PDA

View Full Version : Simple 3D Gear design software recommendation



Leadhead
08-11-2017, 08:20 AM
Could anyone recommend a "Simple 3D Gear design software" please. Would like to generate and machine a couple of bevel gears for 90deg transmission. Not for a strenuous operation, just a mechanical lifter. Looked at Gearotic but that is way over the top. Hopefully open source programme is available.
Any advise appreciated.

m_c
08-11-2017, 11:50 PM
There's nothing simple about gear design, especially once you get more complicated than a basic spur gear.

Best option would be to find somebody with a copy of gearotic who's willing to produce the code you need, or just buy a couple ready made gears.

Leadhead
09-11-2017, 07:04 AM
m-c - Not completely unfamiliar with relatively simple gear design. Just do not want to revise and re-teach myself at this time.
But would like to experiment with a couple of things. This would totally frustrate any third party!
The bevel gear add-in for F360 produces what I want but it is a trade off between graphic presentation and real gear form. But studying this to see if it will suffice in this instance. I need to cut with a vertical axis and radial toolpath, so will not be generically correct in any case. But I will will cut a pair and see. Maybe try printing the same in nylon on a metal core.

komatias
09-11-2017, 02:11 PM
m-c - Not completely unfamiliar with relatively simple gear design. Just do not want to revise and re-teach myself at this time.
But would like to experiment with a couple of things. This would totally frustrate any third party!
The bevel gear add-in for F360 produces what I want but it is a trade off between graphic presentation and real gear form. But studying this to see if it will suffice in this instance. I need to cut with a vertical axis and radial toolpath, so will not be generically correct in any case. But I will will cut a pair and see. Maybe try printing the same in nylon on a metal core.

Could you not just buy the gears off the shelf? They are not that expensive from the likes of beltingonline?

Leadhead
09-11-2017, 04:09 PM
George - That`s not the spirit of it?? - Jigsaw mentality - more fun to do it, than what you end up with.

EddyCurrent
09-11-2017, 08:09 PM
Any decent CAD software has a polar array function, meaning you only need draw one tooth. also it should be possible to get a tooth cross section from somewhere (I haven't looked but feel confident)

m_c
09-11-2017, 11:12 PM
I think you'd get away with a standard tooth profile, tilted so it matches the correct cone angle.

I'd have to dig out the gear design/machining book I've got buried somewhere to confirm that though.

Neale
09-11-2017, 11:23 PM
I think you'd get away with a standard tooth profile, tilted so it matches the correct cone angle.

I'd have to dig out the gear design/machining book I've got buried somewhere to confirm that though.

Trouble is, tooth needs to scale with distance along - tooth width and gap aren't constant. There are some standard ways of getting close using a dividing head and taking three cuts per tooth space but I'm not sure how you would fully automate that.

Leadhead
10-11-2017, 06:51 AM
For clarity, I am aiming at a full 3D file to cut conventionally on a 3 axis CNC. With the axis vertical and if practical a radial toolpath.
I believe I am almost there with a bit of caddery-pokery. Makes one very respectful of Art Fernety and his Gearotic software.

Neale
10-11-2017, 08:09 AM
Out of interest, what kind of size/tooth count are you looking at? Tool choice would be interesting as well - small enough to get into the bottom of the tooth but still giving a decent finish?

magicniner
10-11-2017, 08:39 AM
The bevel gear add-in for F360 produces what I want but it is a trade off between graphic presentation and real gear form.

That is stunningly poor!
I'm Shocked!

m_c
10-11-2017, 09:42 AM
Trouble is, tooth needs to scale with distance along - tooth width and gap aren't constant. There are some standard ways of getting close using a dividing head and taking three cuts per tooth space but I'm not sure how you would fully automate that.

Yeah, you need to take the standard form, then scale it to/from the disappearing point where the centre of the two shafts meets.
As Leadhead isn't hobbing, the variable tooth size doesn't really add to the complexity of the machining, just the designing.

Gear machining is something that fascinates me, especially when you see some of the setups and tools from pre-computer days.

Leadhead
10-11-2017, 10:08 AM
Currently adjusting tooth parameters to be able to rough with a 3mm and finish with a 2mm bullnose. That is, of course, the small end of the small pinion.
Not able to answer you until that is resolved. Will post a screenshot then.

MartinS
10-11-2017, 10:48 AM
The bevel gear add-in for F360 produces what I want but it is a trade off between graphic presentation and real gear form.

I also have often looked at the "pretty pictures" that these cad systems have produced and wondered just what the shape of the tooth is based on. Is it a true involute created from first principles, or is it part of an arc of a circle which just "looks right".

Also, I cannot figure out how you are going to machine a bevel gear with just three axes. (I will muse on it.........).

A previous post mentioned taking the cross section of a tooth and extending it forwards and back. Beware of this approach as the tooth surface shape changes along it's length.

23179

At the small end of the gear the involute tooth surface is generated from a small base circle diameter, whilst at the large end a larger BCD is used. You can see from the pic above that from the root outwards, the radius of curvature of the tooth surface is constantly changing up to the tooth tip. But this point radius of curvature changes from small end to large end also.

With a spur gear the involute shape can be machined onto a cutter (e.g. disk milling cutter to end mill) and the gear form cut with indexing from tooth to tooth. This approach won't work with a bevel.


