Kevin - A V Carroll Horizontal Mill Rebuild

[durableoreo]

(Some of this was posted on Practical Machinist but I eventually figured out that those guys are more serious than I am. So updates will only be posted here.)

Purchased a small horizontal mill recently. It's labeled "A V Carrol" and . I'm getting it ready to make chips. It was fairly cheap but it came with quite a few issues, including lever-feed on the x- and z-axis. There's still flaking on the ways, which is nice to see. The x-axis is a bit worn but I can live with it.

Decided to get the arbor off. Partly to figure out how it all goes together, partly because I want to make another arbor for 1-1/4 cutters. Eventually I figured out that there's a drawbar that was bottomed out in the threads of the arbor but there was still about 1/16 end play. I'll add another spacer when I put it all back together.

Here's the starting point, the mill in all its dirty factory-use glory. It's one of those small single-operation type mills for small parts, so the table is about counter height. It's a lot smaller than I expected, which is perfect because the shop space is about 50 sq-ft and there's already a 9x20 lathe and a drill press in there.

Here's both sides views of the spindle.

Closer view of the working end of the arbor. Note the keyway is not lined up with the set screw hole because I was messing with it. There's a matching flat on the arbor. The bearing is mounted in this little cylindrical mount/enclosure. It's a tapered roller bearing that's tensioned from the back, so the bearing assembly is sealed by that cover but the cover is held on with t-slot nuts---the screws sticking out aren't studs going back to the main casting. The bearing enclosure bears against the main casting and is connected with grub screws. I might add some 1/8 pins because there's quite a lot of play in the shaft, even after pre-loading the bearings.

Closer view of back end of arbor. The arbor has some damage from the set screw but it's a fairly thin-walled tube at this point so it's more of a dent, no burrs that I could feel with a stone. The handwheel came off without trouble.

Not sure about that groove in the drawbar. On the right side, you can see the driven axis, which is a threaded 1-1/8 tube. There's a tapered roller bearing in there and the big lump there (first 2 steps) is a retaining/tensioning nut with fairly fine threads with a heavy bushing (3rd step) that transfers compression from the nut to the inner cone of the bearing . When I removed all those parts, the arbor still wouldn't come out.

Here are images of the make/model. I haven't been able to find information about anything other than lathes that were produced by A V Carroll. I could have sworn there was another thread I started back when I was in rigging mode... Anyway, someone pointed out that machinery companies sometimes re-sell models that they don't directly produce, for business reasons. I noticed that the base is fairly generic and the only other place the manufacturer's name appears is a removable cover---easy to re-brand. So if anyone recognizes the castings, I'd be glad to know more about the manufacturer. It would be so great if I could find another arbor.

The plan... It's too much, as usual. But I'll start with the most important things.

First, I need to re-power the mill. It's running 1:1 with a 1750-RPM induction motor, which is 20-30 times too fast. With a 3" HSS cutter, I need 25-100 RPM to hit reasonable feed rates (20-80 SFM for annealed low-carbon steel, according to my non-machinist reading of the 24th edition handbook). I might use some carbide cutters, so I knew I would need a high and low range. I would love to use a variable-drive type system but I just don't have the chops for that yet. So I'm falling back on a variable speed electric motor. I made a spreadsheet and figured I would need a 1:5 and a 1:10 reduction to get in the ballpark. Those ratios are pretty large and on the verge of impractical for v-belts, especially with the space constraints. Eventually, mostly by luck, I figured out that a 1:15 gearbox and a pair of pulley ratios (1:2.5 and 2.5:1 would be good. Also, I want the 1:15, so I can run the motor at half speed or more. I should be able to do it with 1 belt, switching it between 3 sets of pulleys. I tried to find ready-made stuff but for 1-1/8 bore with keyway, finding a step pulley with 5", 2" and whatever 1:1 is for the same belt length, that's about impossible. Plus, a 5" pulley may not quite fit. It's hard to know because the casting is rough and I don't know how to measure it properly. I think I'll buy blanks at McMaster, bore 1 of them, and do a trial fit. If I need to make it smaller to fit, I can adjust the 2" pulley smaller, too. For 3L belts, a 2" pulley (or a little less) should be OK. I have a 1-1/2 pulley for 3L sitting on my desk. It's extreme but possible. So that's the plan. Got a planetary gearbox on the way from Apex Dynamics (AB115) and materials for the pulleys and idlers will be here in a few days.

