• Hello Guest, welcome to the initial stages of our new platform!
    You can find some additional information about where we are in the process of migrating the board and setting up our new software here

    Thank you for being a part of our community!

dirty, old and rusty: a tale of a $600 meatball

The poly engine/trans mounts are definitely softening up. Still pretty rough vibration-wise at idle but much less so than Friday night. Probably put ~350 miles on them so far. One thing I didn't think about was how the shifter now barely moves at idle or throttle transitions. It is almost eerie how steady it is, I have always been used to that sucker jiggling in the background.

On the LH2.2 tuning front I have figured out the two coolant temperature thresholds for enabling lambda feedback, so I can set them too high and correctly tune the main 16x16 map in open loop. I am going for slightly lean (1.05 lambda) in the cruise areas transitioning to (0.88 lambda) to the top ~2 load rows. I have also found the load based open loop threshold but that seems to be disabled stock for the NA bins as it mainly uses the WOT switch for that. Still need to figure out the axis units. This is the change I made to the map so far, the difference from stock (times 0.88 to scale injectors, posted above.)

I pulled a lot of fuel from the middle/low area, where the car cruises at 50-70 MPH in OD. And got rid of that super weird rich spot on the bottom right.

I've also identified the fuel cut rev limiter, but even maxing it out at 255, the limiter kicks in at about 6200 which is lame. I ended up nop-ing the code that sets the "limiter active" bit so it no longer has a limiter on the fuel side. The EZK rev limiter is set to ~6900. I can't seem to embed videos anymore with the new board software, but here is a link to a video of what the EZK only rev limiter sounds like.

I ended up actually fixing the EZK idle problem I mentioned a couple weeks ago on this page. It turns out the problem wasn't related to the idle signal going low, but it was actually during decel fuel cut, LH2.2 fully cuts the load pulse to the EZK, and the EZK would then go into limp mode timing. I fixed that, too.
Last edited:
I was messing around with this when disaster struck

I am pretty sure I leaned on it while messing with the AMM/intake pipe and broke the radiator inlet. About 25 miles from home. And I usually make sure to do my experimentation really close to home but this time idk why I tried so far away, and on a cold night in the sleet.

There was a CVS nearby so I parked the car and browsed to see if they had any quick epoxies I could try before calling a tow. I found these two and ended up trying the "plastic bonder". It was ~33 degrees out so i tried to mix it fast and apply it before the rad cooled down too much more. I also scratched the heck out of the plastic tank and edge of the barb with a wood file I had in my trunk for some reason. I was able to attach the hose and start driving maybe 3 hours after I applied the epoxy. I used electrical tape to try to reduce the amount of stress on the joint. I also left the radiator pressure cap off.

I got home successfully with the hackery at about midnight and immediately went to bed. This is what It looked like the next day. I then applied a second layer of the plastic bonder. I again tried to reduce the stress on the joint with zip ties. I decided to take the car to work (instead of my perfectly fine truck) after these pictures.

At work, I used the last of the plastic bonder and did a 3rd layer. After it dried it feels stronger than even the lower radiator neck.

I can't wait until this strands me again. Though now I have the original radiator and a bunch of coolant in the trunk ready to rumble.
Of course the radiator started to leak soon after. But otherwise, the globbed on epoxy held the neck on for me enough to keep driving it.

I ordered Northern 209620 radiator as a few people have seemed to have success with it in 240 chassis here. The cats were amped for another large box to destroy.

She fits pretty gud. It needs an ever so slight clearance on the bottom flange so it doesnt rub on the lower rad support. Otherwise the stock lower bumpers fit it nice.

It fits pretty well to seal up the entirety of the useful front space available. I need to find some foam tape or something to really seal the rad to it's supports.

Cheapo 10 psi cap. I hope this is low enough to keep my original heater core from kersploding.

I made some upper brackets with bent aluminum bar.

