Aligning a sketch grid to the global grid.



  • Jacant

    It is best to use the original grid at 0,0 and move the grid upto the point or a distance to where you want to create a section. You can use the 'Lock Base point' and select where you want 0,0 to be. on the new grid.

  • Me Here

    Sorry jacant, I don't understand a word of that.

    What is "the original grid"? A new one is created every time you start a new sketch (AFAIK)

    If you mean, always use last sketch grid every time you sketch, that impossible. I've done dozen of sketches in this model, aligned in all sorts of different ways.

    And this "You can use the 'Lock Base point' and select where you want 0,0 to be. on the new grid" means nothing to me. I've never seen a 'Lock Base point' option, and wouldn't know where to look.

    The bottom line is I need align this grid with the global origin. The global origin is a known point in space, so why can I not just click on it to achieve my goal?

    I can achieve it, by setting a point at the origin (in 3D) then when I enter sketch mode on the plane, I can easily and quickly align the grid to the global origin by using upto that point.

    DSM has all the information to achieve this simple and in my experience, quite frequently required task without me positioning that otherwise redundant point in my model; so why will it not let me use it?

    Of course, the very best solution would be to always align the grid in a new sketch to the global origin, if the plane of that sketch is parallel with one of the ordinal planes, rather than some random point in space.

  • Jacant

    The original grid is based on the World Origin, be it  X,Y  X,Z  or Y,Z on a New Design. When you want a new Sketch plane it is placed at the center of the point picked on an object. There have been many questions raised about this, including that the origin is correct but it is rotated! DSM has failed to provide an answer to this.

    I have found that if you use a 'Plane' on your original 'Sketch Plane' say X,Y that is at 0,0 you can use this 'Plane' to go back to, this helps if you want a new sketch plane directly above it. Or place a Plane on any of the Sketch Planes that you want to go back to.

    The 'Lock Base Point' can be found in 'Sketch options - Dimensions'  this fixes the 0,0 at the point picked on that plane.


    Thought you may find this interesting.


  • Me Here

    jacant:"I have found that if you use a 'Plane' on your original 'Sketch Plane' say X,Y that is at 0,0 you can use this 'Plane' to go back to, this helps if you want a new sketch plane directly above it. Or place a Plane on any of the Sketch Planes that you want to go back to."

    Apart from the pain of remembering to create a bunch of planes "just in case", and having to keep them around; in this case, there never was an original sketch at this orientation.

    The sketch in the OP above, is a project-to-sketch of this:

    onto a plane defined by the highlighted edge, moved forward until it aligns with the tooth tips.

    That edge, and all 3 non-tooth surfaces -- the toe and heel faces and  the back face -- arise from intersecting cones generated by rotating lines drawn in the Z-Y plane around Z-axis.

    (Sorry. That reads like I'm telling you off for your suggestion -- I'm not -- just explaining why it doesn't help in this case.)

    On a positive note. That link is very interesting. I thought that I had probably seen every web accessible technical reference about bevel gears; but I have not seen that one, and it is the first I have seen that discusses the way the machining is performed. Kudos for finding it. (I'd love to see your search term if you remember it?)

    I'd love to read it in the original klingon!

    By which I mean, the translation from Japanese leaves something to be desired; but the pictures do clarify a hell of a lot. I think I shall spend a lot of time digesting that.  Thankyou.

    I'm not sure how much interest you have in gears -- beyond the fun of rising to my challenge:) -- but if you want to know more [insert Starship Trooper joke here], I just found this yesterday:https://www.sdp-si.com/resources/elements-of-metric-gear-technology/ ; but be warned, it has 22 sections.

    The good news, much of that is a stack of tables of standards  info that can be breezed past. The rest is the most thorough, from the ground up making no assumptions about prior knowledge treatise on gears I've yet found.

  • Jacant

    Like I said this has been talked about for years as far back as V1 


    I've already been on that site and downloaded a few gears to look over.


    Here's another link


  • Me Here

    10 hours and counting. I do hope you weren't asking us to call you an ambulance :)


  • Me Here

    Clever! But the links don't work :)


  • Me Here

    I'd looked at some of their spiral bevel gears, but they didn't have CAD available. (I've downloaded cad of SBGs from other places also only to find that they are just blanks with a few lines drawn to indicate the handedness.)

    However, I just looked at a few others and found a matching pair with CAD available:

    It was quite difficult to get the mating right because they knock the cornersoff and destroy the primary reference points. IN the end I had to move them together mm by mm until the interfered and the apart again 0.1 each until they just stopped:

    You just see the occasional tiny flashes of red, which I think means they must be pretty close.

    As and when I've refined (and possibly automated -- external macro-recorder) my method, then I will make a matching set and see how they compare.

    There are some intriguing odd facets on the teeth:

  • Jacant

    The links above now work. Approved.

    One question.

    Why do you want a 30T - 100T spiral gear set?  The ratio is a bit unusual.

  • Me Here

    I don't actually need 100T - 30T.

    I need 4 bevel gears physically arranged like this:

    You'll have to imagine the bearing that sits between the angled part of the axle, and the hub of the combined pair of purple gears.

    Input is via the shaft. The ouput is via the small blue gear. The large cyan gear is ridgidly fixed to the casing that isn't shown.

