[ezakimak]

Then I went ahead and brushed, buffed and polished the parts because I figured that would take ony a few minutes per part.
For brushing, I used these sanding sponges you can get at the hardware store for cheap (150/220 grid, doesn't matter all that much).


WP_20130507_003 by mkloberg, on Flickr

Since the parts still had tap-magic all over it, that make for a neat wet sanding process.
You may want to wear some latex finishing gloves for that, the buffing and polishing, that got kind of messy.
For buffing, I used a felt wheel on my bench grinder with green buffing compound.
For cleaning and polishing, I used regular kitchen towels, white shop rags and a product called 882 Liquid Metal Polish from Duragloss.


WP_20130411_001 by mkloberg, on Flickr

I suppose any chrome/metal type automtive polishing product might do, but this stuff really made the parts shine almost like chrome (available from Northern Tool).
Meanwhile the motors, the belt, the thrust bearings and some other odds and ends arrived at the shop.


WP_20130407_003 by mkloberg, on Flickr

found it, Thanks.
link worked.

[vtxstar]

More stunning planning and work. Bravo Mac!

For the casual readers, the black residue mentioned in Mac's polishing directions is aluminum oxide -- rust.
Aluminum rusts black.
Iron/steel rusts orange-brown.
Humans rusts by wrinkles.

-=Doug

[Mac.CNC]

Thanks Doug :-) How's your machine coming along?

Sorry, this post has no pictures - it's about the controller grounding and shielding strategy and what happened in this machine by accident.
I have a background in electronics and know a few things about shielding cables / basic grounding of things in a setup like this, but I'm still wondering if I'm on the right track.

Status: I think the GeckoDrive grounding and stepper motor shielding cable issue I came across sort of resolved itself, although I'm still unsure if this is correct or not.
Because of the secondary PC style power supply unit I'm using for 5V and 12V, which I also shared the DC- with the Gecko PSU DC- earlier, it turns out that its now common DC- is connected unintentionally directly to earth ground!

I've read a lot of posts that don't recommend doing this, but I guess I did it by accident without having any issues so far.
I didn't expect this at all, the way I discovered it was by placing an ohm-meter between the common DC- and the earth grounding terminals on the machine.
The meter started to beep, indicting a direct connection. First I couldn't figure out where I could have possibly connected the two grounds, because I didn't want to do this on purpose to begin with.
At some point I pulled the power plug out of the small form factor power supply and the beeping stopped.

After doing some more research on late model PC power supplies I learned that pretty much all PC PSU's sold these days have a direct connection between the DC- they output with the earth ground wire of the AC cable that is plugged into it.
Most of the time, the case of the PSU is connected to that too - basically grouding the entire chassis of the PC that way when its mounted inside the box.

More measuring on the cut off motor cables showed that each axis on the G540's motor connector pins 2,3,4 are connected to its DC- input as well, now grouding the shield of the cables to the common DC- and earth ground too.
Strangely, the metal case of the G540 is not connected to this potential.
The premade cables from CncRouter parts have their shields connected to these pins (2,3,4), so that might just work out of the box as advertised.

I also measured these pins at the motor ends, they seem to have no connection to the cases of the motors, which falls into the best practice of shielding the cables at the controller end only - to avoid any ground loops.

I can't say that I'm convinced that this is best practice, on an industrial machine you would never connect any minus potential be it DC or AC to earth ground like this.
I guess I'm just going to see if this will cause problems or not. The alternative is always to find an isolated version of the small form factor power supply I used or make my own 5V and 12V from the GeckoDrive 48V PSU using voltage regulators.

What is your take on this?
--
Thanks in advance for any input,
Mac

[Mac.CNC]

Meanwhile, I worked on converting the low cost x-axis cable carrier track to the IGUS chains.
That went quite well and I don't regret any bit of it, it was quite worth it.

These chains allow for changing out cables witout taking "everything" apart (equipped with a soldering iron), because they flip open and seem to be a lot more stable.

For the Y axis , I made a similar platform as before, raising the game to just above the belt:

WP_20130831_001 by mkloberg, on Flickr

This is how the IGUS chain comes apart, making it easy to feed new cable through the existing chain.


