Milling a simple sensor

[Jay C]

Guys, I wanted to mill something tonight so using Eagle and PCB-Gcode I whipped up a board for a QRB1114 reflective object sensor. The plan is to use it to measure the spindle speed.

"http://www.youtube.com/watch?v=VIV6Mnac1Y8"]YouTube - speed sensor

Jay

[Drools]

very nice Jay.

[kn6za]

Jay,

Way cool setup! Wish I had 16000rpm on my machine. Fantastic results too. I always get a kick out of watching a machine mill out circuit boards.

[Jay C]

Thanks Drew. I like the new look of your site BTW. Do you have a thread on here about your machine. the sample photos on your site look great. The Proxxon spindle will go to 20K but there is a lot of vibration so I keep it around 16K.

Jay

[kn6za]

Jay,

Good to keep the rpms low if you have vibration.

Thanks for the good words about the web site, the tools are selling, so must be working.

I do not have any info posted on the zone about my machine build. I built it from a36 hot rolled steel, 9"x6"x7" travels, hardened and surface ground ways, servo driven on 3 axis plus servo on spindle to 10000rpm. 3 angular contact bearings in spindle assembly, with retention knob style tool holders for quick tool changes. The tools holders are scaled down cat 40 design. The retention system has about 500lbs of spring force, with pneumatic tool release.

No tool changer

I quit my job as a tool maker several years ago, so I no longer have free run access to the kind of tooling and equipment I am used to. Designing a tool changer would be very frustrating to me without heat treating, grinding, cnc lathe, etc, etc.

A guy has to consider his machine finished at some point, right?

[charliex]

Very cool Jay!

[Jay C]

While playing around with the mill this weekend, I decided to see if I could work out the issues that have been limiting my max speeds. Using the advice from Widgitmaster, I applied plumber's silicone grease to the Y leadscrew and rods. I had already done this with the X and it helped but I was still limited to 45ipm.

Well since I had the machine jacked up to get at the Y, I decided to add another 1" piece of MDF under it to help dampen the vibrations that I get. It worked. In fact no more physical limitations. I have the X and Y set to 75ipm and no stalling

To celebrate, I decided to run this circuit as fast as as possible. I opened the top milling files and replaced all F24 with F60. Since I already had a 1/16" ball nose endmill in the collet, I left it. I wanted to see if there were issues, not make a usable circuit. In fact, I used a small scrap of PCB. Due to the short distances, the circuit averaged 23ipm with some straight lines peeking at 55ipm I also milled the circuit in the previously "worst place" on my table. Anyway, enjoy.

"http://www.youtube.com/watch?v=4Rl8k0N-7fQ"]YouTube - Milling anotherther speed sensor AFAP

[Jay C]

The bit used in the video.

Bought it at www.use-enco.com
OSG Tap & Die
404-0625BN
4-flute 1/16"D ball nose

[Eclipze]

I usually mill PCBs around 40..47ipm. I've milled at 55ipm before, however I have to have spindle RPM max'd (loud) otherwise I chip the fine tips off.

Yours is looking good, but your limitation is with your "acceleration"... not max speed. If you're using Mach3, have a look in Config->Motor Tuning... what acceleration have you got? Find out what the max is before it skips steps on acceleration and deceleration, then bump it back to something reliable.

Next, check the General Config and make sure you have Constant Velocity mode as the Motion Mode. Then go to Settings (Alt-6) and make sure CV Distance has a green light and the CV Feedrate light is off (important).

Now machine that same example.... if you were impressed by that, you going to spin out when the acceleration is tweaked

[Jay C]

The acceleration was at 20, now at 80. I use Quantum and the path smoothing feature. Jerk is set to 200"/s^3 with a 2% adaptive. I've long since forgotten why I chose those values I reran the circuit (this time with a V-Bit). It's no faster than the video, but jogging is improved. Again, I suspect there just isn't enough room to go full speed.

Now for a math question (before I go to Google to help): I really like the quality of the ball nose cut on this copper. Can some one tell me the trig equation to find the diameter of the bit at a specified depth of .003"?

Thanks,
Jay

[Eclipze]

The acceleration was at 20, now at 80. I use Quantum and the path smoothing feature. Jerk is set to 200"/s^3 with a 2% adaptive. I've long since forgotten why I chose those values I reran the circuit (this time with a V-Bit). It's no faster than the video, but jogging is improved. Again, I suspect there just isn't enough room to go full speed.

Now for a math question (before I go to Google to help): I really like the quality of the ball nose cut on this copper. Can some one tell me the trig equation to find the diameter of the bit at a specified depth of .003"?

Thanks,
Jay

The math for it is....
Given,
W = width of the cut at the specified depth
D = Diameter of ball nose
z = depth of cut

W = sqrt[D - 4(D/2 - z)^2]

So if you used a 0.04 diameter bit at a depth of 0.003, then the width would be 0.021.

But you also need to take into consideration the thickness of the copper. If you take 50um copper clad board, then your effective cut width reduces to 0.013 (almost half).

[Jay C]

Hmmmm ... the equation or your example is wrong. Also, can you explain why the copper thickness will make the cut smaller versus "wider"?