[mduckett]

Finally finished up the front Y-axis way cover. Bent a piece of acrylic and built up an aluminum bracket and made some gaskets for the front where it joins to the saddle. I still don’t like this setup, but this will do until I can come up with something better. I made the gasket from some 3/16"” and 1/32”" Buna (nitrile) rubber sheeting from McMaster-Carr. A piece of the thick rubber was cut to fit between the bracket and saddle, and it has a pocket cutout to fit around the new X-switch block, so when the bracket is pulled up tight to the saddle it pinches the gasket and seals all the way across. A piece of the 1/32”" rubber is laminated to this to make the seal across the pocket so that it seals against the switch plate. What I don’t like about this arrangement is the switch cable comes down from the switch block and complicates the seal. I didn’t want the wire exposed on top, so this is it for now. I added some angle aluminum side pieces to the bracket to hang down and block exposure at the front of the bent plastic cover, and also added a couple of angles that fit against the saddle and act to channel coolant down from the top of the bracket/gasket safely away from the front Y slides. The old cover allowed some coolant to run straight down the bracket and some would find it’s way onto the front slide area. The plastic cover is a bit wider too this time and a little thicker so it doesn’t sag down onto the motor cover. A nice feature on this design is that the cover can be removed with just two screws, and the bracket doesn’t need to come off to work on the motor/bearing assembly. That thing standing upright in the picture with the gasket is a tool I made to drill holes. Its a piece of 3/8' CRS with a hollow tube drilled/turned on the end. Sharpened the end with a file in the lathe. Chuck this in a drill and cuts perfect holes in the rubber like butter.
Made a couple other new parts that have been on the ToDo list for a while. First is a new draw bar bushing to replace the original that is too loose and allows the draw bar to go off center and causes vibration. The other part is a new cover for the non-driven end of the X-axis ball screw. When I turned that end of the screw I had some excess length to allow for the center drill and tail stock center, and I decided not to cut it off. I thought I could make the old cap work, but there wasn’t enough thickness, so I turned a new cap from some scrap aluminum, primed and painted it, glued in with some silicone. I could have used Delrin but didn’t have any on hand with large enough diameter. It’s nice to have that end sealed, as coolant runs off that side of the table.

[mduckett]

With the new Y-axis way cover back in place I was anxious to do another pocket cut to see if the bearing shims and other changes had corrected the issues with cutting round pockets. I put the vise back on, aligned it, and put some 3/4" ”x 1"” scrap aluminum in the vise. I used the NFS pocket wizard to make the g-code for a .800”" x .375”" deep pocket. The setup used a 3/8”" 4FL TiN endmill, running .020” DOC, 40% stepover, 8 IPM, which I increased to 10 once it got going. Cutting was smooth, flood kept chips moving out. Results: Bad And weird. The pocket is somewhat oblong, and has two sloping sides and two straight sides at the cardinal points. The X-axis direction is worse than the Y, so I took that bearing assembly apart and checked it. It is very tight and smooth, no backlash evident. I put it all back together and will do some more cuts to see if I can figure out how this can happen. Anyone ever see this kind of behavior (see attached cut-away sketch of the pocket and toolpath)?

[Alax7]

From your sketched pictures it seems like your z axis would be the issue. Backlash in the x or y would not cause an angled cut. Maybe the head is loose and when you change directions, it moves and angles the cutter. Backlash in the x or y would give you oval circles not angled cuts.

[mduckett]

Thanks, I agree with you, this is not backlash. I pushed/pulled on the head and spindle, not a quantitive measure but it seems tight. I'll follow up with a dial indicator though. I'm thinking that missed steps/bad control could account for the drift/sloped walls. I've noticed on occasion when jogging I hear a blip or small interruption in the sound of the motor when jogging over several inches at slow speeds. I don' t know if that is a missed step, or a mis-timed step. I did some reasearch and found that sometimes the optocouplers in these systems can delay the rise time of the direction pulses such that the step pulse may occur before the direction change, causing a step to occur in the wrong direction. Apparently the Dir Pulse parameter in Mach3 motor tuning controls the time between when the direction is asserted and the step pulse is issued. I have mine set to 5 microseconds, but will try some larger values to make sure that the timing is not on the edge.
Assuming that this is occurring and the events are semi-random, what could have happened in this case is bad steps occurred for most of the run when travelling in -X, -Y. The entire hole then drifts in that direction, with slanted walls developing behind the cutter. The cutter removes the sloped side in the direction of travel, while leaving the opposite slope, which is sloping away, untouched. The program ends and I'm left with one square side and one sloping side. I checked the part again and there is a small ridge at the bottom that I didn't pay close attention to, so this may be what's going on. I can probably get access to an oscilloscope if needed, but I will try to change timing values first. --md

Edit: I had it backwards in terms of the direction of the hole migration. It appears to be fairly constant in that direction. The important part is that the migrating hole theory does end up with a set of square sides with the opposite sides sloped. Now to find the cause...

