Backlash you'd accept to interpolate a good circle

jools · Jul 30, 2020

Hi All

I'm in the market for a VMC (smaller footprint one) and waiting to see if my finance application goes through to get a new Haas Mini Mill. If it doesn't then the options become narrower and it will be a second hand machine.

So the question is what backlash would be too great to interpolate a good circle?

Backstory
All parts Delrin and some of these circles will be o-ring grooves. Equipment used in diving gear so watertightness is paramount.
I'm just stepping up from my hobby type mill which I've been learning on for the past few years. It has backlash and because of the lack of mass the z axis tends to rattle and give a fairly crap surface finish, faceted rather than smooth. With the considerable mass in the tables of these machines I'm wondeering if I'd possibly have the same problem or if I'm thinking ahead of myself.

Thanks

Jools

Toolroomguy · Jul 30, 2020

I always like using a boring head and a custom ground single point cutter.
You can have the radii at the bottom of the groove, as recommended by o-ring manufacturers, as part of the tool geometry.

jools · Jul 30, 2020

That is a good way of doing it but the oring grooves need to be cut into that machined circle face radially. Other option is a 4th axis but really wanted to program and forget as such.

As an aside: If you have backlash doeas the table not move as the boring cutter engages at the points round the work piece? Or is the mass of the table sufficient for this not to be an issue?

Jools

dandrummerman21 · Jul 30, 2020

Just to be clear, is your o-ring groove on the face of the part, around the hole, or is it inside the hole (ID groove)?

You said "machined circle face radially", "face" meaning "top" to me, but "radially" meaning "inside" to me.

Just about any new or used mill should be orders of magnitude better than any hobby machine you might encounter. I'd say it doesn't even have to be *that* round to get a good seal on an o-ring.

You could easily find something for $10k USD that would cut a groove in delrin with enough accuracy.

Kingbob · Jul 30, 2020

I agree with dandrummerman21. Your hobby grade machine must be in pretty rough shape or incredibly weak not to be able to produce a good oring groove in delrin. O ring grooves come in all shapes and sizes from round to square to any shape imaginable, what's important is the surface finish of the groove itself not necessarily the roundness, the oring can compensate for several thou of eccentricity as long as it is sealing well against the walls. A new or even a decent used mini mill or even a haas tool room mill will do this no problem.

implmex · Jul 30, 2020

Hi jools:
The real question is; what do you consider to be a "good" hole?
For oring grooves, milling is always fraught with potential sealing problems, and the difficulty arises from the "lay" of the machined surface.
A turned surface always makes concentric rings on the bottom and sides of the grooves and the oring will squash its way into the grooves and seal.
Milling makes swirls on the bottom and ripples on the sidewalls, both of which are much harder to force the oring material into reliably.

So in many high end sealing applications milled oring grooves without modification of the lay of the machine marks are not considered acceptable regardless how well they are executed.

The roundness of the groove is irrelevant to that function...so long as the perimeter is the correct length and the lay is good, the oring will work.

Having said that, my vintage 2001 Haas Minimill will still make a round interpolated bore that clocks in around 0.0007" TIR with care.
That's good enough for a lot of things that are supposed to be round...I've milled round mold cores with shutoffs and gotten away with it with a bit of blueing and fitting.
A top of the line Hermle might make the same feature within a tenth.
My Minimill cost forty grand: that Hermle is probably the best part of a million bucks.

For low tech oring grooves 0.0007" is more than adequate and I can get acceptable sealing performance without worrying too much about the lay.
But certainly NOT for any critical applications like high vacuum or high pressure or helium containment or...

Cheers

Marcus
Implant Mechanix • Design & Innovation > HOME
Vancouver Wire EDM -- Wire EDM Machining

jools · Jul 30, 2020

dandrummerman21 said:
Just to be clear, is your o-ring groove on the face of the part, around the hole, or is it inside the hole (ID groove)?

You said "machined circle face radially", "face" meaning "top" to me, but "radially" meaning "inside" to me.

Just about any new or used mill should be orders of magnitude better than any hobby machine you might encounter. I'd say it doesn't even have to be *that* round to get a good seal on an o-ring.

You could easily find something for $10k USD that would cut a groove in delrin with enough accuracy.

Sorry my bad. Orings cut radially, like piston rings on a car.

It doesn't have to be perfectly round as you say, the rubber will deform into it within tolerances. On the hobby mill the cut turned shit at the digonals of the X and Y axis as the x and y motors slowed down. It looked gash

and was to do with backlash and lack of rigidity.

