What's new
What's new

Electronic gearing for gear hobber

This sort of ignores the dynamics of gear hobbing tho. How do you control speed with a stepper ? The process of hobbing teeth is like a very interrupted cut at slow speeds. Maybe at 32DP this won't be a problem but with bigger teeth, the hob motor has to be actively controlled. Mechanical h

Hard to tell from the glitzy video but doesn't look like hobbing at all .
If you watched the whole video you would have seen they roughed the teeth first Then machines the surfaces where they were bolted together Why ?? Because the joint opens up about 15cm when the unbolt them Just then they start hobbing Hobs with special made carbide inserts I have insert on my desk About 15x15x100mm
Then they place the pinion in a temperary set up to see how it bleus Like you see it bleus not all the way The pinion is ground covex to prevent exces side load

Peter
 
Last edited:
There are plenty of Servo/Stepper (I would opt for the Servo Motor every time) control card compatible with a PC Bus or stand alone controller that have Electronic Gearing, Galil being one that I've used often on various projects.

Gear hobbing would be a rather simple application. An example of the use of a Galil 2 axes card I applied to a 3 axes machining centre, the client asked about having their 3 axes, 1986 vintage Mori with a Yasnac control equipped with a forth axis. The machine was not four axis ready and I told the client they could buy a whole other machine for less money than to have a true forth axis integrated with the machine they had.

They needed the forth axis for a few specific jobs, each were only cutting a helical path around the OD of a part. Although we could have piggy backed off of the existing X axis encoder (half the machine would have had to be dismantled to get to it), it was deemed easier to mount another encoder on the accessible end of the X axis lead screw. This would be the Master Encoder.

A stand alone controller was built and connected to the Master Encoder on the end of the machine's X axis lead screw and to the Fourth Axis Encoder, which was the slave. As far as the Machine was concerned, it was oblivious of the existence of the Forth Axis, but when Electronic Gearing was engaged via the stand alone controller, wherever the X axis of the Machine went, the Slave forth axis would follow at the Gear Ration specified

Following is an example of one of the jobs machined on this machine using the Stand Alone Controller and Electronic Gearing. All without the Machine's Control knowing that a Forth Axis existed.

Regards,

Bill

Cam3.JPG
 
Last edited:
There are plenty of Servo/Stepper (I would opt for the Servo Motor every time) control card compatible with a PC Bus or stand alone controller that have Electronic Gearing, Galil being one that I've used often on various projects.

Gear hobbing would be a rather simple application. An example of the use of a Galil 2 axes card I applied to a 3 axes machining centre, the client asked about having their 3 axes, 1986 vintage Mori with a Yasnac control equipped with a forth axis. The machine was not four axis ready and I told the client they could buy a whole other machine for less money than to have a true forth axis integrated with the machine they had.

They needed the forth axis for a few specific jobs, each were only cutting a helical path around the OD of a part. Although we could have piggy backed off of the existing X axis encoder (half the machine would have had to be dismantled to get to it), it was deemed easier to mount another encoder on the accessible end of the X axis lead screw. This would be the Master Encoder.

A stand alone controller was built and connected to the Master Encoder on the end of the machine's X axis lead screw and to the Fourth Axis Encoder, which was the slave. As far as the Machine was concerned, it was oblivious of the existence of the Forth Axis, but when Electronic Gearing was engaged via the stand alone controller, wherever the X axis of the Machine went, the Slave forth axis would follow at the Gear Ration specified

Following is an example of one of the jobs machined on this machine using the Stand Alone Controller and Electronic Gearing. All without the Machine's Control knowing that a Forth Axis existed.

Regards,

Bill

View attachment 379883

You're awesome Bill! You have the best answers to my strangest questions.

So in your experienced opinion, what hardware would be needed to slave the work table rotation and the hob vertical feed to the hob spindle?

Thanks!
 
I had a dream that I was on Practical Machinist.
But it seemed I was in the xxx zone.
I am a big Gail fan using them forever from before the IBM PC invented.
Motion Guru was a Delta-Tau fan and those great also but now part of Ormon. Seems he went away from these.
So much information here. How to sort good from bad?
I have lots of (50) years in motion control but admit to not knowing much.
As time goes by you get that the job is is to train your own assassins. So the new should never be cast aside or ignored.
Bob
 
Last edited:
Kind of, yeah. Basically Gar wants to make a wore-out south bend into a trick cnc lathe.

