oh and speaking of, i've had rabbits this year, no clue where the rabbit hole is!

 

My Rx7s floor is still not completely clean, but it's maybe 95% clean (not including wheel wells, which I have yet to start on). But standing there for hours chipping undercoating and wiping the floor with enamel thinner takes it's toll. At a certain point I decided that it's clean enough for now and I wanted to dig into some bodywork.

You know the quickest way to change your mind when you feel like doing bodywork? Doing bodywork.

I wanted to start with something small, and this little bracket and captive nut for the heat shield is a good candidate. This is one of multiple areas that are better repaired with the floor panel still in the car than with the panel removed, and it's a good starting point so that I can improve my skills before proceeding to anything larger.



Now obviously there's the rust to deal with, but even if there wasn't there's undercoating stuck on the nut and possibly under the bracket that can't be removed easily with it attached to the floor. In this photo (out of focus) you can see the marks I made with the punch:



Those marks are on the two spot welds that hold this bracket on. Then to remove them I used this spot-weld cutter:





Pretty easy to use, although I wish it had a slightly smaller diameter on the cutting head as the spot welds are only about 1/2 the diameter of the cuts in this image. One thing I should have done is made clear marks on the floor to show me where to line the welds back up.

Once I removed the bracket, I was glad I went to the effort of removing this part rather than trying a chemical solution:



There's just no way I could have properly repaired this with the bracket in place. I used a flap wheel to remove all of the rust, but at this point the metal on the right was very thin and I had to cut a section out:



I welded in a patch (16 gauge, since the extra thickness makes it easier to weld):



Unfortunately I had to hold the patch in by hand (magnet was disrupting the weld) and despite looking fine, I actually welded it in about 1mm too deep:



Probably this would be fine, but I can't leave well-enough alone. So I filled the void with weld and then ground it back, doing the same on the left side where the metal had become thin:



That worked reasonably well, but every time I went to grind it down new pinholes would appear in the original metal. This lead to the typical vicious circle: grind it flat, punch a hole, fill with weld, grind flat, punch a hole, fill with weld...

At a certain point I had to decide that it was smooth enough. Too much of this cycle leads to forgetting what the original geometry looked like, and then the metal will be 1/8" thick and the bracket will end up in the wrong place:



On the inside of the car the repair is looking a little bit, uh, tumorous?



But at least it's solid. I checked it for pinholes from every possible angle and then cleaned and hit both sides with weld-through primer. I need to be careful with the stuff, it's really expensive nowadays and I only actually need it for the area that will be under the flanges on the bracket .The rest of the floor can be painted with epoxy primer at the time I paint the chassis.

I'd like to show putting the bracket back on, but I haven't done that yet. Removing the rust on a tiny part like this really needs to be done chemically, and I'm still waiting:



I have limited faith in chemical rust removal processes. They work, but only if they get every last bit of rust and you can never really tell. I'd like to replace this part but it doesn't look like it's sold separately from the floorpan and I don't trust myself to replicate it close enough. Is there some way to search for this type of part based on its dimensions? Otherwise I might take some really good measurements and see if I can have some laser-cut and bent, that way they're 100% the correct size and shape. As a last resort I can reuse the original parts.

Even though it didn't all go quite according to plan, I'm glad I did this work because I learned a few things:

- Make much clearer alignment marks before removing the original part. I operated under the erroneous assumption that I would still be able to see the spot-weld marks after grinding, but I ended up replacing all that metal.
- Cut not just the rusty metal, but enough space around it that there is enough thickness to weld to.
- Keep the welder at the absolute minimum power that successfully penetrates the metal.

There are several more of these little brackets, in addition to some larger repairs that I want to do before removing the floor. So I have plenty of practice ahead of me before the big jobs begin.

Until next time : )

 

this is like an essential truth, lol

if you really needed to do it, i'd think making a piece that is close enough to that would be really easy. and you could use a U nut, or rivnut instead of welding one on
i guess it depends on how concourse you need to get.

and at least you have them! mine is missing a couple, and the rest are tweaked, not sure why, but it could have been when the P.O. of my car drove over another car.

 

The shape of the bracket is not complicated, and welding in a captive nut wouldn't be difficult either. My main concern is making accurate bends in such a small part. My sheet-metal brake is obviously way too big, and going smaller with vise grips or something tends to make the bend go at a slight angle.

