High Helix Angle Crashing Hobbs - Is feed speed my issue? - Page 2

# Thread: High Helix Angle Crashing Hobbs - Is feed speed my issue?

1. Originally Posted by David_M
I worked out the one shown here
Ahhh, I see. I apologize for any confusion. The sheet you just linked is an old one I threw up just as an example of my "knowing" that an approximate 80° helix was possible on these machines because that one had been done successfully before.

I was comparing your math to the one I was doing for my 81.1° helix with 3 teeth at 38 NDP and I the two were not meshing and it was throwing me off.

I guess at this point my rank amateurism is showing. In post #13 for my target 81.1°, 3 tooth, 38 NDP gear, you stated that the C constant was "32." I think I may be misunderstanding the terminology here. As I understand it the C constant is the Feed Constant which is calculated as NCP / (Feed Rate X SIN 81.1°) which, for my math, comes out to 41.84051076. In post #17 you said that C would be 41.8421052631... And I'm trying to figure out that discrepancy.

You stated that my index change gears don't match and you mentioned that I won't catch the error on a double check. Why exactly is that? The only "error" I can find so far is that I had the 7th decimal of PI as a 1 instead of a 2.

Am I making sense? Thank you again for your patience and help.

2. NOTE: I put a reply in earlier but it's pending moderator approval because, I assume, it had a link it it. Feel free to ignore that one if its confusing as I jumped the gun on making that reply.

I think I'm following you pretty well, but I want to clarify some terminology.

You reference the example sheet's "C Constant" as 21.333...

I see where you get that number by following your math. But the number you've derived as your "C constant" is not the same number that my formula generates which is, again, NCP / (SIN Helix X Feed Rate). In the example we are parsing I come out to 39.78164871.

Are we using C constant to refer to two different things? I see how you're breaking down the math and turning 21.333... into the index gear selections, but your C constant is 21.333 and mine's 39.someething

What's the discrepancy here?

3. Cast Iron
Join Date
Sep 2014
Country
UNITED STATES
State/Province
Georgia
Posts
381
Post Thanks / Like
Likes (Given)
286
147
I calculated the 21 1/3 from this example:

You are, I'm sure, referring to the 81.1, 38 diametrical pitch one. I'll sync back up.
Are you interested in a free program that does what you need? If so, what is your computer's operating system? 64 bit or 32 bit? It should be a good learning tool to help with your spread sheet.

4. I would LOVE any program that can help me out. I'm on Win 10 64 bit. Beats the hell out of a cobbled together excel workbook.

That 39.etcetc number was my C constant from the old practice sheet. If I run the math for my 81.1 TARGET gear, I come up with 41.84051076.

NCP / (Helix SIN * Feed Rate)
0.082673491 / (0.987959866 X .002) = 41.84051076
Earlier in the thread you had come up with 41.84210526... That's where I'm still trying to figure out my problem. Even after fixing my Pi error on the 7th decimal, they still don't match.

5. Cast Iron
Join Date
Sep 2014
Country
UNITED STATES
State/Province
Georgia
Posts
381
Post Thanks / Like
Likes (Given)
286
147

6. This is excellent, thank you.

All the numbers it's outputting match my spreadsheet to within 9ish decimals, so I think it really addresses a lot of my insecurities and questions.

At this point, I think there's only two outstanding issues, one of which is more curiosity. Under the "Result" section of the program it lists C as 2107 / 48. Is that the Feed Constant whose formula I expressed above, or is this another constant that is, unfortunately, named the same as the "C Constant" that I expressed above with the NCP / (Sin ψ X Feed Rate)?

As for the last standing issue, which I think is also resolved, I'm probably crashing hobs because I'm going wayyyy too fast. Dropping the feed rate to .002" or less should solve it in your estimation, yes?

Thank you again for all your assistance. From time to time we need gears cut that we do not have the machinery for. Does your business do small runs (10-500 pieces)? I may have a particular item I need quoted soon if I win a particular bid, if you might be interested.

