Lasercutting Spur Gears?

[Tytus Gerrish]

watterjets can do prety much anything , but only one axis at a time, these sprockets had to be file'ed in between the teeth because the tolerances weren't perfect, but never the less its an awesome machine , i wish i had one in my room. prety much anything is possible its completely indiscrimnate with materials you can cut thru anything

[Matt Adams]

Originally Posted by sanddrag

Thanks for all the input. I was just curious since we have lasercutting companies as sponsors who are willing to make stuff for us. The gear I was referring too would be the PIC 0.7 module one that takes so long to get.

Despite that the .7 module gears are such fine teeth, I would imagine that these could be laser cut. Per a slide on some non traditional machining methods, lasers can drill holes as small as .0002 inches in diameter and cut with depth to diameter ratios of 50:1. However, this will largely depend on the quality of the machines they have. In addition, the opperating cost for these machines is low ($1.50 / hour), so assuming that they'd be willing to spare the labor cost (the bigger part of the equation, $20-50 / hr) I'd bet you could make this happen.

Good luck!

Matt

[Bob Steele]

On our robot this year we used several water-jet machined gears.
Specifically we used a hard plastic 6" gear which was approximately 1" thick and was used to operate our arm. The small gear which was attached to the Van Door Motor was also cut on the water jet and was made of aluminum.

These gears performed flawlessly for us after we figured out that we could not use softer plastic gears because of the stress on the teeth. The cog is quite handsome. I would add a picture but I don't know how to attach it here.
The actual gears are skeletized with a webbing of titanium plate that holds the aluminum hub inside...very striking as the plastic is blue.

We were lucky enough to gain a waterjet cutting company as a terrific mentor this year. It opened doors for design that we didn't have before and we are very grateful that he chose to join with us.

By the way, the taper can be compensated for, one of the waterjets that is owned and operated by Rolls Royce has this compensation. The one we were able to use did not. We did not have any problems though with the gear.
I would expect that if you chose to drive a gear more stressfully (IE in a drive train...) that you might see some problems. For our application it worked admirably.

I will post a picture in the gallery if anyone is interested.
One of our mentors also designed omni-wheels which were entirely cut on the waterjet as well. The only additional machining that was necessary were some grooves for the small wheel axles (millwork) and tapping the holes for the screws used to hold the assembly together.

Like any piece of equipment, the waterjet is not a complete answer.
I only know that it opened doors for us this year that helped us immensely.
Thank you ... Mike Trapp and WaterJet Cutting of Indiana.

thanks
Bob

[D.Fahringer]

We have two waterjet machines at NASA Langley, one is 2' X 4' capacity and the other is a 4' X 8' machine. The diameter of the jet stream and resulting minimum corner radius is .020" (when machine is properly maintained). So if you can live with that much radius for accepable clearance in your gear design you are o.k. with a waterjet...something the size of that small gear on the drill motor...forget it.

As for the "funneling" that will angle the sides, we have an articulating head on the big waterjet that completely eliminates that! The machines internal software looks ahead at the part geometry and compensates for stream cone/funneling and lag. So if you send this work out you can specify that type of machine. Ours is a Flow Bengal 4X8.

I have used gears made of 3/8 inch aluminum plate that were waterjetted on the older machine and the fit was great as long as you reverse one so the angled sides compliment one another! (Relativley low rpm use)

The maintainence issue stems from the orifice that the high pressure water (50,000 PSI) and abrasive sand passes through, it is made of ruby but will wear over time and the stream diameter gets larger.

[sanddrag]

I talked to our laser guy and he said if I can draw it he can cut it. He said the finish will not be as good as something you could buy but it should be acceptable for the use. Now I'm just wondering how to draw it? I have no idea how. Also, I was wondering if the gear can be machined on something like a 4 axis CNC. If we put round stock of the proper OD in the 4th axis the machine can index it for cutting the groove between each tooth. However, I have no clue as to what tool (or if one even exists) to cut the teeth in size of gear. Any insight?

Last, I have one more question. Is 3/16 wide steel gear witrh a 2mm keyway (8 mm bore for the Chia shaft) enough to hold the torque it will have on it. Will the keyway strip out or enlarge or should it be okay since it is steel. Now, we could always go to a titanium gear. No joke. The laser cuts it very nice.

Anyway, my end goal in all this is to get another source for these gears besides PIC. If we make them, it would not be much harder to make 20 than it would be to make 2 so we could make it easier for everyone.

