Announcement

Collapse
No announcement yet.

Vorshlag Scion FR-S LSx Alpha Project

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    Re: Vorshlag Scion FR-S LSx Alpha Project

    continued from above

    CAN-BUS TRIALS AND TRIBULATIONS

    To make the LSx engines emissions compliant and to allow the OEM gauges, electric steering assist, HVAC and the rest of the factory Scion/Subaru systems work, there needs to be a "bridge" between the GM engine control module (ECM) and the Subaru/Toyota ECM. The goals were seemingly simple: send some data over the CAN-BUS network from the GM ECM to the FR-S' ECM to allow the stock gauges to work, the HVAC to function, the ABS system to operate, and the electric steering rack to actually have power assist. Every programmer on the interwebs seemed to say "oh, that's easy!" but so far we have worked with 3 different groups and have little to show for it.



    Why is this necessary? Well for a pure race car it isn't - you can use an aftermarket ECM and then just add a digital dash or a separate set of analog gauges. But for most folks, they want emissions compliance, functional OEM gauges and air con to work. The main Subaru/Toyota ECM controls most systems on the body these days - unlike in the 1980s-1990s, where several smaller computers were dedicated to different "body" functions. The main ECM send signals to various systems all over the car via the CAN-BUS network.

    Jason had reached out to several groups and individuals familiar with CAN programming two years ago looking for some help with the dual ECM / CAN integration. This was the solution needed to help "finish" this this type of swap in a street-legal, kit-able form. We have been looking for a CAN-BUS integration for our existing BMW E46 LS swap for 5+ years already, too. No, those solutions aren't 100% there yet either, and these attempts were started long before the 86 swaps began.



    Initially it was two different local programmers that answered our call. They both showed up at different times, met with us, started some work, then went radio silent. A third guy here in the USA - who was working with an Australian group Jason had talked to a earlier - actually flew in to work on the Alpha car. It was a complicated deal, where he was trying to work directly with our customer, even though we had talks with him months earlier. This programmer installed a CAN translator module (shown above) but after many hours of wiring and testing it didn't work as promised. Somehow this install wiped out the custom dyno tune loaded onto the GM ECM, so the car ran poorly afterwards.


    I've done just enough programming to know when a job is harder than someone thinks

    I also met last month with a friend of a friend who worked in the same field I did 15 years ago - controls systems, ladder logic programming, signals and wiring. He did controls on gas compressor systems (natural gas) and I worked in gas fired turbine power plant controls, but those both use similar controls systems, programming, GUIs, etc. So we knew a lot of the same systems, programs, and products. He was also a car nut, and more up-to-date on CAN systems (from Caterpillar diesel systems he works with) than I will ever be. Finally this was somebody that I could talk to and who might be able to understand what we needed to do here?

    I explained the goals of the CAN integration for the FR-S/BRZ LS swap, then he spent about 12 hours over the next weekend researching, looking at Subaru and GM CAN systems and encryption. The next week he came by and had some daunting news - this is never going to be "easy". Why? GM is known to encrypt a lot of their CAN protocols, because they don't want people doing exactly we're trying to do. Same goes for most other car manufacturers. And none of them use the same encryption. The "CAN-BUS" architecture is not universal among differing car companies, either.

    As I suspected, he said this will never be an easy "programming job" where you can hook up a $200 "CAN Sniffer" to both computers' CAN network wires, "sniff" the two different sets of CAN signals, then just make a black box to convert them into the same language. He described it more like a FAX system... where you have two computers talking over a network, but they have to talk back and forth... call, receive, handshake, then talk. When you encrypt the signals at both ends, and they are not made to talk to each other, you get a bunch of "dropped calls". So while a few things might work for a short period, long term communication between two different company's ECMs is going to be "a big challenge".



    As bad as I have been making it sound, and as hopeless as it had seemed - Eureka! This CAN integration box now exists, and it happened very recently. A German company called MRS Electronic, who Jason had first talked to in 2014, seem to have just come up with a marketable, working solution for LSx to 86 CAN integration as of 9/25/2016 (see above). They were the only group we spoke with originally who had "downplayed" the easiness of this integration early on, and yet they are the group who developed a working CAN integration solution for this chassis first.



    They put the call out for testers recently and happened to link up with our old Alpha customer's FR-S. As of last weekend that car has a working set of OEM gauges, power steering and more, as shown in a video posted to the FRS/BRZ LS Swap Facebook group. His car has a GM "E40" series ECM, made for the LS2 engines which used a 24 tooth crank reluctor (also known as "24x"). We used that ECM because the engine chosen for his swap was an older 5.7L LS1 longblock (which has a 24x crank reluctor) with an LS2 intake manifold and drive-by-wire throttle body (which was needed to clear the hood). A more common ECM to use is the GM "E38" which is made for the later LS3 engines with a 58x reluctor. This is the ECM that MRS apparently cracked first, then they worked on the E40 (shown below) for the Alpha customer's FR-S LS1.


    The Alpha build had an E40 ECM, due to the 24x LS1 long block and LS2 intake

    Now that his enigma has been cracked, we here at Vorshlag can stop hunting for this solution and concentrate on making the fabricated driveline mounts, the plumbing and cooling solutions, and leave the CAN integration to others. This is a big deal and should make 86 swaps much easier for all GM LS engines, and has accelerated our in-house FR-S V8 swap build.

