[erikjgreen]

I got some pulleys in from SDP-SI. They look nice, I could make them work... but as a backup idea I ordered a bushing for the top of the ballscrew to fit the old, slightly bent pulley. It looks like it'll fit perfectly, I just need to figure out a way to ensure it and the pulley are firmly attached to the ball screw. Worst case I have to cut a couple of keyways in the bushing and ball screw using a dremel tool. I thought about trying loctite, but I'm not completely sure it would hold. There's a lot of torque involved.

Expecting bearings for the lathe spindle tonight, I have the spindle pulled apart already. I've tentatively picked a motor from my stockpile to run it, and I found the VFD I'm going to use. Hopefully I can get it all together in my spare moments this week. I still have to finish a couple welding projects before next week, too.

I'd hoped to get to work casting a new foundry shell, but it doesn't look like that will happen, either

Erik

[erikjgreen]

Everything's arrived at this point but the metal parts to mount the ball screw. I don't know if I'll get them today, but if I do, I still won't have time to work on things until Sunday afternoon due to family commitments.

Still moving forward as fast as I can, though

Erik

[erikjgreen]

Bleah, I had a good time at the family get-together, then caught a cold. Probably from one of the 10 or so little kids there. They're a lot of fun, but under 10 year olds are little germ factories with legs.

I should get the precision made metal for the mill today or tomorrow, which means I need to get cracking on connecting the I/O boards to the servo drives, so I can drill those holes and put it all together.

While I've been sick, about all I had the energy to do was finish stripping down the mini lathe head, clean it, install a pair of TAC bearings, and reassemble it with the V-belt pulley in place. I did also get a piece of aluminum plate to mount it on, so I don't need to have the motor attached to the lathe bed.

Last week I did win an auction for some bigger lathe stuff. I got an 8 inch 4 jaw chuck, a tooling plate, and some fly cutters. The tooling plate fits the big lathe head I have, a D1-4. The chuck can probably be made to fit with an adapter, or if it's too worn I'll use it for a vise on the mill or something. The fly cutters I may keep a couple and melt the rest down... they're fairly worn. Not sure this mill can swing a 15 inch diameter fly cutter safely anyway

So much to do outside before the snow flies here... let's hope it holds off for another couple of months.

Erik

[erikjgreen]

Well, I've been busy with lots of things, including finishing a welding project and getting my garage ready for winter.

But, I've managed to put in some time on finishing the mill project. I've gotten quite a bit of wiring done, and I'm to the point of finishing the control PC's software load... or I was until the hard drive developed an error yesterday.

I was starting up the "text axis" program in pncconf, which is linuxcnc's configuration wizard for Mesa electronics' I/O cards. I had connected all three axes to my 7i47 card on outputs TX4-TX9, using the first two wires on each output. This let me move the X axis appropriately when testing wiring before I connected all three axes.

When I started the test program, the X and Y axes started moving in the + direction. Because I'm occasionally an idiot, I didn't have the limit switches properly connected to the STOP or ERROR inputs on my DG2S drives. They were instead connected to inputs on the 7i37TA I/O card so that linuxCNC could see when they were active. I have since learned that both are needed.

The X and Y axes moved at about 80 or so IPM to the + stops on their respective axes and stopped, staying there about 15 seconds before I made it to the disconnect for power.

Fortunately, they didn't slam into the stops hard enough to break any metal. However, the servos got hot, and the servo drives overheated (they detected this and stopped moving the servos, flipping on an error LED).

After re-checking my wiring and moving the servos inside the limit switches manually, I tried re-running the test. I had to reboot the control PC, because the hard drive decided to give up the ghost about then. I don't think it was related to the shutdown I did, since I had problems with it occasionally since install.

Testing the Y servo got no response. The drive tuning page showed all ok (the DG2S drives can be tuned via USB connection) but the drive did not move the servo when commanded.

The X axis seems ok, the tuning page still works for it and I can move it back and forth with the DG2S software or pncconf's utilities.

