FC Gears: Feedback thread

Have some feature requests, feedback, cool stuff to share, or want to know where FreeCAD is going? This is the place.
Forum rules
Be nice to others! Read the FreeCAD code of conduct!
User avatar
Kunda1
Veteran
Posts: 13434
Joined: Thu Jan 05, 2017 9:03 pm

Re: FC Gears: Feedback thread

Post by Kunda1 »

freman wrote: Thu Sep 05, 2019 11:47 pm Indeed, my initial comment on all this was that someone may like to update and fix the errors on the FCgear install instructions. Both on the Wiki page and the git readme.
README has been tweaked. Feel free to tweak it further. In my experience, @looo is very receptive and fair with Pull Requests. ;)
Alone you go faster. Together we go farther
Please mark thread [Solved]
Want to contribute back to FC? Checkout:
'good first issues' | Open TODOs and FIXMEs | How to Help FreeCAD | How to report Bugs
User avatar
Hartwig
Posts: 49
Joined: Thu Aug 01, 2019 3:26 pm
Location: Germany

Re: FC Gears: Feedback thread

Post by Hartwig »

Hi,
I am confused about the backlash parameter in FCGears.
I created two involute gears.

First one:
teeth= 15; module= 1mm; backlash= 0mm;
measured tooth thickness= 1.571mm; (This is what I expected.)

Second one:
teeth= 15; module= 1mm; backlash= 0.050mm;
measured tooth thickness= 1.383mm; (This is not what I would have expected.)

usg_zahnrad_backlash.PNG
usg_zahnrad_backlash.PNG (62.67 KiB) Viewed 1681 times

According to Wikipedia I would have expected: 1.571 - 0.050= 1.521mm

usg_zahnrad_backlash_wikipedia.PNG
usg_zahnrad_backlash_wikipedia.PNG (10.26 KiB) Viewed 1681 times

Unfortunately I could not find any definition of the FCGears-backlash. So can anyone explain how FCGears-backlash is defind?
Thanks.
User avatar
looo
Veteran
Posts: 3941
Joined: Mon Nov 11, 2013 5:29 pm

Re: FC Gears: Feedback thread

Post by looo »

Unfortunately I could not find any definition of the FCGears-backlash. So can anyone explain how FCGears-backlash is defind?
Thanks.
This is the current way how it is used:
https://github.com/looooo/freecad.gears ... oth.py#L88

But don't ask me where I have borrowed this formula.

You are right, currently we use the backlash more like a angle-property but I guess the angle should be derived from pitch diameter and the lenght property of the backlash. The only problem I see here is that finding good values for the backlash is quite difficult this way. For example if you have a big gear with a some amount of backlash and you decrease the module, you will get an invalid gear-shape.

I guess we should define the backlash as a value without dimension and multiply with the module like it is done for the shift property. But I am open to other suggestions.
User avatar
freman
Veteran
Posts: 2198
Joined: Tue Nov 27, 2018 10:30 pm

Re: FC Gears: Feedback thread

Post by freman »

if you have a big gear with a some amount of backlash and you decrease the module, you will get an invalid gear-shape.
Why would the the gear-shape be invalid. Surely backlash does not affect the gear profile, just the thickness of the teeth. If you specify a backlash distance and change the gear parameters, that backlash may be excessive or insufficient for the new gear but the profile should still be valid. If it is not maybe you are not producing valid profiles for any gears with your current implication of backlash.

Perhaps you could clarify that comment.

Maybe rather than just telling the user to parse the source code to understand how to apply a backlash as a distance, you could document how to convert it.
kisolre
Veteran
Posts: 4162
Joined: Wed Nov 21, 2018 1:13 pm

Re: FC Gears: Feedback thread

Post by kisolre »

looo wrote: Sun Sep 29, 2019 7:51 pm This is the current way how it is used:
https://github.com/looooo/freecad.gears ... oth.py#L88

But don't ask me where I have borrowed this formula.
In the code here backlash is treated as angle but in the settings is expected as mm. In reality backlash is the minimum distance betweeen teeth when the gear pair is touching in one direction so rotation angle should be calculated based on that distance along the pitch circle.
User avatar
looo
Veteran
Posts: 3941
Joined: Mon Nov 11, 2013 5:29 pm

