Yamazen CNC knee mill

[FPV_GTp]

beautiful work N4NV only wish the items where in my garage


cheers

[N4NV]

My mill has been running great except for one thing. The Z axis limit switch trips at certain RPMs and cutter feed rate. It happens about once every 30 or 40 minutes of cutting. When the switch trips, Mach does and e-stop and I loose my X, Y and Z positions which then requires me to re-zero everything then do a “cycle start from here”, a pain in the butt. The limit switches are high end Omron (about $100 each). I tried varying the debounce interval, but it still did it when I passed 5,000 on the debounce.

Since completing the mill, I started on a Hardinge CHNC lathe. The Hardinge uses Hall effect switches for all limit switches. Once I saw how well the Hall switches worked, I figured they would great on the Z axis of the mill since they would be immune from vibration. I purchased several Hall switches DigiKey:
http://search.digikey.com/scripts/Dk...e=US5881LUA-ND
These switches are only $2.30 each. I made some switch mounts out of Delrin and epoxied the switches to the Delrin. Kirk had given me some 1/8” diameter magnets that are some special material and very strong. I epoxied one of these magnets to the Delrin piece mounted to the Z axis ball screw (see pictures).

The switches are three wires, 0V, +3.5 to 24V and output. I wired two switches in parallel (one for Z+ limit and one for Z- limit) to a CNC4PC C11G board with a 1K resistor across the +5V and pin 12 input. I then set up Mach for pin 12 active low. I lined up the switches so the magnet would pass about 1/8” away from the Hall switches. I ran the Z axis to both upper and lower limits and they worked perfectly!

After getting the switches adjusted I ran some test on the repeatability of the switches. I used a Mitutoyo digital dial indicator that reads to 0.00005” and is accurate to 0.0001”. I was amazed at the results. The Home Z axis repeatability was +- 0.00015”. No wonder Hardinge used this type of switch on their lathe. Just for the giggles I checked the back lash and found the Z axis to have 0.0004”. I kick myself for waiting so long to fix this problem. The total cost was less than $5 and some scrap Delrin, wire and a 1K resistor.

In the pictures of the switches, the white stuff over the wires is and epoxy putty to hold the wires in place.

Vince

[BobWarfield]

Awesome tip on the Hall effect switches!

I'll bet they'll last forever with no moving parts too.

Cheers,

BW

[N4NV]

Originally Posted by BobWarfield Awesome tip on the Hall effect switches! I'll bet they'll last forever with no moving parts too. Cheers, BW

They should last for a while. On my Hardinge CHNC which is 28 years old, there are 19 hall switches for limits and all that I have tested still work properly.

Vince

[N4NV]

The Hall limit switches went so well, I moved ahead and made a speed sensor for the spindle. Up inside the top of the head where the quill shaft goes through the upper bearing, I saw enough room to mount a magnet holder to the shaft and there was a stud coming through the upper housing to mount my Hall sensor. The quill shaft had a keyslot machined into the shaft. I would use this slot to keep the mount from rotating on the shaft.

I drew up a tool path for the magnet mount in AutoCAD, saved it as a DXF file and imported it into Mach LazyCam. LazyCam generated the G-code for Mach. I cut the mount out of a scrap piece of Delrin I had laying around. Here is the Video:

YouTube - Machining a magnet mount

YouTube - Machining a magnet mount

I cut the finished mount off of the block of Delrin with a chop saw, cut it into two pieces with a hack saw, threaded one half for 6/32 screws and machined a small hole for the magnet. The magnet mount has 4 raised bumps on it so that it would not cause a balance problem with the quill. Epoxy was used to secure the magnet.

Installing the magnet mount was quite a task as I could get one hand in the access hole, but then could not hold on to a tool. I only dropped two tools into the quill head. If that don't get your hear thumping, you're one cool cat. I retrieved both tools with a magnet on a string. With the magnet mount in place I could take the measurements to make the Hall sensor mount. That too was made out of Delrin.

I soldered a shielded 4 conductor cable (only used 3 conductors) to the Hall sensor. I held the sensor in place with epoxy putty. After the putty dried I sanded it down so it would not interfere with the magnet and would clamp flat to the top of the quill housing.

I had to remove my touch probe connection to the break out board as I was out of inputs. I used a 1K pull resistor across the 5V and pin 12 and set Mach for active low. When I ran a test it did not work. After removing the Hall sensor I found that I had installed the magnet backwards. The Hall sensor is only tripped with the south pole of a magnet. If you have a compass, the south of your magnet will attract the north pole of the compass.