7. Cast Iron
Join Date
Sep 2014
Country
UNITED STATES
State/Province
Georgia
Posts
381
Post Thanks / Like
Likes (Given)
286
147
Jason said,

"NCP / (Helix SIN * Feed Rate)

0.082673491 / (0.987959866 X .002) = 41.84051076

Earlier in the thread you had come up with 41.84210526... That's where I'm still trying to figure out my problem. Even after fixing my Pi error on the 7th decimal, they still don't match."

I came up with the same C const that you did within a small rounding error.

Since there was no gear set that exactly matched that C const, I had to take the closest gear set and from that point forward use its unique C const. But that's okay because all that will happen is the feed rate will change slightly because it is calculated using the new C and not the original.

edit: I see that in an earlier post I said your C was wrong. That was a poor choice of words. I should have said that your C didn't match the gear set and 41.84210526... does.

8. That makes a ton more sense. Thank you again for all your help. I've got a few posts pending mod approval, but you've already addressed them all.

Going to pour over all my numbers and setups again and take it slowly before I attempt the cut again.

9. Titanium
Join Date
Sep 2005
Location
Oakland, CA
Posts
2,644
Post Thanks / Like
Likes (Given)
370
744
Yes you can hob a 3 tooth gear on a 6-10, but I hope your machine is a triple start machine or you are going to flog the index worm.
Remember the feed value is per work piece revolution, so divide your feed rate by the number of teeth you are cutting and that's the hob advance per tooth of your blank IN THE AXIAL DIRECTION. In that 1 workpiece revolution, the hob has made 3 revolutions- 1 for each tooth. Now take your feed per tooth and divide that by the # of gashes in the hob- that # is your axial feed rate per hob tooth.

Keep in mind that .020 feed is probably a good feed (might be a little light) for a 10-30 tooth spur gear of that pitch. Calculate the chip load for that and you will see why its WAY to high for your helical gear with 3 teeth.

Now go back and read the hobber manual and calculate the actual feed rate for a helical gear and you will find that your federate needs to drop even further because of the high helix angle in your job. From memory- the normal feed rate- the feed rate the hob teeth see- is the axial feed rate/cos helix angle. Whats the cosine of 80 degrees- probably less that .2 or so

From this you should be able to get your gears cut without using up your hobs. Another hint- look at your chips! they must be relativly huge!

10. Cast Iron
Join Date
Sep 2014
Country
UNITED STATES
State/Province
Georgia
Posts
381
Post Thanks / Like
Likes (Given)
286
147
6 starts, i think.

11. Cast Iron
Join Date
Sep 2014
Country
UNITED STATES
State/Province
Georgia
Posts
381
Post Thanks / Like
Likes (Given)
286
147
Originally Posted by JasonGGI
This is excellent, thank you.

All the numbers it's outputting match my spreadsheet to within 9ish decimals, so I think it really addresses a lot of my insecurities and questions.

At this point, I think there's only two outstanding issues, one of which is more curiosity. Under the "Result" section of the program it lists C as 2107 / 48. Is that the Feed Constant whose formula I expressed above, or is this another constant that is, unfortunately, named the same as the "C Constant" that I expressed above with the NCP / (Sin ψ X Feed Rate)?

As for the last standing issue, which I think is also resolved, I'm probably crashing hobs because I'm going wayyyy too fast. Dropping the feed rate to .002" or less should solve it in your estimation, yes?

Thank you again for all your assistance. From time to time we need gears cut that we do not have the machinery for. Does your business do small runs (10-500 pieces)? I may have a particular item I need quoted soon if I win a particular bid, if you might be interested.
I'm glad the program is working for you!

Please read Dan from Oakland's post #29 for his explanation of what's happening with the feed. He has the most knowledge about what you are doing and would be whom I would contact about getting your jobs done. Most highly recommended!!

I have a shop, but I'm not involved in gears. My interest in the formulas and programming is left over from a previous job where they had two G&E hobbers.

