... of my own stupidity.

I had to weld in front nut plates for my bolt in cage, which are basically recessed into the top of the longs.

Took my time and let everything cool, but it appears there was still enough heat to close up the top of my door gaps on both sides.. i.e. the front cowl moved in/down a little.

I can still close my doors, but it's a little tighter than I like. The difference is, since I planned to put a cage in the car, I did not install an Engman kit, so there is no kit for me to cut before doing the whole Porta-power push.

Is my only option here to cut the top of the long, push with a Porta-power, then seam weld them back?

Oops. You have to put a car on jack stands when welding a cage including the foot plates. Door bars are practically required. Was the car sitting on its wheels when you did this?

Yep, it sure was on it's wheels Brant. I thought that I had read through all the threads I needed to and I remembered hearing about putting it on jackstands, but never got it definitively. Dumb, dumb mistake.

I guess that would apply for welding on a long kit also?

That is correct, Bob...

Remove the door latches, close the door and take photos and post them. I want to see the fender to door gap and the door to quarter gap. Also post "F" and "X" measurements from this diagram. Both sides. and both "X" directions.

http://www.914world.com/specs/bodydims.php

Will do. I love being given clear instructions! Thank you for your help. Should be able to snap some when I get home from work.

I'll be out today but I'll check on this later.

How was the car sitting ? On the wheels or on jackstands ? Post your welder setting s when you did this. I never brace anything when I do long kit its which require a LOT more welding and have yet to have a cars dimensions change from that. The only advantage I have is the scissor lift has flat plate on top that support the center of the car.

Okay, here are the dimensions, from the best of my measurements. It is entirely possible I have not measured correctly, especially dimension X but the pictures shouldn't lie.

Dimension F Passenger side - 24 7/8 inches (spec is 25 1/8)
Dimension F Driver side - 24 7/8 inches (spec is 25 1/8)

Dimension X Passenger front to drivers roll bar point - 54 inches (spec is 52 5/8)
Dimension X Drivers front to passenger roll bar point - 53 1/2 inches (spec is 52 5/8)

First three pics are passenger side

Click to view attachment
Click to view attachment
Click to view attachment

Second two are driver

Click to view attachment
Click to view attachment



Scotty, this was done with a Hobart 110 welder on setting "3" so I could get enough penetration on the thick plates. As mentioned before, the car was on the wheels, because I'm stupid and didn't know it should have been on stands.

I should also mention that I had some shrinkage after I did my hell hole/console replacement on the passenger side. The door was tight, but not as tight as it is now, after I did the nut plates.

I know that sitting in the warm sun and drying off has always helped me with my shrinkage problems. Maybe you could roll the car outside and give it a try.

I'm not seeing how jackstands would help. It seems to me, the chassis on jackstands could twist without proper door braces. And if you weld on jackstands... I would think you would want to take reference measurements of the car before the door braces are mounted. Then tension the door braces once the car is on jackstands so the jackstand measurements match the car on the ground reference measurements. It feels dangerous to me, to assume that the car on jackstands is better than on the ground. And putting the car on jackstands incorrectly may be more likely to F your door gaps than leaving it on the ground

All I am saying is, when you weld with the car on the ground, at least it's in its normally loaded state. Once you put it in the air, it is reacting to different loads and tensions than when its on its wheels.

But, I'm not a professional mechanic or welder I just play both in my garage. Can someone elaborate?

If that was my car, I would be thinking about removing the door, measure how far the hinge plate it from the rear door jam, then get a block of wood that fits the hinge plate perfectly, and hit it with a sledge hammer, then measure again.

They do move, but not much.

But, I have been accused of being bush league.

Mental bends, you just need to move it to the right place, with controlled hits from a very large hammer.

Rich

Ya those "X" measurements a screwed up. They are showing the windshield to Targa bar is wiiiiiide but the door gaps don't support that. So the important thing is both numbers are the same and are showing the opening is square. In other words the cowl is square with the targa bar and that's ok.

The gaps are hard to read. If you followed my instructions I'm assuming these were taken with the latch plate removed. This allows the doors to swing from the hinges with no interference and usually lends clues to the method of correction. I often see the door sitting low to the quarter panel at the belt line. (horizontal window seal height) This would indicate a sag at the cowl to rocker. Usually from a front end collision and usually on one side with a "X" measurement that differs from side to side.

You often see the gap at the top of the door is tighter than the bottom but still on the same plane. This often indicates a sag at the rear of the rocker usually caused by weak or rusted longs at the jack point.

Your passenger side looks like the belt line is even with the rear of the door to quarter panel gap being tight but even. The front side looks like the gap is wider at the top. This could have been the result of a prior adjustment.

The driver side looks like the front gap is even but the door is sitting low and tight at the beltline.

If you have done a bunch of welding on the car in the past correcting rust you likely had some issues from that before you started the roll bar plates. Because we have multiplying issues I'm not sure I can help long distance but I'll give it a shot. We'll have to go slow with this one.