[Yan Wang]

With good equipment, waterjetting's funneling shape can be removed. Our team used WireEDM (as Raul suggested) for our purposes.

We drove our treads this year with this custom cut gear:
http://www.chiefdelphi.com/forums/at...achmentid=1992

Which you can see on the robot:
http://www.chiefdelphi.com/forums/pi...&quiet=Verbose

We also used it for a veeery nice gear in our gearbox that I don't have a picture of at the moment. Regardless, WireEDM is the most precise way to do it. It's just painstakingly slow.

[sanddrag]

I was actually able to redraw the one on firstcadlibrary with very little difficulty. So, I will be trying to make the 54 tooth one. Now, simple logic tells me that the actual height of each tooth should be the same between different sizes of the gears. The outside diameter of the 60 tooth gear is 43.4 mm. The root diameter of the 60 tooth gear is 40.25 mm. This is a difference of 3.15 mm. Is this difference number going to be exactly the same no matter what tooth count the gear is? PIC lists the difference between outside diameter and pitch diameter to be 1.4mm. So assuming all this is correct, the pitch diamter of a 54 tooth 0.7 module gear would be 54 x 0.7 = 37.8mm and the outside diameter would be 39.2mm. If we subtract the OD-root difference of 3.15 mm, the root diameter should be 36.05 mm. Is this correct?

Also, I assume the shape and dimensions of each tooth remain constant across the various tooth counts of a particular gear. Is this true?

Thanks.

EDIT: I called an EDM place and they said it would be something like $35-50 per piece and they could do it in a day. I forgot to ask if there is a setup charge. Would anyone here know? Because I don't want to call back because the guy was very hard to understand and he went on and on forever.

[D.Fahringer]

Sanddrag,

With this web utility you will be able to generate accurate DXF drawings of ANY gear you will need. The laser should be able to directly import the DXF file you create and go to town on your new gears. There is also a provision for "offset" that would be the value of the laser diameter of cut.

I have used this, you will like it, and learn gear terminology too!

http://sinfor.lcs.mit.edu:8180/pcbmill/gears

df

[ChrisH]

Originally Posted by sanddrag

.....
Also, I assume the shape and dimensions of each tooth remain constant across the various tooth counts of a particular gear. Is this true?

Thanks.

EDIT: I called an EDM place and they said it would be something like $35-50 per piece and they could do it in a day. I forgot to ask if there is a setup charge. Would anyone here know? Because I don't want to call back because the guy was very hard to understand and he went on and on forever.

Whether or not there is a set up charge is a business by business (and sometimes job by job) thing. But since he was quoting by the piece and didn't mention it, then there probably isn't one. Though it wouldn't hurt to check.

The shape of the involute surface of the gear tooth does vary with changing diameters, but the other parameters (tooth height, addendun, dedendum, etc) remain the same. So you wouldn't want to design a 10 tooth gear and then just copy that tooth for a 60 tooth gear.

The interesting thing is that for conventionally made gears, you only need one hobb per pitch. All gears of a given pitch can be made from a single hobb, which is fairly straight forward to design. One instance where the old way might really be easier.

If your laser guy says he can cut it, and he isn't charging you, then I'd have him do it first. If he really can't produce what you need, what have you lost? But you're sure to learn something in the process.

ChrisH

[sanddrag]

Originally Posted by ChrisH

The shape of the involute surface of the gear tooth does vary with changing diameters, but the other parameters (tooth height, addendun, dedendum, etc) remain the same. So you wouldn't want to design a 10 tooth gear and then just copy that tooth for a 60 tooth gear.

The interesting thing is that for conventionally made gears, you only need one hobb per pitch. All gears of a given pitch can be made from a single hobb, which is fairly straight forward to design. One instance where the old way might really be easier.

ChrisH

Could you please explain to me (or show a picture) of what the involute surface and hobb is? Thanks.

[Tristan Lall]

Originally Posted by D.Fahringer

With this web utility you will be able to generate accurate DXF drawings of ANY gear you will need. The laser should be able to directly import the DXF file you create and go to town on your new gears.

It's convienient, but watch out if you want to cut those on a laser (or a waterjet or an EDM). The .dxf generated by this utility is only approximate--it doesn't contain a proper involute curve, instead relying on straight lines of just about the right proportion. For a large-pitch gear (over 20 or so), this might work, albeit with a lot of wearing-in necessary. For anything finer, I'd avoid it outright, except in an emergency (i.e. you need the 0.7 module gears yesterday, and PIC is out of the question because they take too long!).