For the motor, I'll start with a sewing machine "servo" which is cheap and maybe works. There's a maximum speed dial (350-3500 RPM) and there's a pedal input with brake. No specs, though. I took it apart to figure out what was going on. The motor is brushed. Sewing machines need good torque at low speed so it's probably a series-wound dc motor. There's different speed control methods and I have no idea what it has. There is a switch to change direction and the throttle (pedal input) is a hall sensor. I did some stall-speed torque testing to convince myself that this wasn't a fools errand. I noticed that when the maximum speed was set low, torque was low. So I think I'll set it on maximum and use the throttle to set the speed. I'd like to mount the various controls remotely so I can operate the machine from the front and with a minimum of fuss.

After the mill is re-powered, I'll need to address the x- and z-axes, which were lever operated. It's a good system but a rack-and-pinion situation won't work for me. I want power x feed and finer control of z motion. The x-axis will be fairly easy to switch over to a 1/2-10 acme screw. The z-axis... It's a little harder to imagine. There are features on the castings (pan and knee) that seem perfect for installing a lead screw. So maybe it won't be so hard. Haha. Who am I kidding?

 

[NortonDommi]

That is a very solid looking machine. If you are going to change pulleys have you thought about multi-V ones? Lower profile than V-belts, can transmit more torque, run quieter, they can also be used on smaller pulleys than V-belts allowing a great range of ratio in tight quarters.
Beauty of a rebuild with mods is you can do whatever you want within reason. A gearbox is a great idea! What sort of H.P. are we looking at here?

 

[durableoreo]

That is a very solid looking machine. If you are going to change pulleys have you thought about multi-V ones? Lower profile than V-belts, can transmit more torque, run quieter, they can also be used on smaller pulleys than V-belts allowing a great range of ratio in tight quarters.
Beauty of a rebuild with mods is you can do whatever you want within reason. A gearbox is a great idea! What sort of H.P. are we looking at here?

The multi-groove belts are a good idea. Let me have a look... I might need the low profile and tighter bend radius.

The mill came with a 1/3 HP motor. The motor I'm going to try is rated for 550 W, which is about 3/4 "HP". Not sure if that is an honest number. Barker mills (they still make them) have a 1/3 HP series (PM) and a 2 HP series (AM). I suspect the castings are the same but they put a bigger motor in them depending on the type of work that is to be done.

The gearbox was about 200 $ so I hope it works out!

 

[DiscoDan]

I have a very similar small horizontal Mill. Mine is a Pratt & Whitney from World War II. Mine is powered by a variable speed Craftsman motor. Makes it very easy to set up. If you can find one

 

[durableoreo]

That looks great. You're right, that would make it easy! I've seen some with fancy gearhead motors, too.

I started looking at http://lathes.co.uk/ which has pictures of many historical mills. There are so many of these things... I haven't found anything that's quite the right shape

 

[durableoreo]

Here are some more details about the spindle speed. It's based on Machiner's Handbook numbers for upper and lower spindle speeds.

Here's what you get if you reduce a 3500-RPM motor 15:1 then step up or down by 2.5 to get a medium, high, and low range. I'm suspicious about using the motor at the lower RPMs because I suspect that torque will be lower. But with 3 speed ranges, I can get most speeds in the upper speed range as long as I stay in the box. I can experiment with the lowest ranges but I don't need it for normal mill operations. For really large cutters on a hard shaft or slotting, going below 50 RPM might be useful.

 

[durableoreo]

I also purchased a 3-axis DRO and RPM meter. It was 276 USD. Matching green, for vanity. I seriously considered buying sensors and hacking together an Arduino. Decided against it, in the end, because I'm trying to quit. It's satisfying to do things yourself but it all takes time. The side-to-side travel is the most, at 12" and about 3.25" in the other direction, towards or away from the operator. The vertical axis is 6". Probably could have gotten away with a 2-axis DRO and a caliper but it's almost easier to have a single system

Now, I'll need to figure out how to mount the display. I'll need another panel with various other controls, too, mounted near the DRO. I want it to be sturdy and not retreat from my fingers. I could make a 3-point mount, contacting each of the 3 corners of the table. Or mount it on the back corner so it swings wide and have a hook that catches on the front corner. I want it sturdy but not in the way when I'm changing belts. Normally screens are meant to be mounted so that the top of the screen is eye level. But this mill is fairly small so having the display that high might produce unnecessary moments and look/feel weird. Proportion may be more important than slavishly following ergonomics rules for desktop computers.

