using air intake temp input for wideband datalogging.

[zin]

next year (once winter is over) i plan on installing the ostrich in my LM and the ftdi cable to the diagnostic connector and running both mptune and mpscan at the same time. mptune allows you to make real time updates to the ostrich. so assuming we can get the wideband signal running thru the cars computer i should be able to do some real time tuning. in theory we could set it up like efiLive does with the roadrunner chip so mptune (or a sub-module) would automatically modify certain cal tables through the ostrich based on feedback from mpscan.

That sounds great! It also sounds like it might make sense to merge the two programs, but I digress!

Is/was part of the problem with updating driving the pumpeff to 100% due to a feedback loop or am I not understanding the significance of the pumpeff going to 100%? Seems like it would/could go to more than 100% if the mods merit it.

Mike

[ShelGame]

OK, here's a test version of T/SMEC with a WB2O2 function built in.

Use the new .asm in the .zip file. Copy it to the same folder your working template (.tpl file) is in.

Then, open your .tpl file in a text editor and modify it as follows:

1) change the .asm asignment near the top of the .tpl file to this

;mpassmfile t-smec_v18x_WB.asm

2) copy this over the existing OPTN2

OPTN2_CalCustomOptionConfigFlags2:
    ;mpgrp Setup\Cal_Configuration\
    ;mptbladdr 0x01DF
    ;mptbltype 5
    ;mpsubtype n
    ;mptbllname CalCustomOptionConfigFlags2
    ;mptblcname OPTN2
    ;mpxaxis IN 0 255 X
    ;mpyaxis OUT 0 255 Y
    ;mptbldesc These_are_options_available_to_enable_in_this_custom_codebase.
    ;mpchoices Spark-based_Rev_Limiter  Fuel-based_Rev_Limiter  Suppress_EMR  NO2_Retard  NO2_Switch_Polarity  Alky_Injection  T2_Boost_Control  WB_Input
    .byte 0x02                       ;Point 1
    opt_WB2NB  			 = $%10000000	 ; Internal WB to O2 sensor conversion
    opt_T2BC			 = $%01000000	 ;
    opt_AlkyInj	     		 = $%00100000	 ;
    opt_NO2Pol 			 = $%00010000	 ; NO2 Retard Switch Polarity
    opt_NO2Retard	     	 = $%00001000	 ; Use Nitrous retard routine
    opt_Bit2		         = $%00000100	 ; Suppress the EMR (Emiss Maint Reminder) Code
    opt_FuelRevLim		 = $%00000010	 ; Use the fuel-based rev limiter
    opt_SprkRevLim 		 = $%00000001	 ; Use the spark-based rev limiter

3) add these 3 table entries at the end of the .tpl file (IE, Copy/paste)

AFR2NB_WidebandAFRToO2TransferFunction:
    ;mpgrp Setup\Transfer_Functions\AFR2WB\
    ;mptbladdr 0x1E98
    ;mptbltype 0
    ;mpsubtype n
    ;mptbllname WidebandAFRToO2TransferFunction
    ;mptblcname AFR2NB
    ;mpxaxis AFR 4 19.9375 ratio
    ;mpyaxis O2SensorVolts 0 5 volts
    ;mptbldesc Wideband_to_O2_transfer_function._Should_be_matched_to_the_sensor/driver_output.
    ;mpmapaxis y
    .byte 0x04                       ;Number of points in table
    .byte 0x00, 0x32, 0xFF, 0xF9     ;Point 1
    .byte 0x8C, 0x2E, 0xFF, 0x66     ;Point 2
    .byte 0xD0, 0x05, 0xFF, 0xF5     ;Point 3
    .byte 0xFF, 0x03                 ;Point 4

WB2AFR_WidebandToAFRTransferFunction:
    ;mpgrp Setup\Transfer_Functions\AFR2WB\
    ;mptbladdr 0x1EA7
    ;mptbltype 0
    ;mpsubtype n
    ;mptbllname WidebandToAFRTransferFunction
    ;mptblcname WB2AFR
    ;mpxaxis SensorVolts 0 5 volts
    ;mpyaxis AFR 4 19.9375 ratio
    ;mptbldesc Wideband_to_AFR_transfer_function._Should_be_matched_to_the_sensor/driver_output.
    .byte 0x02                       ;Number of points in table
    .byte 0x00, 0x52, 0x00, 0x9F     ;Point 1
    .byte 0xFF, 0xF0                 ;Point 2
    
NBOFFS_NarrowBandOffsetValue:
    ;mpgrp Setup\Transfer_Functions\AFR2WB\
    ;mptbladdr 0x1EB5
    ;mptbltype 19
    ;mpsubtype n
    ;mptbllname NarrowBandOffsetValue
    ;mptblcname NBOFFS
    ;mpxaxis in_desc 0 255 in
    ;mpyaxis O2SensorVolts -2.5 2.5 volts
    ;mptbldesc Offset_for_NB_lookup_value._Use_to_shift_the_sensor_rich_or_lean.
    .byte 0x80                       ;Point 1

What you'll get is 3 new entries in the template, under Setup - Transfer Functions - AFR2WB
I setup the tables according to your data, it may or may not work as intended with other WBO2 controllers.
Finally, I added an offset value so that you can 'bump' the simulated O2 signal up or down to richen or lean out the mix.

