Open Source LSU4.9 Wideband Controller

Free Open Source Hardware discussion forum. Post your Free Open Source hardware projects here!
joey120373
QFP80 - Contributor
Posts: 39
Joined: Wed Oct 01, 2014 7:06 am

Re: Open Source LSU4.9 Wideband Controller

Post by joey120373 »

Nothing yet, I still have the boards, one assembled, I have a standard form factor arduino uno shield designed, using all through hole parts, but I haven't ordered it. My bad, work has kept me very busy these days( as I type this I'm at work, it's Tuesday at 2pm, and I already have 50 hours on the clock since Sunday morning ).

I want to get back to this project, but I've been spending what little spare time I do have on an SLC-OEM ( www dot 14point7 dot com ) interface and a wireless display module. That project is coming along nicely, I have another thread on this forum about that project.

Now that it's mostly done I'll try to get back to an arduino WB interface. And I should probably order the boards I designed like 6 months ago.....
toc
TO220 - Visibile
Posts: 13
Joined: Mon Jan 11, 2016 11:26 am

Re: Open Source LSU4.9 Wideband Controller

Post by toc »

Did anyone get Hell Fire's design working 100%?

What about joey120373?
Did the PCB end up working with the ground issue sorted, or was Hell Fire's code problematic??
I'm keen to see this through to completion, and confirm it works.
toc
TO220 - Visibile
Posts: 13
Joined: Mon Jan 11, 2016 11:26 am

Re: Open Source LSU4.9 Wideband Controller

Post by toc »

I've knocked together a design based on Hell Fire's schematic and Waltech's one.

Where I struggle at the moment is Waltech's schematic, has the terminal for IP at the bosch sensor plug direct from the output of the op amp if I follow the schematic - output of LM324 pin 7 goes to a 62ohm and then meets RCal, but before the 62 ohm, also goes straight to the IP terminal.

Now, Hell Fire conflicts with this a tad:
His pump cell filter / buffer has the output of the opamp (pin 8!) going to a 61.9 ohm resistor, then this goes on to the bosch terminal and this also goes on to the pump cell current meter amp via a 22k resistor.
The difference between the two is the Bosch IP terminal is either provided through the 61.9 ohm resistor, or direct off the output of the op amp. I'm not sure which is right at the moment.
Which way is right?

Resistor sizes - Between Waltech and Hell Fire some differences here - Hell Fire has used 470k resistors on the signal, so as to not load them down???? Where Waltech doesn't seem to.
Any thoughts here?
Joey120373, with your PCBs, did you have a schematic I could perhaps refer to, to do some reconciliation here, or did you run with Hell Fire's design?
toc
TO220 - Visibile
Posts: 13
Joined: Mon Jan 11, 2016 11:26 am

Re: Open Source LSU4.9 Wideband Controller

Post by toc »

Bump (any thoughts)?
joey120373
QFP80 - Contributor
Posts: 39
Joined: Wed Oct 01, 2014 7:06 am

Re: Open Source LSU4.9 Wideband Controller

Post by joey120373 »

toc, Ill have to look when i get home, I think i based my design off of the hellfire schem
joey120373
QFP80 - Contributor
Posts: 39
Joined: Wed Oct 01, 2014 7:06 am

Re: Open Source LSU4.9 Wideband Controller

Post by joey120373 »

OK, finally got back on this project.

Toc, sorry, the last set of schems i drew up ( for an arduino uno shield ) were based on another set of schems entirely.

I dont want to post them just yet cause im not sure if that is ok to do, and they dont pertain to the Hellfire design.

I wired my small board up to a UNO to do some testing with Hellfires code and i found that my Vground was oscillating at over 10K htz with a 1 volt amplitude!

So that was causing all kinds of issues. I removed C1 and it settled down but the unit still didnt work, my hunch is that the op-amp i chose is not well suited for this purpose.

I am going to go back to the drawing board with this, and build a through hole sheild exactly to hellfires schem and BOM.

Hopefully then we can make some real progress.

Gotta run

Joe
toc
TO220 - Visibile
Posts: 13
Joined: Mon Jan 11, 2016 11:26 am

Re: Open Source LSU4.9 Wideband Controller

Post by toc »

If the Op Amp was oscillating, I'd suggest the load was not enough? A bigger cap / lower value resistors?
joey120373
QFP80 - Contributor
Posts: 39
Joined: Wed Oct 01, 2014 7:06 am

Re: Open Source LSU4.9 Wideband Controller

Post by joey120373 »

I should point out that the op-amp I used for my small board is the same op-amp that Hellfire said he used, however, op-amps being what they are, they can be temperamental. Also should point out that mentioned out that C1 could be problematic.