Start here: Passenger side. reset the passenger door to align with the front fender. Take a small piece of flat bar and grind it to 4 to 5 mm thickness. This is your door gauge. Start with 4mm. This may give you more room at the back of the door. It's a little tight for a 914 but you have issues. Start there. move the door up so the bottom corner of the door matches the front fender bottom. tighten the gap at the hinge side to 4 to 5 mm with your gauge. Be sure the fender and door are flush top and bottom with each other.

Then swing the door shut and re-shoot a photo and post it. Then we'll have a better idea where to push, pull, etc.

Drivers door: Same thing. Check the gap top to bottom on the hinge side with your gauge. Match it to the fender line. The shoot a photo and post it. Let's see where that brings us. Leave the latches out for now.

So what happened? I have a very similar problem over this way...

I took Rick's suggestions to heart. Unfortunately haven't had time to address them correctly. Currently I'm running around like a chicken with my head cut off trying to get the car race-ready for September. Will let you know what I discover.

Let me know if I can help. You may want to get a handle on the chassis as part of the prep. Tough to corner balance a twisted car.

This winter I plan to do a video this very subject. I will do a simple instructional using the Raspberry car. It is as straight as any car I have ever seen. I will demo chassis repair and explain datum and reversing damage without causing more damage.

You can still corner balance this car. It will sit crooked but will be balanced on scales. I ran a race car with the same issue for a dozen years until I built a new car and used jack stands on the 2nd car when installing the cage.

When you welded the plates in you should have put the heat on the plates and flow the welds out into the long in short bursts. (apply heat from thick to thin) Next... Mig welders have a trigger so you can stop too. You had the heat setting correct for thin sheetmetal but not hot enough to get good penetration and heat transfer into the plates which is why the beads are narrow and tall. (proud). The multiple passes look pulled when you should have pushed the gun. Gun was moving to fast where the wire just piled up like traffic on the 405. Not sure what the purpose of the car is but those welds at the foot plates would not pass tech inspection and will raise some stink eye. Lot of porosity. Not trying to be critical just observant.

Take it as a learning experience.

The good news the gap issue is not as bad as one might think. You put a lot of heat into a small area. You also put a lot of heat primarily on the top of the long not the sides or bottom. The top is short and now has a bow to it. What moved? The long did and the quarter panel came forward. Next lesson -door bars or some kind of bracing are not optional. Have to mechanically resist the the shrink and slow way down with the gun. Support the center section of the tub at minimum of 4 corners.

How to correct it.. you could bang on the hinge pillar. The door will sag more if you bang on the lower hinge pocket. If you bang on the upper pocket the gap will actually get tighter at the door/ qtr bottom rear and sit high at the beltline. Beat on both and you might end up with a compromise that works for you.

Push the door opening with a porto-power and the qtr panel/ lock pillar will move outward. I would try that first and see if he door opening can be squared while watching the lock pillar for deflection.

More heat? oh no no no! Celette bench and pull? Not unless you want two 914s.

My thoughts to get it right you will probably need to slice, sleeve and weld the long. Double the gap you are trying to achieve and it will shrink back about 4-5 mm.

If its a race car and the gaps cause you not to lose sleep - grind the welds down at the plates and use larger diameter MIG wire. Do short bursts like plugs that overlap each other half way through the last one. Put the heat into the plates with a bit of overlap into the long. See if the car can be corner balanced without to much issue. If it can bolt in the cage and go.

With plates like that you need the highest setting. The trick is not to burn through the thinnner metal. Not too hard with the long, and this is not where MIG shines I am afraid, but it can be done by focusing your heat on the thick and let it flow into the thin. Good luck amigo.

Jeff,

this is an amazing post. Very informative
can you explain why the tip should have been pushed instead of pulled?
thanks for your time
brant

Pushing and pulling the weld each has a place and reason. Will try to make it short. When pushed will keep the puddle hot and molten longer. The weld will tend to be smoother,lay down flatter and be less porous on thick to thin sheetmetal. Each bead, consecutive overlapping plug or squeeze of the trigger will flow into each other instead of cooling so quickly it stands on top of the last pass. In this case would have actually reduced the amount of wire needed= less gun time =less heatsinking to the long. Practice and technique!

Thanks for the opinions everyone.


I agree Jeff, your advice makes particular sense to me. You also hit the nail on the head. I had traditionally always "pulled" my welds, since most of what I have done has been thin sheet metal. This was my first attempt at something this thick, and the issue was made more complex with the combination of the thick plates and thin longs. Welding dissimilar thicknesses is something I'd really like to improve my skills on.

Here is the dilemma then:
I understand that I will need to grind down the welds and basically start over. But then I'll be putting more heat into an already shrunken length. Do I need to just brace this much better than before and use jackstands this time?

