Given that Eric made 800+wheel from a 6266gen1 I'd say water injection pre turbo works very well. Although I disabled mine for the time being.
I still can't get over 800+ from a old school 62mm. Best graphs ever.
1996 332IS
Built 3.2
CES/Steed TS Precision 6466, spraying a "$π!℅" load of meth.
Technique Tuning 80# tune.
1/4 mile 10.84 @ 136.72
Your 1 and only stop for all your BMW performance needs
WWW.CESMOTORSPORT.COM
The sensor is pre-injection.
What I think is happening is that the sensor body is being heat soaked, then as airflow increases, it pulls the "bulb" temp down.
The good news is that the new sensors I'm using are fast enough to see this. Those Bosch ones are not.
Rather than focus on isolating the sensor, I'm going to work on heat rejection of the entire pipe.
Yes, I'm working on the dyno thing, but my wallet keeps getting hit with other projects. This month is 200 ft of double wide driveway....a built 210 diff...and a new turbo. That's a Dynocom easy right there....
Good to hear your exhaust back pressure issue is resolved.
I'm running 2 different stickers on the outside of my car while getting groceries: "Fire", and "Off". LOL
I've considered it, but I'm concerned about trapping heat on the EGT sensor leads, and also letting the WG cool.
For now I've got all these other heat products to cover the surrounding area.
This was the idea with the GTX3076R....I completely neglected the turbine flow though. I've run out of turbine flow well before I had to chance to run the compressor off the map.
There is a new method of getting the turbo to speed up where a DBW throttle is used to bypass air between the compressor outlet and inlet. It functions like a conventional bypass or blow-off valve, but you can control the flow very carefully. You can then build a turbine speed model based on a launch condition, where you're not having to burn a TON of fuel in the exhaust by retarding timing....the turbine speed will increase without proportional resistance from the compressor.....
I bought a GTX4088R
The easiest way to get comparable data is with "hot lapped" tests where everything is up to temp and saturated. Its not the ideal case for the best performance numbers, but it does show how things work when they are actually working and not under some ideal case.
Even in an engine test cell where we can control coolant temp, and intercooler outlet temp with a state machine and ambient pressure and temperature its still incredibly difficult to get "apples to apples" comparison. For emission tests that require a cold start we let the engines sit for about a full day to let all fluids, block material, ect reach a "known" start point. Its so time consuming that we only do it when we have to, otherwise you just do back to back demo tests or pause for a certain time between tests to somewhat simulate a cold start.
You just roll with what you can get because in reality you can get most of what you need to know from repeating the tests with different settings since you are looking more for trends than absolute numbers.
So...here's a ball buster: Do all of the sensors read the same temp at the "known" start point? All of my sensors have 1% +/- standard of deviation at BEST. And the real question: Is the standard of deviation at the reference temp (Usually 25 deg C), or across the usable range of the sensor curve. I've got a LOT of expensive sensors, and none of them read the same value at ambient temp/pressure. I realize most people have 1 air temp, or 1 oil temp sensor for example, so they don't see the deviation across say 5 sensors in series.......
What is your deviation between the sensors?
What difference is there between the input channels (pull-up, impedance)? Have you put a fixed resistance across the input to see if the issue is with the input channels or with the wiring/sensors?
Are all sensors grounded to the same sensor ground in the harness with the same length cable? Have you tried swapping sensors to see if the error follows?
Verified impedance from sensor to ECU, what's the variance there? How does each sensor measure out @ the ECU plug?
What does Motec say?
Look at the beginning of one of your last logs where you show a full pull, shift, then another pull. The temp drastically rises in that short time period of the shift. While I agree the mass flow will drop, so you will get a larger temp rise in the tube, I'm not sure you're going to see that in the area of tenths of a second.
Maybe you're getting some "splash back" on your pre-turbo sensor? You just get a big temp drop as soon as WI starts going, too much for me to think it's down to just increased mass flow.
I would also put some reflective fiberglass shielding on the intake tube, but the heat shield shouldn't really hurt anything as long as you get the electrical outside of it. It doesn't need to be air tight or anything, so a few holes in it here or there are no problem. You just want to block most of the radiant heat transfer coming from the manifold, which can be really significant at the temps we're talking about.
This is normal. Even $$$ Omega scientific grade Type K thermocouples will be +/- 1-2 deg C at ambient. IIRC, Thermocouples are typically rated per the measurement accuracy (with a looser tolerance up to say 400 C for a type K, and then tighter from there). Most automotive grade sensors I've seen have been rated as a percent of full scale.
Ultimately, sensor accuracy is just something that happens when testing, and you'll never eliminate it. If you're trying to measure something that's in the realm of "good" sensor accuracy, you need a lot of test runs that you average to try to reduce your measurement error. But at the end of the day, if +/- 1-2 C is killing your ability to measure something, it's probably not that important IMO.