Just wanted to make sure if a second hand VMC had backlash I wasn't jumping from frying pan to fire and spending a few £k to possible achieve the same end result. My thoughts are the mass of the table would help but then I know shit about large VMC's :-)

Jools

Ziggy2 · Jul 30, 2020

I would guess that the OP's issue is that he has no backlash compensation on his existing machine. Likely a stepper motor setup with really cheap lead screws and likely encoderless.

jools · Jul 30, 2020

Kingbob said:
I agree with dandrummerman21. Your hobby grade machine must be in pretty rough shape or incredibly weak not to be able to produce a good oring groove in delrin. O ring grooves come in all shapes and sizes from round to square to any shape imaginable, what's important is the surface finish of the groove itself not necessarily the roundness, the oring can compensate for several thou of eccentricity as long as it is sealing well against the walls. A new or even a decent used mini mill or even a haas tool room mill will do this no problem.

Yeah it started out good then went bad

If the finance comes through then happy with the new haas option, if not just wanted to know what level of backlash people thought would cause me an issue for this operation with relation to the faceting I was experiencing.

jools · Jul 30, 2020

Ziggy2 said:
I would guess that the OP's issue is that he has no backlash compensation on his existing machine.

Likely a stepper motor setup with really cheap lead screws and likely encoderless.

Part right, part wrong.

But that digresses from the point of asking for help to remedy the situation with a better setup. My old setup is not the point of me asking for help, any new to me setup is.

13engines · Jul 30, 2020

Look up "Boring and Facing head." Made by Wohlhaupter, Narex and others. If your machine will hold steady on position, this should cut a pretty descent groove. But like anything, you still have the length to diameter ratios to concern yourself with. A facing head takes backlash out of the question... mostly.

jools · Jul 30, 2020

13engines said:
Look up "Boring and Facing head." Made by Wohlhaupter, Narex and others. If your machine will hold steady on position, this should cut a pretty descent groove in. But like anything, you still have the length to diameter ratios to concern yourself with. A facing head takes backlash out of the question... mostly.

Yeah I've seen those. With them though, well the ones I've seen, its a manual operation to make them face which to me kind of detracts from wanting the machine to do everything for me. Not sure how I could cut a groove with them accurately on parts too. The best option for me is a slot cutter and a machine that will interpolate a circle well.

Really looking for an answer on the backlash question though.

jools · Jul 30, 2020

implmex said:
Hi jools:
The real question is; what do you consider to be a "good" hole?
For oring grooves, milling is always fraught with potential sealing problems, and the difficulty arises from the "lay" of the machined surface.
A turned surface always makes concentric rings on the bottom and sides of the grooves and the oring will squash its way into the grooves and seal.
Milling makes swirls on the bottom and ripples on the sidewalls, both of which are much harder to force the oring material into reliably.

So in many high end sealing applications milled oring grooves without modification of the lay of the machine marks are not considered acceptable regardless how well they are executed.

The roundness of the groove is irrelevant to that function...so long as the perimeter is the correct length and the lay is good, the oring will work.

Having said that, my vintage 2001 Haas Minimill will still make a round interpolated bore that clocks in around 0.0007" TIR with care.
That's good enough for a lot of things that are supposed to be round...I've milled round mold cores with shutoffs and gotten away with it with a bit of blueing and fitting.
A top of the line Hermle might make the same feature within a tenth.
My Minimill cost forty grand: that Hermle is probably the best part of a million bucks.

For low tech oring grooves 0.0007" is more than adequate and I can get acceptable sealing performance without worrying too much about the lay.
But certainly NOT for any critical applications like high vacuum or high pressure or helium containment or...

Cheers

Marcus
Implant Mechanix • Design & Innovation > HOME
Vancouver Wire EDM -- Wire EDM Machining

With them being cut radially I think the swirling isn't an issue.

With your Haas does it have any back lash? If so what amount? Those tolerances would be more than enough for me and this could be the answer I'm looking for (unless it has none

)

Jools

Mike1974 · Jul 30, 2020

To answer your question the best I can, a new machine shouldn't have any* backlash, this should be factory set, or worse case the tech takes test cuts after setting up and adjusts it.

A used machine on the other hand is only as good as the worst abuse it has seen. You might find a gem that cuts great and is nice and tight, or you might get a polished turd that has been "rebuilt" aka a bad paintjob. I walked into a job running a beat up Matsurra, which are normally a more higher end machine, but this thing would not make an accurate cut over about 20ipm feed. :ack2:

*any is subjective, everything has a tolerance, and with the right equipment you could surely measure some backlash/play in ballscews etc.