This isn't bad for one bracket, but if a heat shield uses 5 mounting points and they're each out by a few mm then it adds up. Not the biggest thing in the world, but at the end of this I don't want to be stretching or hammering parts to make them fit.

The bigger deal later will be the seat brackets which are also going to need removal and reinstallation. Added to this, they're already bent from the floor bending upwards when the rails are used as a lift point. Obviously that prevents a much larger issue than some heat shields which could just be bent into place.

I also have a few brackets that are tweaked like yours. I'm my case it's from the bolts getting so stuck that trying to loosen them just started twisting the bracket.

 

Lol. I've been in body work hell for a couple years now, it is not fun that's for sure. Looking good though, keep it up!

 

Work on my Rx7 has been progressing. It's quite cold here at the moment, and the days are short. When there are only 10 hours per day of light and you're at work for 8 1/2 of them, it just sucks the energy right out of you. Still, I've been making time to continue working on the floor.

Since I started down the path of dealing with the heat shield brackets and the rust underneath them, I felt like finishing that before starting on anything else. I've learned from my experience on the first bracket and refined my process. I'll use this one bracket as an example:



Obviously the bracket itself is bent, but then there's the rust under it. Out of the 7 brackets I've removed so far, 5 have had some pretty serious rust underneath. This was one of the worse ones:



I used a marker to make note of the positions of the spot-welds for re-installation. Obviously the marker won't stand up to any chemicals or abrasion, but I later made some better scribe marks and I took plenty of photos. I also used a different method to take accurate measurements of the original placement, which I'll get to in a dedicated post later. You'll also see that I've marked this area "D". There are 8 of these exhaust brackets and I noticed that some of the brackets have spot-welds in different places. The welds aren't always centered in the flange of the bracket. So I thought it was prudent to keep each of the original brackets in a separate Ziploc labelled with the corresponding letter. That way I can put the plug weld holes back in the correct place later.

One other thing I learned from the previous repair I made was to cut out a lot more material than I think I need to, and to spend a lot of time getting the shape of the patch panel correct. I made sure to cut the patch slightly oversized, then formed it slowly until every angle matched the original piece. This probably took at least 45 minutes on its own (there are a lot of contours on the floor, and this piece has both a slight curvature and two bends, each at odd angles to the edges (not perpendicular). Then I trimmed the edges till the patch fit neatly into the opening with a slight gap to fill with weld.

I also fine-tuned the welder a bit to reduce the voltage and the number of pinholes:



It's not winning any awards, but it's solid. In one of my recent posts I was thinking about how to replicate the original brackets, and j9fd3s had a good response - "It depends how concourse you need to get". That was a good question, since I hadn't thought about it, and it applies to this entire project.

This is the bottom of the car, so I don't need it to be perfect. But I also don't want to leave anything that I can't live with later. I might be the only one who wrenches on this car, but I know that laying on my back and looking up at the welds as I work will bother me.

So I spent some time with the Dremel grinding down the welds:



And then wire-wheeled it so the finish was consistent, so I can see how smooth it will be:



This will be good enough, I think. Once it's primed and undercoated I'm not going to notice the little marks.

Speaking of the Dremel, I also upgraded my tooling:



I have at least 5 empty kits that used to have rotary tools in them, all of them now empty because the tools wore out. The last one started emitting smoke while I was cutting out some of the rusted areas. This time I paid for the name brand Dremel with the high-torque motor and it's been great so far. No tool lasts forever, but I'm hoping this one at least lasts me the rest of this project.

To be continued

 

When it comes to the brackets, I decided that making them myself wouldn't do. I could do it, but the bends are very small and the value of the time it would take to cut and bend these pieces would eclipse the amount of money spent on just getting them made. So I designed them in Fusion, printed out some samples on my 3D printer, and then when I was happy with it got them laser cut and bent:





Not bad, although I will need to correct the angle of the bends a little bit. For the threads, I purchased some M6 weld nuts:





I used the long M6 bolt and the spacer because I don't have any other clean M6 hardware and I don't want to contaminate the weld. I'll have to order some fresh hardware before I reassemble everything. I cleaned everything with isopropyl alcohol, then welded the nuts on. After the brackets had time to cool I brought them inside where it's above freezing and spent some time deburring, wire-wheeling any slag, and correcting the bend angles on the brackets:



And then cleaned them again:



For anyone doing something similar, rotate the weld nuts 45 degrees from the angle I did. The angle I chose made it impossible to weld the upper and lower sides of the nut since the end of the mig torch won't fit into the bracket. These threads don't see a lot of torque so it won't really matter:







Then I dried them thoroughly and packed them away in a dry location until I can prime them. There's a slight complication in that I'm not happy with the weld-thru primer. I find it doesn't adhere that well nor does it stand up to abuse, and when you try to paint over it the paint runs away (sort of like trying to mix oil and water). A proper 2K epoxy is in order.

