|
Congrats! It's pretty great. 
|
#
?
Apr 27, 2019 04:26
|
|
|
|
| # ? May 1, 2024 01:03 |
|
|
So, unrelated, got some Bilstein and Penske shock pistons, took measurements of them. The point is I want to be able to put a Penske piston that has no Bilstein equivalent, with Penske shaft and clickers, in a 36mm Bilstein strut tube. So I did some reverse and then forwards engineering and voila!    So a Penske style digressive/linear piston, that takes off the shelf shims, fits in a Bilstein tube, and Penske shaft. Next up, Bilstein head unit for Penske seals. mekilljoydammit fucked around with this message at 17:48 on May 4, 2019 |
|
#
?
May 4, 2019 17:40
|
|
|
When you are able to design and machine your own stuff, so much becomes possible. I have *got* to get a lathe and small mill. And probably a shop to put them in. 3D printer while I’m wishing. TIG, Plasma cutter... At least I have a MIG now.
|
|
#
?
May 4, 2019 18:41
|
|
|
Oh yeah, forgot I did this too to be able to mock up tire fitment.
|
|
#
?
May 5, 2019 04:28
|
|
|
 Details not done yet... like, need to get an axle nut to reverse engineer. But built enough room for the rear spindles to go from 0 to 3 degrees of camber without issue. Radial slicks really need the camber.
|
|
#
?
May 5, 2019 18:36
|
|
|
Just sort of thinking out loud here. So, plan is a 1st gen RX7 for SCCA Eprod under the "13B with alternate 5 speed synchro box" spec line... minimum weight is 2127lbs with driver. There's 5.38:1 ring and pinions imported now, so I'm thinking of using a GT86/BRZ trans with the aftermarket 0.86 6th gear and 1st gear lathed off and blocked off. Been doing a lot of research on data systems, downloading the relevant software, etc. I think an AIM logger will do everything I want but I don't like their dash layouts as well. I'm pretty sure I can transmit on another canbus to an AEM CD5 though, who have an excellent dash design setup but don't have ability to log shock pots at enough speed. Win/win. ... except nothing short of Motec or Pi will do math channels I want and I'm not made of Motec money. So I'm thinking... a little raspi with canbus stuff sniffing data could then do math of arbitrarily high sophistican, then transmit some calculated values back over canbus for stuff to use. Besides, learning canbus has job relevance.
|
|
#
?
Jun 17, 2019 20:53
|
|
|
Canbus is cool as poo poo. I can talk ISO 11898 with some degree of competence though the higher level protocols that get layered on it are fuzzy to me. If you need to bounce ideas off someone about the hardware protocol itself though, I'll talk about message boxes, AutoRTR, message ID priority sorting, and message formatting until my face turns blue.
|
|
#
?
Aug 10, 2019 03:15
|
|
|
I need to get some hardware designed and code written to start playing with. The messaging format for all of the peripherals is already set down and I see no real reason to reinvent the wheel; published protocol for the AEM stuff is at https://www.aemelectronics.com/files/pdf/AEMNet%20170116_Public.pdf and there's DBC files for all the normal modules. What I want to do is make, say, a nice convenient laptop based GUI to talk to the raspi thinkybox... feed it DBC files of all the sensors on the networks so it knows what comes in, set up, say, an octave file to do the math, and then have it build a DBC file for other stuff to see the calculated values it's sending out to the network. I've been doing a lot of field testing support for work lately which means not much time for fun stuff. Between that and the adorable almost-4-month-old, progress is kind of slow. May be moving to a new job soon though which will keep me at home and cut the commute by 90%. May also be looking at machine tools... starting off with a toolroom-size CNC lathe.
|
|
#
?
Aug 10, 2019 17:47
|
|
|
Update coming soon - I'm on a new revision of the tube chassis, what I think may be the last version before I start buying tube. I also am trading emails with Hartman Fiberglass who has the original Mazda race bodywork molds - going to see what they want for one, as it's simpler than making a new plug to pull molds from. In the mean time I'm feeling really tempted by this: https://www.ebay.com/itm/223648519403 Vintage racing would be a hoot.
|
|
#
?
