Hurco BMC20 Dynomotion Retrofit

[TomKerekes]

Hi Troy/mmurray70,

The standard screen Feedhold buttons will issue a feedhold to KFLOP then shortly afterward invoke the custom feedhold Action. So I don't think there is a way to prevent having the feedhold button stopping z motion during a tap.

Here are some ideas we might try:

#1 I suppose the tapping loop could also detect a feedhold (CS0_StoppingState>0) and immediately release feedhold (ResumeCoordinatedMotion()) and turn off the Spindle instead. Sort of Kludgey but that should all happen in several hundred microseconds and not be detectable. But then the button won't show FeedHold active on the screen because it was released.

#2 Temporarily removing the Z axis from the Coordinated Motion System should cause Feedhold to have no effect on Z. The tapping loop could then monitor Feedhold state and if detected stop the spindle. It could also turn the spindle back on if Feedhold is removed.

#3 Or add a custom button for Feedhold to conditionally Feedhold or stop the Spindle.

#4 Or we would need to change the main code to somehow detect a tap cycle is in progress and handle things differently.

Are there any other systems in the world that allow Feedhold during a tap cycle?

Regards

[Need TECH Help!]

Hi Tom,
It would be nice to keep the feedhold working like it does and keep the rewind/fastforward buttons.

Could the feedhold button have 2 different states? For example if there is a tap cycle in action, c code in the feedhold button would know this and change to handle feedhold for tapping. And if not tapping the feedhold button would act how it always does. Sorry i dont how to put it in technical terms. But in short the feedhold button would actually be like 2 different feedhold buttons in one. I think this is what you are talking about in option 4 of post?

I have ran systems that will feedhold in a tap cycle, but you can not turn the spindle by hand. The spindle motor holds it from turning during feedhold/paused state.
And other systems simply dont allow you any control what soever besides Reset or Estop to stop motion.
Some older systems i have ran dont allow any feedhold/pause option during tap cycle but do allow spindle to be shut off while still maintaining axis to spindle sync. Would this be a simple way to handle feedhold during tap cycle? By just blocking feedhold but allow spindle to be switched off and on?

Also Murray, thanks for your work on this.
Troy

[mmurray70]

Hi Troy/mmurray70,

The standard screen Feedhold buttons will issue a feedhold to KFLOP then shortly afterward invoke the custom feedhold Action. So I don't think there is a way to prevent having the feedhold button stopping z motion during a tap.

Here are some ideas we might try:

#1 I suppose the tapping loop could also detect a feedhold (CS0_StoppingState>0) and immediately release feedhold (ResumeCoordinatedMotion()) and turn off the Spindle instead. Sort of Kludgey but that should all happen in several hundred microseconds and not be detectable. But then the button won't show FeedHold active on the screen because it was released.

#2 Temporarily removing the Z axis from the Coordinated Motion System should cause Feedhold to have no effect on Z. The tapping loop could then monitor Feedhold state and if detected stop the spindle. It could also turn the spindle back on if Feedhold is removed.

#3 Or add a custom button for Feedhold to conditionally Feedhold or stop the Spindle.

#4 Or we would need to change the main code to somehow detect a tap cycle is in progress and handle things differently.

Are there any other systems in the world that allow Feedhold during a tap cycle?

Regards

#1 sounds like a decent idea

#2 Also sounds good. How would you remove Z from the Coordinated motion system?

#3 We have this working now i think. Just has to be used with a user button instead of the Kmotioncnc pause button. If this code was setup with external buttons it should work really well i think.

Im not sure exactly what to suggest for how Kmotioncnc should handle it. I have never used a machine that allowed a feedhold during tapping. Fanuc doesnt allow it, mazak doesnt allow it and i dont think haas do either. Only way to stop on these machine is E-stop. Almost seems if these guys decided to block it out, then that might be the best way to go. I decided to block it on my machine, just trying to help Troy since hes set on it. I just feedhold and resume a couple times while approaching workpiece, double check code and height when you get close and then let er go. Works for me, and also all the fanuc and mazak users.

Might not hurt to block the normal stopcoordinated motion command if it were an easy fix to do so. This would prevent people from getting in trouble with sync errors or a possible crash. And if somebody was set on being able to stop it they could conditionally stop spindle like Troy is doing.

[mmurray70]

That fixed the restart at second hole issue. What was the change in code to fix that?

Yes I added the rest code to my halt program.

The feedhold and resume ccode programs still cause the spindle to lose sink with z axis when actions are set for factory pause button.

The first time the spindle starts in the rigid tap program it simply sets the CW bit. Your machine runs a little different and needs to clear the CCW bit and then set the CW bit. So i changed it out to match your machine.

Your welcome for the help. I had alot of help from others last fall when i was doing my machine, so nice to be able to give back a little now that i understand things a little better.

[Need TECH Help!]

