[cxixer]

Here is a design I am thinking of using for my rack and pinion drive. My original router design used R&P, didn't work because there was no spring putting tension on the rack. It would work itself loose and screw up my pieces that I was cutting. I am going to basically build an arm and use a spring to press it up against the rack, similar to other designs I've seen around here. This arm will be made of MDF, using bearings around the pivot screw and the motor shaft. In the future I will replace this with aluminum, and may gear it down, but I was happy with the resolution and torque before, so I'm sure it will be fine.

Why do this? To fix the rack setup in the first place I replaced it with 1/2-10 acme screws. Well, now it's slow. Accurate, but very slow - 30ipm max. I'm sure most of this is adjustment, but even if that is perfect, I will not get even 80ipm. Multiple start screws are expensive. I already have this rack setup, and even gearing if I want to gear it down. Live and learn.

Specs: I will be using 1.5" MDF plus another 3/4" to mount the motor so there is clearance for the setscrew so it does not interfere with the bearings. The bottom of the rack is about 5.25" below the bottom of the side of the gantry. The arm will be about 11" long. I am using a 5/16" bolt and the same size skate bearings for the pivot arm (that's what I have laying around). I will be using a 1/2" shaft to connect the motor to the pinion, and the same size bearings to support the shaft. The arm will essentially be pulling up on these bearings.

Please excuse the drawing, I can CAD only enough to sell my products.

My questions -

Is there a more efficient lever design than this? Without actually building/testing, it seems like it will take a significantly strong spring to apply enough pressure.

Will the bearings I plan to use be sufficient?

Any other suggestions? Anything I havn't thought of?

[GaryCorlew]

Check out mechmate http//www.mechmate.com on there they say that the arm should be parallel (or very close to parallel)with the rack to help eliminate the gear from wanting to climb out of the rack.

[LeeWay]

I'm not sure that MDF will even hold up to the torque for very long. If you can cut MDF, you can cut plastics and even aluminum. Steel would be ideal I think. You could even use off the shelf steel parts and design the rest to fit the bar. Steel bar is available in many hardware stores or any welding shop.
This will be a high stress point. The rest of the machine can remain as is I guess, but this is one area (drive point) that you really shouldn't skimp on.

I was building a large 5 by 10 plasma cutter. My mill worked out so well that I no longer need it. I bought steel and racks for it. I plan on seeing how the rack will work on a small steel frame machine with the motors mounted fast and traversing the rack. I think I should be able to get decent speeds, with a little slower acceleration without the gear wanting to jump the rack. It has to be a rigid build though in order for this to work.

My other builds were for production in my one man shop, but this will be testing a design and be more of a little hobby machine. Likely stick it on Ebay when done. I am only in design so far.
Good luck with it.
I know a slow machine is aggravating, but speed and rigidity costs extra. The more you throw at a design, the better it should get.

[cxixer]

In theory, the arm will be parralell with the rack...


Yeah, I can cut plastic, a bit of chatter, but do-able (hence the reason for the upgrade). I could make these out of Polycarbonate, I have that laying around as well. Would I be better off Using .5" or .75" polycarbonate for the arm? I have .25 and .375" polycarb and can laminate it together.

This is actually a very simple design and wouldn't even need to cut it on the router, I could do it by hand.

I do realize polycarb flexes in one direction very badly, but in the other direction, it is very sturdy.

I am re-doing the x axis with vbearings, the z with aluminum extrusion and NSK linear bearings... These rack drives would be the only MDF, and would be short-term... But if polycarb would work long term, I would do that right away.

What do you think?

[Gerald_D]

Have looked at your sketch, and if your proportions are right, that layout will give you some headaches. If the motor must push the car to the right, the pinion will climb out of the rack.

What you need to try and do is to lower the hinge point of the arm. The best place for the hinge point is at the same height as the rack, but it is not essential to go that low. You can screw a plate to the car with a hinge point lower down in that plate.

In effect this will make your arm nearer to the horizontal.


Gerald
www.mechmate.com

[cxixer]

Have looked at your sketch, and if your proportions are right, that layout will give you some headaches. If the motor must push the car to the right, the pinion will climb out of the rack.

What you need to try and do is to lower the hinge point of the arm. The best place for the hinge point is at the same height as the rack, but it is not essential to go that low. You can screw a plate to the car with a hinge point lower down in that plate.

In effect this will make your arm nearer to the horizontal.