I tried powering down and moving the Y servo with a car battery. It moved ok (as slow as I expected) and went in both directions. Powering the whole mill up again, I connected the Y servo to the Z drive, and moved the encoder and drive connections over too. Using the USB utility for the DG2S, I got the servo to move, but the drive faulted out on following error immediately.

I suspect two things are now not working... the Y axis drive possibly has a fried component, and the Y axis servo encoder seems to be non functional.

I e-mailed Balasz (the designer of the drive) to get suggestions on what to try next.

This shouldn't delay me too much since I'm manually operating the Z axis for now anyway (see above). Tonight I'll pull the servo and check its innards, and see if I can test the encoder (I might swap it with the Z encoder or something). If nothing else, I can use the Z drive and servo/encoder to run the Y axis for now.

More annoyance though, especially since it was caused by my own stupidity, or at least not prevented.

In the time since my last post I've had more luck shopping... I picked up a set of four 1.5NM brushless AC servos for a song (I have a good relationship with certain surplus dealers). Each of them has about three times the torque of the existing servos on the mill, and a top RPM over twice the existing servos. I don't think upgrading if the mill servos are working will buy me much performance wise except faster rapids, but it's nice to know that I have options. I'd need to buy AC servo drives though, which won't be cheap unless I get lucky shopping again.

I also picked up some of the first parts for a router table I want to build when the mill is done and the lathe is back up and running. The project list gets longer....

Erik

[erikjgreen]

Yesterday I exchanged e-mails with Balasz, the designer of the DG2S drives at cncdrive.info... outstanding customer service. He gave me a theory about what was wrong and specific instructions on testing for damage on the Y axis (failed) drive.

I've since found out that one of the MOSFETs is blown, and will need to be replaced. Fortunately, they are through-hole chips, so I'll be able to replace it myself. Waiting to hear back on his recommendation for a supplier and exact model of MOSFET to use (they're IRFP260s, FYI).

After I fix that, we'll address other problems with the axis. If it turns out the encoder is broken I have some others I could use, and I completely forgot that I'd acquired some SEM brushed DC servos from a Tree mill I could use as replacements for the current ones. They use a lower voltage, but the DG2S drive can limit PWM to avoid overdriving them.

Last night I checked my wiring... Balasz said he didn't know how the MOSFET could blow because I had the brake circuits hooked up in the right place, with heavy gauge wire, and they were working. A mystery, I guess.

Last night I also played around with tuning the X and Z axes... mostly X, because Z isn't mechanically connected to anything at the moment. I got X to the point where it's rarely oscillating on a move of 10,000 steps, which is a rapid of about an inch or so (if I'm calculating correctly) at a speed of about 100 ipm. I'm still learning, so I don't really know what optimal is for a movement graph. I think what I want is to make the axis respond reasonably quickly without having a lot of overshoot on moves, and no oscillation after reaching the destination.

Tonight I'll try checking the wiring for the Y encoder and make sure it's all still in place, then maybe give the Z drive another try on the Y axis. If it still doesn't work or I don't have a proper response from the encoder I'll pull the Y servo and check it out.

I also started re-installing ubuntu and linuxcnc on the new hard drive for my control PC, that should be done tonight. If I can get the Y axis working with the Z drive I should be able to use linuxcnc to do the sharpie tool test made famous here on cnczone

But first I'm going to triple-check the wiring for the limit switches and test to make sure hitting a limit switch halts all three servo drives...

Erik

[erikjgreen]

Okay, more updates time.

I've been spending a week emailing Balazs, the designer of the Dg2s 160-35 servo drives, about problems I was having. Basically the encoders I've been using are working only most of the time. Jogging in the servo tuning program generally works, and I've been trying to tune the PID loop.

Every so often however, the servos would fault out due to a following error condition. It seemed like the drive wasn't seeing all the pulses from the encoder, so was building up error enough to fault in some longer moves.

I checked with Balazs on the current draw... my Sumtak encoders (the ones that came with this beast) are about 25 years old and draw 140ma each... that's more than eg. a Gecko drive could handle, more than most modern servo controllers could deal with... not the DG2S though, Balazs says it can supply up to about 300ma of current to the attached encoder if it needs it.... excellent drives!