I of course assumed that the inlet/outlets of the radiator would match that of B230 but that was not the case. The B230 coolant inlet and outlet barbs are 1.25" while the upper radiator inlet is 1.5" and the lower outlet is 1.75". I got some of the universal tacky radiator hose with silicone adapters to get it on the road. I hope to find some molded rubber hoses that will fit well enough in the future. I used one of the new coolant jugs as a puke tank until I find something more suitable. No fan yet, but I will figure something out before it gets too warm out. No problems so far tho :e-shrug:

pictureless word salad text book inbound. abandon all hope ye who enter here

aight so i've run into some barriers that keep me from wanting to continue mucking with LH2.2. The first being there is pretty much zero unused program memory for writing modifications. The second being the conversion of the AMM voltage signal to cylinder load is done completely in hardware, so it cannot be modified. (at least with my current skill-set.) I found this out as I put a ton of effort into trying to use a 3 inch -088 HFM5 AMM from a turbo P2, and I could never manage to get the ECU to interpret the air mass value correctly. Either I could get it to work at idle, but once load or RPM changed drastically, it would not be able to correctly calculate the change in air mass in the ECU, or something. It would access cells in the big fuel map very erratically.

I have some info to share about the load calculation I have figured out if anyone is curious: The AMM voltage enters the LH box into a basic voltage divider/op amp circuit to knock it down from 1.25-4.5v to 1-3v ish. This enters the "RCA 91535" DIP16 chip which I cannot find ANY reliable information for, but I presume to be a configurable voltage-to-frequency converter IC. The output of this chip ranges from 500hz to 80khz and seems to have the inverse volt-to-frequency curve to the AMM's airmass-to-voltage curve. These put together effectively linearizes the voltage from the AMM into a pulse count directly proportional to a constant air mass unit passing through the sensor. This is fed into an input on an "8253" timer chip that decrements a counter by 1 for each pulse. Then, for each RPM signal strobe from the EZK, the ECU takes the current counter value, and subtracts the previous counter value saved at the previous RPM strobe. The result is an airmass value ingested by the engine over the last RPM period. AKA cylinder load. On this 544 LH2.2 ecu, it ranges from ~56 to ~360 units, and this range happens to be exactly the load (Y) axis of the large 16x16 12-bit map I was posting and messing with earlier in the thread. The actual data cell to use in the 16x16 map what the ECU uses as an input to the the injection load calculation is chosen by two 1x15 maps that divide the rows/columns into "buckets" to allow for adjusting the resolution of certain parts of the map. If you try to alter any cylinder load parameters, even a little, you too need to alter the "buckets" maps to align perfectly with the new numbers or the ecu shits itself.

tl;dr the LH2.2 ecu integrates and digitizes the airmass from the AMM in hardware between RPM pulses.


Back to the shitbox, I ended up swapping it back to LH2.4 using the 946 box I was using a year and a half ago. It went almost without a hitch, until I noticed the idle control wasn't working right. I had forgotten to change back the LH2.2 IAC to LH2.4 IAC. I fixed that and went for a drive and after a little bit, the car started badly misfiring, mostly at low load/cruise and a little at Idle. I'll be honest, this stumped me for a week. If i unplugged the IAC, the misfiring would stop, so I thought that narrowed it down. I tried my other 946 ecu and that didn't help. I then tried my 561 ecu and that did actually solve(mask) the problem. but why? I eventually found the actual cause of the misfiring was sharing the power source for my ignition coils with the IAC. Once I moved my coils back to being powered from the ignition switch like a stock 240, the issue was gone. But why was this setup never a problem on my LH3.1 572 ecu or LH2.2 544? I looked closer at the circuitry and think I found out why. The 5xx ECUs have some sort of coil flyback damping circuit connected to the IAC driver pin that the 9xx ecus do not. The 9xx actually have the footprint/solder pads for it, but is not populated on any of the ones I have seen. I ended up stealing and re-soldering in the components from my sacrificial broken 556 ecu into one of my 946 ecus, temporarily shared the IAC and ignition coils power source like before, and the issue was not reproducible on the flyback modded 946. I wish I had a good scope to see what it is doing in detail. I also wonder if this circuit change is related to the "radio supression" relay used in 7/9 cars.

Oh and I am still using the LH3.1 TPS with the comparator converter board I made. That aspect of this efi fuckery has been pretty bulletproof.