I need a a switch for the motor, a throttle for the motor, tachometer, and switches for auxiliary lighting and coolant pump. (There's no coolant plans at the moment but I might do that for the future, especially if I want to cut harder materials.) A panel below the DRO seems like a good place for those.

When changing belts, it's best to have 2 switches to prevent accidents. Ideally, unplugging the equipment is the solution. I'm fairly conscientious but I can't get myself to unplug machines when making adjustments. And in industrial equipment, it's not possible to disconnect equipment from the mains. Instead, interlocks and E-stop circuits prevent workers from accidentally engaging the equipment during maintenance. For this application, 2 switches in series seems like a good solution. A switch up by the DRO for normal use and a 2nd motor switch near the access panel for changing belts. My first thought was a toggle inside the main casting. But it would be convenient if it were activated by the removal of the cover... Maybe a reed switch. But the magnet probably can't be mounted directly on the cast iron cover---maybe use a bit of phenolic resin or wood as a spacer. And a reed switch can't carry enough current so I need a relay... Also, I want the motor power switch on the panel to turn off if there's a loss of power, so I need a latching switch there. Maybe the coolant pump needs a latching switch, too. This is getting complicated. I stopped to consider that I'm getting greedy with features. After some consideration, I decided that A DRO, tachometer, and basic safety interlock isn't too much to ask of a mill.

 

[durableoreo]

Well, I'm thinking about feeds. I want power feed on the x axis. Probably z, too, but the x-axis is more important because it will be one of the factors that effects finish. Below is a spreadsheet that addresses cutter RPM and the amount of bite from each tooth. It's based on Machiner's Handbook. I intend to run smaller endmills on occasion, so that's included in the table. It's tricky because feed rate depends on the size of the cut, which depends on materials, diameter of the cutter, number of teeth on the cutter, and the RPM of the spindle. I calculated feed (IPM) for combinations of maximum and minimum spindle speeds and max/min feed-per-tooth rates. There's a lot of room for variation so I plugged in diameters and tooth counts for cutters I have in-hand.

It's hard to get a good idea of what's needed so I plotted all the results. Looks like feeds can be as low as 0.2 to 275 IPM. That's 4 orders of magnitude and I can't afford another fancy gearbox. If I only consider HSS cutters with diameters between 2 and 4.5, the range is 0.2 to 42, which is more reasonable. Oh, I found the manual for the Cincinnati dial-type universal mills. They had feeds from 1/2--40 IPM, in a high and low range. OK, so if the standard in universal milling machines uses a range of 0.5-40 IPM, that's good enough for me. The bad news is that even this reduced range is still basically 2 orders of magnitude, or 80:1. For the spindle, we only needed about 40:1 plus variable speed motor, which is relatively easy. So what to do?

First I looked at commercial power feed units.

There's a unit that Little Machine Shop sells. It's about 400 $ US and I couldn't find any information about it's max/min speeds, torque, etc.

There's an eBay special, 100% Chinesium, which I tried to ignore.

Saw lots of DIY power feed solutions. Stefan Gotteswinter builds a nice one out of a windshield-wiper motor. It needs a clutch because of the worm-gear reduction, which I don't want to make or buy. With the gearbox, typical output would give me 6 in/min. Speed control might give me 0.5 in/min but I doubt the torque would be adequate. Well, those motors usually put out 30--50 ft-lb of torque, so it might be OK. With a 2:1 reduction in a pulley drive... But then you're limited to a maximum speed of 3 in/min, which isn't what I want, either.

MyfordBoy made a nice power feed system using a stepper motor that's rated for 425 in-oz. He runs it on a big vertical mill. I like this solution the best but I can't fit that motor on my tiny mill. There's simply nowhere to put it. I would need to make or buy a right-angle gearbox. Also, it's fairly heavy.

