One of the guys over on the metal shaping web sites has a neat tool he sells called a donut dolly. It works more proficiently in off-dolly shrinking a crowned panel because it supports the bottom side in more than one spot (around the perimeter) as the panel is bumped from the top in the center of the donut dolly. This would seem to be useful in body repair, especially to address high spots found after blocking primer where heat would not be as feasible. A few days ago, someone on another forum asked for advice in removing an outward roof dent. I suggested a low-buck alternative to the donut dolly, simply using a PVC pipe fitting. Where it may not work as aggressively as the donut dolly which has more mass, sometimes slower is better, especially when trying something new. I just today got around to trying this process in the shop, and rather than use the roof of something sitting here, we will use a piece of 18 ga CRS to simulate a roof. Grid layout for the Wheeling Machine, and completed "sample" roof... The first order of business should be to make some profile templates. In the case of an actual dented roof, use the opposite, undamaged side. This will allow you to check your progress as you go, and easily find the remaining high spots. Using a crowned body hammer, a "dent" will be added from the back side, crossing both of our template areas. With the dent added: I chose to use a PVC elbow, it had two different sized ends for two tools in one. As with any body tool, they should be free of any burrs that may mar the metal surface. I also added a slight radius to the edges... Holding the dolly against the bottom of the panel, you can see that due to the dents it only touches the panel at the red arrows. Based on off-dolly principle, the shrink would occur more prominently in the direction of the red arrows. However, the typical dent like this would have more stretch in that direction anyhow, so it actually will work in your favor. Another especially important step in this process is to check your progress with profile templates (perhaps made from a good panel on the opposite side of the vehicle), to better monitor the progress and see where the high spots remain for the next area to address. Here are the tools we will use today: After some off-dolly bumping: Check with the templates Progress....these are light taps only with the hammer. We just want to bring down the high spots, not create craters. Here's where I finished up... I have a little over an hour in making the panel, denting, and removing the dent. Where I still could have gone a bit more, it was about to a point where high build primer should have masked any remaining imperfections. This dent removal could also have been accomplished with heat shrinking using an O/A torch, etc, but for those times where you may not wish to use heat and want another option, this seems to work well. I think with using the donut dolly and it's heftier mass, one could shave some additional time off the repair.
Is that elbow a 1.5 x 2 inch? Since you are using light taps with the hammer, would a heavier mass "dolly" really be required? Maybe just put more of your weight behind the elbow instead?
49rf, look on either metal meet or allmetalshaping, search for donut dolly. Alchemy, the actual OD of either side was about 1.5 for one, 1.25 for the other. Made for 1" pipe. One size up may work better, but being low-buck, this is what I had laying around. I think the accuracy of hitting in the right spot is more important than how hard the swing was. Of course, I was holding the "roof panel" between two legs and hammer and dolly in my hands, so I wasn't applying too much force as it was...
I had some questions on another forum I had this posted on, and felt the discussion would add clarification. First off, lets clarify the on and off dolly again. On-dolly: As the hammer stikes the panel with the dolly held directly beneath, the forces come together (red arrows) and the panel is compressed between the two. With nowhere else to go, the metal is forced outward (blue arrows) in the form of a stretch. Off-Dolly: This time as the hammer strikes the apex of the dent from above, and the dolly rests against the one from below, their offsetting positions will cause the resulting forces in the direction of the blue arrows, forcing the metal back into itself, or causing a shrink. Seldom is the case where we have a one-sided dent. Typically with an off-dolly scenario in fixing a dent, it is exactly as you describe, using off dolly and working around the perimeter of the dent to bring things back where they were. As you surmised, spanning the two sides of the dent, we can off-dolly with more effectiveness, as the forces we are now introducing are more concentrated, or "Off-Dolly on steroids". But lets back up a second and look at the forces at work when a dent forms. As an outside force is applied to the door of your new car (blue arrow), the direct reaction is a stretching of the