Most of the sensors I use are the OEM sensors because the goal of testing is to verify a bad part can be detected with the OEM equipment.
The engines all kitted with additional thermocouples and pressure sensors have variation like you are talking about. We do baseline runs to verify all the sensors are working the same as in the past and haven't drifted. Its just a fact of life you can't look at all the sensors as having exact values, you just know cyl6 reads hotter/cooler by a certain amount.
i don't want the added complexity of having to fill a co2 bottle. Same reason I want to avoid nitrous if possible. I guess it all depends how hard I find it to get on the converter but the way I imagine using it would be to have it enabled when the transbrake is engaged and when manifold pressure reaches 3 or 4 psi. Then it would simply dump the bov reference for 1 second or something. I would just test it increasing the time interval till it gave me the fastest time to full boost.
There are a few other variables.
Temp sensors are wired directly to a thermocouple input on the ECU or E888, and the others get individual 1K ohm pull-up resistors from a 5v reference source. The resistors are 1% spec, but the ones I chose out of the bin are within a couple of ohms of 1K. The wiring harness reads 0.001 ohms or less through all the connections. The thermocouples themselves vary more than the combined values of the wiring harness and resistor that's used for the 5v pull-up. All of the identical type temp sensors read different when plugged into the same connector on the harness.
So I've got 6 different types of inputs:
M150 Thermocouple input
E888 Thermocouple input
M150 Analog input with +5V pull up through 1k ohm resistor
E888 Analog input with +5V pull up through 1k ohm resistor
E888 K type thermocouple
TC8 K type thermocouple
Individual Temp inputs:
Pre turbo air temp
Post turbo air temp
Post intercooler air temp
Pre Throttle body air temp
Inlet manifold temp
Radiator Outlet Temp
Cylinder head coolant temp
Pre-oil cooler oil temp
Post-oil cooler oil temp
Fuel Temp (at FPR)
Fuel Temp (at fuel tank)
Ambient engine compartment temp
Cylinder 1 EGT
Cylinder 2 EGT
Cylinder 3 EGT
Cylinder 4 EGT
Cylinder 5 EGT
Cylinder 6 EGT
Electronics compartment temp (thermocouple)
None of any of the above read the same value, with some connecting directly to a sensor input source. (no wiring harness)
If you look at the derived voltage value, the sensors usually fall within a +/- 1% of each other. However that can mean that 2 sensors are up to 2% different on voltage, and that can translate to a much greater number after conversion to temperature.
I've spent probably a hundred hours trying to chase what I thought was a problem, and then eventually creating custom calibration curves with various methods. Now I'm back to using the default calibration curves, and just accepting the deviation as normal. Where I thought I was going to quantify 1 or 2 degree changes, I'm now looking at trends.
At the end of the day, this isn't a production car calibration, so I only need to take into account the actual sensors I'm using. For oil and coolant, even a 5 degree differences between sensors really doesn't matter. We have that kind of deviation just in vehicle velocity. The fuel temp is about a density calculation, and more importantly, a cooling coefficient of the intake charge. You'd think it needs to be dead on, but this coefficient is tuned in the M1 software, so even a deviation in value between installed sensors can be overcome. The manifold air temp is probably a more critical function, but again, it comes back to what the knock traces are telling me as to how much load I put on the engine with ignition timing. EGT varies a HUGE amount based on mass flow and load. It's also relatively slow to respond compared to an air temp sensor with it's open "bulb" element, but I find with a lot of RPM the Thermocouples aren't that bad with response. Still, I've got a pretty easy 50-100 degrees spread across "stock" sensors. I'd need more dyno time to understand what that really means.
I get the same type of pre-turbo temps with both WI on and off. The temp sensor is also 14" behind the injection point. I've got the injection start point at 4 PSI of boost, so there's already a bit of mass flow.
There are some pretty nice heat-shield products varying from SS to Inconel that I might try to make a manifold shield with. There is a lot of wiring in this area that I'd rather not test the thermal limits of in the future when I run more boost, and eventually nitrous.
Is this the first time you've seen this kind of instrumentation posted publicly?
I heard an interview with French Grimes a few years ago, and he was talking about sensor "Issues" on the dyno. People think that sensors are perfect, but in reality, as you know, they are within an area of deviance.
The sensors we use at work are sensitive enough to use all of a 24 bit A/D converter's resolution. I'll let you work out the math on that one.