Re: FC Gears: Feedback thread

Post by looo »

freman wrote: Mon Sep 30, 2019 4:03 am
if you have a big gear with some amount of backlash and you decrease the module, you will get an invalid gear-shape.
Why would the the gear-shape be invalid.
This is what I meant with invalid gear shape:
Bildschirmfoto von 2019-09-30 08-14-02.png
Bildschirmfoto von 2019-09-30 08-14-02.png (28.49 KiB) Viewed 1623 times
So the question is if the rotation angle should be computed by the backlash value multiplied by the module or if the backlash should really be a length. The current way is for sure misleading as the property indicates a length but the value is used like an angle or a length without dimension.
kisolre wrote: Mon Sep 30, 2019 5:55 am In the code here backlash is treated as angle but in the settings is expected as mm. In reality backlash is the minimum distance betweeen teeth when the gear pair is touching in one direction so rotation angle should be calculated based on that distance along the pitch circle.
So actually for one gear we should only use the half of the given backlash?
User avatar
freman
Veteran
Posts: 2198
Joined: Tue Nov 27, 2018 10:30 pm

Re: FC Gears: Feedback thread

Post by freman »

This is what I meant with invalid gear shape:
thanks for the clarification. That is a crazy amount of backlash before you get and invalid profile. It must be pretty obvious at that point that the backlash needs changing. AFAIK backlash is normally specified as a clearance distance ( eg 3 thou' ) so that is probably what people need to enter in the interface.
So actually for one gear we should only use the half of the given backlash?
It's a good point that in fact the backlash is a property of a pair on gears, not the profile on one gear. It does not make much sense to specify backlash of one gear.
kisolre
Veteran
Posts: 4162
Joined: Wed Nov 21, 2018 1:13 pm

Re: FC Gears: Feedback thread

Post by kisolre »

looo wrote: Mon Sep 30, 2019 6:57 am So actually for one gear we should only use the half of the given backlash?
At first that was my thought but actually you will have half of the backlash per gear and then half of that per side (I assume "rot" is the rotation of one side of the tooth?) so 1/2 * 1/2 = 1/4 is good.
freman wrote: Mon Sep 30, 2019 8:24 am It's a good point that in fact the backlash is a property of a pair on gears, not the profile on one gear. It does not make much sense to specify backlash of one gear.
You never design a single gear. And most of the time axes of the gears are in a fixed positions so to create a clearance you need to factor the backlash on gear creation time. At least that is how I understand it.
User avatar
Hartwig
Posts: 49
Joined: Thu Aug 01, 2019 3:26 pm
Location: Germany

Re: FC Gears: Feedback thread

Post by Hartwig »

Thanks, Looo, for the link and your work.
In my job I design little plastic parts and sometimes gear wheels are among them.
Probably because I am far away from being an expert for gears I like to keep it simple.
I dont think you should couple the backlash with the module or other parameters.
We need to reduce the tooth thickness because we need clearance between teeth of gear wheel 1 and teeth of gear wheel 2. Otherwise functional problems, noise, ...
So why not just offsetting the both sides of the tooth? I wouold be fine with that (usually negative) offset value in the parameter list.
How much offset? There is no simple answer. It depends on tolerances, surface quality, enviroment conditions and much more I guess.

Hartwig
User avatar
looo
Veteran
Posts: 3941
Joined: Mon Nov 11, 2013 5:29 pm

Re: FC Gears: Feedback thread

Post by looo »

Hartwig wrote: Mon Sep 30, 2019 8:46 pm Probably because I am far away from being an expert for gears I like to keep it simple.
I dont think you should couple the backlash with the module or other parameters.
Yes, I agree. "Keeping it as simple" was the idea of multiplying the backlash with the module. But I see the problemsoccurring from this. So my proposal is to change the computation to align with this formula (if we can agree on this):

rotation = backlash / 4 / r_p

But if we treat the backlash this way, we also need to find a solution for bevel gears. Ideas wanted ;)
Hartwig wrote: Mon Sep 30, 2019 8:46 pm We need to reduce the tooth thickness because we need clearance between teeth of gear wheel 1 and teeth of gear wheel 2. Otherwise functional problems, noise, ...
So why not just offsetting the both sides of the tooth? I wouold be fine with that (usually negative) offset value in the parameter list.
How much offset? There is no simple answer. It depends on tolerances, surface quality, enviroment conditions and much more I guess.
An offset is not the right way to treat the backlash. This way the involute is not exact anymore. In my eyes, we should really add the backlash by a rotation. A rotation won't change the involute and therefore also gears with backlash fulfills the constant velocity-ratio which is the basic idea behind the involute profile.
Post Reply