I ran into an issue with the MAF.. Its kind of strange: I get no reading in VAGCOM, but I do get the appropriate voltages and grounds at the sensor. I tested this by measuring the analog output of the Maf, which was about 1.4v on idle and increasing on throttle. I had contact with my supplier and he claims the sensor is not right for my application. I went on the bosch website and found out the original APU sensor is HFM5-4.7, while my current sensor (VR6) is HFM5-6.4. I never took into account there was possibly a different version on the same style MAF, but could it be that this causes my problem? I already ordered another sensor (4.7) so I'll find out this week. Could be a warning to other people who buy larger MAF housings, to not use the sensor unless it is the same type...
Car is running fine by the way, but no difference with maf plugged in or out and I get a code for reading too low.
Took it around the block yesterday, and I can now select all gears with ease! I'm very happy about that. I will post a video of the converting mechanism soon 

I have done about 200km on this engine now, and it is slowly coming together. I ran my flash just once because I lost my pedal control for some reason (plug on the pedal was not connecting properly I later found out) and I wanted to rule the ecu out. I need to adress the timing, up to 0.8 bar everything is fine but when boost goes more up I get all sorts of throttle cut/timing pull also there was something in the engine bay whistling like a trumpet at full boost hehe. Thats not surprising, since half the hoses where not clamped. Might be the dv as well, it was slightly leaking (new unit) and I got sent a new working unit to replace it, but havent done that yet. What a difference between 0.6 bar and 1.2 bar, wow. Feels like a different engine!
I have finished the map adapter and lambda bungs thursday and made the engine bay a little neater today, working on my engine cover.
When the cover is finished I will start to work on a baseline log and then onto the tuning  

I have a returning P1557 error, could it be (besides a possible boost leak) that my maf is reading too high? I get around 4g/s at idle but I have seen peaks above 160g/s on a WOT pull. That looks way too high to me, so I want to scale down the maf a bit. I read about the offset in the wiki, and I have done the following:
Select all 512 values > add 200 > Multiply by 0.95 > subtract 200. Is this the correct way to scale it, or should I have removed 200 and then scale down? This is not really stated clear in the wiki (whether to add or subtract the offset to start with), just to offset it.
** I have done the 2nd option now, this seems to be correct when comparing to the differences of a larger housing scaling with original**
My starting value is original 165.5 and last one 944.3, scaled  for 95% I get 167.2 to 907.1
If I do it the other way around I end up with 147.22 - 887.08 but that is not logical since my MLHFM for 3" maf is 158.7 - 1193.7

When I looked at the log of channel 115, the actual pressure was far far behind on requested pressure. It could have been that at that time I was already in limp mode, because I was only looking at block 2 maf reading before. The increase in maf flow is very smooth, so I guess the sensor is working properly, it might be that the housing is a bit tight or it is not calibrated correctly... Any thoughts?
 

I have made a few runs tonight, and found out I cant read the Nefmoto logs in EcuXplot  
There is still a problem with my turbo not responding to requested load, and a absolute maximum pressure of 1900mbar over all runs. Am I missing something in the software or could it be hardware related? It looks like my wastegate is blown open at 1900mbar, because wgdc is 95% at that time. Anyone care to take a look at the logs via the Nefmoto software? Or tell me how to build a single csv from this xml document, I tried to convert but I get 7 seperate files for 1 xml. Also the definitions seem off by a few decimals and some (like temperature) dont relate to real values at all. I was hoping I could fix this in EcuXplot.
Heres an explanation of factors during the runs:
Log1 - wot 3rd gear on LPG
Log2 - wot 3rd gear on petrol (with faultcode positive boost deviation)
Log3 - wot 3rd gear on petrol - faultcodes erased
Log4 - wot 2nd gear on lpg (possibly with fault codes back in)
Log5 - easy driving through 2nd, 3rd and I think also 4th gear, this way the car feels the most powerfull (lpg)

I think about the following strategy: tune back ldrxn to aim for 1900mbar, if it holds I dont have a boost leak and the issue is either my wastegate or the ecu limiting. Also I see my maf values are still way too high, which is not helping. if i'm aiming for 1900mbar just above 150-160g/s seems more reasonable...
***
Done the above but still no good spool (following stock ldrxn ramp and still underboosting), checked my wg lines and they are not in good condition. I'll fix this first.

ECUxPlot does not support nefmoto logs, unfortunately. Try to get me7l working...

