DIY Fuel injection thread.

[Rookie]

I think you would be much better off with a power meter...

 

[Roman]

When I had the OBD 2 meter there were two types of reading.

 

Instant values - Your fuel economy at that instant, so it changes all over the show.

 

And averaged over a minute, or accumulating over your trip.

 

So one thing to consider is that if you're using the same amount of fuel but going faster, you get better liters per km.

So for instantaneous feedback it would be injector pulsewidth / vehicle speed
 

So something like (injector pulsewidth average over trip) / vehicle speed would show you the fuel economy over the span of the trip.

 

Would give you a number that you could bodge into Litres per km or whatever.

[yoeddynz]

There must be someone out there making such meters? Like in a 52mm standard fitting gauge size.

Like the fact that the bugatti gauge goes to 1001. Never would 1000 be enough.. must have over 1000.

[downtrail]

Our golf has a l/100km display I know it has a wideband so must use some of that data along with injector pw and load it probably does some calculation to figure out where your at

[Roman]

Got my 2nd lot of injectors fitted and all sorted with no leaks - hoorah!



A fuel line from a 20v 4age was the right length for between the rails.

I've set it so the second injectors only activate at 4500rpm onwards, and for starters set it to 90% using outside set of injectors.

Took it for a quick run and bodged some numbers around so it was 13:1 AFR across the board...

But then my fuel map looks whack!

 

I think the peaks and troughs are because there are resonances pushing some of the fuel back out of the runners at those rpms.

So I had a look at the injector timing.

It was setup so that the end of the injection event is 400 degrees before TDC combustion, for everything.

The first problem with having this set statically is that the intake cam advances or retards 45 degrees, so depending on the VVTI it would make sense that this number should be up to another 45 degrees advanced to match the closing of intake valve.
This doesnt matter too much with low cc injectors close to the head, because they're spraying most of the time on closed valves anyway.

But for this it needs to match the injector spraying with the induction event.

so I setup an rpm vs map injection timing table, and added more advance for when the cam advances.

This looks to have smoothed out the peaks and troughs with the fuelling back to normal, so that was probably the issue.

Jury's out though on whether these extra injectors will actually help make any more power.

There's no point in tuning it any further until my IAT sensor arrives though, it falls off tune each time I drive it currently.

They're a 1000cc single pintle and make a racket when they are running so it sounds pretty cool. Hah.

[Roman]

Ahhh so I bought a cheap GM style wideband knock sensor.

However it needs a PITA plug (not included, damn you) that you can only find on a late model LS2 or something, bugger.

 

Just going to epoxy some crimped ends in and run a remote plug end on it.

 

However I bodge wired it up to a Mic jack into the laptop, then as suggested used Audacity to crank it up a bit and output to Headphones.

 

Works brilliantly! It's even better than just using knock sensor > amplifier > headphones, because this way I can record the whole lot as a WAV file and compare notes with the datalogging timestamps once I've filtered the results for the knock frequency.

 

Tapping the knock sensor with a fork outputs a really clear sounding sound of... Well, a fork.

 

So it shouldnt be too much interpretation required to figure out when it's knocking.

[downtrail]

Thats how I do it too cool aye. Your laptop is the amp basically just be wary of charging laptop off car as the ground on the charger is the same ground the mic jack uses. Creates a sweet earth loop and sounds like a million bees in your ears lol

[Roman]

Haha! Yeah it is awesome, thanks for the suggestion.

My laptop cant charge from the car so no worries there.

 

[Roman]

Quick question about shielded earths,

 

Since the standard knock sensor normally has a shielded wire, but is only single wire and new knock sensor is 2 wire.

 

I'm using a spare loom's crank angle sensor wiring instead, as that's two wires shielded.

 

However what do I do with the shielded earth, do I just earth it to the chassis somewhere?

 

Not sure if they need something special.