On a last note my M46 overdrive unit has decided to become progressively worse in the past few weeks. At first it would drop out during hard deceleration on the highway. Then it progressed to dropping out at any speed below 45mph, now it will not work at all below 60mph. I am going to try a hail mary fluid change this week and see if there is anything to see. The fluid in there isn't old, maybe 8 months at worst. I also noticed just this morning that it works much better when cold. I guess I have to start making progress on the T5.

I can't wait until this strands me again. Though now I have the original radiator and a bunch of coolant in the trunk ready to rumble.

this sequence is one of the most potent old-era turbobricks things i've seen in a decade. thank you for doing this bullshit so that i could have a happy little taste of 2009. i genuinely appreciate it and mean this earnestly.
One of my favorite cartoons from the early 80s:

I couldn't find anything on the RCA 91535 chip either. It crosses to a Harris part too. The Porsche LH 2.3 schematics call it a "RCA Timer", but it's odd for a timer or v-f converter chip because it only has a couple analog pins and lots of power/ground tie-offs (if the LH2.3 schematics are to be believed).

When the LH2.4 IACs are disconnected, or not driven at all, they spring to a partially open "limp home" position that allows some idle air to go through.

The IAC electrical filtering results are strange, I wouldn't expect it to cause problems in other circuits. The 7/9 RSR relays are simply a common 4-pin relay with bullet pins. There's no extra filtering components inside them. The radio surpression is simply keeping the power wires in the engine compartment, away from the dashboard radio and wires.
whoops i have a terrible habit of writing responses then not posting them

I couldn't find anything on the RCA 91535 chip either. It crosses to a Harris part too. The Porsche LH 2.3 schematics call it a "RCA Timer", but it's odd for a timer or v-f converter chip because it only has a couple analog pins and lots of power/ground tie-offs (if the LH2.3 schematics are to be believed).
ooop looking at the LH2.3 schematic is a good idea, I forgot about that. I found some info regarding the RCA pinout here in the motoronic2.3.2 data repo: https://github.com/prj/m232/tree/master/Factory data . I'm thinking the chip input voltage just lowers a clock divider of the input clock as the frequency steps seemed to get coarser as I upped the input volt.

When the LH2.4 IACs are disconnected, or not driven at all, they spring to a partially open "limp home" position that allows some idle air to go through.

The IAC electrical filtering results are strange, I wouldn't expect it to cause problems in other circuits. The 7/9 RSR relays are simply a common 4-pin relay with bullet pins. There's no extra filtering components inside them. The radio surpression is simply keeping the power wires in the engine compartment, away from the dashboard radio and wires.
I looked at IAC stuff again and yeah idk what I was thinking, lol. I had thought that little circuit was masking the issue I created in my setup what I think was the ignition coils on the same power supply line as the ECU and IAC causing noise/interference and the misfire. The IAC would also behave erratically, and the little circuit I moved over stopped it until I fixed the wiring.:e-shrug:

Movie of the IAC erratic behavior coinciding with misfire. (I cant embed :-()

This is the difference between 5xx and 9xx IAC driver circuit. I can only assume the flyback spike from the LS coils was going through the IAC into it, or it is doing something else I don't understand.

I was also originally thinking the radio suppression relay separated the noisy things like injector and IAC from the ECU power line. But after spending(wasting) time looking at the available wiring diagrams, it looks like what/how it was used was not consistent at all. I see on LH2.2 7 series, injector and IAC were powered by RSR and that was triggered by the powered system relay. Then on early 7/9 series LH2.4, only the injectors were supplied by the RSR and that was triggered the same way as LH2.2. Then on later 7/9 series, the RSR supplied power to the entire fuel injection system aside from the pumps and O2, and it was triggered directly by the ECU without an intermediate system relay. idk, shit's wack I wish I didn't look into it lol.
Because of the OD issues I tried a juice swap on the M46. IIRC what I put in less than a year ago was 10w40. The last time the old fluid looked about the same as this, with the embedded shininess. It seems like it lost a lot of viscosity though. Nothing settles out so IDK what to think about it. The magnet has a normal (for this particular clapped transmission?) amount of fuzz. My notes say this is 16k miles. :oops:

This is the concoction that went in. The volvo manual trans oil here is just GL4. I laid in a pile of snow which melted under me while I pumped it in. :-(
Last edited:
Back to T5 work. For a refresher, while I was pressing on the 5th driven gear onto the output shaft of the T5 a couple months ago, my janky setup of stuff I was using to support the box fell out of the press and I damaged the "cobra style" inner pocket bearing of the in/output shaft. I have to pull the entire thing back apart to get this bearing replaced.