Before I can weld these back on I need to prime the bottoms of the brackets, as well as the locations on the floor of the car to which they weld (since I won't be able to reach it after they're welded on). I don't want to trap the area on the bottom of the flange against the car without protecting both surfaces from moisture. I have two (expensive) spray-cans of 2K epoxy primer, but the temperature is too cold to apply it in the garage and I never spray chemicals inside the house. Temperatures are a bit warmer later in the week so I'm hoping that between that and the propane heater I can get the garage warm enough to spray the primer and get a few hours of curing in. If not, I'll start on something else while I wait for temperatures to rise enough.

Maybe it's overkill, but everything rusts here so I'm going to take the extra care and get it right. I was briefly considering how to prime the bottom of the weld-nut itself since there's no way to prime underneath it once it's welded to the bracket, but at a certain point you just need to decide you're okay with it. Once the car is reassembled I'll drip some of the Eastwood frame coating in there just to be safe.

Until next time

 

I've realized that the weather is not going to cooperate and allow me to spray primer anytime soon. I was hoping to reassemble some before I took anything else apart (or I'll end up with a bunch of parts everywhere), but that isn't going to happen until things warm up and I don't want to wait until spring to make any progress. I decided to just continue with disassembly for now. Next on the list is this transmission mount:



The rust on the outside is no big deal, but the threads on this weld nut are bad and if I'm able to access the back then replacing it properly is better than using a helicoil.



Plus I get to make sure there's no other rust while I'm in there. Took out the center punch and spot-weld bit again:





Once I had drilled all of the spot welds that I could see, I started working away with the chisel:





I was trying not to pry too hard. Once you're all the way through the area around the spot weld, it should pop free without much effort. Too much prying will bend the piece, and means there are welds still undrilled. There are a few places I found that didn't look like obvious round spot welds, but more like plug welds:



I also needed to find a shorter drill, since the drill I was using is too long to fit inside the transmission tunnel:

 

It popped off without too much fuss. I went too far on some of the welds I drilled, but that's not too big a deal to fix. Once again I'm glad I did this since there was a bunch of rust hiding underneath:





Unsurprisingly there's rust on the back of the mount:



It doesn't look like replacing the weld nut will be too difficult, since there's plenty of access:



Although I'll need to get some new carbide bits for the Dremel.

Lastly, I had to knock the mount side to side a bit to get it free of the car, and I dented the sides in when doing so:



Nothing functionally wrong with it, but I'll spend some time straightening that anyways. Next I'll need to strip and prep both the mount and the surface on the car, as well as remove the other mount for similar treatment. Undecided if I'll remove the large brace that sits under the mounts. On the one hand I don't need to fix what isn't broken, but on the other hand I've come this far and maybe it's worth removing the brace just for peace of mind.

Until next time

 

A belated Merry Christmas and Happy New Year to everyone.

When I last posted I'd come to the conclusion that it wasn't feasible to spray primer in winter temperatures. Since I didn't want to disassemble any further before reassembling some, and I also didn't want to lose momentum, I turned my attention back to cleaning the parts of the car that still have undercoating:





And that's where I stopped, because on December 31 I was skiing and I fell on my back from a pretty good height and with significant speed. Fortunately I managed to not break anything somehow, but I did knock the wind out of myself and it hurt to inhale for a couple weeks afterwards. Still does now three weeks later if I inhale all the way.

Anyways, I'm now feeling well enough to put in some work but it's currently -25C where I live and the forecast says -37C by the morning. I don't mind the cold, but I'm certainly not going to go to the garage and hold metal tools in those temperatures.

So I figured I'd sit near the fireplace and post about something I haven't brought up yet. Over the past 7 months or so I've been incrementally upgrading my tooling. Not the tooling you're thinking of most likely (although there's been some of that too), but computer related tooling. It's become increasingly clear in the past few years that the future of this hobby is at-home manufacturing. Particularly for old cars with out-of-production parts and custom applications. I've posted plenty of 3D printed parts, but we're now seeing affordable laser-cutters and CNC machines hitting the market. I don't have a need for either of those (yet...), but there were a couple of things I wanted to get.