Gerald
www.mechmate.com

Is that what gary was talking about above? I misunderstood. But I get it. Attached is another drawing. I don't have a measurement for the plate ... or a great way to cut the plate... but this should make the arm that holds the pinion about 7-8" as opposed to 11", which will make it sturdier as well. The only other thing I could do to this plate is lower it further on the hinge side to make it under the rack, it couldn't be perfectly in line with it because I have maybe 1/4" clearance and that's not enough for a bolt to stick through to secure the arm onto it. I may have to do this anyway, as the lowest that pivot point could be is about 2.5" higher than the rack because of this clearance. Would I be better off staying as is, or lowering the arm past the rack?

I was thinking I must have overlooked something, and not only will this help the pinion stay on track, but I think it will be more efficient in terms of pulling into the rack itself. Not as much spring tension.

Gerald, do you think the 3/4" polycarbonate will work in this situation? So far (before I hit the machine shop) I've kept my upgrades to about $250, I'd like to not go too much higher, all I still need are two .5" aluminum plates for the z axis, and my ballscrew machined.... I think. If I get these arms milled from aluminum I just threw the cost savings out the window.

[cxixer]

If I brought the pivot point lower than the rack...

[Gerald_D]

Our MechMates all have 45 degree arms - they work fine.

You do not need bearings at the hinge point since the theoretic rotation is nil. We use bolt & nylock nut with some plastic washers for friction.

See: http://www.mechmate.com/forums/showthread.php?t=1055

Gerald

[Al_The_Man]

I used these items on a large gantry for the pivot point tensioner http://www.rostainc.com/pdfs/rostainfo9.pdf very simple.
Here is the example in post#12 & #13, the first was a spring tension method the others were the Rosta method.
http://cnczone.com/forums/showthread...ighlight=rosta
Al.

[cxixer]

OK, I've made my upgrades.. for the most part. But there's a problem. Attached are some pictures of the new router (note it has a brand new MDF table with t-nuts for hold downs not shown). The problem is I'm losing steps somewhere. I went out and tried to cut some rings (gcode attached) and what is supposed to be 13"x13" square comes out to about 12 15/16" x 12 1/2". Now explain that. The circles were also completely screwed up. I cut it twice, once as is, and then I noticed my set screws were not tight on the x axis pinion (shown attached to the router plate), so I tightened them and used loc-tite ... again. The second time the circles got just as screwed up, but the squares are pretty much perfectly the same. So the repeatability is great...yay...


Now everything looks very tight, I do not have a dial to measure backlash, but again, looks tight.

How tight do these springs need to be pulling the pinion towards the rack? Could that be my problem? I can push them down off the rack with finger pressure, but it takes quite a bit of force pushing on an axis to get them to jump off the rack. I don't know their force or anything, I just chose springs that were pretty tight from ace hardware...

[harryn]

Hi, I have been reading about racks, but have not built mine yet - so take this "advice" for what that is worth.

I think I read (somewhere - maybe mechmate) that you need about 80 lbs of holding force to work a cnc router - so I just keep using this number in my calculations.

Try taking the oz- inches of your stepper motor and the pinion pitch diameter and see how close you come to this.

Example 500 oz - in motor (max), 2 in dia pinion

500 oz - in / 1 in (radius) = 500 oz

500 / 16 oz = 31 lbs of holding force (maximum)

So, while it is a strong motor, it probably is not enough torque without gearing down or a smaller pinion.

The minimum pinion size for a given pitch is sort of limited by the point where the tooth interface is no longer the right path (see boston gear - very detailed explanation). In the end, for common tooth pitches, you sort of end up with 1 in min pitch dia. This means it almost takes 700 - 1000 oz in motor to make it work without gearing down. (assuming the 80 lb holding force number)

In theory, you can go to very fine pitch to get around this problem. There is metric pitch rack that is called out as I think M1 or M1.5 - something like this. When you run the tooth strength numbers, it is hard to achieve the 80 lbs numbers this way unless it is a high strength steel (expensive)

Another place to look is the stepper motor driver settings. The drivers tend to reduce the current when the motor motion is idle to prevent overheating, thus - reducing torque even more. I think you can adjust this setting, but I don't know how.

As far as the spring, I would have guessed that you need more like 30 - 50 lbs of force to keep it firmly in the rack, but that is a guess.

[cxixer]

So after reading this post a couple of times I'm still confused. I guess I havn't read that much about racks. Does the cutting force or speed come into these equations at all? It seems like the harder the force on the bit, the more spring force would be needed.

Again, I don't understand your numbers really, but I'm using 425oz-in motors, 16 pitch racks, 24 tooth pinions, direct drive. I do have some belts/gears to gear the setup down 3:1, but it would be a pain to make the parts. It would be more of a pain, though, to purchase absolutely anything else for this money pit, which makes new racks and pinions out of the question. I'd love to just get new springs, but I can make the gearboxes if I need to.