WP_20130902_001 by mkloberg, on Flickr

For the Y-axis I dedcided to stay with the same proximity switches used on the x axis, but I couldn't find a suitable target for these switches.
Going back an forth a few times amongst ideas, I thought the that the belt clamp would make a great target - if it was any bigger.
Making a higher belt clamp worked out perfectly in this configuration:


WP_20130902_002 by mkloberg, on Flickr

The proximity switches to target the new belt clamp are mounted under the support plate for y cable carrier. That made it a little more complicated that it needed to be, but no big problem.


WP_20130903_002 by mkloberg, on Flickr

At some point or another it occured to me that we need to make connections to the limit switches on the y and the z and need to find a place to do that.
This could have possibly been done "inline" with shrink tubes and stuff, but I wanted something more maintainable for repairs.
A while ago, I came across these little boxes at the radio shack store and remebered that. They sell these as "project enclosures" and one of them (the smallest one they sell), was about just the right size to become a y-axis junction box. ;-)
Perfet fit, a standard terminal strip ground to size fits right in there.


WP_20130903_005 by mkloberg, on Flickr

There was a lot of wiring to do:


WP_20130903_010 by mkloberg, on Flickr


WP_20130903_011 by mkloberg, on Flickr


WP_20130905_006 by mkloberg, on Flickr


WP_20130905_012 by mkloberg, on Flickr


WP_20130905_013 by mkloberg, on Flickr

Whoa, this is getting close to let the chips fly (as in cutting something).
This morning, I spent some time on researching the ER16 collet that this spindle has at its end.
About an hour later, I was down a good $150 bucks to get another 5 collets that are needed to actually machine something, in order to clamp some common endmills for the future (1mm, 3mm, 1/4" 3/8", etc.)
Sorry for the rant :-) Probably will ge more costly when it comes to buying the end mils, bits, etc.


In conclusion, I finished wiring the little box and started playing with the machine, the new y limits and established the soft limits for the y.


WP_20130906_002 by mkloberg, on Flickr

Having the X and Y in check now, I felt confident enough to make a little video:

DSCN0040 - YouTube

Next up is to build the Z-axis.
Stay tuned,
Mac

[ezakimak]

Do you have enough clearance when the X axis is fully home for a machine cover to go on?
looks like the new chain sticks out the back a little

nice work though.
Ryan

[Mac.CNC]

It doesn't, but you're right, the chain gets very, very close to cover perimeter.
Good question, glad you ask :-) It took me a while to figure out how to not let that happen, the trick was to choose a chain with a very tight bending radius (1.5" in this case) and position it so that it would not have all that much overlap at the end of the travel. I have to say that a lot of that wasn't foreseeable, because you can't in CAD unless you build a full armature model with hinges, gravity and whatnot of the chain and be able to move it around inside the model. What i'm using is Sketchup and it doesn't do all that, so I just went out and guessed some of these paramters here and there.

Meanwhile, I worked some more on the next item, which was the z-axis.
For the cutter, I'll be using a square, aircooled 0.8kw spindle I got from Ugra CNC.
I wasn't sure how to mount this thing at first, but once I took it out of the box it came in, I found 8 M5 threaded holes on the back side of it.
Cool, that we can work with.
I designed a spacer to clear the z bearings and an adapter plate to mount this thing to the z-axis and made the parts:


WP_20130908_001 by mkloberg, on Flickr

8 M5 bolts with retaining rings are holding the spindle in place, that should never come off I hope.


WP_20130908_002 by mkloberg, on Flickr

While I was prepared to take the spindle back off the z-axis to actually put it into the machine, it turned out that it's possible to just feed the entire axis, with the spindle mounted up from underneath through the opening in the base of the machine, without having o mess with the gantry at all, hehe.


WP_20130908_003 by mkloberg, on Flickr

The only thing left is a limit switch for the z, this will be a proximity switch as well.
I think I might just put it on the left post of the carriage making a custom mounting plate for it.


WP_20130908_004 by mkloberg, on Flickr

Stay tuned :-)
Mac

[cb_designer]

Mac, two things. First, HOLY CRAP!!! My machine now sucks in comparison. Nice work! Second, have you thought about how you are going to tram your machine? It doesn't look like you have added anything to the Z assembly for this, but i guess you could just shim it. Keep up the nice work.

