[RedSquirrel]

I built a table and decided to try epoxy overtop steel to create a level surface for the rails. Some members on the forum seem to have tried it and it looked about right. Now that I have the epoxy down if I put a straight edge on it I can slip a 0.10mm feeler under spots which suggests there are some slightly high spots over the 4' of the straight edge. Using a laser I might be off 1/16" over 20' so for the most part the epoxy did the trick. I'm not even sure if that's far enough off to even be a concern. What sort of tolerances are on your DIY table rails?

Does anyone have suggestions on the best method to shave things down so that it's nice and straight over the length? Should I abandon that idea or sanding it flat by hand and start looking for a thinner epoxy? I've been going at it with a sander and checking with a straight edge but that hasn't been the most effective and I'm beginning to worry about one side ending up slightly different than the other introducing unnecessary twisting of the gantry. Also where I sand there's the issue with keeping the surface flat and not accidentally introducing a slant.

One thing I'm considering is looking for some thinner epoxy to pour on top of the epoxy sections I already have. Any suggestions would be welcome here, I'm not an adhesive expert by a long shot! I used West System epoxy but suspect there might be thinner product with better flowing agents. I'm going to experiment with thinning it down although they caution people not to as it compromises the integrity of the epoxy -- but nothing ventured nothing gained right now.

I've attached a picture of the table and the epoxy I've poured for the curious. It's a 1" strip (which might also play into this, with the surface tension of such a narrow channel) over the length.

[ger21]

It's 20' long?
I'd grind it off, and try again.
How cold is it in your shop?

Use the slowest hardener they have.
Warm the steel before pouring, and warm the epoxy before mixing. Maybe wrap the frame with some heated blankets, to get an even temp?
Warming everything up will thin the epoxy, and make it flow better.

You may need to pour it thicker, depending on how straight the steel is to start with.

[RedSquirrel]

It's 20' long?
I'd grind it off, and try again.
How cold is it in your shop?

Use the slowest hardener they have.
Warm the steel before pouring, and warm the epoxy before mixing. Maybe wrap the frame with some heated blankets, to get an even temp?
Warming everything up will thin the epoxy, and make it flow better.

You may need to pour it thicker, depending on how straight the steel is to start with.

Thanks for the reply Gerry!

The table is currently 19' long (using 20' as an even figure and planning to add a tool rack on the end at some point which will bring it up to about 20').

The shop is usually 65 degrees or better. It's not the warmest but it's a great temp here in my part of Canada. I'm just glad the snow is gone.

I used 206 hardener. I couldn't find 207 in decent quantities locally.

Where I have access to the inside of the rectangular tubing it wouldn't be all that tricky for me to circulate hot air in the frame. I didn't fill it with concrete or anything crazy yet. I am reluctant to grind the epoxy entirely off at this point as it was quite a feat to get a nice seal for the initial strips. It's easy for me to create a seal off the existing strips that are on there now provided they come up to temperature. I suppose I could also keep the resin and hardener in hot water prior to mixing.

Right now I have an experiment with acetone waiting to harden. I have no idea how that's going to turn out but it most certainly turns it to water in a hurry. I don't suspect it'll harden like I want but in another 12 hours or so should have a better idea of that outcome. I'll try to procure some hot water and build some test channels to see if there's a massive change in how it levels itself. I'm unsure how that's going to turn out. I expect it will appear to flow a bit quicker at first but will know more soon.

I assume you have a DIY table of some sort, do you machine wood or aluminum on it? Have you ever taken a straight edge to it and how much of a gap might exist? I don't have anything to compare to for reference here.

[Biggs427]

I used 206 for mine too and West suggested that I didn't apply more than 4mm at a time. This was for 1000mm x 40mm.

Send them an email with what you have available and what you are trying to accomplish. Their support is really good and they usually answer fast.

[email protected]

[RedSquirrel]

I'm going to reach out to them here now and ask. Thanks Biggs427!

As for my acetone experiment it appears to have dissolved the resin completely and doesn't show signs of thickening up yet. Going to also try with lacquer thinner and see how that goes.

[1Jumper10]

You get a meniscus effect about 1/2" from each edge. At least that's what I observed when I did mine. You may need to make it wider to get it flat in the center where you'll mount your rails.

[Biggs427]

I've ordered my West kit from Plastic World.

They have the 209 hardener available which, I think, would be better for a colder shop like yours.

West System 209 Extra Slow Hardener - $49.99 : Online Plastics, Plastic Sheets, Plexiglass, Epoxy, Fiberglass, Carbon Fiber - Plastic World

[ger21]

I still think just warming up the frame will make the biggest difference.

[DonKes]

Warming is a 2 edge sword. Heat makes it cure quicker. I think some experiments in thinning VS complete curing is in order if the mfg can't tell you all about it.

Spreading it evenly helps a lot, but you know that.