[mduckett]

I borrowed a portable scope from work this weekend to check the timing on the pulse train step/direction commands. It didn’t take long to find a potential problem. The posted pictures are pictures of the scope plots. Each has two traces; the blue is the direction signal, the red is the steps. I'm running my setup at 25KHz, so the timing is based upon a 40 micro-second (us) clock. To get the data, I just ran the pocket program and probed the X axis step/dir outputs. I have my controller box disconnected from the machine and the 36 V power supply turned off. Nice thing about steppers is open loop doesn’'t care that nothing is connected to the outputs. Anyway it appears that Mach3 outputs direction changes 1 clock period before starting the pulse train on an axis. I setup the scope to trigger/capture on the rising edge of the direction signal, then repeated with a falling edge trigger/capture. The first two pictures show these. You can see that when the direction changes from high to low, there is no problem, it occurs rapidly and well before the step pulse (red trace) occurs. But for the low to high transition, there is a major delay, nearly half the cycle (20 us) before the direction stabilizes. In a perfect world, this would still be ok because it is occurring before the pulse by about 20 us. However if an interrupt occurs on the PC and delays the output of the direction change, it could make the transition occur AFTER the step pulse, which would result in a step in the wrong direction. It is also likely that some additional delay occurs inside the stepper drivers, because they also have an opto-isolator, so the 20 us cushion against timing errors in the PC is probably much less. The next task was to discover where the delay was occurring. I suspected the opto-isolators, but on my board there is also a Schmidt-trigger buffer chip and there was the possibility that the 3.3V output from my parallel port was not consistently triggering it. However, as seen in the next two pictures the signal through the buffer chip is pretty perfect in both directions. So it must be the opto-isolator. I took a picture of the board so I could zoom it up on the PC, and luckily all of the parts and resistors were identifiable. The opto-isolator is a common part (817C) made by a variety of vendors. Anyway, after doing some research and calculations, the 1KOhm pull up resistor on the output side is a bit too large for the setup, resulting in a sluggish pull up. The proper value computes to be about 384 Ohms, so the value to put in parallel with the existing 1K is about 625 Ohms. I found some 500 and 150 Ohm resistors and soldered them into series. Fortunately, my BOB has a pin header in addition to the wire terminals, and also a 5V pin nearby. I had some pre-soldered pin socket wires from RadioShack, so I split one of the wires and put the resistors in the middle, plugged one end to the step pin, the other to the 5V pin, putting it in parallel with the 1K. Powered on and captured more data. Now you can see in the next two pictures that the low to high transition is much quicker (still not as fast as the other direction, but much better) and this also verified that the fix left the high to low unaffected. I'm pretty confident that this is likely the cause of some if not all of the problems I’ve been having. The plan forward is to rig up two more resistor pull ups for the other channels, put everything back together and do another pocket cut test. If it works, I’ll probably go back and solder some surface mount 625 Ohm resistors on top of the existing 1K’s to make a more permanent fix on all of the channels.

[SCzEngrgGroup]

I don;t know why the people designing breakout boards have so much trouble designing proper opto circuits that don't mangle the signal! They are so bloody simply to setup properly. I've had several that I had to re-bias because they were absolutely destroying the signal quality. One was completely filtering out the Step pulses at about 25kHz. After a few resistor changes, it was producing perfect waveforms at well over 125 kHz.

Regards,
Ray L.

[doorknob]

If your stepper drivers have built-in optoisolators, doesn't that make the optos in your BOB unnecessary?

Putting your step and dir signals through two sets of optos will likely degrade the signal further, even if you have cleaned it up on the BOB.

Perhaps you could run some tests bypassing the BOB optos entirely.

[mduckett]

If your stepper drivers have built-in optoisolators, doesn't that make the optos in your BOB unnecessary?

