AFR controller

[Escort Paul]

has anyone made a DIY AFR controller or recommend an accurate one for use with a EMS?


regards Paul

[Fred]

I put my money on Alan's JAW and sister products for value and if you want a commercial boxed shiny item, LC1 all the way.

Fred.

[Escort Paul]

Yeah I like JAW but have been told they are a bit high on latency...

[Fred]

"and sister products" - Did you check out his other offerings yet? Definitely worth a look IMO. :-)

[DaWaN]

The new PSoC based Wideband controllers from Alan are quite nice, the only problem is that the LC-1 is actually quite cheap, so if you don't flip every dollar I should go for the LC-1.

[jharvey]

Well that's cool.

http://www.innovatemotorsports.com/products/lc1.php

Just wait until the PSoC5 comes out.

[toalan]

I am currently evaluating PSOC3 for wideband, it is quite a nice piece of silicon. Can stuff 2 Wb channels into 1 chip. PSOC5 has almost the same peripheral set as PSOC3, but the core is just much faster. With the PSOC3/5 peripherals I can get 1/100k - 1/1million lambda accuracy, but the sensor is only rated for 1/100 lambda, but that is due to the 1% calibration gas bosch uses to calibrate the sensors from factory.

PSOC5 is overkill, the core is very fast but the limitation is how fast you can sample and feed the PID control algorithms. PSOC1 only offered 3 MIPs and that is enough to keep up with the sample rate. PSOC3 has ~30 MIPs + DMA so you can easily control 2 sensors and have enough power left over to drive a graphics display. PSOC5 should have 70-100 MIPs, maybe you could run uclinux on that ontop of WB control, you have to be doing a really exotic lambda controller to use that type of power.

[jbelanger]

Does it have enough peripherals to control more than 2 sensors? From what you say, it certainly has enough power to drive more and a 4 channel controller would be very nice for sequential injection and setting individual cylinder fuel trim.

By the way, have you performed any formal tests with your controller such as comparing it with a lab grade sensor?

Jean

[jharvey]

Another feature of the PSoC5 is the ARM core. ARM's are known for being very price competitive. It wouldn't surprise me if the PSoC5 ends up being lower cost than the PSoC3.

Keep us posted, sounds like fun. I have the cube, but haven't fired it up since the original PSoC days. Can that program the new gen PSoC's? If so, you might have a beta tester waiting out here.

[toalan]

Does it have enough peripherals to control more than 2 sensors? From what you say, it certainly has enough power to drive more and a 4 channel controller would be very nice for sequential injection and setting individual cylinder fuel trim.

By the way, have you performed any formal tests with your controller such as comparing it with a lab grade sensor?

Jean

I said 2 channels because there is enough analog peripherials to drive 2 of them, more than 2 and you have to start using eternal components which takes away the elegance of using psoc in the first place.

All I have is lab grade gas and that is what I use to verify accuracy.