DIY Fuel injection thread.

[Testament]

yeah if the 560cc injectors have a different opening time the fueling will be a little different, esp in the transition

 

so those early links only do acc enrichment on rate of change of MAP? did not know that.

[Spencer]

Sounds shit.

[Roman]

A workmate just bought a turbo DC2 running one of the eeeaaarrrlllllyyy links.

I'm pretty sure that the scratches on the casing are from when Moses tried to chisel the ten commandments into it.
 

His engine has been running a bit rough so I was all "sweet lets just datalog it"

 

Oh nah, turns out you need a serial board and all this other junk to plug it in, and there's possibility that it's not laptop tunable anyway. Heh.

*back away slowly*

[0R10N]

A workmate just bought a turbo DC2 running one of the eeeaaarrrlllllyyy links.

I'm pretty sure that the scratches on the casing are from when Moses tried to chisel the ten commandments into it.

 

His engine has been running a bit rough so I was all "sweet lets just datalog it"

 

Oh nah, turns out you need a serial board and all this other junk to plug it in, and there's possibility that it's not laptop tunable anyway. Heh.

*back away slowly*

 

Hah, funny that. A friend showed up one evening with Quinn Lee Sang's old BFMP drag car hoping I could pull the tune off it with my laptop/SerialLink so they could do some tweaking. Turns out it was running a G1 V2 so I had to use a hand controller and create a reference map in Excel, bwaaaa bwaaaaarp.

 

Everything about it is stuck in the 90s, from the narrow band O2 to the crude locked dizzy/reluctor setup and the lag monster turbo that only comes alive at 5000rpm... he texted the next day to say he'd found some hand written fuel and ignition maps on a couple of photocopied sheets in a folder he got with the car.

 

Have told him he should really send it back to Link for their museum, they were actively buying back EMX and V1/V2 units for ages!

[kpr]

yeah if the 560cc injectors have a different opening time the fueling will be a little different, esp in the transition

 

so those early links only do acc enrichment on rate of change of MAP? did not know that.

 

Old linkplus could do tps enrich,  but not v1-5 by the looks. 

 

I still got a serial link if  you wanted to access Moses  finest work Roman,    and a hand controller too!

[Roman]

Okay soooooo, I got reminded of a program I heard about ages ago, but I only had OBD2 stuff to play with at the time.

 

This program called Virtual Dyno:

 

http://www.virtualdyno.net/

 

Basically you turn on datalogging on your ECU, and do a run on a flat piece of road/track. (or 3 or 5, pick the result closest to average)

Then export the datalog file into a CSV file (or you can use megasquirt log files directly)

 

Thennnnn you input some data about your car - gear ratios, weight, tire size, etc.

 

From this information, it judges the rate of accelleration of your car to give an estimate of how much power it makes. (So obviously testing somewhere flat and straight is important - a drag strip is perfect!)

 

I previously had the same motor as what's in my car dynoed with factory ECU - 127kw @ 6800rpm. (I've lost the dyno chart though unfortunately, thanks to Imageshack being bastards)

 

So what does virtual dyno show, on this different but largely the same engine from some random road track tune I was doing earlier?


 

 

A peak power result within 1kw, within 200rpm of the dyno results. (Which is perhaps coincidental at best as the two cars had different drivetrain layouts and so on)

 

However it looks to be close enough in order to be useful in terms of optimising ignition and cam timing.
It should also be able to show whether the staged injection earns its keep.
So long as you stay strict with your methodology at least. (And dont lose your license in the name of science)

I cant wait to have a go with this when my car is back together - Just some test runs initially to see how repeatable the results are.

Has anyone else used this program before? Or will consider using it for some tests?

Super keen to see some OS results!
 

[mjrstar]

Yip, my latest mx5 run in 4th is lumpy as hell above 4000 rpm which i assume is probably a bit of wheelspin.

Do you have some vvt point at 4200?

Another good use for this is setting VTEC point, set it real low, do a run then set it real high do another then set changeover point to where the 2 runs intersect. Job done.

You can add afr and timing to the lower plot to tell more of the story too... only problem is its just not safe to tune high powered cars in this manner unless you can get to the track.

[Roman]

My MAF datalogging of various VVTI cam angles shows there's a bit of an airflow peak just below 5000rpm but nothing spectacular going on at 4200:

 

 

I think it must have just been going over a bump or something at that 4200rpm spot perhaps.

This motor when standard has peak torque at 6000rpm. But It's got different manifolds and so on, so it's possible that it's moved the peak a bit I guess. Probably not by that much though.