First, I built a computer last summer:



I've been working from a 2020 Thinkpad for several years. It's perfectly adequate, but for CAD work it's a bit lacking. Also I'm a software developer, so it's nice to be able to open as many IDE windows and browser tabs as I'd like with impunity. It has a 7800x3d processor, RTX 5070 12GB, 32GB of RAM, integrated Wifi, and the case has a carrying handle. All of those choices were deliberate because my real goal with this computer was to be able to run this:



That is a 3D scanner. It lets me image an object in 3D and convert it to a mesh that I can manipulate in the computer with resolution down to 0.2mm. This opens up a bunch of possibilities - take measurements in software, 3D print a copy of something, use the mesh as a template to make a parametric model. The reasons I wanted one of these are threefold:

- Have an accurate reference image of the chassis before I started removing spot-welds.
- Check the feasibility of an idea I had in software before committing to it on my actual car (may or may not happen, I'll give more detail when I make a decision).
- As far as I know, there are no full scans of the FC chassis available. I'm also uniquely situated to take a scan since I already have the car on the rotisserie.

Now this model (Shining 3D Einstar) is the most bang for your buck in terms of the type of scanner at-home users will be likely to use. It's not as good as some of the laser scanners, but I was able to get it gently used for a significant discount and it has so far proven to be very accurate.

This is why I wanted the RTX 5070, the 32GB of RAM, the on-board Wifi, and the carrying handle. These scanners require CUDA cores (and the more the better) so my GPU had to be Nvidia based. The scans eat RAM like you wouldn't believe (more on that in a minute). The Wifi and the carrying handle is so that I can easily carry my desktop computer to the garage from my room, and run it headless while I remote into it from my laptop:



I don't have any pictures while in-use (I was using the scanner of course and didn't have a free hand to take screenshots on the PC), but there are plenty of videos online. Essentially you open the software and point the scanner at the object, then press the play button on the scanner to start preview mode. This lets you see what the scanner is seeing on-screen. Then when ready you press play and the scanner starts generating points. These are points on the surface of the object which form a point cloud in the software.

The longer the scan, the more likely an error will occur. The software makes a best effort to re-align itself if it loses its orientation to the object, but it isn't always successful and when this happens you end up with overlapping geometry. This is a pain to correct. The larger the scan, the harder it is to generate the mesh later too (as I would find out the hard way).

Then you can align multiple scans in the software to create the full object:



Now the Shining 3D software is a bit clunky. I wouldn't go so far as to call it "bad", but it has some awkward translations and it tends to crash suddenly on large scans. Although in fairness, it DID warn me that I was running out of RAM.

Speaking of which, I mentioned that I specifically purchased 32GB knowing I wanted to do this scan and thinking that I could always buy more later on. I realized pretty quickly that 32GB was hilariously inadequate for 0.2mm resolution scans of an object the size of a car:



This was one of my early attempts to combine the scans - you can see that my RAM and page-file totals almost 80GB. For those unfamiliar, a page-file is a way for the computer to offload data that would normally live in RAM onto the disk. This severely impacts performance compared to the ideal solution (which is of course more RAM), but the page-file does allow the computer to keep running when RAM demands exceed the available physical RAM.

At the time I built the computer I paid maybe $120 or so for two 16GB DIMMS for a total of 32GB. I knew I could go all out and get 256GB for $500 and never worry about it again for the life of the PC, but hey, I could always get more if I needed it right? So I decided to save my money.

What I couldn't have known is that the AI boom would drive RAM prices through the roof. The same 2x16GB DIMMS I purchased for $120 are now $600. Upgrading the RAM is no longer cheap and easy. I also increased the page-file size to 200GB, 400GB, and it still wasn't enough.

So I decided to go all out in a different way, and got a 2TB SSD for the second NVME slot I hadn't populated yet and dedicated it entirely to page-file:



I knew this would be slow (and it was) but it was only a fraction of the price of a RAM upgrade.

More VRAM probably also would have helped since my 5070 has only 12GB available:



However the next model up card with 16GB would be a 5070 Ti, which was over $1000 at the time. Other than CAD work I don't really have a need for more GPU power. I rarely game and when I do I think the most graphically demanding thing I play is Morrowind. I decided I would make do with the regular 5070.