Nah. I like your "worn out Southbend" comparison though!

I was thinking more just the electronic gearing aspect. Not really CNC, just replace the gears to the table and vertical feed with servos and drives that can do the same thing. A simple UI for each axis where you just enter the ratio you desire. So the worn out Southbend of hobbers still works the same as it always did, except I guess you don't need gear calculators to finger out your fed gear combinations. You just type in a number and it happens.

That's the scope of the upgrade I had in mind for the old DB supreme.
 
Not really CNC, just replace the gears to the table and vertical feed with servos and drives that can do the same thing. A simple UI for each axis where you just enter the ratio you desire. So the worn out Southbend of hobbers still works the same as it always did, except I guess you don't need gear calculators to finger out your fed gear combinations. You just type in a number and it happens.

That's the scope of the upgrade I had in mind for the old DB supreme.

Speed and feed gears are so easy that you'e not saving anything ... but if you go to all that trouble, you really want to make it do two-cut. I mean absolutely without question, that's the most usefullest thing you can ask it to do.

Then possibly do hob shift. The math is a bitch but continuous hob shifting does wonders for hob life.

Crowning would be useful as well :)

In the year 2525
If Garwood's still alive ...
:D
 
Then possibly do hob shift. The math is a bitch...
On a mechanical machine sure.

Electronically it's easy.

Add distance x to the hob's path along it's axial rotation axis, rotate it backwards or forwards to compensate for the extra rotation, otherwise you turn the gear into a helix of a completely wrong tooth shape.


For a given spur gear, gear turns xx number of times, hob turns xx*gear teeth number of turns. Hob slides on it's axis x number of extra turns, while moving backwards along it's axis x number of turns * pitch distance. Hob moves through the width of the gear blank x distance.

For the open loop system I intend to build, I am limited to a 4 axis drive so the depth of the hob into the gear blank will be adjusted manually.

When setting up the machine, the step count to the steppers is set so that one unit on your control is one turn of the hob.

For the hob shift, you can change the number of steps per unit each time you change the hob, so one unit on the controller is always one pitch distance shift. For the helical gears there will be some math required to link the x distance through the width of the gear blank to the extra partial turn added to the gear blank.


So for a hob shift of 3, a spur gear of 10 teeth looks like this.

0,0,0,0 > 500 gear blank turns, 5000 plus or minus 3 hob turns, x distance for the gear width, 3 units for the hob shift (get the direction correct).

G code looks like 0,0,0,0> 500,5003,3,1

For a helical gear of a full tooth shift, 5000 plus or minus 3... becomes 3 Plus or minus an additional .1 turns for a full tooth shift across the width of the gear on a 10 tooth gear. A 50 tooth gear would be plus or minus .02 turns.


Anyhow the reason I want to do this is because with some extra lines of code and a 5th axis to adjust the gear tooth depth, you can use a single point cutter and use the hob shift to move through each tooth. It only takes on the order of 10 to 20 passes.

I'm thinking to put a 5 to 10 pound flywheel on the stepper driving the hob through a gear driven reduction, for a stepper of a couple hundred watts shaft power. It's not hard to calculate the acceleration limits. So the hob is going to see enough inertia to break it. Rpm would be set for the maximum shaft power, you then work backwards to find the feed rates.

So in my example, to reduce the feed rate by half you would use 1000, 10003, 3, 1
 
Last edited:
On a mechanical machine sure.

Electronically it's easy.

Mechanically it's easy too, but you still have to do the math. I admit, we're not talking calculus here but it is a bit of a head-scratcher for us shop rats.

For the open loop system I intend to build, I am limited to a 4 axis drive so the depth of the hob into the gear blank will be adjusted manually.

That's what I'd suggest Garwood not do because two-cut and crowning are very valuable abilities when making actual gears as an actual business. Two-cut in particular.

Anyhow the reason I want to do this is because with some extra lines of code and a 5th axis to adjust the gear tooth depth, you can use a single point cutter and use the hob shift to move through each tooth. It only takes on the order of 10 to 20 passes.

