Peck cycle for rigid tapping

[Derek]

Hi Balazs.
I know we talked about this a while ago but I figured I'd throw it out there again. It would be so nice if the tapping cycle would recognize Q as a peck depth. It works so well when I hand code it but it's a real pain when I have a lot of holes.

My life would be so much simpler if you could implement this.

Thanks
Derek

[cncdrive]

Hi Derek,

I don't see a problem adding this feature.
Just to make sure that I understand what to implement:

There should be a Q parameter for the G33.1/G33.2 which sets the peck depth which is in other words the depth of the passes,
so if Q is shorter than the distance of the total movement to the endpoint (full tap depth) then the tap is made with Q size passes.
E.g. Z is at 0 and there is a G33.1 Z-3 Q1 code then this will be done with 3 passes with the Z tapping to -1 and reverse to 0 then tapping to -2 and reverse to 0 and tapping to -3 and reverse to 0.
If the Q is not an integer fraction of the full tap depth then the last pass will be shorter in a way that the endpoint, the last pass will always ends at the programmed Z endpoint.

Or if the Q is greater than the full depth or if Q is not defined then the tap is done in a single pass.

Is this what you want?
Is there anything you want different than how I described or anything you want to add?

[Derek]

Hi Derek,

I don't see a problem adding this feature.
Just to make sure that I understand what to implement:

There should be a Q parameter for the G33.1/G33.2 which sets the peck depth which is in other words the depth of the passes,
so if Q is shorter than the distance of the total movement to the endpoint (full tap depth) then the tap is made with Q size passes.
E.g. Z is at 0 and there is a G33.1 Z-3 Q1 code then this will be done with 3 passes with the Z tapping to -1 and reverse to 0 then tapping to -2 and reverse to 0 and tapping to -3 and reverse to 0.
If the Q is not an integer fraction of the full tap depth then the last pass will be shorter in a way that the endpoint, the last pass will always ends at the programmed Z endpoint.

Or if the Q is greater than the full depth or if Q is not defined then the tap is done in a single pass.

Is this what you want?
Is there anything you want different than how I described or anything you want to add?

That's how I think it should work. I would add that if there is no Q it should just do a full pass.

Long ago we talked about implementing an auto reverse or auto peck feature that was based on spindle RPM differential. If the RPM fell below a certain amount it would reverse and back out of the hole. Just thought I'd throw that out there. This would be great on low powered mills. But if you implement pecking I would use it on everything that wasn't a production run so that would solve my current issues.

[spumco]

Third vote/request for peck tapping as you've described it.

Thanks,
Spumco

[cncdrive]

Ok, we've just finished adding this feature, will be in the next test release.

[cncdrive]

Long ago we talked about implementing an auto reverse or auto peck feature that was based on spindle RPM differential. If the RPM fell below a certain amount it would reverse and back out of the hole. Just thought I'd throw that out there. This would be great on low powered mills. But if you implement pecking I would use it on everything that wasn't a production run so that would solve my current issues.

Yes, I remember this conversation with you and I think it is a good idea and that we can add this feature also, but this requires to be implemented in the firmware,
I mean it needs firmware change also so it is a bit more complicated than adding the Q parameter.
I will talk to my collegue and will ask him how complicated will this be for him to implement it in the API.
And if we can add it can you test it on your machine?

[Derek]

Long ago we talked about implementing an auto reverse or auto peck feature that was based on spindle RPM differential. If the RPM fell below a certain amount it would reverse and back out of the hole. Just thought I'd throw that out there. This would be great on low powered mills. But if you implement pecking I would use it on everything that wasn't a production run so that would solve my current issues.

Yes, I remember this conversation with you and I think it is a good idea and that we can add this feature also, but this requires to be implemented in the firmware,
I mean it needs firmware change also so it is a bit more complicated than adding the Q parameter.
I will talk to my collegue and will ask him how complicated will this be for him to implement it in the API.
And if we can add it can you test it on your machine?

But of course:)

If Q is easy then you may want to implement that first. The auto reverse function would be nice for fresh setups when you are not sure how the tap will react. It would be nice if a letter and number in the line of code would designate the percentage of spindle drop before it backs out. A smaller tap would really want a small amount of drop before it backed out. Not sure if you could monitor encoder counts as opposed to index signal since one revolution of the spindle could be the difference between a broken tap or not. Like I said earlier though if I had a peck setting available in the code then I would peck most holes.

[cncdrive]

Hi Derek,

If Q is easy then you may want to implement that first.

See my previous previous post, the Q was already implemented, finished yesterday.

The auto reverse function would be nice for fresh setups when you are not sure how the tap will react. It would be nice if a letter and number in the line of code would designate the percentage of spindle drop before it backs out. A smaller tap would really want a small amount of drop before it backed out. Not sure if you could monitor encoder counts as opposed to index signal since one revolution of the spindle could be the difference between a broken tap or not. Like I said earlier though if I had a peck setting available in the code then I would peck most holes.

Yes, the controller continously tracks the encoder A/B, it has to, otherwise the tapping could not work, so it knows the change of position -> spindle speed.
I thought to implement it the same way as you described, to have a % setting of speed drop. We could measure the spindle speed when the tapping cycle starts before syncronising to the index channel,
the only possible issue I see is if the spindle is not on it's full speed when the tapping cycle starts, however maybe that is not even a problem,
because the spindle should be around the set speed and so that could be controlled with the M3 dwell to be sure the spindle reached the target speed when the tapping starts.
I'm just thinking loudly...
What are your thoughts on this?

[Derek]

Well with a low powered spindle and a large enough tap it will stall the spindle. At that point you have to hit stop and manually remove the tap. At least if it auto reversed on or near stall it would help out a lot. On my mill with a large tap I don't know if it will do it until I try. Even with a peck program if it stalls your done. Of course it the tap is weaker than the spindle HP you are going to break it.

I agree on the tap chart info but different materials can have vastly different results. I'm looking at you 6061:)

There may be too many variables involved to have it stop breaking a tap but if it auto reversed via a value in a G code line then it could be tailored to each machine/job/tap. If it's a blanket setting in config then I would have it settable at the very least. Like I say the auto reverse would be most helpful if it simply reversed out during a stall or near stall.

[Derek]

Just finished 438 holes. Pecked in 4 steps each. No problems at all.

Thanks again for this.

Derek