Sep 3, 2019 01:28
|
|
|
 So, going back to a GT-2 design which means wider wheels and stuff. Dark grey are safety elements, light grey are frame - which means wall thickness differences. There's some similarities to the previous couple designs but other differences here and there - weight is actually down from the previous GT-2 version despite more braces that found their way in on the GT-3 version and there's other subtle differences. 275 pounds so far. Going to start iterating in FEA to see what tubes are important and what aren't - it's getting close though.
|
|
#
?
Sep 4, 2019 17:26
|
|
|
What's your plan for model loads for your FEA, and where? Testing for rigidity or crash worthiness, or both? Do you have any criteria/numbers you're trying to hit or avoid? Just curious how you're going about it. Dunno what your background is as an engineer but how you create the mesh has a big impact on accuracy. Not sure how much has changed since I used to do this 10 years ago (with already antiquated software back then), though.
|
|
#
?
Sep 4, 2019 17:52
|
|
|
So, the goal that I'm looking for is torsional stiffness per pound - how I've done this with the last few versions of the chassis is by fixing either the front or rear bulkhead, then feeding a couple thousand pounds twisting motion into where the shocks will be mounting (vertical up one side, vertical down the other) and figuring out things in ft-lbs of torque per degree deflection. In practice... I'm using baby's first FEA package, the simulation package built into Solidworks which doesn't give me much input on meshing. And I'm not incorporating any paneling and a bunch of other stuff... so at the end of the day I'm basically only looking to sanity check my educated guess on triangulation and tube placement and iterate a few changes, looking at things relatively. Probably all of the actual stiffness numbers will go up if I do things like, say, have decent tabs where the front and rear firewall are riveted in, use solid rivets and so on... not actually semi-monocoque but kinda.  But if I can demonstrate to my satisfaction in FEA that there's tubes that aren't doing anything useful I can cut them out at this stage of things.
|
|
#
?
Sep 4, 2019 18:16
|
|
|
 Because people have asked here and there - there's a screenshot with a camera placed at the human dummy's eye location and about a human FOV. So yeah, the windshield diagonals are visible but not too bad IMO and they add *huge* amounts of stiffness for the weight.
|
|
#
?
Sep 14, 2019 17:52
|
|
|
So started a new job a few weeks ago - smaller company, wider variety of testing, no travel and closer to home. By which I mean a 5 minute commute instead of 50. Also they let me do stuff on the shop equipment after hours. Like so: https://www.youtube.com/watch?v=2I15LI93BXc
|
|
#
?
Oct 18, 2019 22:52
|
|
|
Happy you got a cooler gig. That looks like fun!
|
|
#
?
Oct 21, 2019 01:00
|
|
|
 Final version, or final-enough-for-machining. May cut a little more mass out of the spindle but I'm starting with the hub anyway. Illustrated there at 2.5 degrees of camber, probably about the limit without going to an expensive-to-make CV joint style ball setup anyway. Time to start making chips out of the 40 pound piece of billet I have sitting in my garage!
|
|
#
?
Nov 17, 2019 20:37
|
|
|
That looks absolutely incredible. 2.5 degrees of camber should be plenty, right? Do you have plans of 3d printing all the parts first?
|
|
#
?
Nov 18, 2019 06:00
|
|
|
Huh, clever, took me a minute to see how that all lined up. Is that going to be a splined ball drive?
|
|
#
?
Nov 18, 2019 09:56
|
|
|
The ball is just for mockup to center the dummy axle. There's two outer ends that will work for this, ball drive (basically a CV joint) and crown splines. Short version is that crown splines are a lot cheaper and use standard drive flanges (the red part is a circle track Wide 5 drive flange with 4 holes enlarged) and should be enough camber. And you can just buy crown splined axles from circle track suppliers. My theory is that with an aluminum drive flange and steel axles, the cheap drive flange will be the wear item. I'm just going to start making it; it's a shrunken version of a proven design. Also on the slate, a new front cover. See, the 1st gen puts the motor mounts on the front cover but uses a distributor... the 3rd gen has a crank angle setup but no motor mounts on the cover. So I'm going to try joining two front covers and using them as a casting pattern. Will see how that goes.
|
|
#
?