 There haven't been mayor changes so far, but I have succesfully driven the new 018AE ecu for about 300kms now. I have some hardware issues that I thought was to blame in my tuning, I am still not meeting requested boost. My friend with a MY2000 180hp TT was so kind to guineapig my tune (he has the same 512kb ecu/bootloader version) and in his car I even have seen overshoot (2400mbar max where 2200 req). I have done a little finetuning on his car and now it runs very smooth with acceptable timing corrections (3-5deg under max load). First he did not got any proper maf readings (it maxed at 140g/s) but after swapping his (bosch) maf with one of my spare china maf sensors it is now reading (a twitchy) 165g/s lol. So he will try to clean up his maf in alcohol and if it fails just order a new oem unit. I must say it is fun to see this effect of my own experimenting, the car had wheelspin from 3000rpm while going through 2nd gear on dry tarmac, and that was never before achieved   my friend was already thinking about buying a quattro now, since the next phase is a gt2871 that I talked him into  
But back to my own topic now:
I have yet to calibrate the LPG system correctly, it is running too rich at the moment. I have been researching the possibilities of switchable maps for my goal, and I have found some good topics here and elsewhere. I think it is most important I buy a wideband o2 first and then try to get the lpg fueling right, perhaps with an nb02 emulator out of the wideband controller. My friend will make me an egt sensor with small lcd. The steps to take in the LPG project that I think I need to do:
-Switch signal (generated by LPG ecu on first lpg injector opening) feed into secondairy o2
-customize ecu to switch maps when signal is activated
Maps to be adjusted: timing, target afr, knock based enrichment, egt based enrichment, ....?
-provide emulated lambda signal while driving on lpg to enable afr 15.5

I can calibrate the lpg injectors with the lpg ecu, so it should not be necessary to adjust krkte/tvub when they are properly scaled for my petrol injectors, and the switching is always 100% so there is no need for interpolating values like when tuning for e85/pumpgas mixtures. I have no VVT so in fact KFZW2 could be used for the LPG alone and no extra memory needs to be found for timing.
How should I start, do I need to disassemble the ecu to add commands? I am using a fully defined 512k flash so I can look for grey parts in the hex and write new maps there? I have absolutely no experience whatsoever with (dis)assembly but my TT friend does, if I can get some pointers where to start he can help me integrate this function  

**
I have made a small beginning in the tuning for lpg, I have tuned kfzw for lpg combustion (more timing in low rpm and less timing after 3000rpm) and I have set up KFZWWLNM to correct my tuned timing back to stock-ish petrol (not absolute but at least better than running too much advance) when the engine is under 40 deg C. This is because my lpg injection starts at 25 deg C reducer temperature. As long as I dont run out of LPG with a hot engine, I should be good for now 
I can tune my lpg ecu to correct for the requested AFR too, but this is not how I want it to be. I installed IDA pro and I will take the plunge in reverse engineering 

I had a problem with IDC, but in fact it is logical that I run out of injector so soon with this small injectors. Maybe I will order a set of 440cc units later on this year or put in my friends 310cc's when he upgrades, and most of the time I run on LPG where I have some headroom still. Those injectors flow 60hp per unit @ 1.2 bar but max out on 5 bar. To get a better calibration of my lpg injector times I have adjusted KRKTE for 270cc, which is about my injector flow @ 5 bar. This should help a little concerning maxing out injector/keep injection time more sane. I dont really understand how commercial tunes can claim 200hp on the early aeb without supplying a different fuel pressure...
Currently at 288cc krkte, because I had initial fuel trims of -25% (engine already up to operating temp before flashing). After stft adaption jumped to 8.5% and it was idling just fine. I have adjusted the injector rate up by 7.5% so I should be about right after the next flash. If I get idle to follow requested I'll start logging part throttle first before pulling wide open   ** Is it wise to set TVUB to 1 across the map to get an accurate KRKTE? This seems safer than setting KRKTE too low, but I dont know what is more efficient. Since I am not running at stock pressure nor using the designed injectors for my map I dont have a good set of starting values.
I found this overview on a website (http://injector-rehab.com/shop/lag.html)
VW Passat (98-99) Audi A4 (97-99)       
Flow CC/min / Flow LB/Hr / Ohm /10v / 11v / 12v / 13v / 14v / 15v
------220----/-----21lbs--/--12--/0.82/0.62 /0.46 /0.34 /0.24/ 0.19
or
------240----/-----23lbs--/--12--/1.05/0.84 /0.69 /0.60 /0.52/ 0.43 (this looks a lot more like the tvub I have now)
******
!!! More data, looks like I have to revise my KRKTE anyway  
Injector 0280150447 || EV 1 || operating pressure 400kPa || q-stat@400kPa 194 g/min||q-stat@300kPa 168 g/min || q-stat@300kPa 240cc/min || q-dyn @300 kPa, ti 2,5 ms (g/1000 Imp.) 5,63   Any algorithms how to calculate beyond these pressures and how do they affect OT? I calculate a fuel flow of 279,54cc/min from the heptan value @ 4 bar. wow... So it turns out I totally underscaled my krkte   that will help out with my IDC for sure  

The damos definition seems a bit strange btw. I think its wrong, i'll check with the bam damos. ** > factor was wrong and axis decription as well. changed 40 for 0,0704 and rpm for Volt.

** update 13-1
Car is running fine now. Injector times are back to normal and fuel trims are adjusted 4% down via TVUB. Its running close to 0% idle stft now, so lets see if the scaling was correct when long term fuel trims kick in :-) Power is better too, it must have been overfueling like crazy!   Hope the mileage will recover as well, I was @ 1litre per 7km  
BTW it is @ 5 bar, calculated injector size 312cc/min and scaled krkte to match. (currently 0.1045, it was set as 0.1344   )
Photo of my adjustable regulator with gauge, at last it comes to good use after spending a year in my garage  (manifold vacuum unplugged at photo offcourse)