EDIT - OK, got the .ASM uploaded. I'll add some screen shots of the new values below...

[shackwrrr]

Thank you, I'll try making up one tonight and see what happens. I won't be able to do any wiring until Sunday.

[ShelGame]

.ASM is added above. I forgot to ask before, but you are using T/SMEC V18, right?

Once you get it all installed...

Step 1 - Enable the new feature:



You will also have to re-check any features you want on - when you copied the new OPTN2 definition, features you had selected may not be correct anymore. Please double check.


Step 2 - Actaully, there is no step 2. Maybe just compile and see if it works.

If you want to adjust the NBO2 offset value, it's located here (though I'd suggest leaving it alone at least to start with):



Let me know if you have any issues compiling. If you do, just post the .lst file here and I'll take a look.

You will probably have to use the new .lst file (assuming it compiles OK with your template) to make a new MP Scan setup to capture the new O2 controller values - RawAirFuelInput (same location as previous O2SensorVolts), as well as the O2SensorValue and AFRatio RAM locations.

I'm very interested to know if this works as I want to use it on my racecar as well...

[ShelGame]

Is/was part of the problem with updating driving the pumpeff to 100% due to a feedback loop or am I not understanding the significance of the pumpeff going to 100%? Seems like it would/could go to more than 100% if the mods merit it.

Mike

The problem with it driving the PEFTBL to 100% is 1) the table can't go over 100%; and 2) in the cases we tried, its just not right. The AFR wasn't that far off from the AFR setup. So there was a mismatch in the math somewhere...

[shackwrrr]

Yes v18, thanks ill try it out and see if it starts it.

[shackwrrr]

I got it and it compiled correctly. almost replaced the ;.org line after the OPTN1 section and figured id be best to put the other stuff before the .msg "complete". line. Worked the first time though with no errors.

Now on the tables, I was checking out the options and noticed some stuff that I wanted to make sure on. In narrowband offset table, should it default at -2.5? or should it be 0? Also on the wideband to O2 table, is that the "output" voltage that the system will see after the transfer function? I was wondering why you chose such a broad range for the "center"? From graphs Ive seen the point for .9 volts would be 14:1 and .1 volts at 15:1. I just didn't know if maybe mptune wasn't displaying it right or maybe it goes through another table after that.

[ShelGame]

I got it and it compiled correctly. almost replaced the ;.org line after the OPTN1 section and figured id be best to put the other stuff before the .msg "complete". line. Worked the first time though with no errors.

The ; .org line is just a remnant. I keep them in there just for reference. No biggie. Putting the other stuff before the .msg command is proabbly best, but it really didn't matter. That's just a message printed to the screen.

Now on the tables, I was checking out the options and noticed some stuff that I wanted to make sure on. In narrowband offset table, should it default at -2.5? or should it be 0? Also on the wideband to O2 table, is that the "output" voltage that the system will see after the transfer function? I was wondering why you chose such a broad range for the "center"? From graphs Ive seen the point for .9 volts would be 14:1 and .1 volts at 15:1. I just didn't know if maybe mptune wasn't displaying it right or maybe it goes through another table after that.

I used a NBO2 chart than I got from the internet so it's almost guaranteed that it's wrong, right? If you think the one you posted above is correct, then feel free to change the AFR2NB transfer function.

Basically, what I do is this -

- I take the input value, and immediately run it thru the new routine. If the option isn't enabled, then I skip the transfer functions and save that value to the new O2SensorValue RAM location.
- If it is enabled, then I take the raw sensor volts and covert that to an AFR using the WB2AFR transfer function. The AFR is saved to the new AFRatio RAM location.
- I then take the AFR and covert to a O2 volts using the AFR2O2 transfer function, and then this value is saved to the O2SensorValue RAM location.

The ECU uses the O2SensorValue for all feedback now. So, if it's wrong, the car will not run right.