Looks like you understand it more than a little murray
A 2016 Hurco is the only machine i have ran that allows a feedhold during tap cycle its nice as it keeps the kosher way of controlling a machine. Its more common in a questionable/new operation to feedhold so during a tap cyccle the feedhold would be more of a natural reflex. Been several times during a tapcycle on other machines that i tried to feedhold. But iam up for what is the safest,sure way to handle this for everyone. Which is most likely just blocking the factory feedhold buttons during tapcycle as murray has said.

[TomKerekes]

Hi Troy/mmurry70,

Here is a patch that changes the Interpreter to expose if it is currently in a tap cycle.

KMotionCNC is changed where if FeedHold is pushed while in a tap cycle instead of feedholding it will flag the M119 Tap program (by setting persist Var+0 to -1). It should also show the Feedhold DualButtonState image correctly based on the persist Var state.

It will then be up to the Tap Program to detect the Var set to -1 and stop the spindle. The Resume Button will then Clear the Var to 0 to tell the Tap Program to restart the spindle in the recorded direction.

To solve the problem where the operator pushes Feedhold just exactly before a Tap Cycle and XYZ begins feedholding and then the TAP cycle begins, the Tap Program must detect any motion Feedhold, release it, and trigger a "Tap" feedhold.

Does any of this make sense?

Here is the GCodeInterpreter that exports Tap Cycle State:
http://dynomotion.com/Software/Patch...nterpreter.dll

Here is the KMotionCNC that handles Tap Feedhold:
http://dynomotion.com/Software/Patch...KMotionCNC.exe

Here is a modified Tap c Program that includes monitoring for Tap Feedhold:
http://dynomotion.com/Software/Patch...ld%20Rev%204.c

Could you test to see how much of it works?

Regards

[Need TECH Help!]

Tom,
Tested your files. As soon as Tapcycle is started the spindle shuts off but stays synced. Then if i manual switch on spindle tapcycle finishes for 1st hole and spindle shuts off again after 2nd hole is started.
Feedhold works correctly for the X axis feedmove and is ignored during tapcycle.
Other than spindle turning off all appears to be working.

[TomKerekes]

Hi Troy,

Opps Var+0 was already being set to -1 causing an immediate Tap Feedhold. Lets use -2 to command a Tap Feedhold.

The patch KMotionCNC.exe and Tap C Program have been updated to use -2.

Please try it again.

Thanks

[Need TECH Help!]

Hi Tom,
That fixed the spindle shut off issue. Going to try a few more test scenarios air and then try a test in plastic. Plastic should show any sync issues with breaking tool.
Thanks again,
Troy

[mmurray70]

Nice work Tom, that looks great. I might try the new code on my machine someday too.

Troy, if you want to check how well its synced try tapping a piece of flatbar extended out over the edge of a vise and put an indicator on it. The part will be able to deflect a little and the indicator will show how well spindle is following the tap. My machine was less then 0.005" and i had some slop in Z axis at the time. I was pretty happy with that. Trying plastic or something soft isnt a bad idea first.

[Need TECH Help!]

Ok. Tried tapping 1/4-20 in plastic, alluminum and then C.R.S. all tapped well. And the steel thread had a very good fit. Only had a used hand tap. Will need a new power tap to do an accurate deflection test. But i think its good just by the how the steel tapped.
Thanks again,
Troy

[mmurray70]

Good to hear its working well, your probably fine if taps are not breaking and threads are a good fit.

I peck tapped some 10-32 holes in stainless today and it worked well. One minor issue is it starts pecking immediately from the R plane. I was pecking only 1 turn at a time today so there were several pecks before it cut anything. Ill try and fix this someday when i get a chance, shouldnt be too hard.

[mmurray70]

Actually forget about what i said above. The top of work piece isnt even defined in G84, only the R plane and depth so it has to work the way it is now.

[Need TECH Help!]

Looking for scales for my mill now. Looking at the magnetic scales. Renishaw scales and reader heads are a bit pricey. Dont want to got that route unless i have to. Has anyone used or heard of these scales ? https://www.machine-dro.co.uk/indust...-encoders.html
They have some on ebay sold in the states.
Attached is data sheet of the 5uM. I have emailed them about getting a 1uM or 2 uM.

Tom or anyone see any issues with using these scales with my Kflop/Kanalog/Konnect combo? Obviously want to use them to close the loop at table and head of mill. My max Rapids are 800IPM.
Also is the encoder reading speed of Kflop 1Mhz total for 8 axis or 1 Mhz per encoder?

Thanks,
Troy

[Jim Dawson]

I have Renishaw LM-10 1um on my mill & router, but I put Ditron MR200 1um on my lathe. Same form factor and specs as Renishaw. I have tested them side by side and found them to be functionally equivalent. About $120/axis including the mag tape and shipping.