Thought!
How good is CAD and/or CAM at "topological milling"?

If the gear is mounted with the PCD on the XY plane so it can be indexed per tooth, can the tooth space be roughed out and then finished with a ball ended cutter? It could be somewhat rough if at all possible.


Another way is

23180

Mount the gear as above but via a 4th axis on an inclined plane. Then the tooth can be generated with a straight sided cutter in successive passes whilst rotating the gear. The yellow in the pic is a rack relative to the two rotating gears. Therefore each tooth flank is conjugate to a straight sided cutter (e.g. 20 deg included angle end mill). Therefore this principal can be used to cut the tooth surface!

Hope this helps


PS

You are wanting to make a straight bevel not a spiral?:yahoo:

Martin

m_c
10-11-2017, 12:11 PM
Martin, thanks for clarifying the BCD/scaling issue.

From what I gather, Leadhead is planning on machining the gear, by mounting the blank flat on the machine I.e. so the gear axis and spindle are in the same plane.

If you've not seen it before, you might want to have a play with Gearotic (www.gearotic.com). It's written by Art (the original man behind Mach), and allows you to design/simulate all sorts of weird mechanisms. It originally started as just a gear designer, but has grown. The only limitation without a license, is you can't export different file formats, but if you do license it, it can produce various 3D file formats, or even generate the required G-code using either 3 or 4 axis.

Leadhead
10-11-2017, 01:36 PM
Yes I am doing a simple straight conical form. I just took a form from an internet software. Modified the base form to make it a little deeper so that my 2mm cutter form corner did not foul. Then lofted it as a 2D major circumference 32tpi to the scaled down minor circumference 32tpi for the small gear. This looks quite good. I am then doing the same with the gear form onto the 64tpi. Maintaining the same tooth pitch on the mating major and minor circumferences. This looks fairly promising.

MartinS
10-11-2017, 02:14 PM
m_c:

I'll have a look at the Gearotic software. What I can do with it might be limited as I do not yet have a CNC machine. I have planned to build one since retiring and have decided that I AM going to start one in the new year (I guess I'll be posting many questions so please be prepared to be patient with me!). In my professional career, besides bespoke s/w, I used one from Kisssoft AG http://www.kisssoft.ch/english/home/index.php. It is totally trustworthy. I was using it for calculating gear tooth stress analysis, bending, deflections, optimising shape for low noise/wear etc. As we had many machines designed for hobbing/shaping and tooth grinding, we did not need CAD/CAM for producing tooth surfaces. Have just looked at Kisssoft and it is now owned by Gleason and looks to have gone up-market somewhat. There is a test/demo that will allow generation of reliable data.

Leadhead:

You have challenged me to a project for when I have built a CNC machine!:tiger:. I will be following this thread with interest. I guess that by 32 tpi you mean "32 teeth per inch of PCD"? That would be 32 DP which is quite small (tooth depth of 1.78mm). Heck that is small. If you need any help with gear geometry, just shout.

Good luck

Martin

Leadhead
10-11-2017, 04:38 PM
Martin - Yes an error of scale on the TP.
Thank you for your offer of help. I certainly am in the needy category.

Just selling a medium sized 700x 500 footprint, table top CNC router. Fully complete with Mach 3. Including prepped laptop and smoothstepper.
A much bigger one flew under my radar before I got chance to use it. So as new, plug and play. Not actively on sale till I get round to advertising it. Let me know if it might be of interest. Can do a teach in and demo.

MartinS
10-11-2017, 05:42 PM
Thanks for the early offer but I am hoping to get a lot of pleasure out of the design and manufacture of "my own" machine. So will my wife, she will have the Sky remote all to herself!

Leadhead
11-11-2017, 07:24 AM
Martin - Your wife will still be disturbed by you continually flopping in the armchair expecting consolation for frustrations she could not possibly understand or appreciate. This has to factored into your relationship. I am qualified to make this observation!!!

Neale
11-11-2017, 08:10 AM
Only one way to handle that, and it's to keep the brownie point account in credit. Might I suggest that your machine is capable of turning out, at a moment's notice, wooden templates to use for cutting fabric with a soldering iron (and a couple of these were gear wheels), a range of stacking wooden trays to hold more reels of cotton than you could believe exist, and a small wooden pot to hold a wad of steel wool (don't even ask). Regrettably, none of these projects has needed bevel gears but I live in hope...

MartinS
11-11-2017, 09:52 AM
Leadhead.....dont worry, the wife's disturbed already

but


Hmmm...now you've got me thinking.

In all the months that I have been studying previous bid logs I have seen weld movement, spindle fail disappointments. Dreams turned into success with a bandsaw and hand router (Joe Harris).

So for each one read I should have been creating a Pain vs Pleasure index for the design and build stage.

In fact, we could add the final performance into the formula...

PPP index = Pleasure x Performance / Pain

In fact, I think that Lee should make the PPP index calculation part of each build log

Then noobs like me can quickly weigh up the effect of each design on the pocket, mental health, likelihood of divorce, cat leaving home.....


Oops, sorry Leadhead, the thread is cogs 'n stuff....good luck

Leadhead
11-11-2017, 12:30 PM
Tell the cat what you are doing, so that it can avoid getting kicked!