YarnMaster
We take very different approaches.
I've messed around with microcontrollers, doing some really simple PIC programming in assembly. I was actually first taught assembly programming in high school on Apple IIe's. We loaded programs on pencil marked cards (circa 1982). But that is a very dim, very distant memory.
I regained an interest while messing around with DIY Audio, and wrote a digital PLL in assembly to control a low jitter VCXO. The code measured the difference between two 11mhz clock signals, and then used a digital to analog convertor to generate a control voltage for Voltage Controlled Xtal Oscillator.
Looks a bit battered after knocking around in a bag of junk for 7-8 years but it was classic ghetto electronics and worked quite well.
So I've come to the ECU disassembly with basic understanding of interfacing the "real world" into microcontrollers via ADC's and DAC's.
When I look at the sensors, I know they are a part of a chain that runs something like sensor -> pcb electronics -> analog to digital conversion -> range checking on ADC output -> scaling to convert to correct units -> smoothing -> correction curves (if needed) -> engine management magic....
All of the live data for sensors the Nanocom displays is taken immediately after the smoothing/correction curve stage.
The ECU is configured for each sensor's particular characteristics, and usually limited to the operational range Land Rover expected based on the tune they used.
As an e.g.
The MAP sensor has a maximum output of 4650mV at 250kPa absolute pressure. The ADC range check verifies the result of the MAP conversion does not exceed the equivalent of 4650mV or fall below 500mV.
If the input falls outside the expected range the ECU sets a fault.
If the signal is ok, the ADC result is scaled using values that "reverses" the transformations of the sensor (sensor curve), voltage divider on the PCB, and the ADC conversion to give the value of the physical quantity the sensor measured at that particular moment.
If you substitute a different sensor which has a different curve to the stock curve without changing the "reversing" calculations you get an incorrect reading at the output.
You'll get numbers on your Nanocom display, but they no longer have the correct relationship with the quantity the sensor is measuring.
This is what I mean by a sensor "misreading".
Anyway, this is the basis for recalibrating MAP and MAFs.
The stock MAP has a bit of "hidden" range that allows the limiters and scaling to be adjusted to give a 2.55bar maximum absolute.
It could go a tiny bit further but I'm respecting the lower limit of maximum output voltage range for the sensor.
Using alternative sensors you can go much higher. The MAP I think is the best choice for high boost is a 3.5bar Bosch unit, which has a very different curve to stock. If you use it with an unmodified map, it will read 60kPa when the actual pressure is 100kPa.
With a recalibrate it will read accurately and give a maximum absolute pressure of 3.2bar. This is a limitation of the way the value is stored in the ECU can cannot be worked around. The 4bar sensors that some tuners use don't allow higher readings for this reason.
The stock MAF has a bit of spare range and the configuration can be recalibrated to make use of this. In theory it will give over 800kg/Hr readings but I don't have the ability to test to this level. The curve remains essentially stock below 690 kgHr.
Together the two mods eliminate the need for a boost box on typical stg1 and stg2 road tunes, without further changes.
That said the motivation of the mods was originally to completely change the way Td5 remaps are done but that is another long winded story.
Probably more on point, a UK tuner asked me the other day if I could look into eliminating the delay when the MAF is disconnected.
I think I have found a solution, but it requires modification of a single instruction in the code portion of the engine map, rather than changing a setting in the fuel map.
Not something I'm going to make publicly available because it requires a surgical mod - any error is likely to crash the ECU while running.
cheers
Paul
YarnMaster
I should also add that there is a correction done for variation in the battery voltage on the MAF reading.
If you put a regulator on the MAF 12V input as proposed you are powering off steady 12V while the ECU is correcting for variations in the battery voltage. In other words you INTRODUCE inaccuracies, rather than remove them.
cheers
Paul
YarnMaster
Originally Posted by OffTrack
My bad, I was looking at a half done Defender disassembly when I wrote that, and missed that it pointing to an address just after the BATV which holds the 5V sensor supply reading.
The accurate information is that the MAF is corrected for variations in the 5V sensor supply voltage.
Apologies.