[Mac.CNC]

Thanks :-) As you said, I'll probaby just put some paper/aluminum foil shims behind the spindle mount as Bob did that in another thread.
I was hoping though that I wouldn't even have to do that and could do most of the tramming adjustments through the bearings.
I was thinking that should be possible to minutely tilt the z-axis in any direction by adjusting the bearings in and out as needed.
We'll see how that works out...
--
Mac

[vtxstar]

Mac -- regarding your DC (return) and the chassis ground (earth).
You are correct in that they are two very different items with different purposes.
I have been employed in field service and installation (electrical/electronic) of medical imaging systems for 35 years and we make large efforts to isolate the two.
In fact, part of our annual inspections is to check earth bond (protective) continuity between cabinets and chassis -- different from the DC supplies of our systems.

It may require you to open up the power supply that has the DC and chassis (AC protective) ground 'connected'. Maybe you will find the link/jumper/connection that is making DC - and chassis protective ground connected and separate them. I have seen this be a simple jumper in some cases. In other cases I have had to completely isolate the chassis of the supply from the protective ground to prevent a ground loop.

Will it make your machine go nuts,burn out drivers, or make your LEDs flash/strobe? Dunno.
Is it best practice to keep them apart? Absolutely -- that's a major part of any country's electrical code and you already know they should not be together.
If you were producing the machine for sale, you would have to find and isolate this problem. As a matter of personal use....????

It reminds me of the question regarding AC neutral and Protective AC Ground.
The two are, again, different except they should be connected together at only one place -- the power company distribution point at the transformer.

Regarding my machine -- work got very busy (not so now) and I still find sanding and filling a chore and not so enjoyable. I have sketched up a base/table for the unit as it is still living on my table saw -- which I need to make the base/table for the Momus! I have to walk past the machine to get well into my garage/shop so...shame on me. I have also been decluttering the garage/shop, hauling away stuff that has been laying around and will be taking up space that the Momus will be using.

Build on Mac! She's a beauty!

-=Doug

[ezakimak]

further to what Doug has said, having the negative connected to the ground is also providing another place for any external voltage spikes to enter your electronics. suppose you have your machine running and the microwave/toaster in the house trips the safety switch. this is going to put mains voltage spike on the earth bar for a split second. with the negative connected to ground this will go directly into the control board and computer doing untold damage. a lightning strike with "earth potential rise" would be another issue that you would want to avoid.

it will be interesting to here if you can isolate it.

i work in railway (railroad in America) signalling, un-isolated noise from the outside world into the signalling system would account for a large percentage of faults that cause train delays.

Regards Ryan

[Mac.CNC]

Ryan and Doug,
thank you so much for your input, that totally confirmed my doubts on this topic. It works, but is not best practice at all.
What Ryan said about external spikes on the house ground bar is scary. I've learned about this many years ago in Germany when it comes to house wiring:
Connecting the AC neutral and earth ground in any way, or using one for the other was forbidden in Germany in 1973 (called classic zeroing) for those very reasons.
What I still don't understand is why this type of zeroing seems to be industry practice on the DC- in any PC you can buy today (as in connected to earth ground)...

Anyways, this is not a PC, so I took the auxiliary power supply out of the machine and cracked it open to see what's going on in there.
There were no jumpers, bridges or anything like that. The ground lead from the AC receptacle is directly connected to the metal chassis of the PSU with a short pigtail wire.
The DC- bus is coming out of the circuit board in the upper right corner (the black wires), which is also connected to the mounting pads of the holes in all four corners.
As the board is mounted in the chassis, these pads make contact with the chassis, grounding the DC- bus directly to earth through the chassis, the pigtail wire and the AC receptable.


WP_20130910_001 by mkloberg, on Flickr

I figured the easiest way to stop that from happening was with some 8-32" nylon screws and #8 nylon washers I found at the local Lowes in the specialty drawer.
The orignal screws were metric M3 so I had to enlarge the holes in the board a little bit and cut new threads into the chassis, but that was easy to do.