    WHAT'S NEXT: STAGE 0 PRODUCTION + BETA BUILD!

    We actually began a small production run of CNC laser cut components to make motor mounts and transmission crossmembers last month. We will soon be able to help anyone quickly make a race car V8 swapped 86 chassis using our "Stage 0" parts.



    We just got word from our powder coater that these initial pieces are ready and should be available for sale a few days after I posted this. These were made off of our production fixtures with CNC laser cut parts, which were made off the Alpha car's proven LS driveline brackets.



    Since the MRS folks figured out the CAN Integration we have accelerated the timeline for the V8 swap in our shop's red FR-S. This "Beta Build" will help us develop even more solutions to make this V8 swap easier and more complete, as well as tackle the unfinished business such as our front swaybar solution.


    The Beta FR-S - A New Hope! It was fun but underpowered on the MSR 1.7 mile course

    There is an aggressive schedule laid out and we are acquiring parts now to make this build happen. Even when you are a shop like us, this is not an inexpensive swap, and that's one thing I wanted to write about here - this swap is not ever going to be a "cheap upgrade". There are too many known weak links we will need to address on our Beta Build (which should have 150+ more hp than the Alpha), as we will likely "break a few eggs" in testing. But adding significant, reliable horsepower is never cheap - not even a boosted FA20.



    What might break? Well I was over at "a major Subaru tuning shop" last weekend and talked to one of their top guys there for an hour about FR-S/BRZ power mods and upgrades. He had little good to say about the factory ECM and aftermarket tuning options, and even fewer good things to say about the viability of the FA20 engine. He noted the abundance of low quality supercharger and turbo kits made for these cars - most of which are sold at a "very low price point" and, as always, you get what you pay for. He said there are better aftermarket turbo/supercharger kits, which cost a great deal more, but even those often need some tweaking and re-work by the installation shop.

    Also, whenever you are adding boost to an engine, you aren't making it MORE reliable, you are making it markedly LESS reliable. This is why we tend to prefer LS swaps with larger displacement N.A. engines vs boosting smaller engines to increase power. Boost just adds problems...



    As you might imagine, he noted that boosting a stock FA20 with the factory 12.5:1 compression ratio was a very short term power solution - they all eventually blow up. The right solution for adding boost to the FA20 is to properly rebuild the entire engine to handle this: adding lower compression forged pistons, stronger rods, and maybe throw in some head work - then they don't blow up as easily. We started talking costs on 86 cars they had built "the right way", with real reliability and power approaching 250 whp, and the total price was quickly approaching or even exceeding the costs for an LS swap.


    The factory R180 Subaru differential above is more suited to 200hp than 500 hp... it is not overbuilt

    But then once you get into the 400-500+ hp range, other stuff will likely need to be addressed. Another thing our tuner shop manager friend mentioned was the relatively low strength of the R180 rear differential (shown above) and half shafts, and the somewhat poor function of the factory installed Torsen LSD. Once you add higher strength axles ($1000+) and an OS Giken ($1700 + install) or another good aftermarket LSD, the prices start to climb yet the strength is still limited by the relatively small 180mm diameter ring gear of the R180.



    In the past with BMW LS swaps (a car that does have some overbuilt factory parts), we've put "big power" (500-600 whp) through the medium case (188mm) and large case (210mm) factory BMW differentials, but those are both bigger and heavier than the little R180 in the 86. This extensive "abuse testing" we have done in BMW V8s included standing start events (autocross) and road course events on foot wide Hoosiers. Our Alpha E36 LS1 had 490 whp and 315mm Hoosiers on 17x11" wheels, as shown above. We beat on that car mercilessly for many years without issue on the 188mm housing, but some of our customers were able to break those. The 210mm BMW rear housing was a difficult swap back then, but it is more common now and this is one of those cases where "bigger is better".



    So... maybe it makes more sense upgrading the FR-S to a Ford 8.8" IRS housing from the 2015 Mustang? (shown above) This "224mm" ring gear is inherently much stronger, has dozens of alternate gear options, and many more limited slip differential choice - all at much lower costs. It comes in both a steel case (V8 Mustangs) and an aluminum case (Ecoboost 4, shown above), but both are very strong. We will look at the work needed to adapt this 2015-up Mustang S550 8.8" rear differential to our shop car, as we are shooting for north of 500 whp on the little red FR-S. With an LS3 or later V8 that 500hp goal is actually not very hard to accomplish.



    But we will see how far the R180 can go first. Our red FR-S will still be an air conditioned street driven car, which my wife wants to both race and daily drive. That is toughest possible goal to reach, but that's what most people want (street car + track car), so we're going to build it that way to see what fails and make fixes accordingly.

    More soon,
    Last edited by Fair!; 10-10-2016, 03:04 PM.
    Terry Fair - www.vorshlag.com
    2018 GT / S550 Dev + 2013 FR-S / 86 Dev + 2011 GT / S197 Dev + C4 Corvette Dev
    EVO X Dev + 2007 Z06 / C6 Dev + BMW E46 Dev + C5 Corvette Dev

    Comment

    Working...
    X