metal, causing it to cave in. But we also have an indirect reaction in this stretching as the panel "springs back" slightly and you'll now notice a slight bulge or high spot around the perimeter of the dent (yellow arrows) This is why effective dent removal will include working both sides of the panel. Moving down to the next view, we can see how an off-dolly scenario which supports both sides of the dent should prove to be more effective than the "one side at a time" method shown above in the second picture. Getting back to your suggestion of using a low crown dolly to span the dent, this would work up until the point the dolly made contact with the panel from the hammer striking the other side, as now we are stretching again. And just like the indirect forces in play when the dent was formed, there is also a small amount of springback when hammering out the dent. So to accomodate these forces, a hollow dolly would better fit the bill. Now on to the the problems associated with the hollow dolly: Once our process starts to move the dent, you'll get to a point the outer perimeter has reduced in size. If we don't stop here and regroup, you may end up with the scenario in the upper view, where the dent is now trying to "cave in" on our hollow dolly. This is where a variety of sizes would come in handy, to reduce dolly size as the dent removal progresses to better match the outer perimeter of the dent. When you get to the point where it is only minor movement needed, you'll want a hollow dolly sized the same diameter or only slightly larger than the head of the body hammer. Where a "flat" body hammer in most cases actually has a low crown, this works about perfect to counter the springback effect. Looking at this another way, the dent has formed an arc (stretch) in an otherwise flat panel. We're trying to return the arc to a flat (flatter?) line. Having a dolly made of a non-metallic material also helps to grab or capture the surface of the panel moreso than a metal dolly would.
Well I guess it's about time for a sequel. I had a left over door from a bus repair earlier this year, and as it was damaged the least, we thought to keep it as a starting point. New ones in primer go for just under 1K, so I think we can save some money here.... The last time I used the "donut dolly" was with the low crown body hammer (the flat one) but when I tried it on this flat panel, progress seemed to be rather slow. I think the low crown hammer was better matched to a crowned panel, and with this panel being flat, I may have better luck with a crowned face body hammer. So as we're starting with bent metal anyhow, what's a little trial and error going to hurt? I'll just keep my fingers crossed that the hammering force would not be great enough to pound dents in the other direction. (we're looking for consistency and moderation ) Here's our dented sample today, an aluminum door skin with a nice crease that travels just about the full length of the door. As the donut dolly is very closely matched in size to the hammer diameter, accuracy in locating the devices opposite each other is crucial. Depending on the panel you are repairing, this may be a challenge in itself. To keep the tools properly aligned and thus provide a more efficient shrink, I suggest a grid on either side, matched to identical starting points. I numbered every fifth one just to keep better track of where I was. Some days you need all the help you can get. Also shown is the damage before starting, located at grid 5, 10, and 15, just for reference of our progress. As I am working with a flat panel, I am able to use a straight edge to monitor the progress and see where additional shrinking may be needed. If you are using this process on a crowned panel, I would suggest making a profile template matched to the undamaged side of the car. Also, seldom does the crown on a panel remain consistent from one end of the panel to another, so it is very possible you may need different profile templates to accurately guage different sections along the panel. After about 30 minutes of work, here are the results, again located at grid 5, 10, and 15. And the light reflection shows considerable improvement over the first pictures.... This is now at a point we could strip the paint off the door, and any remaining defects would be taken care of with high build primer.
Thanks for the post, I like the idea. I have a similar long dent running across the rear quater panel of my car to repair so your method will be very useful. Would you use the same method for working with steel too or would the plastic elbow need to be made from a different material?
It's the same plastic elbow used on 18 ga steel in the first post, and it worked there. I think a donut dolly made of steel will have a tendancy to cause on-dolly strikes if you are slightly off in the positioning, resulting in more stretching, not really what you want when removing a dent. The plastic dolly helps to eliminate the chance of that.