What are your thoughts?

Where do I get these different springs?

[harryn]

Hi - well, I am not an expert, so I hope someone else is checking my numbers here.

Here is a link to the boston gear literature section:

http://bostongear.com/literature/index.asp

click on the line for downloading the pdf for "spur gears", about 5 lines down.

Look on page 9 of the catalog for 16 pitch gears.

In the area marked for 16 pitch gears, 24 tooth, it indicates that this gear has a pitch diameter of 1.5 inches.

Imagine that the gear's pitch radius is is the length of the handle of a torque wrench, and you need to have a "force" at this point of "something like 80 lbs" to make your system work. The 80 is not set in stone, but you can imagine that 8 lbs is not enough. (think of how hard you need to hang onto a router to hold it to edge a board)

Now imagine that your stepper motor is twisting at the "socket" end of that torque wrench with 425 oz - inches (divide by 16 to make it lb - inches)

425 / 16 = 26 lb-in.

Now the distance from center of the "torque wrench = the radius = 1/2 of the pitch diameter (sort of the effective diameter of the gear) = 1/2 x 1.5 inches = 0.75 inches

26 lb-in / 0.75 in = 34 lbs.

Ok, not quite 80, but not bad either, so let's assume this is ok for now. Remember though that this is the "peak" - essentially 0 RPM holding torque - and this will drop the faster the stepper motor rotates. This is why the pros like servo motors, the torque is more or less constant with rpm over a wide range, but they are more difficult to tune than steppers, which is why DIY like steppers AFAIK.

[harryn]

Perhaps (and I am just guessing here) we should try eliminating a few things first.

a) Motor microstepping effects
- Consider setting the motors to only take full steps for some testing here
- This will allow them to move with maximum torque and reduce the chances that you are skipping steps.
- This is just for troubleshooting purposes
- The test cuts will likely not be smooth with this, but it is ok for now.

b) Add spring force
- The springs should be strong enough that you have a very hard time moving the spring arm with your thumb.
- Add springs until desired result (arm movement is not obvious even with strong thumb push)

c) Test cuts
- Cut a circle and a square cut small enough to measure accurately
- Use low IPM and measure
- I suggest a light test cut first - perhaps only 1/4 in deep but use a 1/2 in dia bit if you have one handy
- repeat at higher IPM
- repeat at higher IPM and deeper cuts

d) Watch
- Watch for deflection in the spur gear and rack area during the cuts
- Especially watch to see if the spring arm is moving around or deflecting.

Let us know what you are seeing.

[harryn]

So after reading this post a couple of times I'm still confused. I guess I havn't read that much about racks. Does the cutting force or speed come into these equations at all? It seems like the harder the force on the bit, the more spring force would be needed.

What are your thoughts?

Where do I get these different springs?

Yes, cutting force and speed do matter for the springs and design, but we are working with DIY approach here.

[LeeWay]

I was working on a design for r&p on my stalled out plasma cutter project. I often see guys use very long srings and the arms have a lot of room to swing. I think the only reason the pinion needs to move at all it to allow for discrepacy in the mounting of the rack. In other words, mount it solid to start with and then try it to see if you have any loose or tight spots. You could likely check it with calipers along the length ogf the rack. Then once you have this variance, allow only that much sway in the pinion mount. Use large springs to where you really can't move it by hand, but could with a screw driver or something. Compression springs from below would be cheap and easy to get large loads on in a short space.

If you do a good job of mounting the rack, then there isn't ever even a chance it can jump the rack.

I will make a mchine with rack one day and finish that design. It just seems like a more logical way to design them rather than have so much room to sway. Think about when you change direction, if that spring isn't stout enough and you have any good speed, it's gonna skip some teeth.

[cxixer]

Perhaps (and I am just guessing here) we should try eliminating a few things first.

a) Motor microstepping effects
- Consider setting the motors to only take full steps for some testing here
- This will allow them to move with maximum torque and reduce the chances that you are skipping steps.
- This is just for troubleshooting purposes
- The test cuts will likely not be smooth with this, but it is ok for now.

b) Add spring force
- The springs should be strong enough that you have a very hard time moving the spring arm with your thumb.
- Add springs until desired result (arm movement is not obvious even with strong thumb push)

c) Test cuts
- Cut a circle and a square cut small enough to measure accurately
- Use low IPM and measure
- I suggest a light test cut first - perhaps only 1/4 in deep but use a 1/2 in dia bit if you have one handy
- repeat at higher IPM
- repeat at higher IPM and deeper cuts

d) Watch
- Watch for deflection in the spur gear and rack area during the cuts
- Especially watch to see if the spring arm is moving around or deflecting.