Also, I'm trying to conceptualize the loss in length that both Rick and Jeff outlined. If my measurements of the windshield frame to targa roll bar are the same for both sides, how could the top of the long have shrunken without changing that dimension?

The changes in the gaps are perceived to be greater than actual. Heat concentration was low so the affect was down low in the tub. I would bet your windshield header to targa bar did not move at all. Picture the bottom edge of the door as a horizontal lever with its pivot at the lower hinge, the pivot didn't change.. the long below it did.

Upper body stayed and the lower body got shorter.

Imagine a 3 foot long 1/2 inch thick metal rod. Now bend the rod -it's still basically 3 feet long with a bend in it. Not 100% because the inside radius shortened and the outside stretched but close enough to visualize. The overall datum length is shorter but the rod is still 3 feet long at the surface.

Ok, I will have to do mine soon. Should I pull doors, weld one bar at the top one at the bottom of each door, and put on jack stands? I couldn't see anything move then, but I could be wrong. I imagine even just one at the top of each door would be sufficient to keep the car from buckling. Boy I'm glad I read this one, I would have just welded it in.

Brace, support tub and go slowly with the fire control while watching what the chassis is doing. Switch back and forth between the driver side and passenger side to minimize shrink. This will give it a moment to cool down. By tacking from side to side the chassis should not bind up and twist. If you see gaps getting tight STOP, REGROUP AND EVALUATE. GO SLOW WITH THE FIRE.

I don't mean to come off as a know-it-all. I mean no disrespect. I need to say your theory is completely flawed. "Unit Body Construction" is not the movement of one section . The movement of the chassis in one plane effects the car in 3 dimensions.

My background and foundation for my statements: I have attended ICAR training on chassis re-dimensioning. I have worked as a unit body repair tech at three other shops before opening my own shop. I was an owner/operator/ and trainer in the industry as a sole proprietor for 18 years before selling my business and starting in the insurance claims field. I had a Celette in my shop as well as a floor system that we used for light pulls that didn't require a full on engine/suspension removal to be brought back into spec. I have a Celette in my home shop.

I promise I will do a training video that will aid in the understanding of what went wrong and how to think and plan a strategy for proper correction.

Rick,
No disrespect taken. I was refraining from getting into the X,Y, Z axis subjects because that would just throw everyone that is learning a curve trying to visualize what is happening at the longs. I was attempting to confine the subject to heat dynamics and as usual I can go off into other tangents. I agree the entire unibody is affected no different than a simple box structure. I respect your credentials also and they are along the same lines as my own.

So why is it auto claims suck the life energy from the human body faster than it can be put back in?

Would I need to run diagonal reinforcing bars connecting the two door bars to each other? That may help keep it from twisting.

The steel guy we have used on 356 car has a very large piece of plate steel on the floor.
Like 20 by 20

He just welds the car to the floor to prevent twisting.

I did the same thing, but on a bench that is bolted to the floor.
Car did not twist.

For the longs, the factor door bar setup looks pretty complete and much more than any door bar I have seen here....it goes all the way up to the roof.

Rich

I recall foley telling me once that putting in an inner long kit can creat a lot of this type of movement and it might be easier to put on the outer clamshells instead...for that reason of heat and shrinking....dunno

Thanks for understanding. Flat text sometimes comes off as arrogant. I find it necessary to talk in X,Y,Z to prevent the beginner from focusing on the 2 dimensional problem that they "think" they created.

Funny, all the years I spent in the collision business I don't do auto claims. I work as a property storm claims adjuster.

You are probably better off. Especially when it comes to making money.

Would anyone mind if I temporarily hijack this thread and post a few pics of my own door issue for suggestions?

May be better to start a new thread. I'll help if can.

Revisiting this, though it might not be a good time since Rick is going to be traveling down here and may not. Since I had to take the cage out to do some painting, I figured I'd take the chance to get some better measurements. What I found:

F windshield to targa bar dimension = 24 7/8 inches on both sides (spec is 25 1/8)
X cross dimension = 53 1/2 inches on both side (spec is 52 5/8 inches)

So it seems that the good news is if I have had any shrinkage on the longs, its equal from side to side. The bad news is that I'm certainly not very close to factory tolerances.

What I can't understand is that if the distance from windshield to targa bar closed (about 3/8ths of an inch), then how is the X dimension greater now?

Rich had mentioned that it might be possible to pound the recess that the door hinge bolts to inward to gain a little gap at the back of the door. I've considered this, but also thought it might be possible to remove the door, pound the riveted hinges backward a little bit, then tack weld the hinges to the door, effectively moving the door longitudinally forward a little bit. Thoughts?

My door doesn't fit right either.


I wonder why

John


Click to view attachment

It a gain of mm's and a compromise to a proper repair. I will be in Fl. tomorrow but not long enough to travel to St. Pete. I would be happy to help if I could. I may be back in Fl. later in the summer if you get hit by a hurricane.