Watch out for the potential for roundoff error when using this to create metric gears (by inverting the module and multiplying by 25.4)--it reports that my 84 tooth 0.7 module gear has 83.55 teeth (that's measuring at the intersection of the two straight lines on the tooth face--these should be a curve--and taking that to be the pitch diameter, then dividing in the module). Maybe this isn't intended to be the pitch diameter, but since there's nothing else specified....

This is a nice approximation for importing into a CAD system, if you feel the need to jazz up your drawings with actual teeth; bear in mind that I usually just draw cylinders in Pro/E, and mate them at the pitch diameter, or specify a few thousandths open centres (0.005" or so, depending on various factors).

So use it, just be warned that it might not produce "real" gears! (I've been searching for something like this that can put a real involute on the teeth, for exactly this reason--putting it on an EDM would simplify the supply chain for those stupid metric gears infinitely!)

P.S. Now that I look at the 20 pitch gear that it generates, there is more than one segment making up the "involute" surface--and there is no clearly defined pitch diameter--so maybe there is no roundoff error to speak of, and the intersection of those two segments on the gear face just happened to occur near where the pitch diameter was supposed to be. (So don't assume what I just did--that if there are two segments on a tooth face, that their intersection falls on the pitch diameter!) Also, finding several segments on the tooth face reinforces my thought that the accuracy of this tool diminishes with the decreasing pitch of the gear--big gears will have lots of segments, and therefore be better approximations (relative to their size) and small gears will be badly modelled.

P.P.S. So what? Maybe you could design a gear with 7 module, and scale it down by a factor of 10, and get a much improved approximation. Or 700 module, and scale by 1000. In fact, I think that I'll try that! (But careful: I know that CNC plasma and flame cutters don't always like lots of little segments on a toolpath--though they aren't accurate enough for gears anyway--but maybe EDM, laser and waterjet don't want small segments either?)

[sanddrag]

I think I might know what the involute surface is. Is it the front or rear of the tooth? In the firstcadlibrary model it is curved a bit. So this curve changes with tooth count? Is there some sort of formula to determine what radius this curve is based on the number of teeth in the gear? Also, I'm still clueless as to what the hobb is.

Thanks for your help everyone. I am learning a lot.

[ChrisH]

Originally Posted by sanddrag

Could you please explain to me (or show a picture) of what the involute surface and hobb is? Thanks.

Let's start with the easy one first. A hobb looks a lot like a rack from a rack and pinion. I've only seen them in pictures but I'll try and explain them. The teeth on the hobb are trapezoids with the angle on the trapezoid being the pressure angle of the gear. When making a gear, a blank is placed on a rotating shaft and moved up and down in relation to the hobb. The hobb scrapes out one gap, then they rotate the gear one tooth and do it again. They keep it up until the gear is done. Or something like that.

The involute is a particular kind of curve. The easy way to picture it is to place a string on a round object like a can. Put a pencil in the string and unwrap the string while keeping the string tight, The shape you get is an involute. There is mathematical equation for it, but it is ugly. This year I generated a 30 degree pressure angle gear. It was a royal pain and took me several tries to get it right. Even then it was an approximation. I generated several points and then used a spline to connect them. The method I used to generate the points would take me at least a couple of pages to explain. If you do a google seach on Involute you will find tons of info including instructions on how to do a gear in CAD.

If there seems to be a heavy demand I'll do a paper on the creation of involute gears, but not until after Atlanta.

Chris

[Tristan Lall]

So I tried it, and here's the result:

Here's the 0.7 module metric gear (84 teeth, 6 mm bore) that we used on our current gearboxes (you know, the 3-motor, 2-speed, shift-on-the-fly ones... ). It's in AutoCAD 2000 .dxf format (zipped for fun),

Sure enough, the teeth are composed of little surfaces (0.001" long, or so), which approximate an involute. Does anyone have any experience feeding this sort of shape to an EDM, waterjet or laser, and if so, would it work without choking on the tiny little segments?

[sanddrag]

I'm still a little unclear on exactly what the involute curve is on a gear but I think I am getting closer.

It is this:


Or this:


BTW this is the 60 tooth 0.7 module gear from firstcadlibrary.