I ended up having to cut the race out of the output bearing as the 0.80 gear set driven gear is actually too large to pass through the race to remove the shaft the "correct" way. That is why there is a dremel in the picture. Here is the shaft being liberated.

Next comes the cluster gear. I was amazed how tight they made the openings to get these parts in and out of the box. About 1mm of interference of the race in the case and the bearing on the other side prevented this guy from coming out.

Here I am removing the bearing on the other side. It also was preventing me from getting the angle I needed to pull the thing out of the case.

Now that the case is empty, besides the reverse gear crap, I was pulling the "old" 5th gear set as I found a different set i'd like to use. I need to make or get some other lower support for my press as the narrowness of the stock one was really a hindrance here. That 5th gear was pressed on tight.

Here is a picture of the damaged piece. I have already removed the cage as it was mangled.

New pocket bearing pressed on. I also need to get a set of press... dies? Using old sockets are the reason i'm back in here.

The synchros are neat.

New bearings on the cluster gear and case just in case. I used the peelable style shims and used the steel cluster gear retainer reinforcement plate thing. I put a bit more preload on the cluster gear than the service manual wants, it seems like that is common advice from T5 nerds on the internet.
the turd sprung a pretty substantial oil leak suddenly. I took it to the diy power washer and cleaned all the grit off the engine then drove to work and found these possible sources. I think it mostly came from the pressure switch. yeah i know i'm missing a stud :uh:

tho really I was surprised my diy harness took on a full on power washing with grace.
I'd like to imagine this is what dinner at the Samodai's is like
Close. Between Cam and I the conversations are almost exclusively engine oriented but cover a very broad range of topics, such as how smart Chrysler engineers were, how I'd build a 372 SBC if I had another 1969 Camaro, how crappy the Redblock sounds, etc. For those of you interested, this was last night's discussion:


This is from the horse's mouth as to why Dynojets read higher than other chassis dynos. I would have used the info to call out Yamaha, but to each their own. Also, I probably never would have become commercially successful.
The new 0.80 gearset i ordered actually fits on the shaft without galling. 5th driven gear is at home with the snap ring in place.

There is something very wrong in this image which required me to re remove the 5th gear and the shafts to fix. Anyone see what?

5th gear fork and slider installed.

The driven 5th gear is a bit larger than the stock one and prevented the tailshaft from seating fully. A little bit of clearance with a die grinder and it was good to go.

Top cover installed

Now time to shim the input with the new deeworks retainer. The original retainer had a .022 shim installed which when installed in the new one was clearly too much. I played with the shims I had and found 0.015 shim was right at the threshold of zero end play. I installed a 0.020 shim for a little bit of preload. Input shaft feels really solid thrust wise and has about 3/32 inch of side to side play at the tip. Featuring the messy workbench

The bolt head flanges at the front of the cover prevented the adapter from fully seating flush against the trans. A little bit of hand filing and that was taken care of.

I installed the yoke and flange to the 47" coleman 3" aluminum driveshaft and now it is all ready to be bolted in :cool:
Last edited:
you got it lol. i originally noticed because the whole geartrain was notchy everywhere but 1st. and reverse was another neutral. so the entire shaft came back out to put the slider on right. i'm not sure when i had installed it backwards because I didn't pull the main shaft that far apart recently.

the 1-2 slider was not really centered also, after finding it's reverse teeth being interference with the reverse idler in the case. (which caused the nochy-ness) so it was just slightly covering the 1st blocker, i hope :lol:

it all runs smooth through all the gears on the bench so I hope i put it back together right after it was apart for almost 1.5 years :oops: