Gloss streaks when rubbing out finish

[John Hollander]

I have been trying to finish a live edge walnut table for almost a year now and have had nothing but one problem after another.

After having to sand the top down to bare wood for the second time (final sanding at 250 grit) I applied a single layer of boiled linseed oil. I let the linseed oil dry until the top did not feel oily to the touch or leave any residue on my fingers (I believe it was 3 days). I then applied 6 coats of gloss Arm R Seal. Each coat dried for at least a day and received no sanding. I sanded the top lightly with an orbital sander with 300 grit then 600 grit. I had a small spot that looked like I was begging to sand through the finish ( I am not sure how this could be as I only did a few light passes with the 300 grit ). In fear of having to start over again I applied another 6 coats of Arm R Seal gloss. Again drying for at least a day before receiving another coat. I then sanded the top lightly with 600 grit and worked my way up to 4000 grit (wet sanding after 600 grit). Around 1000 Grit I began to notice some odd "lines" They appeared to be glossier than the rest of the finish. I thought that maybe they would go away if I kept working my way up though the grits but they stayed the same.

Does anybody have any idea what this is or what is causing this? If my first coat of linseed oil didn't dry completely is it possible that the arm r seal did not fully adhere to the finish and is separating from the wood? It doesn't feel like these are low spots that arnet getting sanded. I am getting to the point with this piece where its making me want to turn it into kindling.

[John TenEyck]

I feel your pain, John. I struggled with Arm-R-Seal on a live edge walnut counter top a few months ago. So, how did you apply the Arm-R-Seal? Did you wipe it on? I'm guessing you did. Those lines look a lot like you sanded through one of more layers of the finish, and the fact that you let it cure 24 hours between coats means that each coat is a separate layer so if you cut through one layer you end up with those witness lines.

What worked for me was to apply the Arm-R-Seal with a foam brush after cutting it about 25% with mineral spirits. It flowed on and self leveled beautifully. Moreover, it put on a much thicker layer than if I had wiped it on. I had originally put on 3 or 4 wiped on layers and just couldn't get a streak free finish. So I sanded that smooth with 325 grit, then brushed on 3 or 4 layers as described above. I let each coat dry about 12 hours, sanded with 600 grit, then applied the next layer. After the last coat I let it cure 12 days before wet sanding it with 600 - 4000 grit. I had a beautiful satin sheen at that point. I then used an automotive multi level rubbing/polishing compound to bring it to a low gloss. I was very careful (and worried) about cutting through a layer of finish, but either never did or didn't see it.

333059 333061

John

[Jerry Thompson]

I would have applied a one pound cut of de-waxed shellac over the BLO. That would have kept the oil and Arm-R-Seal separate. I always use scrap first as a test. I too have jumped in and used the final piece as the test piece.

[Michelle Rich]

I agree with Jerry..dewaxed shellac after the BLO..

[Prashun Patel]

I doubt the problem is the BLO. You would have noticed that as being a problem if your Arm R Seal dried soft and gummy. After 12 coats, the BLO is locked in there.

I'm assuming you wiped on your coats. Arm R Seal is a wiping varnish and a poly varnish at that. This means that you have a 12 very thin, relatively soft coats (relative to more brittle varnishes or shellac or lacquer). This creates a challenge for rubbing out. You have to work a little more aggressively to polish that softer surface, and you risk burning through each layer if you do so. I think you are seeing witness lines. As you abrade off the thin layer, you are revealing the lower, still glossy layer.

If you did not allow the surface to fully cure - like a couple weeks - it may also not have achieved full hardness, which can prompt one to sand very aggressively as the surface will not level evenly when you try to rub out, prompting you to rub harder than you should.
When this happens to me, I usually get patches - not rings as you have. So that puzzles me. Still I believe this is what happened to you.

Wiping varnishes are designed to make the finishing process easy. By drying fast and thin, the idea is that there should be little need to rub out the final finish. In fact, I would suggest that you apply another 2-3 coats to eliminate the witness lines, then gently wet sand with 600 or 800 grit paper (to remove any nubs, but not to burn through), let it dry, then apply a final thin polish coat to just restore the luster. Then call it done.

If you really want to rub out, then I suggest you BRUSH on two more coats, and wait for a long time before attempting to level and polish it. brushing will deposit about 3x more finish than the average wiped on coat, which will create a thicker layer that can stand up to the leveling stage of rub out.

(edit): Whoops, should have read John's post...I agree with him!

[David Hawxhurst]

to me it looks like you have low spots in the finish. the low spots are the shinny spots. if its low spots you can sand more down or build up more (or some combo there of).