I would imagine that if I did 5 runs with the same tune/settings on a flat road, you'd be able to see the trend to remove outlier runs.

I'll preset some different tunes ready for nightspeed drags that's coming up in a bit, will be a good consistent testing ground.

[Lith]

A peak power result within 1kw, within 200rpm of the dyno results. (Which is perhaps coincidental at best as the two cars had different drivetrain layouts and so on)

 

However it looks to be close enough in order to be useful in terms of optimising ignition and cam timing.

It should also be able to show whether the staged injection earns its keep.

So long as you stay strict with your methodology at least. (And dont lose your license in the name of science)

I cant wait to have a go with this when my car is back together - Just some test runs initially to see how repeatable the results are.

Has anyone else used this program before? Or will consider using it for some tests?

Super keen to see some OS results!

 

 

Yes!  Funnily enough I found out about it about 4 or so years ago after I started building something similar for my own use when I started getting more seriously into tuning and didn't have easy access to a dyno... it got to the point where people started telling me I should develop it into a releasable product as it'd be hugely popular, so started investigating how likely it was that I was reinventing the wheel and unfortunately (or fortunately) I was.

 

I didn't get it anywhere near as mature as VirtualDyno but it was already at the point I could use for meaningful tuning but I did find that I had to jump through some hoops while trying to build a meaningful curve out of what would typically be a 10hz sample rate at best (often the values recorded are effectively coarser than that) and there is no steady ramp rate like a real dyno it was very difficult to produce a smooth curve which also represented the information with enough accuracy to be useful for comparing the results from adjustments.  

 

I had things at a point which worked more or less acceptably but there would be the odd occasion where there was an inexplicable "wave" or random kink in the dyno plot and I've noticed that Brad @ VirtualDyno has clearly run into the same issues and has just given the user the option of cranking up the smoothing to allow them fine tuning of the bias between "smoothing" and accuracy as well as other smarts to try and keep it looking like a legit dyno plot, though even then there is the possibility of getting the odd random bump or tick in a run which isn't necessarily anything to do with how the engine is performing or the road you ran on.

 

Otherwise, I too have found Virtual Dyno (and my own before I flagged the project after finding VD) to be mostly very consistent with the real dyno - up to a point.  That point usually is when traction or at least how quick the car gets through a gear versus the log sample rate gets to a point it can't get a meaningful reading on delta-v, which is kinda critical when you're relying totally on F=ma to build the "dyno curve".  I've only really stopped using it over the last year as now I have pretty easy access to a real dyno for doing tuning now which isobviously far more ideal, so typically the only road tuning I do is just to get things running to a point I can be sure things seem "ready" for a dyno tune so time and money isn't wasted there.

 

My MAF datalogging of various VVTI cam angles shows there's a bit of an airflow peak just below 5000rpm but nothing spectacular going on at 4200:

 

 

I think it must have just been going over a bump or something at that 4200rpm spot perhaps.

This motor when standard has peak torque at 6000rpm. But It's got different manifolds and so on, so it's possible that it's moved the peak a bit I guess. Probably not by that much though.

I would imagine that if I did 5 runs with the same tune/settings on a flat road, you'd be able to see the trend to remove outlier runs.

I'll preset some different tunes ready for nightspeed drags that's coming up in a bit, will be a good consistent testing ground.

 

So that bump was ruled out as an outlier by doing multiple runs?   If not, see my above ramble

[Roman]

Hey, thanks for post!
Yeah, my car is in pieces so I cant do any runs currently haha.

Previously I've just been trying to suss fuel for the better part, and I've used the MAF and some Toyota documentation to figure out VVTI settings.
But I've been conservative on ignition timing as I've had no feedback loop for that.
All of my existing datalogs have had no effort put into getting logs from a straight flat road or whatever, because it's been irrelevant.

Which is why I'm skeptical about that 4200rpm bump being legit, I could have run over a hedgehog or something haha.

 

Most of my datlogging is from Hampton, where there's a lot of up and down so it makes things largely incomparable.

With the link I can set a trigger so that it starts/stops recording for WOT only, at 100hz internally.

 

So I'll get lots of small files which is perfect for this, assuming I can get some repeatability.
The drag strip will be as good as it gets so that's where I'll initially test with a few runs on an idential tune, and see what variation there is.

Then if it seems good, adjust some settings and go from there.

 

I do realise the limitations of something like this compared to using a dyno, and I'm trying (failing) to not get too excited about using it.
 