[m_c]

I built a small digitiser/probing machine, using a 3040 as a base, a KFlop + KStep for control/drive, and closed the loop using 1uM scales from Machine-DRO.

They are pretty generic Chinese scales, so if you're across the pond, I'd look at buying direct from china. The same scales can be found on aliexpress. Although Machin-DRO at the time happened to not cost much more than buying direct from China, and had what I wanted ready to ship.

To get 1uM, they need different reader heads, and if you're wanting to use them with the Kanalog, I'd be asking to get the differential output option as well.
The 1uM option isn't actually any more accurate than 5uM, as they use the same optical scales, however they're still more accurate than magnetic tape systems. (I just rechecked that, and the Renishaw tape system only has positional accuracy of +/-20 to 40uM, and even the standard M-DRO optical scale only varies from +/-5 to 10 uM depending on length).

[Need TECH Help!]

......The 1uM option isn't actually any more accurate than 5uM, as they use the same optical scales, however they're still more accurate than magnetic tape systems. (I just rechecked that, and the Renishaw tape system only has positional accuracy of +/-20 to 40uM, and even the standard M-DRO optical scale only varies from +/-5 to 10 uM depending on length).

I wondered about that. What is the point of close resolution if the accuracy is so much larger? I did use glass scales one time from SINO that had an accuarcy of 5uM. But was on a small RongFu mill. They worked really well, but feed was only 100IPM.

[Need TECH Help!]

I have Renishaw LM-10 1um on my mill & router, but I put Ditron MR200 1um on my lathe. Same form factor and specs as Renishaw. I have tested them side by side and found them to be functionally equivalent. About $120/axis including the mag tape and shipping.

Where did you get them? Alibaba is the only place i have found so far.
Thanks,
Troy

[mmurray70]

Looking for scales for my mill now. Looking at the magnetic scales. Renishaw scales and reader heads are a bit pricey. Dont want to got that route unless i have to. Has anyone used or heard of these scales ? https://www.machine-dro.co.uk/indust...-encoders.html
They have some on ebay sold in the states.
Attached is data sheet of the 5uM. I have emailed them about getting a 1uM or 2 uM.

Tom or anyone see any issues with using these scales with my Kflop/Kanalog/Konnect combo? Obviously want to use them to close the loop at table and head of mill. My max Rapids are 800IPM.
Also is the encoder reading speed of Kflop 1Mhz total for 8 axis or 1 Mhz per encoder?

Thanks,
Troy

Im surprised to see all you guys using scales. Seems like a headache to setup and potential things to break with adding scales IMO. Especially Chinese ones. If you have some accuracy problems now why not spend the money on new thrust bearings and rebuilding ballscrews? Just seems so much simpler to me. And a machine without scales can still be very accurate as long as theres no slop in the axis and you have decent encoders. Your hurco should be very accurate with those dmm servos if there are no mechanical problems.

Each axis can take 1Mhz input. The previous owner of my machine actually set my Z axis speed set at 1.25mhz and it works perfectly, still running it this way. So theres a little headroom there. Probably safer to stay below 1mhz, im sure Tom set that maximum for a reason.

[Jim Dawson]

Where did you get them? Alibaba is the only place i have found so far.
Thanks,
Troy

I purchased directly from Ditron. I have attached a pdf that is not available on the web.
Contact Info:
Jaeger Hou
E-mail: sales@dcoee.com

Im surprised to see all you guys using scales. Seems like a headache to setup and potential things to break with adding scales IMO. Especially Chinese ones. If you have some accuracy problems now why not spend the money on new thrust bearings and rebuilding ballscrews? Just seems so much simpler to me. And a machine without scales can still be very accurate as long as theres no slop in the axis and you have decent encoders. Your hurco should be very accurate with those dmm servos if there are no mechanical problems.

You are absolutely correct, the machine should be as mechanically perfect as possible. You should not rely on electronics to overcome mechanical problems, but it is possible.

The reason I went with the mag scales originally was because I needed to replace my encoders during an upgrade and thought I would try the mag scales to see how they worked out. In the case of my mill I have had Renishaw LM-10, 1um units installed for about 4 years, they have been trouble free and are extremely accurate. I bought them cheap on Ebay when there were hundreds of them available. All of those are gone now. Shortly after I did the mill upgrade, I installed them on my 4x8 router also, and they work fine at up to the 600 IPM rapid speed of my router. I have no way to test them at higher speeds. Since then I have installed mag scales on 3 other machines with great success.

I have some other projects the I am going to CNC (is that actually a verb?) and want to install mag scales also. With the used Renishaw units no longer available, a bit over a year ago I asked Ditron if they could duplicate the Renishaw units, and about 4 months later they had them in production. I have tested the Ditron units against the Renishaw units and have found them to be functionally equivalent. I have found both to be very accurate and with a tight machine actually able to machine parts to single digit micron accuracy. Close enough for me