Mike1974 · Jul 30, 2020

jools said:
Yeah I've seen those. With them though, well the ones I've seen, its a manual operation to make them face which to me kind of detracts from wanting the machine to do everything for me. Not sure how I could cut a groove with them accurately on parts too. The best option for me is a slot cutter and a machine that will interpolate a circle well.

Really looking for an answer on the backlash question though.

You wouldn't necessarily need a facing head, a regular boring head should work too, you could grind a tool to the proper shape, set your diameter and feed slowly in Z (if I understand what you are doing). Or a homemade trepan tool, depending on the diameter (s?) of the o-rings....

13engines · Jul 30, 2020

You'll notice implmex says "with care". That steers the outcome more then backlash. How little backlash is acceptable? As little as possible. Look up your best case scenario, the Haas Mini-Mill you mentioned, and use that as a baseline if you get stuck with buying used. Almost every machine you run into that isn't a hobby machine is going to have backlash compensation built in. Some even have separate compensation settings for Rapid and Feed Modes. That combined with ball screw error comp is about the most you can hope for. The "with care" part of this has to do with work holding, tool choices and feed rate. The last one there is able to give you a broad range of end results. If you have backlash or squareness, lead or servo mismatch errors, all of those start to stare you down the faster you try to run the feed.

Whether you are able to test a used machine at someone's shop to the extent you can reliably determine it's backlash state is hit and miss. Your best bet is to have any used machine that you buy laser shot and ball bar tested once you get it set up. These two things will get any used machine into it's best possible state of both being and knowing what you're working with. Then you can then begin "taking care" to get the holes and O-ring grooves you're after. At that point it's all about the application of tools and knowledge, with the machine only mimicking the best it can to your commands.

Hardplates · Jul 30, 2020

Obviously it depends on your application but personally I'm looking for no more than a couple tenths. On a rigid box way machine you can often turn a couple tho into a couple tenths with backlash comp. I'd imagine a light or linear machine would allow the cutter to push and pull the table around more.

I've had to interpolate holes to a very tight tolerance and was able to play with backlash comp to achieve acceptable roundness. Due to screw wear it was a few tenths off in other locations on the table. If I did more work like this I would have a brand new machine.

gregormarwick · Jul 30, 2020

If your parts aren't too big you can make or buy a hollow body grooving tool. Horn make them, although maybe not standard range.

Sort of like a trepanning tool, inserts facing inwards, goes over the top of your part and interpolates around it. Similar idea to thread whirling. Much better finish and must faster cycle time, backlash matters much less because the actual movement of the cutter blends out sharp changes in direction.

For example:

implmex · Jul 30, 2020

Hi jools:
Gregor has it if you're looking for the best oring grooves you can make on the circumference of a circular part with a mill.
I say this because the most critical surface of an oring groove is the bottom of the groove...the sidewalls don't matter nearly so much.

If you just saw them with a tee slot cutter, you have ripples across the bottom from sidewall to sidewall which is the worst possible orientation of these machining marks.
But if you swing them out as Gregor suggests, the machining marks are almost as good as turned, even though you've circular interpolated them just like you would with a saw.
The whole difference is that when you swing them, the arc of engagement of the cutting edge is very long compared to the arc of traverse of a saw tooth (which will make little scallops) and it's oriented in the same direction as the floor of the groove.
Imagine how the tool tip moves in each instance and it will become immediately apparent.

BTW, you don't need a super fancy tool like Gregor posts for this; a single lip tool knocked together from bits and pieces will do the job just as well, especially in Delrin.
Even a boring head with the tool tip pointing inward will do it, and the code is simple.
Make the tip of the tool far enough away from center that you can align the spindle over the stub and drop it down without the tip touching.
Interpolate your arc, re-center the spindle over the axis of the stub and withdraw the tool.
Easy Peasy.

Cheers

Marcus
Implant Mechanix • Design & Innovation > HOME
Vancouver Wire EDM -- Wire EDM Machining

gregormarwick · Jul 30, 2020

implmex said:
BTW, you don't need a super fancy tool like Gregor posts for this; a single lip tool knocked together from bits and pieces will do the job just as well, especially in Delrin.
Even a boring head with the tool tip pointing inward will do it, and the code is simple.

Exactly, I have an old Wohlhaupter boring head that I keep for exactly this purpose. Long superceded by my other boring heads, but makes for a fantastic OD tool when required.

Backlash you'd accept to interpolate a good circle

Aluminum

Cast Iron

Aluminum

Stainless

Stainless

Diamond

Aluminum

Stainless

Aluminum

Aluminum

Stainless

Aluminum

Aluminum

Diamond

Diamond

Stainless

Stainless

Diamond

Diamond

Diamond

Similar threads