EDIT: So, these are the O2 sensor limits:

O2SensorLimits		
	
O2HIGH_HighLimitForImmediateSwitchToClosedLoop	0x26	0.742
O2LOW_LowLimitForImmediateSwitchToClosedLoop	0x0f	0.293
O2TLMH_HighLimitForDelayedSwitchToClosedLoop	0x1f	0.605
O2TLML_LowLimitForDelayedSwitchToClosedLoop	0x0f	0.293
O2MIDH_HighLimitForInMiddleDetermination	0x1c	0.547
O2MIDL_LowLimitForInMiddleDetermination	        0x12	0.352

Using those, I think your NB table is right and mine is wrong. This will put the 'in-middle' values (which are the toggle values) right around 14.7.

[zin]

it's sounding like it may be possible to run closed loop all the time if the allowable range on the O2 inputs can be altered?

What I'm thinking is that if the parameters that control when closed loop is allowed were to be altered to allow it to run under more than just cruise/light load, the WB could be used to provide "block learning" to self-tune to a particular a/f ratio and/or simply maintain a given a/f ratio for a given condition.

Would that require a new table? One that would hold the target a/f ratio/voltage for a given load?

I may be getting a bit off on a tangent here, more food for thought at this point I imagine...

Mike

[shackwrrr]

I had another thought, another use for the charge temp input that could even be used as is with only table modifications, or made better with an addition of a timing factor based on charge temp volts too. My idea is to use an E85 sensor like this. That sensor outputs a 0-5v scale that relates to the percentage of alcohol in the fuel. We know that at 85 percent alcohol you need around 40% more fuel to keep the car running right so multiplying current fuel maps by 1.4 will account for the ethanol. The charge temp table is already set up to do this, As the voltage rises the factor by which the fueling is modified rises. All we would have to do is match sensor output with how much fuel is needed to keep ideal fuel ratios. This would keep fueling correct no matter what percent of alcohol you happen upon or mix up in your tank. If it is possible, another charge temp table could be added that modifies spark advance. There is already a table that controls boost by charge temp so higher boost could already be added with e85 use.

Hopefully I made it clear, Its pretty clear in my mind but getting that in a format that is understandable is pretty hard sometimes.

[ShelGame]

Did you get a chance to try this? Or were you waiting for something else from me?

[shackwrrr]

Ive just been really busy. Every time I go to work on it something else comes up or I waste the whole day on something else lol.

Ive been looking for some decent wire lately, The parts stores just sell that crappy shiny primary wire that melts if you look at it wrong and Ive pretty much picked all my wiring harnesses clean of any wire longer than 3 feet. A nice assortment of GXL spec wire is needed. I think I have enough to do the o2 sensor wiring though. Ill try sunday since I work by myself and it will most likely be pretty dead.

[wallace]

If you end up having to break down and purchase some wire I've used Waytek. they do have the automotive wire I think its txl?

[shackwrrr]

If you end up having to break down and purchase some wire I've used Waytek. they do have the automotive wire I think its txl?

Thanks txl and gxl are pretty much the same thing, just gxl has thicker insulation for underhood use.

[zin]

Always being one to enjoy the "inside baseball" kind of stuff, I'd love to learn more about these wire designations, links, etc?

Mike

[shackwrrr]

Always being one to enjoy the "inside baseball" kind of stuff, I'd love to learn more about these wire designations, links, etc?

Mike

Here you go this describes it fairly well. The normal stuff you buy at the parts store is GPT, it has a PVC insulation that doesn't take to soldering well. When you heat it the insulation shrinks back and swells.

An easy way to tell if wire is _XL wire or PVC wire is when you strip it the _XL wire's insulation will be white under the coloring. Our stock wire is a PVC based wire, the reason why the injector harnesses crack and get brittle, and why the insulation pulls back with age.
http://www.kayjayco.com/catPWireSelect.htm

[zin]

Many thanks!

Mike

[shackwrrr]

sorry about the long wait, I got the wiring done for this and re set up my MPscan layout(PITA). I will do testing tomorrow, I tried it earlier today but I didn't work but I didn't set up MPscan yet so I could see what it was doing. I may need to set up my wideband as I may have changed the output settings. Also the output has a ground ref that needs to be hooked up and I dont know if hooking that to the O2 ground is satisfying it, I may need to switch to the 10bit DAC that doesn't use a ground ref.

[shackwrrr]

It works! Just flashed it to the ostrich and let it warm up and started working just as it should.

The problem I was having was I had changed the output of the wideband to a different scale, I set it back and it works perfect. I can get a log of it working in a bit.

[ShelGame]

It works! Just flashed it to the ostrich and let it warm up and started working just as it should.

The problem I was having was I had changed the output of the wideband to a different scale, I set it back and it works perfect. I can get a log of it working in a bit.

Sweet, good to know...