That's called tangential hobbng, it's useful for a one-off wormgear but for making parts numbering one or more ? It's crazy. A 30 tooth gear will take you 30 times s long as using a hob. You could call it the Rip Van Winkel Method :D
 
So splain me more about this two cut...

This is fancy gear speak for rough and finish pass or am I misunderstanding? Purpose is better final finish? Less wear on the hob?
 
So splain me more about this two cut...

This is fancy gear speak for rough and finish pass or am I misunderstanding? Purpose is better final finish?

Exactly. Being able to do it automatically is wonderful. It's so wonderful that most manufacturers designed ways to do it mechanically, before electronic machines. It's a very desirable feature.
 
That's called tangential hobbng, it's useful for a one-off wormgear but for making parts numbering one or more ? It's crazy. A 30 tooth gear will take you 30 times s long as using a hob. You could call it the Rip Van Winkel Method :D

You can call it what you want but in terms of number of facets or scallops that the gear tooth is going to be cut by, a single point cutter will replace a hob
provided you take 8 to 16 times as many passes, it has no relation to the number of teeth other than the fact that fewer teeth is less metal to remove. suppose you use a computer to calculate the depth to make the cuts equal area so the machine can cut the teeth out quicker, taking deeper cuts later in the program. the tooth can still have an involute profile approximated by 16 sides of the polygon, corresponding to a good quality hob. cheaper ones might be only 8 sides. yes the gear will supposedly have 3 or 4 points of contact with the hob at all times.. but will it really?

it takes a certain amount of pressure for a cutter of given sharpness to shave metal off of anything. if the gear cutter isn't infinitely sharp, you can only assume it will have 2 points of contact, not 3 or 4, cutting the gear blank. if the hob is a spiral and segmented 8 times radially, its no better than 8 single point flycutters when it is dull.

anyone climb milling with a gear hob? why not?

i will admit that even with initial roughing passes you still need a full 16 passes to approximate the involute curve so you're not completely wrong with the idea that it would take 30 times as long, but if you're going to argue that, then you're going to run the machine slower and hope the gear ends up more accurate.. when instead you're still geometrically limited by the number of cutting teeth on the hob and friction to remove metal.

with regard to friction.. we got grinding machines for that. leave the hob for the roughing operations.
 
Last edited:
with regard to the cnc gear hobber i want to make.. rather than take an 8 radial segmented hob and make the gear supposedly more accurate by slowing the feed rate a lot.. instead you can run it at an average speed, finish the gear, return to zero, slide the hob one 1/16th of a tooth (1/8/16th of a turn) and run it again from the beginning. i suppose in theory that hob shift will continuously take care of this problem, but it will still result in a helix angle where the scallops of each tooth of the hob will be in a different place on the gear face. if instead you start the machine over again with the hob set half a tooth off, you'll get the most reduction in ripple so to speak,

all of the scallops will now be cut in half. the problem of course is you won't actually remove any metal on the second run because there will be so little to remove that you'll never overcome the minimum cutting pressure and the hob will just rub on the gear. maybe you could climb mill on the second run, or run without coolant, so the coolant doesn't get in the way and push the cutter away from the work. but changing the direction reverses the forces in the entire system which means you have to zero the axis from the other direction.

and yes i'm really talking about reversing the relationship of the hob to the gear tooth by 1/30th or 1/16th of a turn (of the hob) and running the machine "backwards".

don't bs me and say a stepper can't do that. if i'm going to make a stepper driven hob, i have by default 200 steps per turn of the stepper, reduced at least 1:10 to get to the hob.
you could run the machine again 1/200th of a tooth off and selectively take some metal off of just one side of all the teeth if you want. this wouldn't increase the depth of the tooth but it would decrease the pitch diameter and clean up the face of the tooth you were driving the hob into. the other side of the teeth wouldn't be touched.
 
Last edited:
You can call it what you want

It's not what I want, it's what machine tool manufacturers have called it for at least fifty years. Gould & Eberhardt even built machines specifically for cutting worm gears designated TWG-30 (for example). That's tangential worm generator - 30".