Anyway, we're still debugging this, although I think the solution may be to just switch encoders. I had a different model of equally old encoder sitting around in the same form factor, so I stuck one on my Y axis, and presto, everything works.

I've also gotten in some MOSFETs and an oscilloscope to repair my blown servo drive... just got to get time to solder that sucker in, and I still need to buy Oscope probes, my used scope didn't come with 'em.

As part of my servo testing I re-wired the limit switches so they all work in series to hold a relay in place. When one of them is triggered, the circuit is broken and all three axes' reset pins are triggered, stopping the mill. I still need to ensure a latch/reset behavior and put in my e-stop switches, but this setup works pretty well for safety.

In other news:

I have been talking with folks in another thread here about the backlash in my machine... now that the servos are sort of working I found out that when I reverse directions, there's about 8 thou of backlash in the ballscrews. Probably I could work with this, but it'll make circle interpolation annoyingly oval, and I'd have to be careful when doing anything at all to make sure to compensate.

I looked up a couple excellent references for narrowing the cause of backlash in CNC mills, and found out testing with a dial test indicator is the best route to go. There are diagrams out there, but basically you put a DTI against various mill parts and try to move the table by hand with a lever.

I botched this the first time I tried it and missed the actual problem, so I thought I might have to replace my ball screws. I ended up re-testing in a less superficial manner (read: took some stuff apart to do it right) and found out that my ballscrew support bearings are worn.

Last night I took out the Y axis bearings and the outer one came apart in my hands... the other one is still in one piece, but my testing showed the ballscrew moves about 0.01 when the table is pushed, so these bearings are allowing that much movement. I priced out new ones, ouch. Cost will be about $175 each (Barden L078H or equivalent) and I need four of them.

At least once installed this thing will basically be new accuracy wise on the X and Y axes (still have to repair Z). The ballscrews and nuts on X and Y are in great shape(Y ballscrew and nut here, pic taken from up under the knee):



For reference, they're NSK 32mm ballscrews on X and Y.

As part of my backlash debugging I've decided to take the table off the mill and clean up the innards. There's a bit of grit and other junk on the Y ballscrew, and the X (being exposed) is dirtier and has a few rust spots near the ends. The ways on the other hand look perfect

I'm building a small crane to do that with.

Soo.... I'll probably order bearings next week, and depending on what Balazs says I should get my servo drives working by then, although I have to wait for the bearings to do tuning.

Once the X and Z axes have the bearings in and the drives are tuned, I can carry out my plan to drill the new Z axis bracket and mount the Z ballscrew, at which point I should have one very accurate CNC mill working.

Assuming I don't find anything else wrong

Erik

PS: The seller of this machine said "I think they only used it to cut plastic"... what he DIDN'T say is "That's all it was capable of at the time"

[will gilmore]

Keep up the good work! You are getting deep into this project. I had a few thou of backlash in x and y. I just compensated for it in software but I should probably check out the support bearings. I'm not prepared to spend money on a regrind or reball for the screw but if I could eliminate the backlash with the support bearings I would probably go for it. I'll be watching to see how it goes for you.

[erikjgreen]

Will -

You can check to see where exactly the backlash is coming from, so you know whether replacing your support bearings will help. I can't find the reference link at the moment, but if you:

1)attach a dial test indicator to your mill frame or a solid floor stand
2) put the tip on your ballscrew end on the supported side of the screw
3) Use a 3-4 foot lever to push the table hard in one direction towards the ball screw supported end (protect your pry points to avoid damage, and don't use precision surfaces like the ways to do this).
4) Zero the indicator and make sure it's still in place
5) Use the lever to try to move the table the other way, then check the DTI for movement.

The table may move or may not - it takes some muscle. If it moves, it will be because of backlash in the ball nut, ball screw, or support bearings.

If the dial test indicator shows movement then at least some of that backlash comes from play in your support bearings.

My dial test indicator showed about 0.01 inches of movement total, moving from one direction of pry to the other.

Erik

[erikjgreen]

Bunches of updates here, in no particular order.