[John TenEyck]

By golly, you might be right David. It does look like that.

John

[Steve Schoene]

Sure looks like cutting through varnish. Problems: using power sander. Running through too many grits. Start with 600 to remove gross defects then shift directly to 2000. Get to gloss with polishing compound.

Apply coats in set, spaced just several hours apart or as soon as finger doesn't lift varnish. After set of three let cure 24 hours touch sand with 320 or 400 and do next set.

Sets let coats link together giving thicker effective coat to avoid cut through.

After last set allow several weeks to cure before rubbing out.

[Alan Lightstone]

to me it looks like you have low spots in the finish. the low spots are the shinny spots. if its low spots you can sand more down or build up more (or some combo there of).


That's what I was thinking when looking at it.

[John Blazy]

Steve, John and Prashun are correct. This is why when I spray (or brush, if I brushed) I always recoat within the melt-in period. Lacquers always allow melt-in due to solvent welding the interface, but most other coating systems (WB, polys) and especially 2K systems require melt-in to previous coat, which is essentially a wet on wet build. I always do some degree of rub out, so melt in as one homogenous layer is my goal. I have had witness lines during rubout before (WB, 2K urethanes, conversion varnish, etc), and that is when I determined to get intercoat melt in. Hot days will lessen melt-in times to just minutes with certain coatings.


On the molecular level, one has to understand that the top 2 to 10 microns of the crosslinked matrix of the polymer is many times denser than the rest of the 2 to 5 mils of build within that layer, under the top surface (each "coat" that you layed on and allowed to dry). This is easily seen when you are buffing a conversion varnish (which is nearly impossible to buff due to this very aspect) because many coatings have slip agents and hardening agents that migrate to the surface after lay down, and require physical wet sanding to "punch through" this surface layer before buffing - if it will buff out.


So your pictures are clearly showing the dull, less crosslinked lower layers of that givin coat, which rub off the surface of the highly crosslinked surface of the previous coat.

[Scott Holmes]

If the shiny lines move location as you are sanding, then they are witness lines (sanding through the different layers of finish) If they are truly slight depressions that don't change location; then they need filling, with more layers of finish.

I think they are witness lines.

[Steve Van Kirk]

Hi John,

I have felt your pain. I ran into this same problem a couple years ago refinishing a conga drum with poly. I agree with Scott - I think these are witness lines. Sanding them won't help, and - at least in my case - neither did re-coating over them & trying again more carefully. After re-coating, the witness lines were less visible, but still present. To completely eliminate them I had to go back to bare wood.

My problem turned out to be four-fold: working with poly - a reactive finish which builds discrete individual layers rather than one thick layer like lacquer or shellac; refinishing an inherently curved surface (a drum - very easy to sand through a layer); pores not being filled well - they weren't flush with the surface, leading to low (shiny) spots in the poly, which led to more sanding to to get the sheen to match, which led to sanding through layers, which led to witness lines; and finally, trying to do the whole process with an electric sander. :rolleyes:

In the end, what I ended up doing was this: sanded off the poly; reapplied 2-3 coats of spray shellac (Zinsser dewaxed from Lowe's/HD - to provide a thick enough layer to level the surface); sanded the shellac flat (i.e. leveled the surface - no shiny spots from pores); brushed on a few coats of poly (onto the now level surface); & hand-sanded from there. I'm not saying this is the best way to solve the problem - it was a lot of work and I'm sure there are much better approaches. But it seemed to do the job.

Some day I'd love to try John's method regarding re-coating within the burn-in period. I was kinda wondering at the time if something like this might work. But I hadn't yet read of it & thought it was probably just wishful thinking. But if it's as slick as it seems to be, you'd end up with a single (relatively thick) layer of poly and little (no?) chance of witness lines. I just didn't know enough at the time to give it a go (timing of coats, etc.). And I probably still don't. :) Plus I was brushing. Not sure whether that would affect John's method or not.

The bottom line is that whether you're brushing or wiping, if you're following the likely mfg's directions & letting the poly dry between coats (say, 12-24 hrs), then what you have to work with when you rub out is still only as thick as the last coat. You could put on 50 coats of poly & let them cure for 6 months, but when you try to rub it out, you've still only got the final coat to work with - sand through that & here come the witness lines. Grrr.... (The final coat is thicker if brushed, of course, giving you a little more room for error than wiping.)

Again, not saying what I did is THE method. Just saying it's what I did.

I wish you well! If you have any questions, holler.

Steve