But on the whole I'm pretty excited about having a few new tools in the toolbox for when the car's running again. Which give me the confidence to tune the ignition timing, now that I've got some feedback to measure from. (Knock sensing setup, and this if it works alright)

 

[Bangbug]

It's where you need a nice long hill to provide resistance.
Great if you can max out in 4th but probably going to be 3rd or something.
Get her rolling at low rpm and into gear then just brick it till you run out of rpm.

Make changes and repeat

[mjrstar]

If your car won't max in 4th you need more horsepower.

 

I wonder if a minor trigger error (rpm signal)can cause the "waves" in the plot.

 

I have the super wavy plot from the latest runs, which could be wheelspin (or even perhaps clutch slip would skew results) I can post up but it's more of a Virtual dyno fail than the conservative timing maps at the same boost level which were much smoother... might have a tinker and report back..

[Lith]

Previously I've just been trying to suss fuel for the better part, and I've used the MAF and some Toyota documentation to figure out VVTI settings.

But I've been conservative on ignition timing as I've had no feedback loop for that.

 

While it's inexact, there is a heap to be said from calculating the airflow through the engine to work out how things have changed after making a physical change to the setup - obviously cam timing (or intake/exhaust mods etc)... I rely on all the data I can get my hands on when tuning on the road or dyno to try and build a picture of whats going on and where possible, how to make sure the engine is getting and doing what it wants.  "More tools in the toolbox" is a great way of putting it, the dyno is a tool.... not THE tool.   You can't get all the data you need to do a tune with a dyno, but it helps build the picture.

 

Since you mentioned tuning VVTi - this seems vaguely relevant.  I've done i-VTEC tuning but tbh after building >25 maps (no shit.  I am pedantic, and with a K-Pro  there are maps for different cam lift and cam timing states as well as maps for defining the states :/) I was all out of fucks to give to build graphs like this, so here is a much more straightforward and still entirely relevant comparison I did when tuning a Toyota with variable cam timing.  

This is from tuning a 20valve 4AGE with a mild cam upgrade - I fully tuned it with VVT off then rebuilt the maps with VVT on, whenever I am doing this kind of thing I track the changes in IPW versus AFR and MBT (where possible) at different loads just to make sure everything seems to be lining up, but when the difference with this car was so interesting I decided to record it all and make graphs because it was quite interesting.  

Firstly the dyno graph - green is with VVT on and red with it off:
 

 

 

Now the fuel requirements of the engine through the same rev range (obviously directly related to air flow):

And then "optimal" timing curves for the same thing:

It shows how much the volumetric efficiency lines up with the torque curve, and also how much the required timing is tied to cylinder filling as well - strictly speaking if you have a fuel reliable reference points for timing and fuelling it's not that hard to get a lot of the map very close to correct working from the trends you start building as you tune... and then future runs can be a bit more confirming that all is safe and well if road tuning, or if you have a dyno or virtual dyno just to confirm it's doing what you want and adjust if needed.  I rely on a lot of this kind of thing, especially when "road tuning" as really the less time at full load, the better

[Bangbug]

It's more 3rd on account of how fast you'll be going in 4th

[mjrstar]

So blue line was the run in tune with no timing above 4500.

Green line is with the new deeper compressor wheel, and a little bit of timing. Same bit of road.

16 psi no tweakage of the boost controller.

I do doubt the 280kw figure though?

[eke_zetec_RWD]

Tiny bit off track but I'm looking for a new set of injectors, I currently use these 880 ish cc injectors but they are too small. I run e85 350 whp.

http://www.siemensdeka.com/product/80lbh-siemens-deka-4-mototron-high-impedance-short-style-ev1-fi114700/

Does anyone know a direct replacement about 1400cc?

[Testament]

either you are going to Bosch(ID1000/samesame) 1000ccs with higher fuel pressure or to bosch 2000ccs if you want to stay high impedance. you can find china versions high impedance 1000cc on aliexpress for about $30 each so far they seem ok I think roman has used some? and I also have a set but havent used them.

 

the old indy blue is there at 1600cc but low impedance

 

 

 

the only others I know of are the rochester disc style might have something around 1600cc but not sure if they are high impedance or low

[downtrail]

What motor is it? 880s in a 4 cylinder should be ok for that power witg e85

[mjrstar]

Dont be afraid of big bosch injectors, my mate runs a set of 2200's on his 2j set to group fire and it idles like a gem.. the number of people i know that started on id1000's or equivalent and now run either id2000's or the bosch 2200's.

Go big first up and save yourself some hassle.

[eke_zetec_RWD]

That's my zetec at 0.8lambda. But it's at 100% above about 6800 rpm. Already have high fuel pressure too.