As posted in another thread, I located and bought some new sets of thrust bearings for X and Y... the axes should be very accurate once I install them, the ballnuts don't seem to have any backlash at all.

Before I install them however I'm working on cleaning and painting some parts of the mill that will be hard to re-work once the bearings are back in. Specifically, I've done the ballscrew mounting blocks (big pieces of metal with the bearing pockets in them) with Old Ford Blue paint... about the closest match I could find to the mill's factory blue. At some point I'll repaint the whole thing with a unified color in some tough industrial paint, but for now I'm just using plasti-kote primer and paint.

I'm also going to do the shroud that covers the front of the table... this is a piece of sheet metal right at the front of the mill that covers the limit switches, X servo, and extends out over the ways and way covers. It's actually attached to a T slot that's in the front of the table, with T nuts pointing horizonally out. If I didn't need the shroud there I'd leave it off, so I'd have more options for mounting things. You can see the part I'm talking about below, it's got the big "Lagunmatic" sticker on it:



Alternatively I'd love to make a new one of these, maybe out of stainless steel or something. I also need to seal the gap between its edge and the table, either with a rubber gasket or silicone, to keep coolant from ending up on the ways and servo. Possibly I can attach an enclosure there too, I'm still designing.

I've got on order a pile of stuff, some of it came in Saturday. I ordered three AMT102 encoders with differential line drivers to replace the Sumtak and DRC encoders I've been using. The Sumtak encoders didn't want to work with the DG2S drives for some reason, and the DRCs worked but are kinda big, so I decided to pick up the AMTs, which are capacitive and selective resolution, I can pick from 30ppr up to 2048ppr. They were about $40 apiece.

I decided to replace a lot of the worn looking cap screws on the mill parts I'm working with. I ordered alloy steel replacements from McMaster-carr in several sizes and lengths, plus dowel pins for aligning parts like the bearing mounts when I reassemble things.

I found after some work some new nuts to fit the ball screws - the old ones still work, but they're mangled. At least one other person before me installed them using a screwdriver and mallet, so I wanted new ones. Measuring them, they're M15x0.8 nuts... not a normal size. They also don't match KM-2 ballscrew nuts or the standard nuts that come with the FK15 series mounts. I did find that McMaster had a pretty close match in an imperial nut size... they're 0.586"-32. Re-checking the ballscrew with my thread gauge, the -32 is a pretty close fit, and 15mm is 0.590551". Hopefully the nut will fit ok even though it's 45 thousandths under the nominal size.

If not, I'll have to either make my own nuts or re-use the old ones after some cleanup. I also need to make a proper wrench to put them on, the bearings need about 175lb of preload.

So, a summary of my current status on this mill: Painting and cleaning then reassembling the X and Y axes, replacing encoders soon which will make the servos and drives work under linuxcnc. After X and Y are working I still need to make the adapter hardware for the Z ballscrew. When that's done I should have a working mill. Then I can start adding things like coolant, etc. There are also a thousand little things to do like finishing the replacement of the air lines and valves, more wiring for I/O and things like the oil level sensor, etc.

Erik

[erikjgreen]

Okay, you know how sometimes you say when you're taking something apart "Well, now that I've got it apart, I may as well......"?

I'm the poster child for that.


So, I took the mill way wipers off to clean them, decided to check the gibs and clean as much of the ways as I could, then took the table off. Here's what the X ways on the saddle look like:



I like the curly oil grooves, and also the fact that there's not much wear - the scraping originally done in the factory is obvious. I took this off to clean it out, and I found some pools of oil in the casting. I also found that rather than the pressure fittings I expected on the end of the lube tubing, there are pretty much just empty tube ends.

The way the auto pump for the ways seems to work is that it slowly pushes out a set volume of oil over time when the mill is powered, rather than a set volume when pressure is applied via the piston like the spindle lube pump.

Most of the ends of the tubes are either pushed up into small holes in the castings (like for the oil grooves) or siliconed in place over where they're supposed to drip, like over the Y ball nut.