It's just normal nomenclature in the industry.

if the gear cutter isn't infinitely sharp, you can only assume it will have 2 points of contact, not 3 or 4,

No, they cut in a line. The whole edge on each side of a hob tooth cuts.

anyone climb milling with a gear hob? why not?

Pfauter loves climb cutting. Everyone else is like, "use it when it makes sense." Conventional cut almost always works, climb is like on a mill; sometimes it's good, sometimes the situation is too shaky.

i suppose in theory that hob shift will continuously take care of this problem, but it will still result in a helix angle

The point of continuous hob shift is cutter wear, not accuracy or anything else. Several manufacturers promote the practice for this reason but as you point out, it results in a spur gear becoming a helical. So you have to treat the simple spur as a very long lead helical of the opposite hand for the part to come out correct.

Continuous shift does work well; shifting in increments between parts does not do as good a job but most people avoid continuous because it makes a simple job complex. Also most hobbers don't have it. It's most common on automotive or other high-production machines, for this reason.

all of the scallops will now be cut in half ...

That's not generally a big concern.

EG... walk away. :)
Much messy.

It's okay, not too concerned about the mess, they'll do as they please anyhow but there's some general info that could help which we can get into ... I think it's not the best use of Mr Garwood's time but it's his time so what the heck. And at least he's not starting with a drill press or lathe or ancient k&t or some other abortion.

Besides, I'm in prison for seven more days and there's nothing else to do ... except wonder how that stick can go from the back of the nose directly to the brain. I'd rather think about hobbers :)
 
Last edited:
@Garwood
Let me write a few thoughts about turning a hobbing machine into a CNC hobbing machine.
1. Making a CNC machine from a non-CNC machine is a very, very thankless task. It seems to me that this only makes sense for some rare, unique and / or very expensive machines.
BUT! This applies mainly to conventional lathes or milling machines. The hobbing machine is very different: the processes are slow, interpolation as such is practically non-existent.
2. New gear machines are expensive. Very expensive. According to my information, the cost of the new Liebherr, which can machine gears in various ways, is approaching a million dollars. So the modernization of the gear-cutting machine, in principle, can be relevant.
3. You need to think about why you need to turn your machine into a "CNC" at all. Or in CNC. Making parts faster? Get rid of a lot of manual operations so that the machine works by itself? To give the machine the ability to make cool and complex parts that it cannot do now? Cut costs?
4. Even to build a very cool gear hobbing machine that can change mills itself, a full-fledged CNC system is not at all necessary - a good PLC is enough.

I have dealt with several hobbing retrofit projects. At a minimum, as a result, the time for equipment changeover has been reduced - it is no longer necessary to install 8-10 gear wheels. As a maximum, the owners received a more accurate machine.

I have never seen a project where someone would make a machine "like the new Liebherr". One team was developing mechanization for tilting the milling caliper, but they did not seem to have reached the implementation. No one even thought about the direct drive of the rotation of the table, as well as the direct drive of the rotation of the mill.
That is, usually people leave the rotation of the main spindle from a conventional asynchronous motor, and the table rotation drive - using a worm gear.
You can install some kind of controlled motor to rotate the worm of the worm gearbox of the table drive. And you can install the same motor to turn the vertical movement screw of the milling carriage.
As a result, you will already have a machine that does not require the installation of gears. You can cut any number of teeth, any angle of inclination of the teeth - restrictions are only on the geometry of the tool.
I will say what else I saw in the implemented modernization projects.
-Encoder on the table for precise rotation of the part. Let's say you can calibrate the error once, enter it into the correction table and not install the encoder.
- A linear encoder to control the movement of the milling caliper and another linear encoder to control the radial movement of the table, this was in the same project. Let's say it might be cheaper to put a ball screw in place of a conventional trapezoidal lead screw and ditch the linear scales. The result, together with precise control of the table rotation, is a very precise match of the tooth angle to the drawing.
- Mechanization of movement of the spindle with the mill along its axis, for uniform wear of the mill. Movement does not occur during cutting.
- Encoder on the milling spindle for machining worm wheels with a single cutter. For this, in the pre-CNC era, "tangential calipers" were needed.
 
Last edited:








 
Back
Top