WP_20130910_002 by mkloberg, on Flickr

Testing the PSU with an ohm meter after mounting the board back into the chassis showed no evidence of a connection between DC- and earth ground anymore.


WP_20130910_003 by mkloberg, on Flickr

After reinstalling the PSU in the machine, it stayed that way - no connection at all between the machine ground terminals to the DC- terminals. Yay! Problem solved :-)

So, thanks again guys for giving me input and confirmation on something I thought wasn't quite right to begin with.

What remains a little strange is that the shield in the molded connector stepper cables from CNC Router Parts is connected to DC- through the G540.
While I don't think this is correct (shields should be earth-grounded at the controller end only), I'll probably leave this alone for now and see what happens.

Next up will be installation/power-up of the VFD unit and getting that connected to the spindle through the two cable carrier chains (that line will be a noise spewer and has me worried). Initially, I thought about running that cable inside the vacuum hose that will be dangling off the back of the gantry at some point to avoid that cable to be anywhere close to the stepper cables or the limit switch control cable. Since all that doesn't exist yet, we'll have to run it throught the chains for now and see what happens.The cable will be shielded of course, and have a maintenance disconnection point at the gantry.

Doing some more research lately on the amperage that cable needs to carry and the really tight bending radius I'm dealing with now (~1.75"), I had to find a very flexible and shielded 4 conductor cable for the spindle.
I selected one directly from IGUS: Chainflex CF10-UL-05-04
IGUS doesn't really have a click and buy store and you'll have to get a quote first, talk to an engineer and then call their customer service to place the order with the supplied quote. Still, going through all that, everyone was really helpful and friendly, even for the few feet that I needed, so thank you IGUS! :-) The cable should arrive at the shop tomorrow.

Meanwhile, a basic set of precision ER16 collets, some inital ballpoint, engraving and end mill cutter bits arrived. The chips will be flying very soon, whoho!! :-)


WP_20130909_001 by mkloberg, on Flickr

Stay tuned & thanks again for all the help :-)
--
Mac

PS: Doug, I'd be happy to send you a Sketchup model of the Momus workbench I built (standard construction lumber), you'd just have to shorten the design I guess. PM me if interested...

[vtxstar]

Mac,

A+! Good find on those mounting screws.
I have seen nylon/plastic screws in those type of supplies -- I am addicted to taking stuff apart, been doing it since I was a kid.

This supply obviously would not pass the electrical standards/codes and may exhibit what I call 'BILe' -- Brother-In-Law effect.
(The assembler may have had a brother-in-law in the fastener business who had too many of the wrong type but right size fasteners -- it is a joke but...??)
Or it could have just been a error on the assembly line.
Another item I see is the use of foam rubber type material rather than a heavy paper composite to prevent board flex causing chassis contact by the pcb.

Welcome to the world of statistical quality control.
Another possibility is an error on the circuit board that put the DC return on the mounting pads.
You could have cut that trace but your solution got the job done well enough.

In the interest of full disclosure regarding working on the Momus, I did participate in a contest on Instructables. That took a couple weeks of focus away from my build and that is all it takes sometimes to have a project slip. HERE is that project.

I will take you up on the offer for the bench -- if I can figure out how to PM you.

-=Doug

[Mac.CNC]

Hi Doug,
I did trace the circuit board of the PSU I'm using and the connection between DC- and earth ground is fully intentional by design. This wasn't a fluke on the production line I think and that's just what I kept referring to that I don't understand.
To me it seems that all PC's out there right now are wired that way, which goes completely against what we learned in this thread.

You don't find much information on this topic on the web which made me even more curious, so I took a multi-meter, pulled a brand new power supply and two old PC's from the shelf at work and checked this on them.
Sure enough, same thing: Direct connection between the DC- bus (the black wires coming out of the PSU), the entire PC chassis and earth ground on the AC plug.
I even popped a Dell PowerEdge 1850 Server we had on the shelf open, same thing there.


WP_20130911_001 by mkloberg, on Flickr

How strange is that?

Apparently this is actually common practice with personal computers, which explains some of the noise problems you read about everywhere when connection audio and video equipment to a PC.