The eBay power-feed offering claims 150 ft-lb of torque and 0-200 RPM, which sounds good; that's 0-20 in/min. The problem is that I don't believe that the torque is enough at low speeds and I'm not sure that the system will run at 0.5 in/min. Then, by luck, I found a review on the Build Something Cool channel, (YouTube) where Dale tests out this cheap item compared to 4 others. The slowest this unit will go is 0.150" in 10 seconds, which is 0.9 in/min.

Maybe I can live with 0.9 in/min. If you look at the table, most of the less-than-unity speeds are in the Min/Min category. It account for about 12% of the cases. Honestly, I don't have the experience to say if these are important. I suspect that I will be giving up milling very hard materials---shafts that are so hard that I need slow spindle speed AND because of inadequate stiffness in my machine, I need to have slow feed. Or perhaps buy the right cutter, one with smaller diameter and fewer teeth.

There are 27% of cases that are too high to accommodate with the Ding-Wang unit. Almost all of these are in the Max/Max category for carbide tooling. That is, maximum spindle speed with the maximum feed rate. I do not anticipate needing this particular category because I'm limited in spindle power and rigidity.

So, it's a compromise but it only cost 125 $ US. It will arrive in a few days. Unfortunately it weighs 14 lb, which is a lot for my little mill. And it's a chunky monkey. I hope it doesn't end up looking too weird.

Here's the cases that the Wing-Ding unit covers, if connected 1:1 to the 1/2-10 leadscrew for the x-axis. They're in green.

What if I stepped it down a bit? As I reduce the lower speed, the upper speed is also reduced, so it's not clear if this will results in more or fewer cases being covered. I chose some ratios and counted the number of cases that are covered (feeds between upper and lower limit). As the ratio decreases, the number of cases covered goes down. If I step it up, the coverage is better but I lose all the Min/Min cases for HSS tools. So I think direct drive (1:1) is probably the way to go.

Finally, the most basic aspects of the mill are determined. If I had bought a Grizzly horizontal mill, this would all be taken care of. Sounds, nice, actually. But from the reviews, I've heard that the G0727 doesn't have low-enough speeds, so maybe it wouldn't be nice at all. I think they had to make some sacrifices in the cartridge spindle to allow larger cutters and smaller endmills.

 

[durableoreo]

Now, it's time to make pulleys. Turns out, the specifications are fairly simple. First, if a belt fits a pulley, it does not touch the root of the groove in the pulley. It can protrude beyond the outer rim of the pulley. For small pulleys, under 2-1/4 inches in diameter, included angle is 34 degrees and the width at full diameter should be 0.494 for A belts. This is from Machinery's Handbook, 24th Ed. As pulleys get larger, the angle increases so that pulleys over 4-1/4 are angled at 38 degrees. After reading through all this stuff, I'm impressed with the complexity of the mechanical world. I had previously though of belts as obvious and old tech that was simple. But if you get into the shape of the belt, finding the pitch line, the details of the pulleys, how the belts age and the effects on efficiency of the drive---it's more complex than I thought. And there's also flat belts, narrow v-belts, and ribbed v-belts to consider.

The casting around the spindle does not give me a lot of clearance. I considered switching from 3L (3/8 wide, a size below A) to a grooved belt. They don't have the wedging action of a v-belt so they can be run at a higher tension. Also, the profile of the pulleys is less and smaller pulleys are possible. Also, I think they would be easy to make with a form tool. I wouldn't have any reason to get out the terrible compound that came with my lathe. After searching for what's available, I found that I was going to need to make the pulleys wider than the stock I had purchased. Has to do with how much horsepower per rib the belt can support. So I'm going back to 3L. Those belts are available at McMaster in 1-inch increments. In sizes around 27" they're available in 1/2-inch increments, which is convenient.

So how long of a belt do I need? I took some measurements and made a spreadsheet. Instead of calculating the exact length, I'm assuming that half the diameter D/2 at each pulley, except the idler on the tensioner, which was D/3. Then use the distance between pulley centers to estimate the belt spans between pulleys. Then I made a table in which the first row represented a 2.5:1 pulley ratio (5" and 2" pulleys) and the corresponding belt length. Then I calculated the belt length for 5" and 5" pulleys, then 4.75" and 4.75, etc, decrementing pulley diameter by 0.25" for each row. Eventually, I found that a pair of 3.5" pulleys requires the same belt length as the other steps in the pulley. So now I can make my custom step pulleys---5, 3.5, and 2". This is all depending on being able to get a 5" pulley in there.