YarnMaster
So IMO giving a stabilised feed to the MAF will not hurt though
I completely agree with the MAP recalibration for higher boost rather than using a frigging boost box but on the MAF anything higher than 650Kg/hr REAL air flow can lead to turbocharger overspeed and premature turbo failure cos the turbine must spin very fast in reality to achieve such suction... also it's not normal to be such high suction at a REAL boost of 1.3 bar at factory wastegate setting.... i'll improvise to fit the anemometer(when it'll get o me) on the intake somehow and make some live measurements while driving.... IMO even if it's not stated in any document that cut-out at 680Kg/hr was made so on purpose especially for turbocharger overspeed protection or it's some very strange coincidence cos i suppose that the builder didnt presume that the vehicle's owner will start messing with the wastegate ...that's my presumption
my blower was too weak to be able to force the MAF above that limit on the bench so what i want to ask you is that the cut-out is made by the ECU or eventually the MAF's electronics cuts the output at that value
Discovery Td5 (2000), manual, tuned
YarnMaster
So IMO giving a stabilised feed to the MAF will not hurt though
I completely agree with the MAP recalibration for higher boost rather than using a frigging boost box but on the MAF anything higher than 650Kg/hr REAL air flow can lead to turbocharger overspeed and premature turbo failure cos the turbine must spin very fast in reality to achieve such suction... also it's not normal to be such high suction at a REAL boost of 1.3 bar at factory wastegate setting.... i'll improvise to fit the anemometer(when it'll get o me) on the intake somehow and make some live measurements while driving.... IMO even if it's not stated in any document that cut-out at 680Kg/hr was made so on purpose especially for turbocharger overspeed protection or it's some very strange coincidence cos i suppose that the builder didnt presume that the vehicle's owner will start messing with the wastegate ...that's my presumption
my blower was too weak to be able to force the MAF above that limit on the bench so what i want to ask you is that the cut-out is made by the ECU or eventually the MAF's electronics cuts the output at that value
The cutout is always in the ECU unless the sensor is faulty. It would be a very bad design if the sensor was cutting voltages before the ECU as standard.
There is a limiter at 4950mV which corresponds with roughly 685Kg/Hr.
These are actual ECU readings for the mV not calculated. There is a slight skew between readings so the two do not line up exactly - there will be a ms or 2 between the MAF and the MAF mV.
The diagnostics actually allow single or small groups of sensor readings for diagnosing sensor issues. You can read the MAP, a corrected MAP value, MAF and MAF mv in a single request, so you can log the values of just these sensors 20+ times a second - the Nanocom logs once every 1.2 seconds. LR docs mention the ability of Testbook to read single sensors for diagnostics purposes.Code:MAF MAF mV 5843 4752 5788 4752 5658 4741 5901 4736 6061 4741 5931 4779 6464 4913 0 4946 0 4962 5788 4762
cheers
Paul
YarnMaster
Thanks for that sir.... i'm coming closer to the gist of it post by post... i'me very interested in your oppinion about that presumption of mine about turbocharger overspeed protection though....cos i'm 100% certain that a REAL air flow above 680kg/hr caused by suction without wastegate alteration can be dangerous... albeit i'll stand corrected in case of a valid argument against that
Discovery Td5 (2000), manual, tuned
Fossicker
SupporterSent from my XT1039 using AULRO mobile app
YarnMaster
Again it is not possible to talk about Td5's in general. EU3's take very specific measures to guard against overspeed based on MAF. EU2's seem to restrict the boost to keep below a specific pressure ratio. The stock turbo map has a maximum pressure ratio of something like 2.45, so assuming a 1bar atmospheric pressure, that translates to 2.45 bar. Going past that means reduced efficiency, increased heat into the intercooler and high turbine speeds.Thanks for that sir.... i'm coming closer to the gist of it post by post
It's probably more dangerous and less reliable than staying stock, but I haven't seen reports of an epidemic of blown turbos as a result of boost boxes and wound up wastegates.
It is definitely less than ideal, and the real solution is to go to a better sized turbo, or VNT.
cheers
Paul
The problem with MAF limit is that it is NOT defined by the voltage range. It's somewhere in the settings. I have found that it is anywhere from 650 to 675 kg/h. Here is how. I am running 1400 kgh Pierburg maf insert in 82mm ID body. 650 signal is below 4V. Thus I should have no problems, right? Wrong. Once 650 or higher (seen 820) shows up on Nanocom I get shaking and power loss. No faults recorder. I made bunch of maps cutting maf table at 645, 650 and so on up to 750 and then spent hours running them on the road. The verdict is that 645 map is smooth as babys arse. 650 starts jerking when cold and 675 and above gives problems all the time. Butchering the map table to read 900 at around 3V I managed to see this number steady while parked and revving the engine to 3000 rpm. This limit is a "dynamical load dependent creacher". Another funny thing I noticed ... the magnitude of the problem depends on fuel temps. Once it reaches 50C there is no jerking up to 675 and with the stock maf engine does not notice the limit - NANOCOM shows maf default at 0, car does not jerk and easily pulls to 5000. Same story with running without maf - once at 50C on fuel and 40C on intake air motor runs as if maf is there. Weird. I am thinking of ditching eu3 map in favour of eu2, but I really need this 25CR ID table that is absent in eu2 map. Currently running 1.3 bar boost and 70cmm ID.
Paul or anyone has any idea what's going on here? Thank you in advance. Stepan
YarnMaster
The MAF sensor range is 100% defined by voltage.
There are a lot of limiters that you can run into once you go beyond stock operating range - I can think of at least two you are probably hitting based on what you've posted.
It sounds like you are doing a duration mod based remap seeing you say you "really need 25CR ID". If that is the case you might as well clamp the MAF with a boost box.
Posting Permissions
| Search AULRO.com ONLY! |
Search All the Web! |
|---|
|
|
|
Reply With Quote
Bookmarks