A friend of mine uses CO2 for the throttle stop on his bracket race cars. He says one filling lasts him a season. He has 2 cars though, both in the 8s, and probably does 50 passes on each. What I'm told is that it lasts longer than you think. I'd probably put a big bottle in my car though with a WG...as those suckers leak more than people know....
favorite thread....and build principals....a few of you probably knew that though.
where do i read more on the quoted? Sounds allot like some concepts i thought possible a couple years ago. If you remove load from the compressor "carefully" with rotor shaft speed management you should be able to build compressor velocity, stay in effcency islands longer, and retain more speed between gears. A possible limitation I considered is the instantaneous loads that woul be introduced and how happy the comressor wheel and shaft will be with uncontrolled slugs of load.
With a built 210 diff on the way, I had to figure out how to mount it. I bought an E46 M3 subframe and support bracket, but I'm undecided on what to do with the bushings. For subframe to chassis interface, I think I'm going to pull the solid bushings out of the 188 subframe, and put them in the 210 subframe. This will help the suspension mounting points stay stable. For the 4 diff to subframe bushings, I'm actually thinking of sticking with stock ones to reduce shock loads and vibration.
Thoughts?
Please tell me that your new 200 ft of driveway has FIA curbing.
'97 M3, Estoril blue, 2 dr, euro 6-spd, EFR 9180 divided T4 .92 IWG, RK tuning, CP 8.5:1 pistons, Eagle rods, Schrick cams, L19 11 mm ARP studs, O-ringed block, Supertech stainless/inconel valves, Supertech springs & Ti retainers, ported head, S54 oil pump/pan, 80 lb. injectors, OBD1 intake manifold, Steedspeed twin scroll T4, 3.5" SS exhaust, eBoost2 EBC, HFS-4 W/M injection, AEM Failsafe, Zeitronix data logger, Racelogic TC, OpenOBC w. ethanol %, Ireland Eng. engine mounts, UUC black tranny mounts w. enforcers, UUC twin disc feramic, ARC-8's, MCS 2-ways, Z3 rack, Rallyroad strut bar, X brace, Eibach sway bars, Ground Control LCAB bushings, Bimmerworld RTAB's, Powerflex subframe bushings, 210 4-clutch LSD, Stoptech BBK, titainium shims, steel braided lines, brake cooling ducts.
Seems I've proved this calculator close:
http://www.22rpd.com/boostcalc.php
8.5:1 compression
3.014L engine
91 octane fuel
I discovered the transmission was leaking oil shortly after installing it.
Initially I was concerned that I could be from the input shaft, but on cleaning up the bottom of the car, it seemed to be coming from the back of the transmission. There was oil accumulating on the drain nut, so I pulled that off, sealed it with teflon tape, and put fluid back in. Same result, although there was definitely oil above the output flange spraying around. It was hard to tell if it was because of airflow under the car, or fluid being flung by the flange. So another clean-up of the bottom of the car, a quick drive through the gears, and I came back to see fresh fluid on all the same spots. It wasn't a huge amount of oil, but it was enough to leave a drip on my garage floor.
I cleaned up all the surfaces again, and this time let the car sit overnight to see if I could find weeping out of the shift selector shaft seal, or output flange seal. The next day, there was nothing.
Looking at the breather on top of the transmission, I realized it was different from the ones I normally see. It's a little cap sitting on top of tube, which is held on by a chain, that can lift up a little bit. I felt like this could be the source, where fluid was being flung inside the case and pushing the cap up to spill out. I grabbed some shop towel, folded it over a few times, and wrapped this around the cap and zip tied it on.
No more transmission oil leak.
Any wiring that runs by a heat source got at least one layer of heat shielding on it when I did my build. I'd cut open a reflective fiberglass tube sleeve and then zip tie it on the wiring bundle. Easy to service in the future, and really keeps the peak temps down. Of course, hold the wiring harness every ~5-6" (no longer than every 10-12") with lots of cable ties/cushioned adel clamps.
On the data front, not really. I spent 6 years in an aerospace test lab running thermal/fluid tests, many at these sorts of flow rates and temperatures, so it just jumped out at me as you either have a TON of heat dumping into that intake tube, or there's some evaporative cooling of the sensor bulb. Since you're sure it's not evaporative cooling, you definitely need to do some heat shielding STAT in that area.
It will be really interesting to compare the inlet/outlet temps before and after the heat shielding. Being that my inlet manifold temps are really good, it's not critical for engine life, but I'd rather not be putting that heat in there in the first place. Kind of like putting a BOV after the intercooler....where you're dumping air that's been cooled, and the intercooler is shedding that heat for no reason.
The main problem with a high turbo inlet temperature is that the compressor temp rise is based on that already very elevated temperature, so you add even MORE heat in an absolute sense when compressing it. It's the same as running a restrictive exhaust on a turbo car and getting a really high EBP value pre-turbine due to the turbine pressure ratio.
Bookmarks