Don

[RedSquirrel]

Thanks for the response everyone! I can tell I'm going to love this site. The West System guys wrote me back and suggested attacking it with the sandpaper once again but using a jointed board to help get the flatness I'm after. I'm going to get things as close as I can once again with a laser (by hand) and then run the length of it with a jointed board for awhile and see how that turns out. I'm hoping with consistent pressure it'll take care of high spots.

[NIC 77]

I briefly had a look at some posts on the UK CNC site. I believe some people reported deviations in epoxy of 0.003" to 0.007" if I remember correctly, not including the meniscus, so 0.1mm or about 0.004" seems right to me. That's still a long way off from 1/16" or 0.0625"!

I can see being off 1/16 in thickness at the ends due to the table being higher at that end, but after you'd milled down your spoil board (assuming this is a router / mill and not a plasma) that should match. Over 20 feet, is it not hard to see a difference in 1/16" when leveling the surface prior to pouring? That's about 0.015 degrees off from level.

For a short span, I'd say to find a really straight, flat, heavy and tall (so it doesn't deform much because of it's own weight) piece of metal, wipe on some mold release very thinly and evenly, and use very thick epoxy, perhaps mixed with metal powder. Shim the ends of your table in four places using stainless steel shim tape or something similar. Put down the epoxy, then put the flat metal piece on top of the shims and it should squeeze the thick epoxy to the right thickness. The result would be flatter with no meniscus. You can also use half a shim here and there to make sure the rails will not only be flat, but square to each other as well.

But of course, that idea is useless for such a long span unless you can find something that is really flat and won't deform much under it's own weight, and is 20 feet long........Of course if you can find such a thing.......

I don't see a problem with sanding the high spots, but how will you know where the high spots are? And how will you keep it consistent between the two strips? That's the big problem I see with sanding. At the end of the day it might be easier to do it over completely.

I hope that rant was helpful.

[louieatienza]

Thanks for the response everyone! I can tell I'm going to love this site. The West System guys wrote me back and suggested attacking it with the sandpaper once again but using a jointed board to help get the flatness I'm after. I'm going to get things as close as I can once again with a laser (by hand) and then run the length of it with a jointed board for awhile and see how that turns out. I'm hoping with consistent pressure it'll take care of high spots.

In the famous MadVac build, he used piano wire pulled very taught to check for flatness. Another way would be to source out drop-offs from a place that cut granite countertops. The stones are polished using an overhead gantry machine with diamond wheel hones, so they're reasonably flat even though they're not machinist grade test surfaces. You can attach peel-and-stick strip adhesive to it, mark a section with chalk, and sand using only the weight of the granite. When you blow the surface off, all the low spots will have chalk. You just proceed till one swipe with the sandpaper can remove chalk along the entire length, then the surface would be reasonably flat.

[RedSquirrel]

That's still a long way off from 1/16" or 0.0625"!.

The 1/16" was just my best guess at how far off it might be in some areas. I've been using a laser level to project a beam across both epoxy pours and using that as my baseline for how much material to remove. Unfortunately, the beam width it projects isn't very consistent. It's a DW089LG if you're interested, supposedly +/-0.125 IN over 100 FT which gets me mostly level but the more important thing just being a straight line that can be cast 360 over everything while I work.

The 20' steel straight edge is completely out the window. I can't fathom what such a thing would cost. Yikes!!

I don't see a problem with sanding the high spots, but how will you know where the high spots are? And how will you keep it consistent between the two strips?

Well this is more or less the dilemma right now. If the epoxy had settled like water everything would be good. But it didn't and I had to sand a bit which means I'm back at square one trying to get it flat. I don't really have a way to know where the actual high spots are except using a straight edge and feelers. Once I mess with it by sanding there's also no real way to make sure they're parallel other than going off the laser. I think my next attack method is going to be getting them close using the laser and then a large 6' or so jointed board with sandpaper run across them until the straight edge doesn't show any gaps.

[NIC 77]

The 20' steel straight edge is completely out the window. I can't fathom what such a thing would cost. Yikes!!

Yes, I agree, I don't see this as feasible at this juncture. I think you'll get it sanded to where you can live with it. What kind of machine is this that you're building? Are you making boats? Do you mind if I ask what city you're from? I'm in Southwestern Ontario.

I didn't mean to imply that the straight edge had to be made of steel. It could be made out of aluminum, or even a built up wood frame, so long as it was really flat on one side and didn't deform very much under it's own weight. It only needs to last for two pours. 20' though, is a challenge. So I agree, I don't think it's feasible.

[pippin88]

Acetone should not be used to thin.

My understanding and experience has been you need a thick pour to get past surface tension. I did a thin pour that did not self level.

You need a slow as possible hardener.

I used West system. I'm not sure if lower viscosity epoxies are available.