[downtrail]

Earth the sheild with your ecu earths at ecu leave other end not connected at plug end but take it as close as possible to plug. All it does is act as a small antenna to drain and noise to earth before it gets a chance to mess with your signal wires

[flyingbrick]

Roman

http://m.ebay.com/itm/Knock-Sensor-Connector-Pigtail-Harness-LS2-LS3-LS7-Camaro-Corvette-Wiring-2-Wire-/201013928662

Does that help

[Roman]

Hey thanks,

 

Yes I saw those but problem is that I want to make sure my engine isnt knocking before Nats trackday and possibly Trackday this weekend too.

 

So didnt really have time to wait for Ebay stuff. I had a scout at Zebra, but even if there were any there, knock sensors are the worst thing possible to try and get at haha.

So I depinned a factory loom plug that slotted onto the pins, then epoxied the bejezus out of the whole lot. Just like a bought one... Hah.

 

Stoked that I've got a way to check for knock now though, I can have a bit more of a play with the timing as it's been on the conservative side.

 

[yoeddynz]

I'm keen on this. My factory mazda knock sensor has only one wire I think. I'll check. I've left it in place in case I wanted to use it. It's mounted right down in the V.

Have a picture of the knock sensor you're using dave?

[Roman]

Yeah most are single wire because they just earth to the block.

 

Which makes sense, but to use with headphones etc I think you need the earth coming back with the main wire.

Using this one:

 

http://www.aliexpress.com/item/GM-Knock-Sensor-Ignition-Knock-Detonation-Sensor-for-2013-Chevrolet-Malibu-12623095/1738801957.html

I dont have it bolted on the motor yet, but it pretty much works like a stethescope when you put it up against things.

 

It's Interesting to hear the injectors fire or you can hear the intake sucking through the side of the plenum and things like that.

Most of those noises couldnt be heard from the side of the block where the knock sensor normally goes, which is a good thing I guess!

[yoeddynz]

could I earth the wire for headphone jack at other end? So long as it earths?

[downtrail]

you could try but suspect you'll pick up all sorts of noise

[CorollaGT]

could I earth the wire for headphone jack at other end? So long as it earths?

 

Tried this when tuning my car but found it didn't work as well as I hoped. Gave up and bolted the sensor with the Link Knockblock onto the motor instead.

[Testament]

Ahhh so I bought a cheap GM style wideband knock sensor.

However it needs a PITA plug (not included, damn you) that you can only find on a late model LS2 or something, bugger.

 

Just going to epoxy some crimped ends in and run a remote plug end on it.

 

However I bodge wired it up to a Mic jack into the laptop, then as suggested used Audacity to crank it up a bit and output to Headphones.

 

Works brilliantly! It's even better than just using knock sensor > amplifier > headphones, because this way I can record the whole lot as a WAV file and compare notes with the datalogging timestamps once I've filtered the results for the knock frequency.

 

Tapping the knock sensor with a fork outputs a really clear sounding sound of... Well, a fork.

 

So it shouldnt be too much interpretation required to figure out when it's knocking.

man i probably have the plug on my cut up loom/too late lowl

[Roman]

This is a recording of the knock sensor audio through the rev range for those that were curious:

 

Picture unrelated but Youtube wont let you upload just a WAV file.

Will need to get someone else to drive for a bit and introduce some knock at low load / rpm to get an idea of what frequency it happens at and how it sounds.

One problem with using my laptop though... There's a noticable lag in the signal as it's software processed and my laptop sucks. Bugger!

So I might end up going back to an electronic stand alone amplifier instead as that should result in no latency.

[ProZac]

I've bene part of Canterbury's FSAE team the last couple of years, and last year we were sponsored access to Andre Simon's High Performance Academy series of videos.

 

They echo alot of the information in this thread, and are really worth the watch if you're starting out with Tuning. However, they are pretty spendy.

 

I would reccommend paying for the cheapest series of videos HPA does (I think it's 'Understanding Air Fuel Ratio'), as that account will then give you access to all the webinars and forums, which are pretty damn good viewing.