Even with all of the above, it took two days to export the combined mesh in the Shining 3D software and then it crashed when I tried to save it. Also my GPU now has a seemingly permanent coil-whine. Not very loud or anything, but noticeable.

While disappointing, I did find a better solution which was to export each scan as an individual mesh, simplifying it by 60 - 80% in the Shining 3D software beforehand to strip out excess details that aren't needed. Then take the meshes and import them into MeshLab, which has a pretty good alignment tool. I had to play with the settings to dedicate all of my VRAM to prevent MeshLab crashing, but I was eventually able to export an acceptable mesh of the floor:



This is not a finished scan, just a proof-of-concept I used to test what my hardware can do and the accuracy of the scanner. I need to do it again, after cleaning the undercoating off of the wheel arches (you can see the texture in the scan). But my goal is to have at least some large-scale meshes like this of every part of the chassis, and ideally a single mesh of the entire chassis if I can get it to actually export.

I also made an upgrade to my 3D printing equipment this winter:



Spider2k recommended I pick up a Bambu printer a few years ago when I was having problems with my Ender 3 V2. Although really, when was I not having problems with my Ender 3 V2?

At the time it wasn't feasible for me financially so I kept babying the Ender along, but after its most recent failure I couldn't convince myself to spend any more money on it. My dad also wanted some parts that would be too large for the Ender build-plate. I sold it with a large pile of spare parts and purchased this instead - an H2D.

Coming from an Ender 3, this thing is nuts. The H2D is to the Ender as the chainsaw is to the butter-knife. It prints at 300C out of the box, not requiring mods like the Ender did. It's enclosed, so the drafty basement area it lives in doesn't warp the prints. The enclosure is temperature controlled. It's quiet. It has dual-nozzles, so I can print support interface layers in a different material with minimal waste and have no difficulty removing them from the part. With the AMS 2 it can auto-change four filaments (per nozzle, per AMS, although I only really need one) and with the smaller AMS HT it has an 85 degree dryer which can dry even Nylon filament. It even has a vented outlet for which I've created a window plate so I can dump the fumes out the window rather than into my house.

And the feature I care about most of all is that it just works. I think I've had maybe 1 failed print that wasn't user error, and I've probably printed 100 things since I picked it up. It doesn't waste my time constantly like the Ender did by failing every 5 minutes and it when it does it has AI print failure detection that works pretty reliably to stop the print and prevent wasted filament.

The scanner + the printer put me in a position to scan something before I take it apart, have a reference in software for what it looked like beforehand, and then (if necessary) even 3D print a copy of it so I can hold it. And while I can't 3D print metal, I can make a model using the scan as the template and then send that out for machining. At this point the biggest limitation is my skillset.

That's where it sits for now. As soon as it warms up I'm going to continue chipping away at undercoating. Only three wheel arches and the spare tire well left...

Until next time

 

i think 3D printing is the future, and its basically here now. especially if you're doing something like a T2 swap, where a lot of little Knick knacks are NLA/or expensive

 

The availability of filaments that are robust, heat resistant, and UV resistant has been a goldmine for replacement interior parts. That plus the increasing affordability of high quality printers. Little things like the dash vents and the mounts for the cluster hood are easy to print and strong enough to last. Not to mention cupholders.

I've been hoping to find a use for PPA-CF filament soon. From what I can tell it's crazy strong, but also it's expensive enough that I don't want to buy it unless I have a specific use case for it.

 

i don't know if you saw this, but check it out https://www.rx7club.com/1st-generati...n-3dp-1172266/

 

if we are encouraging 3d printing, here's a good FC goldmine...

https://www.rx7club.com/2nd-generati...hings-1157452/

 

I just wish we could CAD in mesh like the implicit modeling of the near tomorrow. Going from mesh to CAD to STL is stupid. One day the printers will take native CAD and by then we will all be using implicit modeling which can model things live in real time that you would never be able to 3D model since even the CAD/3D modeling software's are subtractive in nature rather than generative.

great work!

 

I've been meaning to take some scans of the FC interior parts, but I have all of them buried in various bins. On reassembly I should take a scan of each part before it goes back onto the car. The only thing I don't like about 3D printed interior parts is that they don't have the correct surface pattern to match the rest of the interior, but certainly a 3D printed part is better than something broken or missing.

Your understanding of this subject definitely exceeds mine, but I think I'm picking up what you're saying.