Putting your step and dir signals through two sets of optos will likely degrade the signal further, even if you have cleaned it up on the BOB.

Perhaps you could run some tests bypassing the BOB optos entirely.

I agree, it's not a good idea, but should be workable; as Ray points out these signalling rates are not that fast. When I bought the electronics I really didn't know that much about it, and went for the isolated BOB, not knowing the drivers would also be isolated. I suspect a more expensive version would have a better setup. Keling has a wiring diagram with my exact system setup, which I followed, so I assumed it would work. I would rather have all the isolation in the BOB, closest to the PC, instead of in the drivers, but I'm hoping this will work out with the fix. If not, at least I know where the problems lie and can try as you suggest bypassing some of the isolators. --md

[mduckett]

Tried to get things set up last night but found out after hooking everything up that the X and Y axis limit switches were off line. Arrrrrgh. I was able to play with the motors though and right off noticed that there is a definite change in sound. I guess I’d say smoother, but something is different in the sound. I bumped up the velocity because it seemed that the movement was slower when jogging. Not sure I understand that, but anyway it seems nice and smooth now, and rapids are very smooth. It was too late to do much investigation on the limit switches, but I ohm’d out the cable and found everything was good on the mill side. Was worried all day that I popped something when I was playing around. Unhooked everything tonight, opened it up and found that the common ground connection shared by the X, Y switch inputs had popped out, hard to see unless looking for it. Re-connected and everything was back on line. Loaded up the metal block from the last test and re-ran the pocket program. The interpolation is still noisy, but less than before, again seems to have smoother transitions. The picture shows the two pockets side by side, left is noticeably more round. The bottom of the other is scuffed up from measuring squareness so ignore that. The pocket is supposed to be .800” diameter x .375” deep. The new pocket measures .798 “(Y) x .7995” (X) and .380 deep. I’m not surprised that it’s out a bit in the Y direction as I haven’t calibrated that since fixing the bearing slop. Little surprised by the Z, and that depth error is very repeatable with the other pocket. I also know that the tram is out about .002” in the X direction, and .001” in Y, so maybe some small contribution to the depth error from that. Overall I’m very happy, way better from where it was a few days ago. I’ll pull out the BOB soon and make permanent fixes for the pull up resistors on all the channels.

[mduckett]

I've been making steady progress on a number of items, but have been negligent on posting regularly. First, after success with the temporary resistor changes on the BOB, I got around to ordering the proper 625 Ohm (1%) surface mount resistors and succeeded in getting all of the output resistors updated (13 in all I think). The process involved putting a tiny amount of solder paste at each end of the existing resistor, placing the new resistor on top with tweezers, then carefully heating one end to tack in place then finish tacking the other end. Not too bad once I got the hang of it and it took about 40 minutes. I confirmed the resulting resistance values for each with the multimeter. FWIW the first picture shows the row of output resistors with their added resistors. Sorry it's not easy to see unless you zoom in though.
On the spindle speed/direction controller project, with enough of the prototype circuitry finally built up, and after some software debugging, the PWM to Voltage conversion function is working. I set up Mach3 spindle control and wired a cable over with the direction, PWM outputs and ground. For this prototype testing the isolation chips were not installed. Instead, just jumpers across the pads and connected the grounds for each side of the isolation. While it's taken a long time to finish the prototype, it was well worth the effort as there were additional problems remaining in both the hardware and software that needed to be found and fixed. Second picture shows the prototype, not pretty but it works. All of the problems discovered have been addressed in the next revision of the schematic and layout, and this design was then modified to include the plug connections to allow (in theory) this board to accept all of the cable/plugs that currently go to the stock speed control pot board. The new boards were ordered last week, and should arrive in 8-10 days. Most of the speed controller design features were described in a previous post, but briefly, this next board will receive 5V power, ground, PWM, direction, enable, and spindle index pulse signals from the PC side (BOB), and the PC side signals are connected to the mill side via high speed isolators. On the mill side of the isolation, 5 Volt power comes from the main connector, and the circuitry develops the voltage and direction signals and feeds these signals into the main connector for spindle speed and direction control. The spindle motor controller fault signal routing to the fault LED is preserved, and is also buffered and sent back to the PC side via isolator so spindle faults can be detected in Mach3 if desired. An add-on feature will produce a spindle index pulse from one of the motor's Hall sensors, similar to Don Bird's design, and this will also be routed out to the PC side via the isolators so Mach3 can use it for spindle speed monitoring and closed loop speed control. More details for this feature in future posts. This controller will attempt to generate "smart" motor controls based upon the enable, direction, and PWM outputs, for example instantaneous direction changes are not allowed, but will require the PWM command to be reduced to zero, and if available, the detected spindle speed (via index pulse) to fall below a threshold before allowing the direction to change state. The existing motor controller logic may already have these features, but I'm not sure as reversing the spindle isn't part of the stock mill setup.
On the mill enclosure, I did some more sealing with silicone on the back corner where the hopefully last leak was occurring. Also added some braces to connect the sides with the back panel to stiffen the enclosure. In order to keep it easy to disassemble, I riveted some 1/8" plates to the under sides of the side and back panel angles, and drilled/tapped them to accept a single bolt to hold the braces in place. With a few more mods, the enclosure will be able to be removed in 3 sections (doors, sides, and back) very quickly.
Plans: Hope to get some run time in this week to confirm operation with the BOB resistor changes, re-calibrate all axes, measure and apply backlash corrections to all axes. When all of that is complete, it will be time to start the next project, which is a spindle light assembly made from a white "angel eyes" LED ring.