Nov 18, 2019 15:19
|
|
|
More and more stuff to post when I have time to do an effortpost... this weekend's projects, print out a bunch of header mockup blocks. See, there's this company called IC Engine Works that came up with this really cool idea - these little blocks in various mandrel bend centerline radii and diameters. You snap them together, rotate them around and figure out header routing. Well, someone on thingiverse came up with their own version to 3d print! Except he did it in 22.5 degree increments instead of equal 1" centerline length - I'd prefer it to be where you know X number of blocks means X primary length in inches. So I redesigned it. 
|
|
#
?
Dec 14, 2019 17:28
|
|
|
That's a really drat cool idea, not gonna lie. Seems even better than ultimateforce's pool noodle idea from years ago.
|
|
#
?
Dec 14, 2019 18:52
|
|
|
It's not my idea per se - I just tweaked stuff to my preferences. The original commercial version is https://www.icengineworks.com/ but they're like, $7 or something per individual piece. https://www.thingiverse.com/thing:1831000 (and all the other sizes he has) is a guy who remixed it as 3d printable, but I had my own preferences, so, remix. I think my cost to print them is like, $0.60 or something. Since I'm on a roll today...
|
|
#
?
Dec 14, 2019 19:01
|
|
|
Ahh yes I've seen those click together things when I was visiting a race shop. Neat idea.
|
|
#
?
Dec 15, 2019 00:02
|
|
Have you seen my apex seals? I seem to have lost them.
|
 Now to just use up a bunch of plastic.
|
|
#
?
Dec 15, 2019 00:27
|
|
|
Neat!
|
|
#
?
Dec 17, 2019 00:22
|
|
|
Poking at bearing catalogs, as one does, I may have just found a holy grail of mine - a cheap way to get Miata brakes and stuff onto the FB. Will report more after playing mix and match with parts tomorrow night.
|
|
#
?
Dec 20, 2019 00:51
|
|
|
What's special about Miata brakes? Just looking for something low weight? I know when I had an NA a bunch of track folks would upgrade to wilwoods due to the stock ones not being great in regards to heat or when you had more than stock power.
|
|
#
?
Dec 20, 2019 12:03
|
|
|
The key is that there's off the shelf stuff for the Miata. You have your choice of a couple different rotor sizes, a bunch of big brake kits, stock calipers you can get anywhere, pads that are available in any compound you want, so on and so forth. And since this would be converting to 4x100, say the same about wheels. There's no good cheap options for the FB RX-7. So think of it not in terms of "oh the NA Miata has better brakes than the FB RX-7" although it might, but "hey, now I can bolt any brake setup that bolts to a Miata to this car"
|
|
#
?
Dec 20, 2019 14:00
|
|
|
mekilljoydammit posted:Poking at bearing catalogs, as one does, I may have just found a holy grail of mine - a cheap way to get Miata brakes and stuff onto the FB. I am very interested in this, though my SA has '84 struts, so I can at least use FC brakes with just a machined bearing spacer and a caliper mount. Unless I just swap in the whole FC subframe, of course, which is a distinct possibility.
|
|
#
?
Dec 20, 2019 20:45
|
|
|
Forgot some things about the Miata hubs... so "not easily" is the consensus there. Have to think more about if it's worth messing with. In other news, some more work on this - found a commercially available FB body, cut a lot of unecessary parts off of it, reduced the polycount and scaled it. Good to see that it pretty closely agrees with my not-very-good attempt at bodywork where the stock contours are. Also raises questions about if I should just CNC rout a whole plug to pull molds from or use one of my spare shells for the stock bodywork - in the grand scheme of things I'm not sure how much it matters either way.  Finishing up a first draft of suspension geometry next, then maybe some messing around with paneling and ducts and stuff.
|
|
#
?