Let me start a fresh post to explain about the C constants...
Last edited by David_M; 03-26-2020 at 02:45 PM.

12. Dan,

Thanks for breaking it down. When I had first set the feed rate I hadn't really stopped to consider the full implications. I've found some similar gear sheets in my archive and the lowest I've found so far is a .00375. Which I'll use as my starting point for this one, but I'll do the math as you explained first to ensure that I'm not still too high.

My BC 6-10 is a six threaded index worm as David's image shows.

Thanks!

13. Cast Iron
Join Date
Sep 2014
Country
UNITED STATES
State/Province
Georgia
Posts
381
Post Thanks / Like
Likes (Given)
286
147

Barber-Colman refers to the C const that is produced from:

as the theoretical value of C.

You can round it off to a whole number, or search for the closest set of gears (chances are against you of finding an exact match).

at this point, my method is to run this formula on the returned set of gears:

(This version of the formula only works if the Index constant is: 6; Hob threads: 1; and Number of teeth: 3)

where x is the product of your drivers and y is of your driven gears.

This is from a quickly written program:

It shows the proof that the new C const is exact for the index gears. Take the new C const with this formula:

to get the feed gear ratio (divide the above by the feed constant) :

But, this only returns 5 decimal points of accuracy, just better than BC's cutoff of 0.000015.

This is 4 decimal places better:

This program compares every conceivable C const and returns the best result.

Details:

The error that you got in the index gearing is: 0.0000006221332978
That is comparing apples to oranges (imho). When the C const is changed to make your index exact it puts the error in the feed gears. An error of: 0.00000203234487. This is what I was talking about you not getting an accurate 'double check' (poor choice of wording on my part, again).

BC says you can have an error in the index (most people frown on it, though). I just don't know how to quantify it. Plus, it is too easy to always make the C exact.
Last edited by David_M; 03-29-2020 at 12:52 AM.

14. Titanium
Join Date
Apr 2018
Country
UNITED STATES MINOR OUTLYING ISLANDS
Posts
3,969
Post Thanks / Like
Likes (Given)
0
1794
Originally Posted by David_M
BC says you can have an error in the index (most people frown on it, though). I just don't know how to quantify it. Plus, it is too easy to always make the C exact.
An error of one or two in the fourth place is fine. You can even stretch that, but it's a good goal. And they don't frown on it, it's in the manuals. There's more manufacturers than just Barber-Colman.

You're still barking up the wrong tree. As Dan knows, feedrate isn't everything

However, in this case, due to the narrow angle between the hob and the workpiece, the first tooth is slamming in there full-depth, which is not normal for hobbing. A quick look at those photos looks like the hob needs to be shifted farther out of the cut so that it can ease in and out of the workpiece, at least a little. It might be nice if you could infeed first, then kick the longitudinal feed lever, to at least get most of the material out of the way with more than one tooth of the hob doing the roughing.

And the arbor sucks. No matter what you do it's going to want to quiver, and when it does you lose teeth on the cutter.

This is really a part for cnc turning or a thread mill. A thread grinder would gobble these up and give nicer parts, too. Doing it on a #3 is about the most difficult way you can think of.

15. Cast Iron
Join Date
Sep 2014
Country
UNITED STATES
State/Province
Georgia
Posts
381
Post Thanks / Like
Likes (Given)
286
147
Originally Posted by EmanuelGoldstein
[T]hey don't frown on it, it's in the manuals.
Are you saying that the Index gearing has to be exact or that it can have an error and that it is acceptable if it is in (a) tolerance?

Yea, if I had to do it like the OP I would have the hob made with the first several teeth looking like the first teeth on an Acme tap. G/D/R

16. Titanium
Join Date
Apr 2018
Country
UNITED STATES MINOR OUTLYING ISLANDS
Posts
3,969
Post Thanks / Like
Likes (Given)
0
1794
Originally Posted by David_M
Are you saying that the Index gearing has to be exact or that it can have an error and that it is acceptable if it is in (a) tolerance?
Can be off .0001 or .0002 without noticing a difference. I've done even more if it was unavoidable but don't like to. On a helical I always tried to wind them up tighter because they unwind in heat treat.