[mduckett]

The new PCB's have arrived, so the speed controller is one step closer to completion. The picture shows the format for this version of the boards, left side pretty much the prototype layout, right side has the connectors. I installed the connectors (not soldered yet) to show how this board will be plug compatible with the original potentiometer board connectors (that's the replacement pot LMS sent me during a recall on that part, but my original seems to be ok). Hope to start stuffing this board soon, but realistically it will be at least 3 weeks before it will be ready to install into the machine. I'll be posting progress.
I borrowed my co-worker's wireless pendant, which is one of the Chinese $100 units that uses the ShuttlePro driver, to evaluate. I want to get a pendant, and wanted to try out this type, to compare with the VistaCNC units which appear to be a better product according to reviews, albeit at a higher price. The Chinese unit actually performs well, however it does have some quirks that are annoying. The jog mode does not always cycle between Continuous to Step to MPG, instead it goes between Continuous and MPG unless you press the button to select the step size, at which point it switches to Step mode, but also advances the step selection, so you have to cycle it back around to where you wanted it. At times, the unit also seems to miss steps when turning the hand wheel, but that may be a setup issue. On the positive side, it was extremely easy to get up and running, just drop the driver .dll into the Mach3 folder, plug in the USB transceiver, power on the unit and it came right up. One other criticism is the display shows 3 decimals, and the step selection shows the same value for thousandths and tenths. Overall, it would be worth the $100 if budget is tight, compared to $155 for the comparable VistaCNC unit. I really enjoyed the pendant for setting up the work and doing ''manual" cnc operations, so will need to plan on getting something in the future.
Decided to give the machine a workout to see how well the fixed BOB resistors helped on the step and direction signals. First step was to calibrate Y and Z axes, then did some backlash measurements on the Y axis. At the middle of the travel, backlash is < .002", farther away it goes to about .0035" repeatably. Didn't turn on compensation for this run though. Moving on, I had a scrap of aluminum that could be sized into a 1.5" cube, so I made up the tool paths for a 1.5" turner's cube. This requires six repeated setups, and since all it does is circle pocket cuts, it is a good test to see if the machine can maintain it's positioning. First step was to face the block to size, got it to within .002 or less in all dimensions. Next, put it in the 3" vise and aligned it to the edge of the vise to establish a repeatable position after changing to the next work face. Used the touch probe to set the work zero point at one corner, touched off a brand new, 3/8" 2FL HSS end mill, and started the run. The first three operations required about 11 minutes each, using .025" DOC and 25% step over (feed 10 IPM, plunges 1 IPM), followed by a full depth (.250") finishing pass at .002". In the remaining 3 setups all feeds were reduced to 50%, 50% and then to 40% for the final passes in order to reduce the loads as I was concerned that the metal might break away from the corners. Thankfully it all stayed together and it took over 2 hours to complete the machining, not counting the facing operations. The part turned out pretty nice, the only negative is the end mill marks at the bottoms of the pockets. A couple of quick measurements shows the machine cut nearly perfect circles. The flood system worked great moving out the chips, but once again another leak was uncovered in the long operation. The fix for the previous leak held, but this is a different leak. After figuring out where the leak was occurring, a piece of flashing over the area to funnel the coolant directly to the drain stopped it instantly. Easy to fix but annoying, maybe this is the last one...ha. Sorry about the marginal picture quality, my cell phone was the only camera handy, i.e. I'm lazy. One other potential issue to address will be some active cooling, i.e. a fan for the spindle motor. After running for over 2 hours straight, it got too hot to hold a hand against it. I used a large box fan to provide some air flow across the top of the enclosure, but a small fan directed at the motor will help out greatly. Stepper motors held up fine, got pretty warm but within expected levels. Spindle light is still the next project, after fixing the flood leak.