Dec 24, 2019 19:57
|
|
|
So, been busy at work, which is good but haven't gotten as much done outside of it as I'd like. Lessee... thing I'm working on lately is an extremely bad CMM machine. OK, extremely bad by CMM standards. Basically converting an old 3d printer to a different control board intended for plain cnc stuff and setting it up with a probe. Then I wander off for a while as it accurately (ish) digitizes things. I'm not the right sort of patient to figure out photogrammetry and I'm not sure it's accurate enough anyway. Short term intent is to output to 3d printable files for casting, longer term I'm looking into the cheap Chinese mills converted to cnc for porting and the like. I'm really tired of grinding rotary intake ports.
|
|
#
?
Jan 22, 2020 19:39
|
|
|
That sounds cool.
|
|
#
?
Jan 23, 2020 00:10
|
|
|
So the origin of all of this... I've wanted, for a while, a 13B front cover with the 1st gen mounts and with the FD crank angle sensor. I kinda figure I might not be the only one who wants one but nevertheless, I feel like doing a casting. And the more I look at it the less confident I am that I can use hacked together stock parts for a pattern. If it's digital, it's much easier to resize, and then modify for dry sump stuff and the like. I'm all of like, $40 into the CNC digitizer project so far. I figure I'll test it with some smaller stuff and then expand it later on if it works. If it doesn't welp, $40.
|
|
#
?
Jan 23, 2020 00:20
|
|
|
That's a pretty good way to cheaply digitize things. I'll have to mention it to one of the guys at work that's been struggling with his super budget photogrammetry attempts.
|
|
#
?
Jan 23, 2020 12:36
|
|
|
It's unproven and to do digitizing quickly I'm probably going to have to write a python script - I'll document what weird hacks I do in the meantime. The duh-simple way you can do in a normal "write G-Code program" is go to a Z higher than everything, come down until it touches, then go up to where it's safe, move a bit in X or Y, and repeat. But all the Z moves makes it slow. The faster version is semi-closed loop and only goes up a bit on Z after every probe, but also has to detect if it runs into anything on the X or Y moves and I'm not sure if any of the cheap G-Code interpreters handle conditional statements.
|
|
#
?
Jan 23, 2020 15:02
|
|
|
Any reason to sticking with the FD trigger? I guess if they work they work.
|
|
#
?
Jan 23, 2020 15:30
|
|
|
Aeka 2.0 posted:Any reason to sticking with the FD trigger? I guess if they work they work. I figure if I ever sell any whatsoever I want to be able to direct people to a trigger solution that's off the shelf. RX-8 cover puts the sensor in the area where the FB motor mount goes, while saying "OK the timing is an FD now" gives already proven solutions where I don't have to make more parts.
|
|
#
?
Jan 23, 2020 15:50
|
|
|
Fair enough. I think im going to scope my FD triggers when im done to see how clean they are at 8k RPM, haha. I have my own engine harness as well so the wires are shielded, drained, and twisted.
|
|
#
?
Jan 23, 2020 17:09
|
|
|
I'm probably just going with the FFE setup so I don't have to reinvent that wheel too. Really I'm starting to want to make some front covers that break compatibility though - like, start over with the legacy stuff (distributor gear for example) removed and shrink the package down, have mount lugs for an external dry sump pump stuff like that. Maybe go to Trans Am style "here's lugs to bolt to a chassis plate".
|
|
#
?
Jan 23, 2020 17:41
|
|
|
|
| # ? May 1, 2024 01:03 |
|
|
When I needed to CMM something I ended up putting it on an old shipping crate in one of the laser cutters about to be shipped at work and used the machine vision system to locate the center of each feature. Got me within about a tenth of a thou on each one.
|
|
#
?
Jan 23, 2020 18:22
|
|