I'm excited to see that there is work happening to make parametric CAD models automatically from meshes. I took a scan of a part from my dad's Quattroporte's door card to model it and add a speaker mount, and modeling a very irregular part like that is a big challenge. Probably someone who does it for a living would find it trivial, but for me it took several hours and it still isn't quite right yet. If you could automate the job of replicating a mesh in CAD it would save loads of time.

I saw a company advertising that they were using AI to do it (although everyone says that these days, so they may just mean ML and not AI necessarily). I'm not sure if their product is in testing yet but I was thinking to myself that a semi-automated process where you provide the mesh as well as a description of the item would be great. For example, a 3D scan of the screw posts on the back of the door card piece I mentioned shows some softness on the edges of the posts and only shows the screw holes as small depressions. Really they should be more cylindrical and with a properly sized screw hole. If you could upload the mesh into a software and then provide a conversational description of the item eg. "This is a part from an automotive interior. The screw posts on the back are cylindrical, and they accept a self-tapping screw of 5mm diameter." then you would add context that can be used to not only generate a model that matches the mesh, but also corrects for some of the common artifacts that scanners might pick up and make it closer to the real part than the mesh ever was.

 

It's slowly getting warmer here, so it's getting easier to work on my Rx7. I'm still been putting in an hour here and there after work to chip away undercoating in the wheel wells. The rears and the spare tire well are still pretty much untouched. Fronts and floor are mostly clean, although the fronts are at the stage where using enamel thinner is the quickest way to make progress and it's still too cold for enamel thinner to be effective.

So here's some bike related work I was able to do without going outside (much). Amongst the many missing parts on my bike was the neutral / turn signal housing, which is also the ignition barrel housing. I recently picked up a lock and key set, but that means I need to have someplace to actually install the ignition barrel. It also means I'll need a wiring harness, but I'll deal with that later...

Anyways, the part is NLA. There were some listed on eBay at absurd prices (I think one was $300?). Instead of purchasing it I figured I'd rather put in the time to make something that would work. Here are some of the photos I used as a reference:



The first thing I noticed is that this piece is actually fairly complicated. The top part for example protrudes a couple millimeters further than the ignition barrel, but the bezel has an inside chamfer down towards the lenses. Looking at the next photo makes the chamfer a bit clearer:



Then looking at the side profile, you can see that the hood part extends downwards further than the mounting surface:



There's also a cutout by the ignition barrel, I think to make it easier to remove the barrel. Look at the other side and you'll see that the hood also doesn't extend as far downwards, because the surface it mounts to near the speedometer is shaped a bit differently:



I did have a reference photo of the bottom of the piece, but it's an absurdly small thumbnail so it isn't very illustrative. After a lot of modelling, printing, test fitting, and revision I ended up with this:





Hopefully the bottom view illustrates what I meant about the hood extending past the mounting surface and the different lengths either side. The mounting surface is stamped, and on the left hand side it's rounded so it needed to have a contour as you can see.

Anyways, here's the finished part printed in PETG / Carbon Fiber:



I selected PETG-CF for the strength and impact resistance (in case I drop the bike) as well as because I will need to be able to glue to it. Nylon CF would have been sturdier, but I have never had any luck gluing Nylon. It needs to be compatible with glue so I can install these translucent PETG lenses:







The "TURN" and "NEUTRAL" lettering did end up having to be a bit larger than original, and I couldn't find a good way to replicate the diamond pattern in the lenses without scattering so much light that the 6V bulbs would be too dim. So I did have to compromise a bit with flat lenses. However, the finished piece fits perfectly and feels very solid. I was also able to replicate the aforementioned chamfer, which was important since I wanted this piece to be as close as possible to the original.

Until next time

 

Tackled the driver's rear today:



This was the easiest wheel well by far. With the chisel and hammer I was able to make quick progress:



For much of the wheel well I didn't even use the hammer. I just used the chisel like a scraper. After a couple of hours this was the result:





You can see that a lot of it came up in large patches. Definitely easier than the front wheel wells where it came off more like sticky dust. The front will need a lot of cleaning with enamel thinner to get the gunk off, but the rear looks like it'll just need a once over to lift any residue.

Of course I exposed a couple of holes. This one I knew about and had stuck a small patch over years ago:



Makes perfect sense since there was a prior repair on the outside of this area before I purchased the car. This is opposite the 1/2" thick Bondo bulge that you can see in some of my earlier photos.

