Homemade NON dithering A/F meter
08-22-2003, 04:38 PM
#1
Drive-by provocation guy
Rennlist Member
Rennlist Member
Thread Starter
Join Date: Apr 2002
Location: NAS PAX River, by way of Orlando
Posts: 10,439
Likes: 0
Received 0 Likes
on
0 Posts
Homemade NON dithering A/F meter
08-22-2003, 05:57 PM
#3
Addict
Rennlist Member
Rennlist Member
I see two issues with this circuit.
1. The resolution with only 10 LED's, about 0.1 V per LED, is going to be very poor. Basically one LED will cover the entire useful part of the rich region from about 13:1 A/F and lower. Another LED covers the entire lean range from 15:1 and higher. Even the Autometer A/F gauge gives you 20 LED's.
2. The circuit which smooths out the dithering will also make it slow to respond to rapid changes in A/F ratio, when you suddenly go to WOT from cruising or vice versa, for example. It will always be lagging what's actually happening.
1. The resolution with only 10 LED's, about 0.1 V per LED, is going to be very poor. Basically one LED will cover the entire useful part of the rich region from about 13:1 A/F and lower. Another LED covers the entire lean range from 15:1 and higher. Even the Autometer A/F gauge gives you 20 LED's.
2. The circuit which smooths out the dithering will also make it slow to respond to rapid changes in A/F ratio, when you suddenly go to WOT from cruising or vice versa, for example. It will always be lagging what's actually happening.
08-23-2003, 01:12 PM
#4
Racer
Join Date: Aug 2002
Location: San Diego, CA
Posts: 457
Likes: 0
Received 0 Likes
on
0 Posts
Doesn't dither? Not quite how I'd put it...
The input filter does two things: (1) it buffers the input with (technically) an infinite impedence, so there is no load on the sensor (i.e. the meter doesn't interfere with the computer's measurement from the sensor). (2) It passes the O2 signal through a low pass filter with a cutoff frequency in the neighborhood of 0.5 Hz. This is just a 'noise supressor' which will remove any AC component from the signal, and consequently not let it change at a frequency greater than 0.5 Hz which is still plenty fast in the analog domain. If you want to disable this 'feature', you would build the thing without the capacitor and resistor (and perhaps OpAmp) after the first OpAmp that the O2 signal connects to. No big deal. I'd leave it in, personally. If you were to actually build this thing and run it through a signal generator, you'd notice that it would still work well.
The thing that I don't like about this home-brew approach are the astethics. That thing is ugly.
If you want more than 10 LED's, you can chain another LED driver on to the one that's there (the datasheet tells you how to do this) and run as many LED's as you want (each of the drivers has a 'carry out' line, essentially). Note that care will have to be taken with the last red LED since it is not driven off the driver, but rather it is controlled with a current sense circuit so that it will go on if none of the other LEDs are lit (since the LED driver will require a minimum voltage to turn the bottom LED on).
If you wanted to make this in to a Knock sensor, just make a full wave rectifier type circuit and hook it up to the LED driver instead of the O2 signal (this is how the KnockLink is designed)... or you could use an integrator circuit. The only change would be that you would want to use a bandpass filter instead of a low pass filter, tuned to around 2kHz - 18kHz... and some tinkering would be necessary to get the gain right so that it lit the LEDs appropriately (the KnockLink uses a simple voltage divider with a variable resistor, but you could also use a negative feedback OpAmp conviguration with a variable resistor)
If anyone wants just the signal conditioning part of the circuit, I can make you one for the cost of parts (about $5 for the chip, resistor, capacitor, and something to solder it all to)... let me know if you want one.
The input filter does two things: (1) it buffers the input with (technically) an infinite impedence, so there is no load on the sensor (i.e. the meter doesn't interfere with the computer's measurement from the sensor). (2) It passes the O2 signal through a low pass filter with a cutoff frequency in the neighborhood of 0.5 Hz. This is just a 'noise supressor' which will remove any AC component from the signal, and consequently not let it change at a frequency greater than 0.5 Hz which is still plenty fast in the analog domain. If you want to disable this 'feature', you would build the thing without the capacitor and resistor (and perhaps OpAmp) after the first OpAmp that the O2 signal connects to. No big deal. I'd leave it in, personally. If you were to actually build this thing and run it through a signal generator, you'd notice that it would still work well.
The thing that I don't like about this home-brew approach are the astethics. That thing is ugly.
If you want more than 10 LED's, you can chain another LED driver on to the one that's there (the datasheet tells you how to do this) and run as many LED's as you want (each of the drivers has a 'carry out' line, essentially). Note that care will have to be taken with the last red LED since it is not driven off the driver, but rather it is controlled with a current sense circuit so that it will go on if none of the other LEDs are lit (since the LED driver will require a minimum voltage to turn the bottom LED on).
If you wanted to make this in to a Knock sensor, just make a full wave rectifier type circuit and hook it up to the LED driver instead of the O2 signal (this is how the KnockLink is designed)... or you could use an integrator circuit. The only change would be that you would want to use a bandpass filter instead of a low pass filter, tuned to around 2kHz - 18kHz... and some tinkering would be necessary to get the gain right so that it lit the LEDs appropriately (the KnockLink uses a simple voltage divider with a variable resistor, but you could also use a negative feedback OpAmp conviguration with a variable resistor)
If anyone wants just the signal conditioning part of the circuit, I can make you one for the cost of parts (about $5 for the chip, resistor, capacitor, and something to solder it all to)... let me know if you want one.