Anyways, back to the Momus - the IGUS cable arrived and I'll be trying to spin up the spindle this evening. :-)
--
Mac

[ezakimak]

Interesting Mac, explains why I've lost so many power supplies after big storms. Will have to have a look.

We run dc traction return on the trains here and we do run electrolysis cables from other infrastructure to prevent stray currents from rusting them away. Gas and water pipes would not like stray dc current

[vtxstar]

Well then, that is puzzling but it also may explain why so much industrial work has opto isolated signals from OTS (off the shelf) PCs and servers...or the practice is to opto isolate somewhere in the drive system.

In the case of the G540 (same thing I have), it obviously is isolated.

I will check the frame supply I bought for this condition. It is not a PC supply but similar to those sold by the CNC vendors.

-=Doug

[vtxstar]

It is official, the frame supply in my project does have isolation between DC return (-) and chassis protective mains AC ground. I have no continuity between the safety ground and the -V on the terminal strip.

Just looks like we need to be aware that not all suppliers of component/parts adhere to industrial codes of electrical safety.

Attachment 200760

-=Doug

[Mac.CNC]

Hi Doug,
I can confirm your find on my main PSU that does the 48V for the Gecko and the steppers, no continuity between DC- and earth (I got mine from CNC Router Parts, btw).
Looks like you're in good shape on that one... :-)

I guess it would be a good idea to check the DC- bus and AC ground for direct connection once in a while while building a machine like this. This can sneak up on you where you'll never expect it during the build.
Good example, I'm grounding my maintenance connector shells for the steppers between the x and the y, to give them a place to unplung. Turns out that the leads on the grounding pins were touching the shell of the connectors, making this loop to DC- again (through the G540), fixed it by cracking these connectors open again and putting some shrink tubing over the three shield leads in there.

Meanwhile, I've been trying to figure out where to install the DELTA variable frequency drive I got from UGRA-CNC into this machine (VFD-M).
First I was a little scared of this little box, because its manual is about the size of 1/3rd of a phone book!


WP_20130911_002 by mkloberg, on Flickr

It's got 157 programable parameters to mess with... 8-)


WP_20130910_005 by mkloberg, on Flickr

It's got a display panel and buttons for programming, so it'll be nice to have this at least on the operator side of the machine.
Originally, I thought I could just tell it what to do and never worry about it (treat it like a black box, a SuperPID perhaps), but it turned out that there is much more to get a VFD working properly.
Anyways, the first task at hand was to where to put it, so far I was planning to keep it under the machine somewhere in the workbench - treat it like a black box power converter basically.
The other day I got a little curious and explored the possibility to mount it directly inside the operator panel bay.
The problem was that this thing is higher than the 3" we have available in the Momus wall, but after some goofing around it turned out that it would actually fit if we'd route a .5 clearance pocket into the door of that bay and it miiiiigth just fit in there then, lol.


WP_20130912_001 by mkloberg, on Flickr

Making the operator panel a seperate part from the door paid off at this point, because I was able to remove the door itself for modification without disconecting a single wire.


WP_20130912_003 by mkloberg, on Flickr

This opening had to be perfectly square, so I used a guided jig saw setup instead of the router to cut it out.


WP_20130912_002 by mkloberg, on Flickr

Routing the clearance pocket in the door for the VFD:


WP_20130912_006 by mkloberg, on Flickr

Subtracting the 3/4 for the outer skin, this will actually fit in there now! Mounting it this way pretty much violates every instruction of the operators manual but since the whole bay is ventilated, it shouldn't matter.


WP_20130912_007 by mkloberg, on Flickr

The hardboard mounting plate for the controls was temporary anyways, the final engraved aluminum plate needs to extend to the left by a few inches to cover up the VFD operator panel.


WP_20130912_008 by mkloberg, on Flickr

At this point, I'm still struggeling with controlling the RPM of the spindle accurately via the G540.
It kind of works, but I'm loosing untraceable control voltage between the Gecko and the VFD input somewhere.
That problem is probably home made by now, because the gecko is on the right side of the machine and the VFD is on the left (picking up noise from the VFD and everything else going on in the machine).
It's pretty strange, when telling MACH3 to run the spindle at 24,000 RPM (max), the G540 outputs 10V as it should, but as soon as I connect that to the VFD input, that voltage drops to about 6.5V.
I've been fighting with this for the past three evenings, tried pull up and pull down resistors, it just doesn't make any sense so far.
The wiring should be correct, 10V supplied by the VFD to the Gecko, one wire coming back. Everything is shieldedd to e-ground...
That has to be a noise problem, of some kind - although I don't see anything strange on the limit switches or the stepper motors.

