The TJ dash is a great leap forward in jeep design.  In incorporates an excellent heating and ventilation system, puts the gauges in front of the driver where they belong, and is a home to the airbags.  Unfortunately, it is also a bit bulky and makes placing a traditional roll cage problematic.  Because of this bulk, a roll cage placed outside the dash ends up taking up a lot of valuable real estate in the cab.  To minimize this loss, the front legs are often bent in several directions to closely follow the dash profile.  These extra bends are natural weak points, and by closely following the dash, the air vents and speakers may be partially or even fully blocked.  Finally, and in my opinion most importantly, by placing the legs outside the dash, it is difficult or even impossible to remove the dash without first removing the roll cage.  Having to pull the entire cage to change a bad heater core, for example, is simply unacceptable.

    The other way to build a cage is to run the front legs is through the dash (actually between the dash and the firewall).  This method causes almost no loss of useable space in the cab, doesn't block the vents or speakers, and most importantly, it doesn't make dash removal difficult or impossible.  Unfortunately, until recently, the only through the dash cage I'd seen required removing the dash to install and also required modifying the vent duct work to clear the uprights.  Additionally, because of the positioning, there was a potential for limiting visibility from the driver's seat.

    So, I waited for inspiration.  And it came in the form of Raymond Woo's cage.  After a little experimentation, Ray discovered that he could pass a 1.5" tube behind the dash without removing the dash itself.  Ray's cage, featured on his home page at The World of Woo, was the first I'd seen that passed behind the dash but didn't require dash removal or duct surgery.  I had finally found the answer...almost...I really wanted to use 1.75" tube.  The reason for this was that I wanted to be able to use the factory door surrounds, which are designed to fit onto 1.75" tube.  I could have used 1.5" tube for the uprights and 1.75 elsewhere, but I'm too much of a perfectionist to go that route.  So, being the intrepid experimenters that we are, Dave Bandel and I decided to try to use 1.75" tube...and it worked!  We built Dave's cage and then a few weeks later we built mine.  To make the cages, Dave and I split the cost of some tube, a bending die, and a notcher.  Dave built a stand for the bending die and we were in business.

    The buildup of my cage is shown below in a series of pictures.  Some of the things that aren't readily apparent are the reasons behind some of the design features.  One fairly unique design feature is the diagonal stringers from the main hoop to the front and rear spreaders.  Most cages use two stringers about 12" apart near the center of the cage.  This is a great design if you need a place to hang a CB, but doesn't do much for a rollover where the corners hit hard.  I already have a good place for my CB, so I decided to have a single stringer fore and aft and stringers on either side going from the center of the main hoop to the corners of the cage.  This gives excellent strength for diagonal rollovers while offering more strength for straight endos than 2 bars going straight fore and aft.

    Another unique feature are the side stringers between the main hoop legs and the rear legs.  Often, cage builders will place these on a diagonal.  This adds a lot of rigidity to the cage for a roll back.  But, it also blocks some of the visibility out the sides.  I also needed a place to mount my rear rack in the back.  Since I have two Delta Hoppers mounted in the tub, I had to work out a way to mount the rear rack higher than the legs that came with the rack so I could open the bed boxes.  By making the side stringers horizontal, I added a measure of rigidity to the cage while providing a place to mount the rack where I could still have full use of my bed boxes.  It's a small compromise in terms of strength, but I felt it was worth it.

    Also, it is quite common for cage builders to add a spreader from side to side just above the dash.  This adds a measure of resistance to collapsing during a side rollover, but can cause an obstruction to visibility as well as an impediment to dash removal.  Since the front legs on my cage go through the dash and also since they are bolted to the windshield, I felt that there was adequate resistance to folding, so I omitted the dash spreader.

    Finally, the cage is made from 1.75" x.120" DOM tubing.  DOM, or drawn over mandrel, tubing is seamless and is about 40% stronger than ERW, electro-resistance welded, tube which has a seam.  Cost wise, it is about 15% more expensive, but since my total expenditure on metal was low, the added cost was minimal.  DOM tubing is stronger because it is work hardened when it is formed.  Some of this extra strength is lost when the tube is welded (in close proximity to the welds), but even in those areas, it is stronger.

    By the way, this cage would not have been possible without Dave Bandel.  Not only did Dave act as guinea pig by building his cage first, but he also did all the welding.  I can weld well enough to save my life, but Dave is a true artist.

