Can you set a DRO to add knee and quill?

[durableoreo]

I am trying to combine the knee and quill z positions on my DRO. I need it to store a reliable z offset for tools in the DRO's library. The idea is to speed up the drudgery with a Haimer and DRO. (Probably would be better to buy a Haas but I'm 50k short and don't have a place to put a Mini Mill.)

TL;DR I'm installing an import DRO on a G0755. I'm considering combining A+ and B+ signals from two different read heads to sum the column and quill positions

In a little more detail: The DRO is a ToAuto brand Asian import but the 4-axis type. The manual is of the usual quality. There's some hints for adding axes in lathe mode but it uses the PCD-mode button so it seemed like a dead end.

I've looked at the boards in the read heads---they're surface mount with 2 pots, an LM339, and a AM26LS31C on the board. That last chip is a differential line driver but there's only 4 wires and the cable shield connected to the board. Two of those wires are labeled A+ and B+ on the silk screen.

I'm considering electrically joining the A and B quadrature signals from the column and quill. Probably with a multipole switch so I can use the scales separately or in combination. I know this sounds crazy but the output of the LM339 is a open collector NPN type circuit so it might work. The comparator output relies on the wire being pulled to 5 V all the time and pulls it to 0V when the comparator is triggered. So putting 2 LM339 outputs on one line might be OK.

Just wanted to do a sanity check before going any farther. Could it work? Is there another way to achieve the same end? Some other DRO I should have purchased?

 

[Mitch Alsup]

"So putting 2 LM339 outputs on one line might be OK."

a) a differential line driver is not going to like a second differential line driver on the same circuit.
b) if you could get both read heads routed to 1 line driver, and use diodes to isolate the pull downs, you could get it to work.
c) you probably want the knee and quill to subtract (rather than add) as they both move the tool deeper into the part.
d) this subtract (reading=knee-quill) function might already be a capability in the DRO

 

[JimDawson]

I am trying to combine the knee and quill z positions on my DRO. I need it to store a reliable z offset for tools in the DRO's library. The idea is to speed up the drudgery with a Haimer and DRO.

When I set up tools for a job I take the longest tool and use that as the master position, setting the knee Z 0 to that position. Then all other tools are zeroed by raising the knee and writing the knee position on a piece of paper. That way the quill Z 0 position is fixed for the job, and the tool offsets are input by adjusting the knee to the correct value.

In my case it would be very easy to combine the knee and quill values, I just have not found a good reason to do it yet.

 

[durableoreo]

"So putting 2 LM339 outputs on one line might be OK."

a) a differential line driver is not going to like a second differential line driver on the same circuit.
b) if you could get both read heads routed to 1 line driver, and use diodes to isolate the pull downs, you could get it to work.
c) you probably want the knee and quill to subtract (rather than add) as they both move the tool deeper into the part.
d) this subtract (reading=knee-quill) function might already be a capability in the DRO

I had been assuming that the line driver wasn't used but after reading your comment, I see the traces on the PCB. Looking at the datasheet, I think the driver can be used in single-ended mode. Also, my assumption about a simple pull-up and open-collector outputs were not correct.

Considering an Arduino Uno to manage combining the signals. Combining signals (subtracting) should be easy enough in software. I wasn't sure if it would be fast enough. I have 5 um scales. If rapids are 0.1 in/sec, that's 2500 ticks/sec. To detect a square wave, you need more samples---probably 4x or more, so let's say 10 kHz minimum. But the digitalRead() function is reported to execute in about 5 us, which is about 200 kHz. There are pin-change interrupts on the Uno and the interrupt handler doesn't have much to do so I probably won't run out of time there, either.

BTW, after sleeping on it, two LM339 outputs won't play well together. If one read head is pulling the line down, the other read head's output will be squashed.

 

[durableoreo]

When I set up tools for a job I take the longest tool and use that as the master position, setting the knee Z 0 to that position. Then all other tools are zeroed by raising the knee and writing the knee position on a piece of paper. That way the quill Z 0 position is fixed for the job, and the tool offsets are input by adjusting the knee to the correct value.