Of course normally on a spur you're not going to be doing that, index gears should be exact. But a little bit off is not the end of the world in practice for stuff that's less than AGMA 10-ish. There are so many other errors going on that that tiny bit is going to be buried.

More info would be in the Cleveland manuals (cash) and maybe Pfauter machines with 'minch'. That's where it is common.

Yea, if I had to do it like the OP I would have the hob made with the first several teeth looking like the first teeth on an Acme tap. G/D/R
Tangential-feed worm gear hobs are made that way.

But if I had to buy a hob, I'd just send the parts to a thread grinder. Or a cnc lathe.

You could probably make a nuclear submarine on a Bridgeport, too, but why would you want to ?

17. Titanium
Join Date
Sep 2005
Location
Oakland, CA
Posts
2,644
Post Thanks / Like
Likes (Given)
370
744
Just a few comments. Not sure, but I would imagine that 6 start Barber Colman 6-10s are pretty rare beasts.
One thing that I did not mention and EG mentions it above, with a standard hob and that high helix angle, the 1st tooth on the hob is pretty much doing all the roughing and actually all but a tiny portion of the finishing as well. Having tapered hobs made would be best, as it spreads the roughing over multiple teeth but this is not an inexpensive or timely solution. Because of the helix angle I think you would need a hob about 12" long to be able to shift the cutting zone off the end the hob and even if you could, the hob approach time would be kill you.
Just another reason to lower your feed rate on jobs like this. Keep an eye on that 1st hob tooth as you run the job and keep the hob sharp.
Regarding index gears, I simply make it a point to always select a feed constant that gives me an exact index ratio and then minimize as much as possible the feed gear error.

You don't mention if you have a right angle hob head for your machines, but thread milling is certainly a viable method for that fine pitch. You would need to make an work arbor that could be indexed for the 3 teeth.
If all you have is your 6 start machines, the challenge would your index ratio as the machine would need to be geared for about 300 teeth. In this case, the index ratio controls the chip load and the feed gears are calculated for the lead of the tooth. Then you simply use a space cutter in place of the hob. Hope this helps. Gotta go sneak to the shop and cut some gears for a few hours.

18. Originally Posted by David_M
Very helpful, thank you. I fell like I'm truly starting to grasp the whole process mathematically. Maybe...

Originally Posted by Dan from Oakland
You don't mention if you have a right angle hob head for your machines, but thread milling is certainly a viable method for that fine pitch.
Unfortunately, what I have is what you see in the pictures I posted. I have one more 38 DP left handed hob left and 5-6 38 DP right handed hobs that have never been used.

Also, please correct me if I'm wrong, but from what I gather hob shifting requires a hob mount that allows for movement of the cutter laterally along the axis of the work. I don't think I can do that with the machine I'm using. Or am I woefully under-educated about my BC 6-10?

These comments are all extremely helpful, so thank you again. As I said, I am NOT an engineer or machinist of any kind. I'm just a self-taught upstart desk jockey who's in over his head here.

19. Titanium
Join Date
Apr 2018
Country
UNITED STATES MINOR OUTLYING ISLANDS
Posts
3,969
Post Thanks / Like
Likes (Given)
0
1794
I can assist with this part ...

20. Titanium
Join Date
Sep 2005
Location
Oakland, CA
Posts
2,644
Post Thanks / Like
Likes (Given)
370
744
Jason- I don't see any pics you posted so. . . . David posted a pic of a right angle hob head in post #8. Regardless, because of the high helix angle your hob spindle is almost parallel to the work axis, so you will gain very little with any hob shifting. I would not try to cut your left hand parts with a right hand hob at this helix angle as any backlash in your index will kill you, as you would be essentially climb milling, and I don't think your setup would be rigid enough for that. Can you try to post a picture or two of your setup so we can review?