    When the axle rotates, the combined action of the forces acting at the purple hub, and the large purple gear acting on the large cyan gear, cause the purple gear to nutate. The small purple gear then drives the blue output gear.

    There are just under 3000 different permutations of the 4 tooth counts, between 20 and 100 that produce the gear ratio I am after. For practical reasons, 30T is about the smallest gear I can use and still fit a bearing between it and the angled portion of the axle whilst making the axle strong enough.

    Hence, whilst I work out how to produce spiral bevel gears reliably and consistently, I chose to mate 100T with 30T. If I can do both extremes, I should be able to do everything in between.

    The ratio I am after, is 100:1 reduction. This arrangement of 4 bevel gears -- known in the literature as a "pericylic gearbox" or "pericyclic mechanism" see https://ntrs.nasa.gov/api/citations/20180004734/downloads/20180004734.pdf ) -- can produce extremely high reduction ratios, eg. [ 98 99 100 99] & [100 99 98 99] both produce a reduction ratio of 9801:1; though they tend to get quite inefficient when you go that high.

  • Jacant

    So the gears that you want to mesh as a set have the same number of teeth.

    What is the 'angle' between the gears?

  • Me Here

    Jacant:"So the gears that you want to mesh as a set have the same number of teeth."

    No. Either set of the two meshing pairs can have equal numbers of teeth, but the other set will not have. And it can be that both sets are unequal.

    Here's a roughly representative sample of the 1782 permutations that produce exactly 100:1 (the other 1200 produce 100.something ratios):

    N1 = (large) fixed meshes with N2 = (large) nutating
    N3 = (small) nutating meshes with N4 (small) output.

    ratio formula: 1/( 1- ((N1*N3) / (N2*N4)) )

    ratio [ N1*N3 / N2*N4 ] { N1 N2 N3 N4 }
    100.0000 [ 9900 / 10000 ] { 100 100 99 100 }
    100.0000 [ 5643 / 5700 ] { 99 76 57 75 }
    100.0000 [ 2772 / 2800 ] { 99 100 28 28 }
    ... ~600 ellided ...
    100.0000 [ 2475 / 2500 ] { 75 25 33 100 }
    100.0000 [ 4950 / 5000 ] { 75 50 66 100 }
    100.0000 [ 7425 / 7500 ] { 75 100 99 75 }
    100.0000 [ 7425 / 7500 ] { 75 75 99 100 }
    ... ~600 ellided ...
    100.0000 [ 3564 / 3600 ] { 44 40 81 90 }
    100.0000 [ 2772 / 2800 ] { 44 28 63 100 }
    100.0000 [ 3564 / 3600 ] { 44 36 81 100 }
    100.0000 [ 4356 / 4400 ] { 44 44 99 100 }
    100.0000 [ 4356 / 4400 ] { 44 50 99 88 }
    ... ~600 ellided ...
    100.0000 [ 1980 / 2000 ] { 20 100 99 20 }
    100.0000 [ 1980 / 2000 ] { 20 25 99 80 }
    100.0000 [ 1980 / 2000 ] { 20 20 99 100 }
    100.0000 [ 1980 / 2000 ] { 20 40 99 50 }
    100.0189 [ 5248 / 5301 ] { 82 57 64 93 }

    jacant:"What is the 'angle' between the gears?"

    As yet undecided, as it depends in part upon the chosen permutation. The literature suggest nutation angles of between 2° and 8° are "best".

    For my purpose, smaller is better because it reduces the axial space consumed, leaving more room for the rest of the mechanism, reduces the forces on the teeth and bearings and reduces the dynamic moment of the nutating element.

    But there is a trade off in that the smaller the angle, the less room there is for the bearings.

    It is my intention to start testing with a nutation angle of 5° and adjust from there depending what comes out of the FEA.

    There is also the matter that I have played with the layout somewhat. Normally, the teeth of the nutating pair are on opposite sides of the nutating element;; whereas I have put them on the same side -- for reasons to do with the rest of the mechanism. This means that the outer diameter of the smaller nutating gear has to fit inside the inner edge of the larger one. 

    Which suggests that this set of values might be a good starting point.

     100.0000 [   5643 /   5700 ] {   99   95   57   60 }

    Given the number of possible permutations, you can see why I would lke to automate the production of the models.

    The key to constructing the pericyclic mechanism, is that the apexes of the pitch cones for both pairs must be coincident.

  • Me Here

    BTW. I forgot to mention this above and  I'm not sure if it is of interest to you, but ther eare several advantages to having the N1 == N2, and subject to all the other constaints, and testing, it is my intention that they should be so.

    Having them the same means that the nutating element ONLY nutates, and does not rotate as well. This has the advantages that it reduces the magnitude of the dynamic moments which allows smaller bearings, and greatly simplifies the calculation/simulation of those forces -- no centrifical force to account for.

    Another advantage is that it if the N1/N2 mesh forces and their teeth prove to be the limiting factor strength-wise, you can switch from (say) 100T+100Tx1m to 50T+50Tx2m  or even 20T+20Tx5m without changing the ratio.

    It also has the benfit of making it simply to adjust the ratio by only changing the tooth numbers of N3/N4.

    eg. with N1 == N2, moving to 50:1 allows N3/N4 = 49/50; or 40:1 39/40; or 80:1, 79/80 and so on.