Epoxy or No Epoxy Under Filler
by
Serge


The Saga

Many discussions arise around whether or not to put filler over epoxy. Some say, epoxy manufacturers actually support this claim, that filler over epoxy is better. The reasoning behind it is that the filler/epoxy/metal bond is stronger and more flexible (epoxy has more flex than filler…) than the filler direct to metal…

I wondered in earnest which method was better. Someone on the Autobodystore news group suggested that someone does a controlled test….well I did.

The Test

I was doing filler work on a truck one night and I decided that while the filler was curing I would go ahead and prepare the test. I cut small coupons of 22 ga, satin steel. The coupons are about 7 by 9 inches…I prepared the coupons as follows:

Prepared with a 36 grit Roloc disk (my usual way of prepping filler)

Prepared with 80 grit, hand sanded 3M stickit gold.

Prepared with 180 grit, hand sanded 3M stickit gold.

Prepared with 80 grit + one coat of epoxy.

Prepared with 180 grit + one coat of epoxy.

I then prepared a single big batch of filler and slapped about 1/16 in on each prepared coupons, pressing the mixture on the coupons well…All of the coupons were ‘filled’ while the filler was still in it’s liquid/plastic state, I was far from the hardening point. The epoxy was cured ½ hour with the help of a medium wave UV lamp. I set it up such that the coupons were warm to the touch, not hot…I wanted to make sure the epoxy was still in its recoat window (7 days at 70 deg f…). Also note that I did not scratch the epoxy prior to putting the filler on it. I followed PPG’s instruction of putting filler on the epoxy within the recoat window.

The coupons were left to cure inside my house for a little more than one week with filer and all.  
 

The Products

Epoxy is PPG’s DP90-LF with DP-402 catalyst. The epoxy was given a 20 minute induction time even though it does not require it. The pot life of this epoxy with the 402 catalyst is 8 hours, so the 20 minute induction time is not a problem…

The filler is Rage Gold.

The metal is 22 ga, satin steel.
 

The Torture

To gauge the adhesion of the filler, I did 3 destructive tests with the coupons. I tried to figure out a way to simulate the worse in car hit for a filler repair. I figured it was an inside out hit. Hitting the metal from the inside forcefully. So I found me a table corner I could spare and gave each coupon a good whack from the back…the results were interesting…

Then I proceeded to bend them. In this bend test, I supplemented it by trying to remove filler pieces with my fingers and removed anything that peeled easily…

Then I finished with a screwdriver. A flat screwdriver scratching the filler and trying to pry the filler off the metal….
 

The Results

The whack test…

I tried to give each coupon a good but equal whack on the table corner…

The metal was not kind to my work bench corner…but it was all for the advance of science…So here is how the test went, blow by blow…

80 grit + epoxy

Whowa! The first whack a piece flew off…So much for the extraordinary bond of epoxy…Notice that the bond failed in the epoxy, not in the filler. The filler piece still has the epoxy bonded to it (black backing of the removed piece).
 

180 grit + epoxy

Again a piece flew off...Actually 2 pieces…was predictable in view of the 80 grit example…note that the epoxy is still attached to the filler…

36 grit direct to metal

Decidedly, whacking it from the back was quite a torture. The 36 grit prepared coupon failed but not as badly as the epoxies. It is not obvious from the picture but the break is not a clean one. There is still some filler stuck to the metal in the area where the piece is missing. The filler failed here, not the bond to the metal…Interesting…

80 grit direct to metal

 

Decidedly, I am lucky I did not bet the farm on the result of this test! The 80 grit did better than the 36 grit. Nothing flew off…
 

180 grit direct to metal

This is the surprise of the bunch…it’s the one that did the best on the whack test! Go figure…I would have never predicted that…The funny thing is that the results are gradually better, 80 did better than 36 and 180 did better than 80…
 

If I can add my general observations on this test is that the direct to metal did much better. The bond with the epoxy looked like the epoxy broke in two. There was still epoxy in the scratches on the metal and epoxy on the pieces that delaminated. This confirms my intuition that the epoxy was the more flexible of the two solids but also the weakest.

The filler that failed direct to metal failed in the filler itself. It did not delaminate cleanly like the epoxy did. It left a layer of filler stuck to the scratches in the metal…

My conclusion is that the filler is much harder than the epoxy which was the weak point in this test…

The bend test…

The bend test is pretty easy. I just bent them in two to about the same angle…

 

80 grit + epoxy

Well, the epoxy did not do better here than with the whack test. This large piece of filler I delaminated with my fingers…Once the bond was compromised lifting with my finger would take large filler pieces off….

180 grit + epoxy

This one, predictably, did worse than the 80 grit epoxied coupon…Again, note that the epoxy to filler bond is very good…

36 grit direct to metal

Well, this one did very good. I just could not snap any pieces off with my fingers. I had to take a screw driver to pry some pieces off…The bond was very good…

80 grit direct to metal

Again, the bond was much better than with the epoxy but not quite as good as with the 36 grit…but still very strong. Notice that even when some pieces delaminate, they leave some filler behind in the metal scratches….

180 grit direct to metal

This is a surprise. It is surprisingly strong. I would even go as far as saying “STRONG ENOUGH!” Gee, this filler does not need much to hang on…

The bend test simply confirmed what the whack test revealed. The epoxy was surprisingly easy to delaminate. One thing this test confirmed in my book is the ability of filler to chemically bond to DP epoxy. All of the pieces of fillers still were attached to the epoxy (they had the black epoxy backing). In this test, I did not scratch the epoxy at all, the filler was applied directly to the epoxy within the recoat window as per the PPG P sheet…

The screw driver scratches…

The screw driver scratch test simply confirmed what the other tests had revealed. The funniest thing is that I could remove most of the filler over the epoxy…not difficult to do with a screw driver…

That’s pretty bad…

 

In contrast, the 36 grit roloc, direct to metal prepared coupon was tough as nails.

In conclusion…

Well, that settles it for me…I will continue to put my fillers direct to metal. The bond was much tougher, even with the 180 grit scratch which was surprisingly tough. Actually, I would say that the 180 grit scratch was tough enough! I don’t believe that it would ever delaminate in a normal auto body situation. It did very good (the best actually) in the whack test and was very difficult to delaminate with the screw driver…

I believe that the weakness of the epoxy would have been multiplied with two coats. The thicker epoxy would have been just a thicker layer of weak material.

It was interesting to note that in both cases, it was not the bond that failed but the material itself. Both the epoxy and filler were properly adhered to the metal. When the epoxy delaminated, it left quite a piece of itself in the metal scratches. It did not come off clean. This tells me that the epoxy layer was ripped apart. The same was evident with the direct to metal, but interestingly, the layer of material that was left stuck to the metal was much thicker with the direct to metal filler…

To me, proof is in the pudding…no wonder I never get a filler failure when applied direct to metal…that stuff sticks!

Serge is a regular contributor to the Autobodystore.com Forum and has had extensive experience with both body repairs and painting.  His results should be recognized as a reflection of this experience and his ability to evaluate the materials and process he used.  Different materials and processes could result in different or similar results.



Disclaimer: The ideas and methods described in this web site were developed under unique situations. Since these situations cannot be duplicated, you may get different results. Use and application of any of the site's content is at the user's own risk.

All information on this site is the property of Stuart's Autobodystore.com and may not be reproduced without prior written permission.