Let us know what you are seeing.

I'll do some test cuts and see. I watched the motors and axes for an in-sync movement in the past to 'test' for backlash, and thats how I figured out that the set screws were loose in the first place.

I did think of something just now though.. I mentioned that one axis was off about 1/16", the other was off 1/2", now both the x and y use the same drive mechanism, motors, pinions, etc. The only difference is the spring. Stronger springs are on the y axis, which was only off 1/16".. I have extra springs, I'll go test. And I'll measure and go buy some heaftier springs...



To LeeWay: Yep, you got me. The entire point of the spring setup was not to make it look cool, it is because I tend to have a difficult time doing anything perfect with my collection of metalworking tools (angle grinder and a harbor freigt welder). I'm a woodworker... I don't even really have proper measuring tools to work on metal. So, the springs compensate for the rack being slightly skewed in some places. I originally had this router setup with the same racks and pinions, without the springs, it failed horribly. I've learned to build in some adjustment - self-adjustment where possible (eg. springs). Also, you mention changing direction at a good speed, this could also be adjusted out using mach3 acceleration settings.


And "to measure accurately" I'll need to go get some batteries for my digi caliper..

[cxixer]

OK, I tested. Sorta. I didn't go get the battery yet, so I am left with my 1/32" tape measure.

I noticed the other two set screws on the x axis were loose. (Yesterday it was the ones to the pinion, today it was the ones to the motor) So I pulled it apart, loc-tighted them, and twisted as hard as I thought I could without stripping it. Then I added two more springs onto the x axis. Through the testing somewhere I added an extra spring onto each y motor as well.

I cut some circles about 2", turned out perfect (as perfect as I can measure). I did some square pockets with mach3's deal, that's cool by the way, and they turned out pretty much perfect - I set it to do a 2x2" pocket and it did a 1.75x1.75, I may have set it with the wrong bit. Then I started cutting what I tried yesterday and it was much closer than before. Still off by just a hair, less than 1/16". I'm guessing that with proper monitoring of the set screws and a few tighter springs, it should be good enough for me. Which means within around or less than 1/32" in any direction.

Another note, I don't know exactly how much pressure is being applied by my multitude of springs on any motor, but it is very close to the limit of the motors. I tried jogging the x axis while pushing up on the swing arm a smidge, and the motor stalled. Again, not very scientific, but you are correct about gearing down if I need too much tighter springs.

I will go pick up a battery and some springs tomorrow and make some more test runs. These were all done at ~30ipm, except the piece that was off yesterday, it's set to 80ipm.

[cxixer]

Updates. Got the new springs, they're just tighter, I don't know what they are rated as far as force goes but it takes a good amount of finger pressure to get them to move off the rack. Picked up a battery for the digital caliper, made some 2" wide, .25" deep test circular pockets using the Mach3 wizard. Results are as follows:

Again I'm looking to beat 1/32", closer than that and it's too close to tell with a tape measure anyway. For those of you metric people, that's .03125".

I recorded speed: accuracy around the circle (time it took m:s)

20ipm: +-.022 (2:50)
40ipm: +-.014 (1:28)
60ipm: +-.019 (1:02)
80ipm: +-.017 (0:49)

Well that is good enough for me. I will be cutting at 80ipm at least until I feel the need to go faster, I've just upgraded from topping out at 20ipm even jogging, that's painful.

I did notice that the accuracy was closer on the y axis than the x, those +- numbers were always short of 2, not over, and that was the maximum around each circle that it was off. But if it is this close, I'm ok with it.

I did notice a pattern though, the faster the feedrate, the closer the accuracy got for the most part. I figured it would be the other way around. Any comments on this?

And I made several test circles just to be sure. Neat little programs, those wizards are.

[harryn]

Those results look pretty good to me as well. I hope someday to have similar results, so when I get stuck, I will drag you into the conversation.

How deep are you cutting at this point ?

While your machine is working so well, perhaps make up some replacement Al parts to work your machine away from MDF.

Gear rack is a funny beast. It is easy to buy, but hard to get a spec on how accurately it is actually built. The metric stuff can be specified like this, but it gets expenseve fast.

Just curious - did you use 14 or 20 degree pressure angle rack ?

As far as cutting accuracy vs speed - I would be just guessing, but I have noticed that when I use my hand held router to edge a board (with a bearing bit), there are certain feed speeds that give better results than others.

For my bosch router and a whiteside 1/2 in bit, it was around 5 - 6 IPS (300+ IPM) in 3/4 plywood. Kind of a scary pace with a DIY cnc router.