Stay tuned :-)
--
Mac

[cotang]

Sound like a short to me.

[ezakimak]

Approximately 2/3rd voltage available, What configuration have you got the spindle wired. Is the spindle itself earthed? Might be some funny emf monitoring circuit that's back feeding through 1 phase of the spindle.

That's all I can think of at the moment. If I think of anything else ill let you know

Ryan.

[Mac.CNC]

Thanks Ryan and Cotang, for your pointers - never mind though, I figured it out meanwhile :-)
It was just noise coming from the VFD and putting a 4.7µF capacitor between GND and the AVI 0-10V spindle control wire to dampen this stuff out at the VFD end cleared it all up.


WP_20130917_003 by mkloberg, on Flickr

After that I just had to adjust some parameters on the VFD to get the spindle to track the RPM coming from Mach3 almost perfectly.
What was a little funny about all that was that when I had the idea with the capacitor, the local radio shack was already closed, so I had to salvage the capacitor from an old Sound Blaster board I had laying around, lol.


WP_20130917_004 by mkloberg, on Flickr

To validate the RPM on this rig so far, I was thinking about ordering an optical tachometer, but I found in one of the threads on this board that harbor freight sells one for around $30.
To make this work, all that needed to be done was to place a reflective strip of self adhesive tape on the spindle shaft to get a reading from this thing.


WP_20130916_003 by mkloberg, on Flickr

I don't trust it all that much, the tolerance is 0.05% but at 24000 rpm that this spindle makes, this could be quite off.
Nevertheless, that gave me a better idea on how this all works.


WP_20130916_007 by mkloberg, on Flickr

Checking speed readings and frequencies of the VFD, it turned out that the RPM of the spindle is perfectly linear to the frequency the VFD outputs and displays on the front panel.
400hz = 24.000rpm, 200hz = 12.000rpm, etc.

The setup of this in the smoothstepper config and Mach3 motor tuning and pulley configuration dialogs was more than obscure to get it right, but I won't go any deeper on this at that point.
Ping me if you want to know more.

Anyways, having the rpm issue out of the way, it was about turning the spindle on and off.
One could rely on the rpm signal only, but having all this noise in the machine now I didn't want to get my hands hurt while doing tool changes, just by relying on the Gecko to output zero RPM.

Some kind of on-off relay was needed. I had a solid state relay around that I tried first, but that for some reason introduced even more noise into the setup.
Very strange stuff happened at that point such as the spindle turning on but never turning off and keeping to turn at a low RPM, although there were no signals to do that coming from the Gecko.


WP_20130915_002 by mkloberg, on Flickr

That solid state relay was designed to switch 50V AC, clearly the wrong component for this application. Eventually, I soldered together my own little relay board and put a LED on it to be able to see if it's off or on.
I left a lot of board real estate blank, for who knows what other challenges come up with controlling the spindle.
Perhaps we want to run it in reverse down the road to be able to tap threads, which would require another relay to tell the VFD the direction, etc.


WP_20130915_008 by mkloberg, on Flickr

At this point, the spindle worked perfectly. Telling Mach3 to run it at 24000 rpm resulted in a frequency of 400hz on the VFD display, likewise to run it at 50 power (12000rpm), the frequency went to 200hz.


WP_20130917_006 by mkloberg, on Flickr

I did some readings with the tachometer while at this stage and the spindle really turned around the targeted rpm's, just about 100 rpm below the targets.

Switching the spindle on and off now works great, but I still have some concernes about safety.
Just relying on the Gecko outputs to keep the spindle still during tool changes with your fingers on the spindle collet gives me the creeps.
I can't do much more at this point because my e-stop button only has one contact.
Perhaps I should get a relay with a lot of contacts and wire the button to that relay instead to be able to switch more things on and off when the e-stop occurs.