I had to remove the magnet mount, pry the magnet out of the epoxy, re-drill a hole for the magnet, re-epoxy the magnet and re-install the magnet mount. After all that it worked great. It reads to the top speed of my mill at 4,000 rpm and down to 60 rpm. Below 60 rpm it reads 0.

My next project is to make the same part for my lathe.

Vince

PS, the black tube in front of the quill housing is my fiber optic light pipe.

[johnmb]

hi vince
how do you wire them up in parallel and how do the three wires attach on the board.
im a newby at this and would appreciate any help.

john

[N4NV]

Originally Posted by johnmb hi vince how do you wire them up in parallel and how do the three wires attach on the board. im a newby at this and would appreciate any help. john

I assume you are asking about the limit switches. All three leads just parallel. The ground leads tie to the ground terminal on the breakout board, the VCC leads tie to the +5V and the output leads tie to input terminal 12. I had to use a 1K pullup resistor between +5V and terminal 12 because the Hall sensor uses a logic low. On my CNC4PC board, each input terminal has a +5V terminal next to it which makes for easy wiring.

Vince

[N4NV]

Well, I thought I had my problem with the limit switch tripping solved when I installed the Hall limit switches, but no, it still stripped. After a lot of testing I determined that all my limit switches were causing a fault. I decided it had to be a noise problem so I installed a 33uF cap (the only size I had on hand) across all the limit switch inputs. I ran my test program 3 times in a row without any trips (previously I never made it through the program)! Time will tell if this solves the problem for good.

Vince

[mkenney]

I'm trying this setup and struggling so far. I have the same sensor's and swear I thought I have things wired correctly, apperently not. I have +5v going to pin 1 (VDD), pin 2 to the input on the break out board and pin 3 goes to the ground. I have a 1k resitor between the +5v and the input on the breakout board. I wave different magnets different ways and get nothing. What am I doing wrong, been one of those days?

[N4NV]

Originally Posted by mkenney I'm trying this setup and struggling so far. I have the same sensor's and swear I thought I have things wired correctly, apperently not. I have +5v going to pin 1 (VDD), pin 2 to the input on the break out board and pin 3 goes to the ground. I have a 1k resitor between the +5v and the input on the breakout board. I wave different magnets different ways and get nothing. What am I doing wrong, been one of those days?

Try different values of resistor. With my CNC4PC card, some of the inputs required only 300 ohm. On my new PMDX board, no resistors are required. You can also hook up the VDD to +5V, pin 3 to ground and a volt meter between +5V and the output pin. If it is working you will see a voltage change on the meter. The magnet has to be perpendicular to the sensor and it only works with the south pole of the magnet.

Vince

[mkenney]

I had two threads where I was asking the same question, my fault! I'm moving over some info from the other thread to this thread to combine for both myself and other's to benefit from:-)

Other Thread:

http://www.cnczone.com/forums/showthread.php?t=73017

I don't which thread is better to have the info in. Mine was on home and limit switches so easier to find searching but this is the thread that really started the discussion on hall effect switches. I yield to the admins and those who know better to tell me what to do:-)

Originally Posted by cadmonkey I assume you're not really using application circuit 12.1 just what you have listed above correct? I would wire up the circuit in 12.1 exactly, hooking grounds to the ground on the BOB and the Vcc to the 5V on the BOB. Before hooking up the Vout to the BOB, do you have a multimeter or better yet a scope? Personally I test on a breadboard on the bench (electronics, not machining) before it goes near the controller.

Originally Posted by mkenney Yep, just using what I listed. I ordered the 470pF capacitor that I did not have. Was trying to get away with just a pull up resitor as the orginal poster of this was able to do, not my luck I guess:-) I do have some breadboards and a multimeter and orginally bread boarded it and was never able to get any change in the output?? I assume when they say that the output is active low that it turns into a ground when the hall effect sensor sees a magnetic field, is this correct? Thanks for the help! Mark [mkenney]

Mark [mkenney]

[mkenney]

Originally Posted by N4NV Try different values of resistor. With my CNC4PC card, some of the inputs required only 300 ohm. On my new PMDX board, no resistors are required. You can also hook up the VDD to +5V, pin 3 to ground and a volt meter between +5V and the output pin. If it is working you will see a voltage change on the meter. The magnet has to be perpendicular to the sensor and it only works with the south pole of the magnet. Vince

Still working on it. When off the output pin is at .535Volts and when on it is at .022Volts. Any chance you could tell me what yours are showing. Starting to pull my hair out on this one:-)

Mark [mkenney]