In my case it would be very easy to combine the knee and quill values, I just have not found a good reason to do it yet.

Sounds reasonable. I'm a beginner, though, so I frequently end up moving the knee. Usually it's a tool that I didn't anticipate using. I've started writing down my plan to avoid surprises. But also, I have heard you should work with the quill retracted, if you can. Maybe I'm weighing that advice too much.

Thanks for the input, guys. I'll let you know how it turns out.

 

[JimDawson]

Sounds reasonable. I'm a beginner, though, so I frequently end up moving the knee. Usually it's a tool that I didn't anticipate using. I've started writing down my plan to avoid surprises.

In the normal course of machining the knee with a vertical turret mill or in your case with a G0755 bed mill, the head, is moved up and down perhaps several times in a single setup. It is not just a beginner thing. Tools are different lengths and you need to adjust for that. If you are using standard R-8 collets, then presetting tools is impossible. If you are using something like ER tool holders, then it is possible to preset tools. The only time I preset tools is if I am doing a multi-part run where I have the tools laid out for the job. If I am doing a one off part, then I just touch off to zero for each operation, it makes no sense to set up tool heights for a single part. The exception to this might be where you are using the same tool in different operations that have different heights, where you are doing other operations between actually using that particular tool. This is normally a pretty rare case.

Now I get the opportunity to get up on my soap box and give a speech. You should machine the whole job in your head before ever making a chip. As you gain experience, this will become second nature. Order of operation is important as is work holding, this is probably 50% of the job. When I started doing CNC work, this really became apparent. In many cases, I spend as much time working out the details of how I'm going to do the work as actually making chips.

But also, I have heard you should work with the quill retracted, if you can. Maybe I'm weighing that advice too much.

That is correct. You normally want to keep the quill as retracted as possible for the job. For milling operations, full retraction and locked is the best where possible. In non-CNC operations, the quill is normally only used for drilling type operations.

 

[paradox_pete]

For what it is worth, Yuriy's TouchDRO system should be able to sum a 4th scale with the Z scale. I intend to do this myself, but I don't have the knee scale installed on my mill yet. (I do have the X,Y, and Quill scales installed and working). There is a setting regarding how to display the 4th scale and summing it with the Z is an option. Unfortunately it is likely to be a few weeks before I get the knee scale installed on my mill, but I should be able to just hook up the scale and verify that this does work later today.

It's possible the scales you have would work with TouchDRO. I gather TouchDRO is just a hobby project for Yuriy, but he does sponsor a forum here on HM.

-Pete

 

[durableoreo]

If you are using standard R-8 collets, then presetting tools is impossible.

That's a good point. I wonder how endmill holders repeat in Z. According to the internet it's decently repeatable but you pay for it with longer stick out and reduced rigidity. I'm chicken when it comes to getting close tolerances. I usually end up stopping to check critical dimensions. So if I can get tools heights repeating within +/-0.002, that would probably be good enough.

Now I get the opportunity to get up on my soap box and give a speech. You should machine the whole job in your head before ever making a chip. As you gain experience, this will become second nature. Order of operation is important as is work holding, this is probably 50% of the job. When I started doing CNC work, this really became apparent. In many cases, I spend as much time working out the details of how I'm going to do the work as actually making chips.

That is correct. You normally want to keep the quill as retracted as possible for the job. For milling operations, full retraction and locked is the best where possible. In non-CNC operations, the quill is normally only used for drilling type operations.

Thanks for the advice.

 

[JimDawson]

I wonder how endmill holders repeat in Z

I would expect better than +/- 0.0001'' . I have no machines that use R-8 collets, all use #20, #30, or #40 tool holders. I have had no problems using tool holders, even on my machine about the same size as yours.

But using holders like these will minimize sacrificing rigidity, if they ever get back in stock. https://www.shars.com/catalogsearch...21&q=r-8+er+32+collet+chuck+set&shank_size=R8

 

[Illinoyance]

Cheap 4 axis DROs from China have a Z axis summing feature.