Robert, you are wise beyond your years. I like your grid, i have seen you use that for different things and it is very effective. The part i am having a problem is the leading edge of the dolley is on the high part of the crease while the lagging edge is flat on the now repaired surface, so if we gave the crease a north to south axis and we were repairing from the north and going to the south then the pipe dollies' leading edge would be south and high on the crease. The lagging edge of the dolly is north and flat against the repaired metal and that would leave east and west both high above the metal. When you strike with the hammer what does the east and west do? Do they make contact with the pipe dolley? I am assuming you are striking in the center of the pipe dolley. Thanks, oj
OJ, all your north and south confuses me, I prefer parallel or perpendicular to the crease. Hammer goes on the high side of the crease, in this case, the back of the panel. Dolly on the low side of the crease, or the front of this panel. When looking at the long crease on this panel, the majority of the stretch is perpendicular to the crease. Although I did not take a picture of it, if the straight edge were held parallel to the crease, there was little to no deflection in the panel. If we did indeed have more stretch in this direction, I would think it would be wavy. Thus, the way I read this, the majority of the stretch is perpendicular to the crease, and the majority of the shrinking we need to do is therefore perpendicular to the crease. So where the contact points with the donut dolly is aligned also perpendicular, this is where the majority of the shrinking will take place.....right where we want it. Depending on the formation of the dent, stretching results may vary. But typically with this process, shrink the high spots, check with your profile guage, and repeat. When it gets to where the dolly makes contact on all sides, you'll shrink in more than one direction. That's my story and I'm sticking to it.
Hey Mac, Another top drawer repair demo ( per your usual form) complete with all the trimmings Using this repair method, had the paint not been broken, say the damage was just caused by a rubber bumper strip, and you used a plastic or rubber hammer with the pvc dolly, you may well have been able to just polish out the marred paint, and sold the repair job as a ''paintless dent repair''. A ''putty flinger'' would still be slathering on a second or perhaps a third coat of ''fender-in-can'' and still not be finished with the repair in the time you've drawn out your grid and followed your repair plan Way kool! " Life ain't no Disney movie "
nice, very nice. great Idea I really like the grid.Now I need a pvc pipe elbow. I just have a hard time remembering all the awesome stuff I find on this site. Thank you
Great info and pics! About the only thing I could suggest is to use different color pens (grab a selection instead of all Black) for the grid, and maybe have a mirror and light mounted on a swivel so you aren't bobbing your head above and below every few hits.... Might make it easier to spot the Red/Red grid or the Red/Blue spot...
I was wondering who was going to pick up on that. Where this panel could have been finished with a torch, a shrinking disc, or many other methods, this process works especially well to minimize damage to an existing paint finish. Had it not been for a gouge in the paint, this could have been buffed out and call it a day. Well, except for all those sharpie marks...
Thanks for the post MP&C. As a beginner at panel beating and metal shaping I'm still trying to get my head around some of the concepts and how to put them into practice. IMO, Your explanations of shrinking and stretching are the easiest to understand. I have no trouble knocking up patch panels on the bench top but when it comes to removing a dent from the centre of a fixed panel I struggle to understand how and where to hit and place a hammer or dolly. Thanks for the good info this has really helped. My next few weeks are now going to be full of smashing old panels and trying to knock them straight again!
Great info MP&C - thanks heaps. My initial thought was to use a reducing bush or socket instead of a bend, but thinking about it, the bend is easier to hold in your hand as you move dolly and hammer around. Those high pressure PVC bends are available in a range of sizes - I noted that you suggest using 1&1/4" x 1", but there are plenty of other sizes available for little dollars, so a variety of donut dollys will cost no much bucks. Now, if only knowing all this stuff meant I could actually do any of this cool stuff. Ho hum. I'll go now.
Sorry, i got stupid there - a long day. I thought about it this morning and it was one of those classic 'what was i thinking' moments and when i read the above i wanted to thank you for not rubbing my nose in a dumb comment.
I wouldn't say a dumb comment, just gets me to further explain something I likely didn't give enough detail on. I think I said it standing in your shop...having to explain something will help you to understand it better yourself. ......and I missed answering part of your question. When two opposite sides of the dolly contact the metal, it should shrink between them in the direction of the contact path. If you have three sides touching, the third side should react like any other off dolly operation. It will still shrink, but less aggressively than when supported on two opposite sides. So the majority of the shrink will be between the two opposite contact areas, and a lesser amount of shrinking, between the third side and the hammer strike. I think that answers your question..
The Jalopy Journal
