what makes a stepper a stepper as opposed an alternate coil arrangement that exhibit those characteristics - it didn't it have holding power when energized like that? One characteristic i remember about steppers is that their maximum torque holding is at a standstill, ie should have more holding power that rotational power, but alas I've exceeded my knowledge on the inside of disk drive motors

 

[Kludge]

The motors that I've pulled out of HDs have multiple leads going to them, usually 4 though I've seen more.

Been thinking upon this. Have you done continuity checks between the wires? If you have two separate windings, it will show as infinite in two cases and whatever the winding resistance is in two other cases. On the other hand, a standard Y-wound brushless motor will show continuity no matter what with the center of the Y showing the lowest resistance to the remaining 3 wires. There are other configurations using more wires but the Y-wound with no sensors is the easiest (hence cheapest) to make.

Just a thought on a hot day in Hawaii ...

best regards,

Kludge

 

[ksouers]

By the way, two windings kind of sounds like a bipolar ...

Yep, that's what they called them! I couldn't remember exactly, though I'm sure some of the guys were rather bipolar themselves.

The step was quite large, somewhere around 30 degrees or so. I had a brief moment some years back when I thought about adding CNC to my Sherline and thought those motors with some kind of reduction drive might fit the bill. Then I got back on my meds.

Mcgyver,
It had plenty of holding power when energized, surprisingly. Took a relatively hefty twist to knock it off it's step. Had a few minutes of fun trying to make it spin around just by pulsing the electric.

 

[ksouers]

Been thinking upon this. Have you done continuity checks between the wires? If you have two separate windings, it will show as infinite in two cases and whatever the winding resistance is in two other cases. On the other hand, a standard Y-wound brushless motor will show continuity no matter what with the center of the Y showing the lowest resistance to the remaining 3 wires. There are other configurations using more wires but the Y-wound with no sensors is the easiest (hence cheapest) to make.

Just a thought on a hot day in Hawaii ...

best regards,

Kludge

I did the continuity checks at the time but I don't recall what the outcome was. This was back around 2003 and I'm at the age where I can barely remember breakfast. Sorry.

Now I just salvage the bearings out of dead drives and toss the frames in the scrap bin for castings.

 

[Kludge]

let me see if i can get the head spinning at the same time

Gee, now THERE is starting torque. ;D

this'll be tough if you're new to gears and how they are generated.

Just new to doing it this way. Cutting them the old fashioned way using an index and a single disk gear cutter is familiar. But that's from how it was done in watchmaking and how I've done it (with sometimes less than ideal results - ie, spurs should not wind up with 1/2 teeth anywhere), not with the big boy toys.

Study a picture of a hob.
...
Well a gear hob is just taking that rack tooth form and wrapping it around a cylinder in a helix to make a cutting tool (gashing and relief etc is add, but I'm keeping it simple(ish).

After reading this and taking a pass at Wikipedia (a good source for quick knowledge but not in any real depth) I got it! (Miracle do happen!)

Somewhere in my sinful past, I remember seeing how to use a common tap (or an uncommmon one if you prefer) to create both a worm and a worm gear.

In the case of the worm, the blank was turned while the tap was allowed to ride against it cutting a single start worm down the length. The process was repeated until the desired depth was reached. Oh, yeah. The cross slide to which the tap was firmly attached was free of the lead screw so it was pretty much on its own to cut the worm.

The worm gear (And I think this would work with a helical gear as well though I haven't really thought about it much) was formed by turning the tap in the headstock with the gear blank held to it and allowed to rotate freely so the tap could form the teeth as it turned. Again, it was advanced into the tap until the desired depth was reached.

In any case, it looks like I've gotta get me some of those hob things 'cuz cutting one tooth at a time for a spur is a serious drag.

Best regards,

Kludge