nom

Sorry about the strange title...
What I'm trying to do is to work out how to feed wideband sensor output into a Link. The Link doesn't do anything with the signal (not to control anything anyway!) as it runs open loop in all situations, but it's 'expecting' input from a narrowband, ie 0-1v, not 0-5v, & I'd like to be able to log the info.
So... there appears to me to be a couple of ways of doing things. If the Link can manage a 5v input for the lambda, then it can be wired to the linear output from the wideband's processor (from which it's very easy to get the AFR) & then scaled down within the Link's software. Or, if the Link can't take the 5v input, the same linear output can be reduced 'somehow' to a 0-1v range... the 'somehow' I'd be a little stuck on as electronics is not my thing
Or has anyone else got some fabulous ideas here? I'm guessing some have tried to do this? Or have found a way around it, anyway!
Ta

john banks

http://www.e-insite.net/ednmag/index...cleid=CA260066

http://www.google.com/search?sourcei...p+circuit+gain

Use a non-inverting op amp with gain of 0.2. You might need a dual op amp with an inverting buffer and use an inverting op amp to get gain under 1. Needs some research as I can't remember or never knew

Do some reading up through google and have a look at Maplin for suitable op amps. They are pennies. You will find something in an 8 pin DIL package to do what you want easily.

The other way - just buffer the signal and then put a potential divider as a variable resistor across the output. Use say 10k - the input impedance into the ECU will be far higher.

[Edited by john banks - 4/6/2003 11:21:29 AM]

Bob Rawle

If you are cute enough to do it then Johns sugestion may work but ..... Link is designed to use the factory narrow band sensor, and I will say, does a good job with it as well, it cant cope with a 5v sensor without some external help. I use a Motec PLM with separate display.

bob

nom

Well, with a whole mass of cables running about the place, I can log the info without the Link. It just seems to keep things tidier using it...
I think I'll give John's method a go & see how well it works (or rather, how well I work ) but keep the 'mass of cables' in mind as well!
Thanks John, & thanks Bob!

Denmark

iTrader: (2)

It will be nice with and update ones you have tried if it works,
as i would also like it to work that way.

Skassa

nom

Thought you might - I saw your request some time ago on - I think the MRT forums? Not a lot of response there! I'll speak to a mate who seems to have stacks of electronics piling up his place. He'll probably be able to get all the parts together easily, & we'll see how it goes... (I am a bit of a cretin when it comes to physically putting electronics together - theory's OK, but that's it!)

Denmark

iTrader: (2)

Are You done jet

Pavlo

nom

what version of the techedge do you have? You have the 1.1 with the linearised output? or just the older one with the output to the display?

I think even the version 1.0 had space on the PCB for an Atmel controller that could output a corrected voltage etc.

if you just want to scale the 0-5v to 0-1v then a potental divider would do it. The voltage output to the displat is buffered so you could take it from this no problem. could be as simple as a 5 or 10k pot.

the newer techedge 1.5 has a simulated narrowband output that you could use.

Paul

nom

Hi Paul - was wondering when you'd pop up
It's actually the 1.5 with the fangle-dangled outputs - the simulated narrowband is great for simulating a (what I think is probably perfect) narrowband sensor, as in it can be set to switch from 0v to 1v at 14.7:1. It outputs nothing but 0v or 1v, and the actual switch value can be changed (so it's possible to set the cruise value to 14.9:1 or whatever instead), but that's not very handy for what I'm trying to do! I'm actually just trying to log the wideband output without a zillion cables running about the place. This seemed to be a simple fix - a single cable from the linearised output from the 'black box' across another black box (to drop the voltage) and then into the Link, which does all the logging.
I have no idea what an Amtel controller is
I think I'm going to have a go with the potential divider. Should do the job - just need to check it's doing it accurately, but that's easy enough. Was going to use vastly vast resistors too, as I don't know what the resistance across the Link is... All I need to do is find 5x1 mega-ohm resistors somewhere! Something I don't usually keep kicking about the place

Pavlo

potential divider would be easiet with 5k ohm or 10k ohm linear pot. Then just calibrate using a voltage source and voltmeter.

Alternatively, you should be able to change some values on the PCB of the main unit to change the 0-5v output to 0-1v, no black boxes required.

If i weren't flat out this week (oh god give me more than 24 hours in a day) I would be happy to help. Wont be free til sunday though.

Paul

nom

Um, ok.
That's nice

nom

Right - an update! And I think I'm there...
I went for the simple option: taking the linearised wideband output from the 'black box' & running it across a potential divider (1M ohm & 250k ohm split) into the Link's AFR input. Aside from the initial confusion of it all being upside down (richer is lower numbers rather than higher now), I get values beween around 10 (hopelessly rich) & 95 (lift-off - beyond 25:1 - no fuel!), correlating very nicely it seems with AFR. The 'useable' area uses between '20' & '65' (approximately 11.5:1 to 15.5:1) so it's clear where the ratios are, being linear, etc.
Hope that helps someone out there...
(this is all with the TechEdge 1.5 wb unit & Link ECU, in case I didn't say earlier can't remember )

David_Wallis

have just ordered the tech edge for my new project..

is the Nband output backwards to normal or is it the link thats confusing things or am I half asleep..

Where did you get the NTK Uego Sensor from?

David

nom

The NB output is good - it can be adjusted to flick across at whatever AFR you want, so can go a little leaner than 14.7:1 or whatever - but it flicks rather than drifting about in the middle like the normal NB, making it useless for anything other than an ECU to cruise at!
The NTK EUGO is a bit of a £$%^*£%^& to get hold of. I got mine expensivly from Honda - seems that Honda is the only place it's possible to get them from now, directly or indirectly. UK (suprise suprise) is hugely more expensive than other parts of Europe. There's a bloke in Finland selling them at the moment - Nope, can't find it! They come up here if they come up anywhere! Ah - found the Finnish bloke too here.

Pavlo

scrub the linearised marrowband output for now, looks from the schematic that it's a analogue setup and not a lookup table in a micro controller that I originally thought.

looking some more, the best bet I think would be to frig the 0-5v output in the box after the PCB. A better way for the more technically minded is to put some feedback on the U3 output buffer to scale down the voltage. A few resistors and calibration pot on flying leads, glued to the board somewhere.

Paul

[Edited by Pavlo - 4/8/2003 3:51:13 PM]