    One of the things we discovered on Dave's cage was that with 1.75" x .120" DOM, we could re-use the windshield brackets for the stringers that go over each door.  This makes lining up all the parts much easier, looks clean, and adds a degree of rigidity to the cage.  Here's a picture of the bracket cut off the factory door stringers.  You might be able to see that the factory tube necks down from 1.75" OD to 1.50" OD just before the bracket.  Once I cut the bracket off, I used a bench grinder to take down the lip so it would fit into the end of the new tube:

    We decided to start the bending with the only bar in the cage that required two bends.  This bar was the rear spreader bar.  The rear spreader can be a straight bar, but this would have made sliding tall objects onto the rear rack difficult.  Bending the bar offered two challenges.  First, the bends have to be in the same plane.  With a little patience and some careful marking, Dave and I were able to get the bends in the same plane.  Second, the bar has to clear the hard top.  Luckily, we had already built Dave's cage, so we had a good idea how high we could go.  Here's the rear spreader lightly tacked in place.  It was tacked up so we could accurately determine the angles, lengths, and fishmouthing required for the tubes that would eventually be welded to it:

    With the only multiple bend on the cage completed, we were able to cut, bend, and fishmouth the stringers for the rear half of the cage.  We then started on the second most difficult part of the build-up...the front legs.  We started by cutting the dash, rubber gasket, and firewall flange (the last done with a plasma cutter) to allow the tube to slip behind the dash.   We then worked out the angles and with a little experimentation, we figured out the best angle to bend the legs to allow minimum reduction of visibility.  We put the bend as low as possible to keep the tube very close to the windshield.  Here's a picture of the dash cutout.  In the future, I plan to put a rubber gasket on the rough edge of the dash cut to give it a clean appearance:

    With the front legs bent and positioned, we made pads for the legs to sit on at the floor of the tub.  We used pieces of 1/4" steel plate welded to match the angle of the floor.  I drilled two 1/2" holes in the pads to bolt it down later.  My intent was for the bolts will go through the double floor area and into the Currie rocker protectors, but the protectors only go about 1/2 way over the hole.  I plan to hog out the hole in the protector to allow the bolt head to sit flush with the protector.  Here's a shot of the pads with a sheet of paper behind the pad to make it easier to see:

    With the front legs in place, we positioned the door stringers and cut and fishmouthed the front spreader.  As with the rear spreader, we tacked the pieces together to allow precise fitment of the remaining stringers.  Here's a shot of the junction of the front legs, door stringers, and front spreader.   Note how closely the leg follows the windshield and how the factory windshield bracket is re-used to ensure correct positioning:

    With all the pieces cut, bent, and fishmouthed with the notcher, we were ready to quit for the day.  I took all the bars and laid them out on the floor for this shot.  Note that the rear spreader is not the one we made, but a Hanson Enterprises piece.  Originally, I had planned to use a few of the cage parts from Hanson Enterprises and ordered the rear spreader and door stringers (he calls them sport bars) complete with welded on windshield brackets.  I'd assumed they were DOM tubing, but when they arrived, I discovered they were ERW tubing.  After explaining that I wanted to use DOM, Wayne happily accepted them back for a full refund...he's an excellent guy to deal with.

       The next day, we got an early start with Dave beginning the welding chores.  In addition to welding in the new tubes, Dave also circumfrentially welded the seams of the factory main hoop and fill welded the spots where the factory bolts were:

        Welding the cage took quite a while.  While Dave welded, I cleaned up the pieces in preparation for welding.  Here's a shot of me working on the disc/belt sander:

    Here are a few shots of the cage after welding.  This first is looking forward from the tailgate.  Notice the soot on the seats...a blanket used to protect them caught fire, but damage was minimal:

    This shot is looking back from the driver's door.  The wiper motor housing just barely clears the rear spreader tube.  Initially, it had a minor interference when closing the glass, but after a week, the plastic had warped slightly allowing the glass to be opened and closed without difficulty:

    Here's a shot looking through the driver's door.  Again, you can see how closely the legs follow the windshield.  You may also notice that the latches for the hardtop have been trimmed slightly to clear the front spreader.  Also note that the stringers are bent to allow more headroom.  We could have bent the spreader instead, as we did at the rear, but this would have been weaker.  I am was willing to sacrifice a little strength at the back (since I don't carry rear passengers), but not up front:

    In this shot, you can see the rear side stringers.  They will support the rear rack, as well as providing rigidity to the cage:

    In the last two shots, you can see the view from the driver's seat.  On the passenger's side, there is no loss of visibility, even with the door removed.  On the driver's side, there is a slight loss near the bottom of the windshield.  The obstruction is very minimal and leaning very slightly forward (less than 1") allows me an unobstructed view:

     Overall, I think Dave and I met my goal of building a strong, functional, and good looking cage.  It took nearly two full days, but this was mostly because I'm a bit of a perfectionist and because of the complexity of the design.  Since the build-up, I've painted the cage black (the legs were painted before being passed through the dash), drilled the holes for the feet and for the factory door surrounds, and welded tabs to the rear side stringers for the rear rack.