[wizard]

The 1/16" was just my best guess at how far off it might be in some areas. I've been using a laser level to project a beam across both epoxy pours and using that as my baseline for how much material to remove. Unfortunately, the beam width it projects isn't very consistent. It's a DW089LG if you're interested, supposedly +/-0.125 IN over 100 FT which gets me mostly level but the more important thing just being a straight line that can be cast 360 over everything while I work.

The 20' steel straight edge is completely out the window. I can't fathom what such a thing would cost. Yikes!!

Just imagine how much it would weigh!!!!!

Well this is more or less the dilemma right now. If the epoxy had settled like water everything would be good. But it didn't and I had to sand a bit which means I'm back at square one trying to get it flat. I don't really have a way to know where the actual high spots are except using a straight edge and feelers. Once I mess with it by sanding there's also no real way to make sure they're parallel other than going off the laser. I think my next attack method is going to be getting them close using the laser and then a large 6' or so jointed board with sandpaper run across them until the straight edge doesn't show any gaps.

You might want to think about how jointer planes work. You still have the issue of keeping both sides in the same plane though.

I suspect that your best bet here is to start over. I know that doesn't sound too appealing but epoxy leveling can work under the right conditions.

I'm a bit confused as to what your real overall error is. One entry implied that you aren't off much more than 0.004" which might not be a problem depending upon your goals and how fast that deviation happens. The thing with linear rails is that they don't tolerate rapid deviations from the plane they are supposed to be mounted on. Your machine is rather large so I have to wonder what you are looking for in overall tolerance and what you can accept.

One idea for local low spots would be to get a surface plate to use to flatten out some filler materials. The idea here is that the surface plate would bridge the high spots forcing the filler epoxy to flow out of the gap and flatten out. This might be workable for a few minor low spots. Do use a release agent and ideally a Saran Wrap type barrier, otherwise your surface plate will become part of your machine.

Another option is to talk to the people handling Moglice. In the USA that is Devitt Machinery. I don't believe (do check though) they have self leveling epoxy but they specialize in the use of Epoxies for machine ways, alignment and so forth.

[1Jumper10]

Products Page

The steel casting epoxy with the anti-bubble pigment is what I used. You would probably pay less to Mike at precision Epoxy than getting Moglice from Devitt Machinery. I checked with Devitt when I was researching leveling techniques and products. IIRC the price I found from them was ahem; steep.

[louieatienza]

Acetone should not be used to thin.

My understanding and experience has been you need a thick pour to get past surface tension. I did a thin pour that did not self level.

You need a slow as possible hardener.

I used West system. I'm not sure if lower viscosity epoxies are available.

You can use MEK... but it's nasty stuff.

[wizard]

After posting the reply above I got to think about what is possible measurement wise here, and frankly I' not to sure that laser levels too informative. Laser levels are great for construction work but I hardly see them as useful for determining anything beyond 0.020". Even if you have really good eyes. For one laser beams aren't perfect and will diverge a bit over distance. Second how exactly are you measuring the beams position?

I know tools exist to do this with some precision but your laser could be off by 0.025" all by itself, assuming my back of the napkin calculations are correct. I imagine by that spec DeWalt means off from true level and probably measured from the center of the spot. You might get better results by rigging up a target with which you keep the beam centered upon and with an additional gage to measure the vertical displacement to keep that beam centered on its target.

[the_canuck]

I had the same issue with the epoxy leveling of my 8' table rails. To solve I ended up sanding/scraping the epoxy beds for final leveling. To do that accurately as I could I picked up a used 48" Sterrat straight edge and a couple of 15" long granite parallels. I used the straight edge starting on one corner of one rail and worked my way along using the granite parallel as a sanding block to take the high points down. I used a machinist level to make sure I was also level across the rail. Next I then worked the second rail, getting it flat and then checking it against the first with the straight edge and machinist level.

It took a couple of weeks working nights and long days on the weekend to get it to where I was happy with the results. It was a ton of work but I did get it within 0.002" from one end to the other checking with feeler gauges against the straight edge and sliding the straight edge along the rail. Rail to rail was harder to measure with the machinist level and the straight edge. I was using a 0.0005" level on 12" and I got it to the point I couldn't see the bubble change much measured several points along the table.

The hardest part of the process was not putting a twist in the rail from one end to the other. Key was work slow and check often the progress to make sure you weren't introducing anything you didn't want. I ended up with a bit of twist on the second rail which was uniform from one end to the other so I resolved by shimming the linear bearing rail with some very thin shim material.

If I ever build another table I would seek out a shop that can machine the tops of the rails flat. The cost would be worth the time and frustration. Of course I found out after I was done a local shop did indeed have a machine large enough to handle the job. They could have also drilled the bearing mounting holes saving even more time. I did mine with a mag drill and lots of patience. The straight edge was really useful for this part of the process as well. 36 tapped holes per rail in 1/4" steel was not so fun by hand. Only broke one tap and the holes were aligned so that the rails turned out straight and parallel.

Andrew