I did the same analysis assuming I need to cut down the largest pulley to 4.5", which requires 1.8" to get a ratio of 2.5:1. The 1:1 pulleys for that length of belt is 3.25. Here's the table for that:

I also need a tensioning mechanism. I'd like an over-center mechanism that uses the belt tension to hold it in place and the operator can feel the tension directly. As much as I'd like something fancy, a tiny masterpiece, I should probably resist the urge and do something simple. Making speed changes easy is part of the psychology and I want to stick to that. Moving the belt should not require tools, fiddling, or brute strength. Well, whatever the mechanism, there's another hole in the back of the mill that I plan to use. Not sure what it was for. The original tensioner is that bar on the left, which does not rotate because it was set with millwright strength. It's held in position by a 3/8 bolt and nut. Despite a 7" lever arm, it holds tension. Maybe they were only taking very light cuts... looks flimsy to me.

 

[durableoreo]

Plans have changed. I was reading about the 15-$ DC-motor speed controlers on eBay... they don't control RPM so the speed will dip under load. Maybe I'm being fussy but I just want steady speed. Can you imagine setting the RPM during a pass only to have the speed double when you complete the pass and the mill has no load? Can you start the approach to your next pass with the RPM zipping along so the teeth first encounter the work at high speed? Doesn't seem like a good idea to me.

So I purchased a 1-HP three-phase motor, used. It runs at 1750 but I was careful to get an induction motor so I should be able to spin it at 2x, which is the top speed for the transmission I designed. I also purchased a S70 VFD, which was amazingly cheap. Maybe I should have done this in the beginning without bothering with gears and pulleys. I just didn't know enough to decide. I also didn't realize how fantastically inexpensive this type of setup could be. But I'm going to stick to my transmission design for now because I don't know about torque at very low motor speeds. Perhaps it is possible to run fairly slow but perhaps heat or something else is an issue. If I can run the motor near its rated speed, I'll get 300, 120, or 47 RPM, depending on the belt position. See the color coded boxes in the chart below. The box is what I expect to use most of the time. Overspeeding the motor is probably OK---the practical machinists say 150% to 200% is find for induction (squirrel-cage) motors.

I cut some pulleys today. There's nothing like doing something new to realize how much you don't know. I bought 1/2" rounds from McMaster and thought the rest would be fairly easy. But it turns out that my chuck can hold nothing larger than 3.9" so I had to figure out a work-holding strategy for the 4" and 5" stock. And then I discovered that the inside jaws on my chuck could grip the 1.25" IDs of half the pulleys but not the 1.125 IDs of the other half.

For the 2" pulleys, I was able to grip the work in the chuck and bore out 1.25 and 1.125" holes. Then using those holes, grip them from the inside and cut the groove. For the small pulleys, it was 32 degrees, or 17 degrees per side. I did the setup with a protractor. Testing the belt, it's fairly important that the belt rides at least 2/3 deep in the groove and that it not touch the root of the groove.

The larger parts I had to bore on the faceplate. Instead of using a set of clamps, I drilled and tapped 1/4-20 holes in the face plate and drilled clearance holes in the blanks. Then, secured to the faceplate with some machine screws, I was able to bore out the center of the 3.5" and 5" pulley.
I was able to grip some of the pulleys (1.25" ID) in the chuck. Then it was the usual operations---face off, true the pulley, break the edges, and cut the angled groove.

The other pulleys won't fit on the chuck so I think I'll make a stub shaft with a tapped hole in the end. Haven't decided but I'll show a picture of what I end up doing.

I think that took me 6 hours between trying to decide what to do and doing it. My lathe is marginal, at best, and I'm not an experienced machinist. If I had paying work, I would love to replace it with something less janky. I know it can be improved but I don't have the patience to thread the belt this way and that, to change gears, etc. But, in the words of AvE, you've got to **** with the dick you have.