‘Technically you shouldn’t but practically you probably can’ – falling into the same category as ‘fiddling’ the full depths / OD’s to achieve the engagements / clearances you need – which more-or-less results in the same thing.

A couple of ‘prototypes’ and your inner monitor will tell you whether the idea has legs or not..

Depending on the tooth form and application…

I made a clock train where I achieved what you want to do by using different moduli on two pairs of gears.  However the gears were profiled out by CNC from dxf files generated by this web page:

https://hessmer.org/gears/CycloidalGearBuilder.html

…where you can specify what module you want for each gear pair.

On 26 February 2025 at 08:27 HOWARDT Said:

Reduce the pcd rather than increase to achieve to required centres. Reducing gives a better tooth form, do not reduce by more than half a the module on the pcd.  I design gear trains on a daily basis for thirty years of my working life, mostly simple spur gears and never increased a pcd.

The wise voice of experience is duly noted … thanks.

MichaelG.

You would not want to make each half a tooth smaller or for that matter make the other pair each half a tooth larger. You would alter the PCD which will be proportional to the number of teeth in each gear

So as the OP rightly says the gear will in effect be a MOD1.008 so each tooth becomes 0.008 larger irrispective of the tooth count of the gear

Same applies to the other pair when making smaller alter the PCD and all teeth will be proportionally the same.

On 27 February 2025 at 06:57 JasonB Said:

You would not want to make each half a tooth smaller or for that matter make the other pair each half a tooth larger. You would alter the PCD

The tooth reference is a useful shorthand, no more. The explanatory power of thinking of the problem in those terms can be useful.

To run at common centres, the discrepancy in total tooth count between the pairs is one tooth. It may be an unconventional measure, but it is legitimate nonetheless.

Somehow that one tooth discrepancy has to be accommodated by modifying minimum one, maximum four of the gears in the train. What you actually do to the gear or gears (cutting the correct number of teeth on an altered PCD) results in a one tooth correction at the end.

That one tooth required correction could be divided into four and 1/4 of a tooth’s-worth of correction be made to each gear.

The option suggested by HOWARDT is half a tooth’s-worth of correction to the 56/73 pair (making each PCD smaller by a number that equates to half a tooth at that module).

The original question would be half a tooth’s worth of correction to the 22/106 pair (making each PCD larger by a number that equates to half a tooth at that module).

Which of the three options is (a) easiest to do; (b) diverges least from the standard situation?

Half a tooth’s-worth of correction (and an increase at that) distributed among 22 teeth is much larger proportionally than half a tooth’s-worth of (decrease) correction distributed among 56t and larger again than one full tooth’s-worth of (increase) correction to a 106t gear (which could be option 4).

The closer a gear is to a rack, the lower the effect of any correction, so with acknowledgement that it is not the ‘best’ situation, but is a quick and easy solution that is likely good enough, cutting 106t on a blank sized for 107t might be pragmatic.

On 27 February 2025 at 12:35 DC31k Said:

Half a tooth’s-worth of correction (and an increase at that) distributed among 22 teeth is much larger proportionally than half a tooth’s-worth of (decrease) correction distributed among 56t and larger again than one full tooth’s-worth of (increase) correction to a 106t gear (which could be option 4).

 

 

You are still not getting it. You do not alter each gear by half a tooth.

The alteration is that you need the total PCD to be 1 tooth different so total pcd is 128. and you nee dto make it equate to 129

Therefore alteration per tooth is 1/128 = 0.0078125 or rounded as above to 0.008

This will mean a small less than 1% change to each tooth on the small 22T gear and also a similar small change of less than 1% to each tooth on the larger 126T gear.

Now as suggested the other pair are reduced then the change per tooth is 1/129 = 0.0077519 so you are effectively cutting a MOD0.992 gear

So you turn your blanks to a diameter calculated as No of teeth +2 times MOD

So 56 + 2 x 0.992 = 57.536mm and the PCD of that gear will be 56x 0.992 = 55.662mm

and 73 +2 x .992 = 74.4mm and PCD of that gear will be 73 x 0.992 = 72.416mm

Double check the new total PCD will be 55.662 + 72.416 = 128.078 which when you allow for my rounding is the same PCD of teh 106 and 22 pair at 128mm. Half that value for the shaft ctrs.

Is this a metric to imperial conversion gear train as it looks like it works out at 1mm = 1/16″

Profile shifting not only moves the gear tooth relative to the original PCD it also changes the shape of the tooth. For a positive shift, making the addendum bigger than the dedendum, the shifted tooth shape is thickened at the root and narrowed at the top. This is equivalent to the tooth profile of a gear with a higher tooth count than the gear in question.

A positive shift is often used where undercutting would occur. The shifted profile is equivalent to the tooth on a higher count gear, so the undercutting is eliminated. The shifted profile tooth is also stronger in bending.

It is important to note that profile shifted gears are usually manufactured using a generating process, normally hobbing. Making the hob cut deeper automatically generates a shifted profile. Cutting deeper with a standard involute cutter does not create the proper tooth profile, although in practice it might be acceptable for low speed, low power applications.

Julie

Ummm. there seems a of complifying going on here so I apologise if I have I missed something….

Is the ratio so critical that the 73T wheel can not be 72 instead? Or one of the first pair increased by 1 tooth?

Either way will then give equal tooth-sums, which is what you need.

Having had a quick look at teh video it would seem the best option is to download the new and old versions of Gearotic and look up the spec for those particular gear pairs and see if he altered the MOD or used shift.

Though with the longer tooth of the clock type gear form I wonder if you could get away with the gap as the drive is always in one dircection and it is certainly not a heavily loaded or fast application.

Please forgive the digression

Three short videos … well worth watching

https://www.youtube.com/playlist?list=PLVBMUkXqu-olm5aY1snFgTsAliVQ_8hI0

MichaelG.

Nothing wrong, profile shift does not alter the theoretical value of the PCD of the gears. The addendum and dedendum change and hence the outside diameter of the gear. Also the close mesh centre distance can be changed with profile shift.

Neil

For the 22 to 106 tooth gear a profile shift of about +0.51 applied to one gear or shared between both gears should result in a centre distance 64.5mm, the same as the 56 to 73 pair. Someone might like to check my calculation just in case I can’t remember how it’s done, it has been a long time since I needed to do such things.

I’ve just noticed that the calculations that Jason did in the Gearotic were for an epicycloidal tooth form. My profile shift figure was calculated for an involute tooth form which will give different dimensions.

The PCD or reference diameter as it is sometimes call, remains the same.

Neil