What I'm noticing is that when I power the pump and pull the piston to manually inject lube, it doesn't come out everywhere. The grooves don't get much oil at all, and I'm not sure some of the tube ends are in the best spots. Probably they contributed to the pool of oil in the casting (most of which I already mopped up in the pic above).

So I decided to cause myself another delay and order replacement tubing from McMaster-carr. It's 5/32 nylon, and rather than replace it with 1/4" polyethylene from a local store I decided to do an exact match. Once it arrives I'll disconnect the lube hose from the pump, blow it out in reverse with compressed air (I cleaned the pump earlier, but I'll check it again) and replace the tubing. I'm thinking I'll use shoe goo or a similar rubber adhesive to put the oil tubes in better spots this time.

I'm also working on cleaning the table.. I think it has silicone residue on the front edge because it was sealed with it there (against the shroud) so I might have to get a special solvent for that. I'm using naptha and a citrus degreaser and that's handling most of the problems:



There's a little rust on the dovetail from storage... I'm probably going to have to clean that as best I can with some 600 grit sandpaper and check it with my surface plate. Suggestions for cleaning without ruining the accuracy are welcome.

The bottom of the casting looks ok, but I'd like to touch up the paint, which will mean cleaning what's there well enough to make that work. Lots of solvent and wiping :\

My biggest roadblock at the moment are the cap screws holding the Y ball nut to the yoke. I can't remove the saddle without getting those off, and someone seems to have put them on with an impact wrench.



I made an extension for my impact screwdriver (the kind you hit with a hammer, great tool) but no dice. I don't have my air compressor ready so I can't get them off that way. I'm going to try more elbow grease I guess, and I might try PB blaster. I don't think they're corroded, but you never know. Just had a thought... maybe someone loctited them on?

Lots of parts on the floor at the moment, that's the casting for the X ballscrew floating end bearing soaking in citrus degreaser, and the X servo getting its encoder replaced.



I'm planning on scrubbing and repainting the servo mounts visible in the pic (you know, as long as they're off . I'm using "old ford blue" for now, it doesn't really match the factory color, but I couldn't find that anywhere. At some point I'll repaint the whole mill with a unified color, but for now it'll do ok and won't rust.

I'm trying to keep all the parts sorted out and obviously not doing the best job... if this drags on for more than a few more days I'll have to ziploc and label everything so I know where it goes.

Other stuff... X ballscrew is in great shape:



I'll wipe it off before re-installing and coat it with light oil (I've been using marvel air tool oil to keep things from rusting while working).

That's the X gib too, also looks nice.

Servo drives are repaired, tested, and remounted:



They should work well with the new encoders once those are mounted.

I got in all my orders for parts pretty much... I now have new cap screws to replace old/messed up ones, dowel pins for alignment during reassembly, washers, etc.

I thought about replacing the way wipers, but after cleaning them they're in ok shape... they're silicone rubber under plastic covers, which I would think would degrade, but other than a few scratches and such they're ok.

My next steps are to get the Y ballscrew off and remove the saddle to clean off the Y ways, then touch up the paint and start reassembling.

I want to thoroughly clean up the table and ways, check them for flatness, etc, plus get the lube tubing in place before reassembly goes too far though.

Once the cleaning and paint touch up is done, then reassembly can proceed to the point where the electronics are in and working and X and Y should be able to move with new thrust bearings

One step forward, two steps to the side, one step to the other side, and a step back followed by looking around wondering how I got here... but I keep asking myself if I'd rather be done faster or do it right, and I keep coming up with the same slow answer

Erik

[Al_The_Man]

I also found that rather than the pressure fittings I expected on the end of the lube tubing, there are pretty much just empty tube ends.

The way the auto pump for the ways seems to work is that it slowly pushes out a set volume of oil over time when the mill is powered, rather than a set volume when pressure is applied via the piston like the spindle lube pump.

Most of the ends of the tubes are either pushed up into small holes
Erik

I have never seen a multi-lube system that was not metered?
I am guessing that maybe there was metering originally over time get plugged, maybe someone got lazy and did not bother to replace them, with open ends you are going to get one point of low resistance that gets all the lube.
Al.

