Technical Woj's turbo project

Currently reading:
Technical Woj's turbo project

A new clip/spring set might help pull the shoes back. Could be a badly manufactured set of shoes?

Shoes are Ferrodo Premier, so I doubt that's their fault. But indeed, probably replacing the complete mounting kit would be good idea, but then replacing the shoes would not hurt while at it.

Tomorrow I will be driving it and doing more ECU tests, will see if refitting the drum and cleaning stuff helped.
 
It seems that just wobbling the brake two days back helped it. The pedal was pulsing a bit during the first braking, but the noise seems to be gone, at least for now.

More importantly - I spent the last two evenings implementing a somewhat proper AFR based fuel correction and it now works, even though it is not PID controller based, a simple correction, but this time properly implemented. Tried it so far only in stationary conditions (see the clip), and it will probably need some small tweaking, but it works. It is not as stable as I was hoping for, but it might be also because of the AFR module readouts. First, I still have horrible problems calibrating the reading table to get the right AFR values in the ECU, they constantly seem to be ~0.3 AFR off. I would probably get much better results using the linear output of my AFR module (right now I use another, non-linear one), but that one is taken by the narrow band signal simulation that I really badly need. Alternatively I might need an anti-noise filter on the connection in the ECU.

Going back from work I will try it in the driving and high load more.
 
Drove a bit with this correction on today, in principle it works, no singns of it not doing its job, also on boost and rich regions. But there are some quirks. The correction itself is not yet as stable as I would like to be, perhaps I am looking at implementing a PID controller afterall. But it might be also the AFR input I get from Zeitrnix - it seems very unstable in the ECU comapred to what the Zeitronix shows on its display and log. I read about it a bit and aparently I am not the only one having problems with the module's non-linear output. The concensus is that it is better to use the linear output, but this I cannot do, at least not without getting a narrow band O2 sensor first. Plus I still cannot calibrate it right, the correction becomes a bit pointless when the ECU wants to set fuel to AFR 11.5 on boost but because of the AFR readings it is corrected to 12.4 instead :( I might be bothered to write Zeitronix about this actually, they always had very good customer service.
 
I did not write Zeitronix in the end, which was a good thing, cause I would make a fool of myself :D. Something new to learn everyday :eek: I was about to start fiddling with the AFR module's outputs, getting new wires in the bay, etc. etc. at which point I decided to first check where in the Zeitronix software I should change the output type from narrow band to linear AFR to use the narrow band wire as a linear AFR input to the ECU. I ended up on the Zeitronix page that explains the linear output settings and read there, tada, that 2010 onwards version of my module (which I happen to have) provides linear output on the wire that I was using and thinking it was non-linear, and that there is no non-linear output, period. Looked at my calibration tables and suddenly everything has falled into place. Have not tried it on the car yet, but looking and the new linear calibration values and the results I was getting before it should be spot on now. And it also explains unstable readouts over certain AFR point, this is a clear values taking off because of the linear and non-linear curve getting serisouly apart from each other.
 
It is mega sweet (it would have been even sweeter if my virtual Windows session did not decide to crash badly - apparently it is not a good idea running two different logging programs on two different COM ports at the same time :D In any case, a carefully designed 10 minute running log went to bust).

AFR is now nicely calibrated, see the first video (it may seem there as the values are ~0.1 off, but the resulting correction I get tells me I should not try to recalibrate it, in the stoich load sites I get long stretches of stiff 14.7). And the closed loop correction also works very well, see the second video. It also worked nicely on boost and high loads, but the log I wanted to show is in nowhere land. But did one without correction only logging AFR errors to see where my VE table is, that's in the picture. Evidently some corrections still to be made.

I also feel that my ignition is a bit too optimistic, I think I have heard signs of knock (I was running with knock control disabled, it gets in the way of the fueling), I will have to drop it a couple of degrees. Also, the ECU is very clever, I took it to the load site where I had ~5500 RPM and mid vacum and just wanted to press the throttle a little to raise the load, but it just chocked on me, apparently (by the feel of it) some sort of protection fuel cut.



 

Attachments

  • AFR_error_log.png
    AFR_error_log.png
    16.4 KB · Views: 94
All good things must come to an end :(

(This is not a selling ad, not yet at least).

I will be breaking my car for (mostly turbo) parts this week as the poor thing needs to go. The parts will be offered in complete functional kits (as in oil plate with cooler and lines, complete turbo kit for 1.2 MPI, or complete AFR/EGT setup). If you feel like you are going to be interested in any of the bits give me a shout through PM.
 
Let's do some post-mortem posting, another thread on dyno runs reminded me that I have long over due videos to publish and link here. These are two fragments of my dyno tuning runs back in 2011 after installing new cam. Unfortunately I do not have vids of the final power run.

The first video gives the full run of the 2300 RPM tune, first goes the fuel setup through the whole available load range, then we played with ignition advance at 94kPa and full available boost on that RPM, and towards the end there was a swipe going up and down with load over this RPM cell to see how it behaves and that there is no knock.

The second vid is merely setting up fuel at full boost at 3300 RPM. This was problematic as I remember because the AFR was having its own life because of the small intercooler not catching up with the continuous high load and IAT fuel correction getting in the way.

The changes to the ECU map were happening live, the screen shows the dyno's visions of them.

(The engine is not really audiable (turbo is :D) because of the huge dyno fan).

And yes, I know, this not really very exciting to watch, you don't have to tell me, I have over 2 hours of this.



 
I found a simple(ish) explanation of the functioning of the BOV in Mark Warner's Street Turbocharging (HP Books).



So, correctly fitted and functioning, the BOV will only vent when the throttle is suddenly closed: it doesn't sense boost at the wastegate. So, no "psssht" if the throttle is open. Or, if you like, the BOV protects the turbo from power off overboost, the wastegate from power on: if the wAstegate sticks or can't vent enough, the BOV can't save you. Emerald recently ran an Audi TT motor with an undersized wastegate (BOV fitted): it overboosted.

Lots of stuff in the Warner book about seperate(d) and remote wastegates with good pictures.

Fingers, just seen this post from a while back....Slightly off subject but did Emerald have the Audi TT motor fitted into a Fiat?? Was thinking of an Audi TT engine and drive train in a Seicento....

TIA.
 
Back
Top