There's also this, which appeared to be a repair over an existing repair:



I don't think this one was me, but I could be wrong.

Then there's this guy, which will be a lot more work to fix:



How it looks now:





Hoping to tackle the passenger side tomorrow (and maybe the spare wheel area as well).

 

Picked up where I left off yesterday and did the passenger side today, as well as scraping lots of dirt from the little cavities where the side markers live. Speaking of which, I'd forgotten to remove the markers:







Another hole here:



Moving to the passenger side, I started removing the fuel fill tubing. The metal tube in this photo is one I made a couple years ago to replace the original:



That was pretty simple, but these two nuts here did not want to come out:





Fortunately I've been regularly drinking PB blaster to improve my ability to improve stuck fasteners, and it seems to be working:





Then the whole assembly can be lifted outwards:



And if you rotate it 180 degrees, it comes right out:



This gives me much better access for these holes, which I was already aware of:



There will be more scraping and cleaning in the coming days

 

A friend of mine printed me a cup holder for my FC as a side project while he was learning to use his. I noted the same thing but feel this is a surmountable problem. Depending on your choice of "concourse level" I was thinking there might be a couple options. For my cup holder it fits in the arm rest, I think it's dark enough and hard enough to see that if I used something like finishing puddy or hi-build primer that could get a pretty smooth surface. For some of the textured areas I was thinking about trying to make a mold or replicate a similar texture pattern to apply to 3D printed pieces.

 

It's not too difficult to get a smooth surface with putty / primer. One issue I found is that the plastic doesn't sand well, since it gets hot and starts to clump up. I think fiberglass filler + wet-sanding is the way to go, then dry thoroughly and finish it with glazing putty before high-build.

I find smooth surfaces sometimes look out of place though. Everything in the FC interior is textured. I spent some time trying to figure out how to replicate the OEM texture but never thought of a great solution, so I just orient whatever side of the print I care about most towards the build plate so it has that PEI texture.

 

Ran into an interesting video on youtube . It is a free app by the author of the video "CNC Kitchen" that applies different customizable textured skins to your models. Might be worth checking out to get a closer oem like surface finish without having to post process the print which can be very time consuming.

 

I made a resolution not to post until I had something more to say than "I scraped away more undercoating". The undercoating is 99% gone now, mostly confined to the little nooks I can't easily reach. I'll get the rest while I perform the metal work. I have a very minor update today.

It's a long weekend here in Canada, and we've had beautiful weather. My dad and I have been building a loft in our garage for extra storage space, which means my Rx7 had to be relocated to the driveway while we are working. I had a few free hours today so I spent them continuing to build the temporary structure inside the car before I remove the floor. This is what I put in a few months ago (not sure if I posted about it):



I'm well aware that the fitment between the tubes is terrible. I made an error when cutting the birdsmouth shape. This isn't a permanent structure so it doesn't bother me too much, but I will be grinding out and redoing those welds and I'll probably cut a small filler piece for the big gap visible in this photo.

In my defense, I had never built this sort of thing before. My work today is much improved.

The purpose of the x bar is that I need some sort of structure in the hatch area to which to weld the long tubes that will prevent the car from flexing when the floor is removed. The other end of the tube needs to go to the front of the car, ideally an area with strength behind it:



This little round protrusion is a natural choice:


(I ground the remaining primer away after taking this photo - before any welding)

Not only does it sit right on the other side of the front frame-rail, but it also provides a very convenient way to center the tube.

The cuts to the tube weren't very complicated. One angled cut to fit flush to the firewall:





And then after cutting it to length, I added a birdsmouth shape on the other end:





As you can see, I've significantly refined my process since my first attempt on the x bar a few months ago. I'm really happy with the fitment:



Then a few tack welds all the way around to keep everything secure:



All my welds will need a once-over before I call them solid. However, this should be strong enough to prevent chassis flex when taking out the floor. Otherwise the weight on either end of that wide unsupported space might cause it to try and splay outward, and lining it back up to put the floor back in would be a nightmare. And while the rocker area would still provide some structure, I wanted to be extra sure.



If you're wondering why I put the tubes at an angle, it's because putting them parallel to the floor would make it very difficult to maneuver the floor panel out. Angling them gives me more space, and should still be plenty strong especially when considering the rockers will still be in space.



Once we're done building the loft I'll have a bit more space available, and I think I'm going to proceed with removing all floor attachments next (including the frame rails) so I can straighten the floor.

Until next time