[InMesh]

This is my router build ,1850 mm x 850 mm area or there about after 8 weeks it is kind of finished nice and rigid removable table segments ,now for the electrics going to have a fourth axis running 34 nema stepper motors all will be good.

[mduckett]

Nice build. Did you mean to post here, or were you trying to start a new thread? To start a new thread, go up one level to Benchtop Machines and click on Post New Thread. If that isn't visible, click on the Title/Thread Starter on the border to get the dropdown for starting new threads. --md

[cncadmin]

Subscribed

[mduckett]

There has been some interest in plans for my LMS conversion, so I'm posting some of the requested drawings for the X and Y axes parts. These drawings should be very close to what has been presented in this build thread, but no guarantees.

Status on the speed controller and spindle light projects: Both are moving forward at a snail's pace, hope to make some progress that is worth reporting before Thanksgiving. --md

[Brass_Machine]

Thanks for posting those! Working on my refit of my currently CNC'd X2 and I like your design.

Eric

[mduckett]

Thanks for posting those! Working on my refit of my currently CNC'd X2 and I like your design.

Eric

You're welcome. A couple notes on some things I noticed in the drawings: 1) The 10mm ID x 12mm OD spacers used on either side of the bearing can be made, but I bought those off the shelf from a skateboard shop (Daddies Board Shop, Landyachtz Aluminum Wheels Spacers 10mm x 8mm Set of 4), and 2) the optional bearing seal is incorrectly sized in one of the drawings, it should be 22mm x 12mm, and 3) the Y-axis bearing block drawings is missing the distance from bearing center to the lower mounting bolt center, which is .798" down. This dimension is shown on the lid drawing but omitted on the bearing block.
When I get some time I'll add the Z-axis bearing block/drive and update the drawings. --md

[mduckett]

After a long pause I'm starting back to work on the mill. In the interim I've used the machine to make some small parts for a few projects and even for some wood working where I needed some accurate holes drilled repetitively on multiple pieces of material. I've made no progress on the spindle controller, though it is still on the list of things to finish. During this time I've also discovered that one of the frustrating things about converting the mill to CNC has been the loss of the ability to just walk up to the machine and use it for basic milling/drilling without the need to connect and boot a computer. I don't want to leave my Mach3 laptop computer in the garage, so the setup time has kept me from using the mill as much as I could have. Some people have left the hand wheels installed on their CNC machine axes for this purpose, but with my enclosure it wouldn't be practical. I hate to start a new project with others waiting to be finished, but I see this as a fundamental capability that will help to make the machine more usable. So, the solution I'm planning is to build up a standalone, 3-axis Manual Jog Controller (MJC) (see attached rough CAD models) that will provide "manual" mill operations on X, Y, and Z via the stepper motor interfaces without the need for Mach3; rather it will interface directly to the parallel port BOB, with the stepper motor movements controlled by step and direction inputs generated by the controller. It's basically a narrow box with 3 hand wheels, a display, and an assortment of switches and buttons for the various controls and a microcomputer to pull it all together. This controller will attach to the front of the mill stand, convenient for reaching the work for setups and touching off cutters. This is going to take a while to build, and I plan to cover this in some detail, so it will be broken into small segments. The next posts will cover the overall features, plans, and design for the controller, followed by build progress.