[erikjgreen]

It seems odd, but it definitely was designed that way.

You can see in the pic of the ways the tubes go into small holes in the iron to reach the various locations on the ways, in the grooves, etc. No room in those holes for fittings, the tubing itself is friction fit.

Is it possible there's metering/flow division stuff in those little brass manifolds? From the parts list below, it seems there's different meter levels.

Here's a diagram of the lube system from the manual for reference. The pump in question is a Bijur TM-5, the spindle pump is not a TM-1 as indicated, but rather is a non powered pump with a single knob to pull. Maybe that got replaced, or maybe it was the default when the mill was made (the manual page is from a newer mill version)



Erik

[erikjgreen]

I think I understand.


After looking up specifics on the pump and those fittings, the metering units are those little "extra" things screwed into the distribution manifolds between the brass bar and the compression fitting.

They're type C Bijur control units, FYI.

Erik

[Al_The_Man]

Yes, the meter units appear to be screwed into the manifold block and are labeled:
Unit, Meter #, where the # is the orifice number.
BTW, they are pretty hopeless to clean, they clog the filter in them over time.
Al.

[erikjgreen]

Yes, the meter units appear to be screwed into the manifold block and are labeled:
Unit, Meter #, where the # is the orifice number.
BTW, they are pretty hopeless to clean, they clog the filter in them over time.
Al.

Any advice on testing them? Should I just look through them and see if I can see light?

I've got some replacement ferrules on order for the fittings, plus the tubing should arrive soon, so I hope to start reassembly of the lube system this Weds.

Erik

[Al_The_Man]

I don't think you can see through them normally because of the filter.
An oil pressure test might show if they are completely blocked, but there is no easy way to check the volume through them.
If they have been on for anything like a couple of decades or more then if you have the machine down it pays to change them,IMO.
Al.

[erikjgreen]

Well, I decided to bite the bullet (again) and order parts. I've got $225 worth of new bijur metering units and fittings on the way now. I guess a few more days' delay in getting things back together won't hurt.

I've got the lube system pulled out now, ready for the parts. I'm going to use shoe goo instead of whatever silicone based glue they had before to put the ends of the tubes where they belong.

I found that the end of the tube hanging over the yoke was supposed to be embedded in the lube hole in the Y ball nut. The ball nut had the snapped off end of a tube stuck in the hole. I don't think it had been too bad off, the ball screw and nut look/feel good and I didn't find any slop there.

After cleaning, the table looks good. I used my little surface plate and a ground straight edge to check for flatness. It has some dings and two major gouges from tool crashes, but no spots sticking up, so I think it'll work out fine.

The saddle ways are mostly in great shape with the original scrape marks all over them, but one end of the X (top) has some parallel scratches in it, which can be seen in the right hand corner of the pic above. They don't stick up either, so I'm guessing they are from some debris that got into the ways... no corresponding marks on the X table ways, so maybe they're some kind of casting defect?

I'm about ready to touch up the paint on the saddle and powder coat the accessory metal parts.. the ballscrew mounts, the servo covers, etc.

Other than that I'm waiting on lube parts. I should have all the nuts, bolts, dowel pins, etc I need to reassemble everything. I'll start with the lube system and test it, then put the saddle and table back on with the ballscrews, followed by the servos and their new encoders. A quick functional test and if all is well I'll bolt on all the accessory tin except the Z axis stuff, which is the major remaining project.

I'm also joining a local shared workspace group, so I should have access to their shop soon. They don't have a CNC mill, but they do have a bridgeport in decent shape, so I am considering just using that mill to make the adapter parts I need for the Z axis.

Erik

[jsantos]

Hey Erik,

It's been sometime since I visited here and talk to you. Looks like you're still hard at work getting your machine running.

You're welcome to use my mill anytime you want to. Just shoot me an email or give me a call.

[erikjgreen]

Thanks Joe... I've been thinking about it. We should get together and catch up sometime soon.

I'm just about done painting and replacing my lube system, should be getting this thing back together soon...