I'm thinking about hitting the e-stop button during tool changes, is that ok? Will the Gecko hold its axis positions during e-stop condition, or will all axis have to be referenced again coming out of that?
Just wondering what your experiences are... :-)

Having the spindle sub-project somewhat completed, I turned to laying down the epoxy to build up the machine bed.
I did the calculations with the info that Bob provided in the manual and no matter how I looked at it, for the machine at this lenght, simply mixing half a gallon of epoxy and a quart of the hardener should be just about right for a Momus at this length.

The first step was to get the machine perfectly leveled, if you followed this thread we're talking about heavy by now. Really heavy. I don't really know, but I can hardly lift the machine from one end with both my hands by now, I would guess it weighs more than 400 pounds by now, including the long workbench that it sits on.
Very early, I built furniture leveling screws into the workbench, so at this point it was simply a matter of disengaging the retractable casters and engaging the leveling screws.


WP_20130919_005 by mkloberg, on Flickr

I work for a manufacturing company (Liebherr), so I was in the lucky position to be able to borrow a precision box level from work (thanks Chris! and Cort!) :-)
This is a box level with a resolution of 0.02 mmm/m, or 5µ per 250mm. Putting a level like this on a wooden structure is of course total overkill, but I thought it would at least give me a good idea on where to go with the leveling.

Getting ready:


WP_20130917_008 by mkloberg, on Flickr

Bob recommended to heat the epoxy to about 90deg F before pouring. I simply put the two bottles in the kitchen sink and filled it with heated water, checking the temperature for about two hours with my Thermapen meat thermometer.
Whenever the temp dropped below 90, I let some water escape by pulling the plug and aumented that with some more hot water from the tap.


WP_20130918_005 by mkloberg, on Flickr

While that was going on, I leveled the machine using the precision box level and the leveling screws under the workbench.
That took a while but eventually I got the bubble somewhat centered in both the Y and X directions across the base.
This was mostly averaging, but it worked out.


WP_20130918_008 by mkloberg, on Flickr


WP_20130918_011 by mkloberg, on Flickr

While the epoxy was still warming up I started to build the dams, so that the epoxy can't escape.
I reached into the cabinet where I keep the painters tape and saw a roll of HVAC reflective aluminum tape underneath.
Immediately I thought that that might just work better for that kind of thing, because it'll come off easier on the epoxy, is plyable/bendable, is much more adheasive and more workable into the corners to prevent leaks.

The stability of this type of tape also lends itself to counteract the tendency of the epoxy to creep up on any vertical edges it's poured on.
Bob mentiones that in the manual and I used that knowledge the build "outside angled" walls with the aluminum tape to get around that, basically nullifying this effect where I could.000


WP_20130918_012 by mkloberg, on Flickr


WP_20130918_013 by mkloberg, on Flickr

I just tried to bend the tape as far outwards as possible, without ripping it.


WP_20130918_014 by mkloberg, on Flickr

Ready to mix and pour (I got a little nervous at this point):


WP_20130918_015 by mkloberg, on Flickr

The blue bowl now has half a gallon of resin it it and a quart of slow setting hardener, I'm stirring it carefully but really fast, trying to not get any air bubbles into the mix.
I don't know, but if felt like the more I stirred, it began to harden because it resisted the stirring more as I went along.

Time to pour.


WP_20130918_021 by mkloberg, on Flickr

I had some leaks, dripping on the garage floor, just a few, but nothing major and walked away from it.

Leaving this alone for about a day (wasn't easy, lol), I pulled the tape and used a block plane to shave some of the major overflow stuff away.


WP_20130919_001 by mkloberg, on Flickr

Whops! I think there is a BUG in the machine!
Turns out that after I poured a resin, a Moskito fell into the resin somewhere near the right wall in the middle of the base and found an untimely grave there. Sorry, buddy ;-)
It will be remembered as such, forever be encased by resin.


WP_20130919_002 by mkloberg, on Flickr

Next up, adjusting the x-axis rails parallel to the resin base plane. I have an idea already, involving the markup gauge to measure things.


WP_20130919_003 by mkloberg, on Flickr

Stay tuned :-)
Mac