[mduckett]

A while back I evaluated a friend's < $100 pendant controller and found it to be nice but with a few shortcomings. Some of the other pendants from VistaCNC, at almost $300 are very nice (from reviews, never tried one myself), but I started thinking that maybe for that cost I could make something that would work better for the way I want to control and use the mill.
What I want is a controller that gives me a manual operations interface, i.e. two hand operations for milling (X,Y table movement), with a third Z control. With the stepper motors in the control loop, very precise positioning control, as well as auto feed would also be possible. DRO displays for each axis are mandatory, and some other programmed functions would also be nice but can be added over time. I still want to have access to CNC operations too. Here's the feature list so far:

Hand wheel jog controls for X, Y, Z axes, with individual calibrations, step scaling (.001, .010, .100, etc.), backlash compensation, individual axis on/off controls (simulates locking an axis, and also to prevent inadvertent movement from bumping a hand wheel).
DRO displays for X, Y and Z positions, with ability to set zero on each
Auto feed on any single axis using a cruise control style interface to set/resume/cancel the feed established using the hand wheel. Feed could be set via display interface too. The Resume function would use the direction set by moving hand wheel and pressing Resume button.
Support for functions such as peck drilling cycles and bolt circle/patterns, maybe circular and rectangular pockets, etc., again to support rudimentary operations.
MPG inputs for Mach3 when operating as a CNC machine. I'll need to add more inputs to support the 3-axis MPG, though I probably have enough inputs to support a single MPG input.
Spindle speed control (via display interface/controls; this is future implementation once spindle controller is finished)

There will also be a USB interface for re-programming software and maybe assisting in setups that are too complicated for the onboard micro to handle through the built-in display, such as calibrations and backlash determinations (remember I generally don't want to hook up a computer for most operations).

[mduckett]

This post is an overview of the electronics and mechanical workings of the controller. Each hand wheel is connected via a shaft to a rotary encoder inside the box. I picked 24 step encoders with detents at each step. To keep costs down, these are mechanical encoders, but they have 100k revolution life rating so they should be ok for longevity, and are cheap (< $5 each). The shaft is supported by a small ball bearings so that the encoder is not subjected to any radial or axial loads; only rotation torque. The encoder shaft slips into the end of the hand wheel shaft and is secured with a set screw. The encoder end of the shaft is threaded on the OD so the hand wheel can be pulled tight against the support bearing using a nut and spacer. The encoders are panel mount types, so a U shaped bracket supports the encoder body over the end of the shaft. The encoders each produce two square wave outputs that are offset by 90 degrees (in quadrature). These signals go into a microcontroller for scaling and can also be used for speed and direction determination, or ignored if the enable switch for that axis is turned off. The encoders can also be routed out as MPG signals for use with Mach3 in CNC support mode. The top of the box is angled at 45 degrees and has a rather large 4x40 LCD character display, which is also interfaced to the microcontroller. To the left of the display are three buttons that will be used to zero the X, Y, Z DROs and maybe support other functions such as axis calibration. Further left is the auto feed "cruise control" buttons: Set, Resume , Cancel. On the right side of the display are either 2 or maybe 3 buttons for menu navigation/selection. To the far right is an eight position rotary switch to select the jog mode and other functions, such as changing to Mach3 support mode. So far I have spent about $65 for parts including shipping. The electronics came from Digikey and SparkFun, and bearings from VXB. I already had a microcontroller board (also from Digikey), so adding another $20 for that, I'm at $85. There will be another $20 or so for wiring and other stuff before it's over. The enclosure is going to be 1/2" birch ply and 1/4" sanded ply panels, all scrap from other projects. Fasteners and incidentals will add a bit more, so the final estimated cost should be $125 or so. The controller will mount to the front of the mill stand using some kind of sturdy but removable bracket TBD. The hand wheel arrangement is: far left is Y axis, far right is X axis, and center is Z. Distance between X, Y hand wheels is about 10 inches. The interface to the BOB will be a standard 25 pin printer cable, exiting the bottom of the box. When operating as a Mach3 MPG, there will be another smaller connector, probably one of the round metal connectors, forget what they are called, with 8 to 10 pins. Attached some pics of the electronics. Note the red button switches, those are momentary normally open type. Digikey had the little rubber caps that will make those waterproof. The three silver buttons have green LED lamps and are latching type switches. Those are the enable buttons for each axis and will light green when the axis is enabled. The three silver parts wiht flat shafts are the encoders, and the rotary switch has the long handle and an aluminum anodized knob. The large LCD display is 4x40 characters with a nice white backlight. It should have enough space to allow displays of X, Y, Z DRO, plus spindle speed, and support a menu area for function select. I decided against a keypad, at least for now. I started making some parts, so the build posts are next.