Scope Adventures in the Machine Shop

[robrenz]

Don't miss the embarrassing story of professional tail chasing at the end.

I built an over the wheel dresser for my surface grinder many years ago. It used a hand lever to move the diamond across the grinding wheel and I was fairly good at getting just the right velocity for the type of dress I was after. But my son works for me now and he convinced me I needed to make some way to make the dress velocity adjustable and consistent. So I opted for driving the dress slide with a air cylinder and controlling the speed with a Kinecheck hydraulic stroke control. That works fine except the speed scale on the Kinecheck is on the 20mm diameter body and has insufficient resolution for the fine adjustments needed. So I made a large dial sleeve out of PET plastic and a indicator bar with a 95mm diameter for greatly increased resolution. But now I need to calibrate the dial in Inches per minute. I figured a linear pot would be the easiest feedback so I made one from some resistance wire I had that was 3 ohms per inch and a few pieces of Vespel, copper tubing, and brass shim for the wiper. The dress stroke is 1" and I made the pot stroke 1" with the wiper just off the resistance wire at the start of the stroke. This makes whatever voltage is across the pot wire equal to the 1" stroke and I can use the slope measure of the 1054 to measure inches/second that I can convert to inches per minute.

Kinecheck with the large white dial disk added.  The little graduations on the blue sticker were the original setting scale. The left micrometer thimble is the diamond downfeed and the right micrometer thimble tilts the whole dresser to get a dress that is parallel to the chuck. The black Vespel block in the middle with the fine wire is super glued to the moving part of the dresser and has the electrical connections on the back side. The other block has a .1mm brass shim bent into a wiper and is soldered to the two copper tubes that are visible on this side.


Operator view of the new large dresser dial and the linear pot glued in place. notice the copper posts for the electrical connections to the pot.


Power Designs sitting in the grinder coolant pan supplying the voltage across the pot and the differential probe connections to the pot.


Setup on the grinder table with the 1054 fed by the Preamble 1855 diff amp. 48GX in the bag for the calculations. The 87 is reading the voltage across the pot for a sanity check. I hooked up the scope direct to the pot first but the switch mode ballasts only a meter away were giving me some spectacular Oscillofun displays so I had to break out the 1855 to get rid of the all the common mode noise.


Setup was DC coupling, 1X gain, /1 attenuation, 100kHz filter, 100meg input impedance.


1054 showing the rate measurement at the bottom in V/s.  At 400mV equaling 1" of travel and needing to calibrate in inches per minute I multiply my desired inches per minute by .0066666 to get the volts/second that I need to adjust to. I just keep adjusting the dial setting until I get the correct volts/sec. I used single shot mode to take each test.


Once I get the dial position correct to get the V/s I am shooting for I marked the dial in pencil with the corresponding Inch/min value. I did this for the entire range from 5 to 160 inches per minute. Notice the severe nonlinearity of the scale at both ends.


I then milled the graduations and engraved the numbers and filled them with epoxy filled with carbon black. After it cured I sanded of the excess leaving a crisp dial.


Now its time to confess how I chased my tail for about an hour and a half with the scope and diff amp setup. I started with 1 volt across the pot set by the PD power supply. I fumbled around trying to get 1V pp on the display of the 1054 but it would only read 984mVpp and only when I set the probe factor to 2. now I am thinking I have found another bug that was caused by doing the 1054 upgrade. I decide to put the 87 across the pot as a sanity check and it reads what you would expect, full voltage at one end of the pot and close to zero at the other end. What's up? So I start dropping the voltage on PD to see if I get a change on the scope pp reading. Nope just stays at 984mV, now I am convinced there is something wrong with the 1054. Finally (after a few more laps of chasing my tail) I remember one of the main features of the 1855, it limits the output to the scope to roughly 500mV so the scope is never overdriven into its possibly long amplifier recovery time.  So I drop the PD supply to 400mV and set the 1054 probe setting to 1X and the world starts spinning on axis again. What a trap for young players like myself.

[Lightages]

Thanks for the fun post robrenz. I do have one concern, that the hydraulic stopper is stable with temperature. Hopefully you have a climate controlled shop. I don't, not yet.

[rx8pilot]

Drooling over the Preamble 1855.........

[Paul Moir]

This is exactly what happens when you presume to become the king of all trades. 

[robrenz]

Thanks for the fun post robrenz. I do have one concern, that the hydraulic stopper is stable with temperature. Hopefully you have a climate controlled shop. I don't, not yet.

Yes climate controlled shop and the Kinecheck uses silicone based fluids which have very low viscosity change with temperature.  I did notice that there was a small droop in system air pressure when actuating the dresser because the regulator response is not fast enough for impulse loading. This showed up as some inconsistency in the repeatability of the velocity. So to add some output capacitance to the air system I temporarily added about 90 feet of tubing between the regulator and the directional valve to act as a small accumulator. That made the velocity much more consistent so I will add a small accumulator to the system.

[robrenz]

This is exactly what happens when you presume to become the king of all trades. 

"When you think you are the Jack of all trades, you don't know Jack!" (robrenz 2015)