[erikjgreen]

Well, I've been spending the last three or so weeks working on the lube system, and since that involved removing the table and saddle, I decided to clean up and paint the parts involved.

While waiting for my lube system parts I disassembled and completely cleaned the TM5 lube pump too.

Here's a couple pics of the table... it cleaned up nice, and I painted it with dykem and checked flatness with a small surface plate and a straightedge. Aside from a couple burrs that I removed, it's flat.


That's the saddle standing on end in the background, by the way.



I cleaned it up with some 600 grit sandpaper using WD40 as a lube/solvent. Most discolorations came out nicely. I used a clean rag to wipe away any grit, then degreased, wiped, WD-40, and another wipe. The top surface that I used dye on also got cleaned with acetone.

Then on to painting.

I tried painting the parts with Valspar Anti Rust armor, and it looked "ok", but even after baking in my powder coat oven it came off fairly easily and didn't look all that great. Possibly if I'd been able to spray it things would have gone better.

I also tried some chrome powder coat from Columbia coatings. It went on ok, but the masking and required clear coating afterward made it annoying, and it didn't look all that good.

Since I'd decided to use the chrome powder coat on a few parts, I went looking for a good silver paint to use that would hold up... I decided on the Rust-Oleum hammered silver. Quick and easy, and looks good.

About the time I got the painting with Rust-Oleum started, lube parts started arriving. Replacing the lube system is actually pretty simple once you have access. I saved the parts I was going to re-use, which included compression nuts, two hoses, the manifold blocks, and the compression bushings.

I cleaned out the manifold blocks with WD40 and a small brush. I then inserted the metering fittings. The first block to go in was the one inside the knee... it feeds oil to the knee jack and the column ways. It just hangs in space inside the knee, so I used a couple zip ties to hold it steady. The tubing was protected with a couple sections of small flex conduit outside the knee where it meets fittings for the column ways, but they were in poor shape, so I replaced them with a couple snips of air tube.



Once all the hose inside the knee was connected I attached my cleaned and partly refilled (and tested) pump to the feed, with a brass plug to hold pressure in the hole in the tee that would have otherwise fed the rest of the system. I let it run for a couple hours, and saw no leaks in my hose or fittings. Oil went to the right places.



A quick side note here - if you're putting these fittings on a machine, they're delicate little things. I sheared one off trying to tighten it. I thought I might have to order a new manifold for a half hour or so.

You only need to tighten the metering fittings until their tips make contact with the sealing surface inside the manifolds. Only tighten the compression nuts about a half to three quarter turn past finger tight, they'll seal.

After testing the knee tubing, I started on the saddle, which I had cleaned and repainted. I installed tubing in the holes for the ways and gibs, holding it at the appropriate depth with a few drops of wood glue. It should hold the tubing in place ok but be easily removable if needed.



While I was waiting for the glue to dry, I cleaned and re-installed the sliding covers over the top hole in the knee casting. The bar you see in the tool holder was still there from my runout testing.


I then installed the manifolds in the saddle and placed it on the ways (I took a last moment to wipe the ways clean one more time and check them).

Then I hooked up the lube pump again, and while I was running that I re-installed the yoke that holds the ball screw nuts, using pins to align it with the alignment holes in the saddle.



Everything checks out ok. The two pieces of tubing not connected to anything go in holes in the ball nuts. I put the Y gib in place and tightened it by feel. I'll re-do that once the whole thing is together using a dial test indicator for as little wiggle as possible.

The new color looks a little funny since I haven't painted the base or most of the column, although I have done the head, but the blue and silver go ok together.

I also re-installed the Y ball screw and the bearing mount (again using locating pins). The bearing mount is powder coated rather than painted. There's a tiny bit of surface rust on it at the moment. I should have been quicker to oil it, but after all the prep I went through to clean it for painting I wasn't eager to get it "dirty" again



Next up on the to-do are finishing the ball screw installs and connecting the lube tubing, moving the lube pump back to the rear of the machine, installing the table, installing my new bearings and nuts, re-installing the repainted parts of the table and head, and then come the servos.

Still going...

Erik