I still need to remove R4 and re-test, not sure if Hellfires code revision accounts for R4 being in the circuit or not.

The more I look at wide band controller schemes, the muddier the water gets. I'm not sure why the Ip drive current is being generated in the controller via a PID loop ( I think this is what is happening on both Hellfires design and with Alan's stuff ) to me a simple op-amp would do this job much more elegantly, this is why I drew up another schem to play with. My first Wideband was completely analog, and while it was not perfect, it was very, very fast, and that was extremely usefull. In my feeble brain, the ideal WB would be entirely analog, with a micro controller supervising just the heating element, however that is a discussion for another thread I suppose.
toc
TO220 - Visibile
Posts: 13
Joined: Mon Jan 11, 2016 11:26 am

Re: Open Source LSU4.9 Wideband Controller

Post by toc »

Isn't the IP current an AC current?
Or is it AC just by virtue of the sensor going lean / rich?
joey120373
QFP80 - Contributor
Posts: 39
Joined: Wed Oct 01, 2014 7:06 am

Re: Open Source LSU4.9 Wideband Controller

Post by joey120373 »

Yes, as i understand it, the current is AC with reference to VGND.
this, however is completely dependent on weather or not the O2 sensor is in a rich or lean environment.

This is why there is a VGND.

As i understand a WB O2 it sensor ( and please understand i am no expert ) is nothing more that a standard O2 with an extra circuit added on, called a nernst cell.

So, to understand how it works, one must look at a standard O2 sensor operation. A standard O2 sensor, when heated to a certain temp, generates a specific voltage relative to the amount of O2 in the exhaust gasses versus the amount of O2 in free air. At a lambda of 1, a standard O2 will generate a voltage of 0.45v.
A rich mixture ( less O2 in the exhaust ) will generate a voltage of > 0.45, due to the difference in O2 between the exhaust and free air, and a lean mixture will generate a voltage < 0.45v because the ratio of O2 in the exhaust is closer to that of free air. The problem with a standard O2 sensor is that it is only accurate a a lambda of 1, or at least very near 1. Rich mixtures will result in a higher voltage, and lean mixtures a lower voltage, but there is no way to accurately determine how rich or how lean the mixture is.

Enter the wide band 02, it adds another layer (cell) onto the standard O2 ,the nernst cell, that tries to keep the standard O2 cell generating a voltage of 0.45v (where it is most accurate) by feeding in, or removing electrons ( current ) to compensate for the exhaust gasses. by measuring this current, positive or negative, ( generally by measuring the voltage drop across a 61.9 ohm resistor ) an accurate representation of air fuel ratio can be measured.

because current must be pumped into or out of the nernst cell, it is referenced to a "VGND", or virtual ground. As we are talking about a 5 volt system, lets assume we will put VGND at 2.5 volts . It can be higher or lower than this, but for ease of discussion we will set it at 2.5 volts.

This "VGND" is the "ground" for the standard O2 part of our wide band O2, such that when the standard O2 is at a lambda of 1, the output of the standard cell is 0.45v + 2.5v relative to actual ground(2.95v). because the nernst cell shares this VGND, we can "pump" current into the cell with a voltage greater that 2.5v, or "pump" current out of the cell with a voltage less than 2.5v.

I'm sure i am butchering this, but bear with me.....

If we can pump current into or out of the nernst cell to keep the voltage coming from the standard O2 cell at VGND + 0.45v ( 2.95 volts ) and measure the current, positive or negative, we can then determine with a degree of accuracy how rich or lean we are.

This is where an op-amp comes in so elegantly. An op amp does whatever it can to keep its inputs at the same voltage, so if we feed a set voltage of 2.95 volts into its positive input, tie its negative input to the the output of the "standard O2" part of our WB sensor, and tie its output to the pump section of the nernst cell, it will do everything it can to keep the voltage on its negative input ( from the standard O2 cell ) at 2.95 volts by pumping in or pumping out current to the nernst cell. In this way the WB o2 sensor is part of the op-amps feedback loop.

If we put a 61.9 ohm resistor ( as per the bosch data sheet ) in series with the op-amps output pin, we can measure the current into or out of the nernst cell to determine how rich or lean the exhaust gasses are.

I hope that i have explained this correctly ( i am definitely NOT an expert ! ) , or at the very least have not completely missed the mark on how WB-O2 sensors work.

joe
Post Reply