Tuner Studio and O2 voltage mismatch

[markvwguy]

Couple weeks ago I pulled out my Megasquirt setup from boxes. I sold my DBL cab, but kept all my MSQ stuff. Sitting for packed for 3 years. I put about 10,000 miles on this setup to prove to myself Megasquirt was solid. It ran great very low oil temps and smooth and powerful.

What I have is a voltage mismatch between TS and the Innovate LM-1. All the way to the AD5-1(pin 28) I have Higher voltage in to the CPU than what TS is displaying on front bank voltage 1. My AFR reads richer than my LM-1 also.

This is the data,
The LM-1 is set for 2.34v = 14.70:1 AFR crossing.
The Jimstim set for 2.34v = 14.71:1 AFR
Both yield the same voltage at 14.7 to the CPU (AD5-1) 2.34v EXACTLY

Tuner Studio shows 1.50v = 14.7:1 AFR
I see this difference with the engine running. Before measuring everything; I predicted .5 to .8 volt differential. I guess I was correct.

Why? Any ideas?? It can't be resistance mismatch this is at the CPU pins and two separate source devices. Has to be software??
Can I custom make my slope to make sure my 14.7 crossing is 2.34v in TS?

Thanks,
Mark

What I have is:
TS MS Lite v2.6.19
v3.0 board MS I (extra code 029y4)
Gen wheel direct drive dual coil ; set for 0-5v Innovate LSU4.2 wide band
German plenum Vanagon 2.1 injectors

[miniman82]

I've been out of the tuning game for a while, but if I'm not mistaken you need to set TS to accept the correct signal by telling it which wide band you're using. I am using a techedge wide band, so I tell it that's what I'm using and it works fine.

[markvwguy]

Yeah, me too I took a few years off of the tuning thing too.
I am telling the ECU I have an innovate wide band. I did see the a selection for "custom linear wideband" which I have selected. Didn't seem to change the reading.

I know its only MS1.... but it should read close. Given I'm seeing the correct voltage in the ECU input leg at the ECU.

I just did some Analyze Live tuning today after changing my Spark and Fuel target tables. I tell you that Analyze Live is the best investment!!! So awesome!!
The fact here remains Tuner studio reads the AFR voltage .8 volts off(lower). I can rescale the whole table to compensate, but I want TS to read what is on the input leg. 2.34v NOT 1.50v at 14.7.

My next Megasquirt build is to start soon in the buggy, excited about that!!

Thanks for the response!

[Clonebug]

If I remember correctly...my Innovate LM-1 was off a little bit too.

It finally died so I bought a PLX AFR with dash gauge instead and it reads much better. The fact it has a gauge in the dash instead of the big box I had to mount before was a great selling point in itself.

For the $179.00 I paid I couldn't even bother trying to get the Lm-1 fixed.

[markvwguy]

Or reduce the Required Fuel ( instead of scaling the VE table) to mask the voltage differance.
The down side is TS will show the AFR in the red about all the time. The Innovate LM-1 would show correct.
From TS to LM-1 AFR the voltage difference is about 2% i.e. 12.5 = 14.5 ish from what I see.

That makes a hugh difference in tuning! Still stumped

Mark

[Piledriver]

Nonononono....

ANYTIME you have an analog input it will have some difference.
A common ground is a MUST, the heater ground should go elsewhere, but even then you are likely to have some error.
Its not a flaw in the wideband controller
its not a flaw in the Megasquirt
it is simply a case of how stuff really works---a very common problem in electronics, and why the "custom linear wideband" setting exists..

I thought this was documented somewhere, but thinking about it maybe I'm the only one explaining it.
This will only work for linear output widebands.

This is a lot easier to do than explain...

Alan To (14pont7.com) added a 2 step "cal signal" (producing two known AFR voltages to cal against) start up routine in his recent widebands after I asked about/explained the issue and my method on his forums awhile back. He came up with a Better Way, (even his Spartans have it) but here's mine:

Doing this from memory, so I reserve the right to come back and edit this:

If you have any electronics background, all you are doing is scaling the zero and span of an input, digitally.
That's what the "custom" WB numbers are for.

Selecting "custom linear wideband" is the first step. for ANY analog input wideband.(if you want it accurate)

Your wideband may be 0-5v or 1v-5v or 1v-4v---doesn't matter, the custom entry will allow you to scale the zero and the span to match its outputs at its AFR limits, whatever they are.

I did this with an old JAW 1.03 (10:1-free air, also has serial comms) and a newer 14point7.com SLC-OEM EVK analog or I2C,(10-20:1) it will work on your setup.

For the record, Megasquirt 1/2/3 whatever is "done" on AFR input at 25.4:1 as its an 8 bit real number (0-254). It will see/target no higher, and weird may occur if you try.(it will probably just rail at 25.4, but no promises)

In the "custom" config box,
the top line is zero
bottom line is span,

Hmmm. Have to be hooked to the ECU to see anything useful.

Short explanation:
1) You have a WB controller with serial and analog. YOU ARE GOLDEN.
All you have to do is use a second serial port to see what the digital reading is, drive the AFR (via twiddling the idle VE or whetaver) to the lower and upper AFR limits, and tweak the numbers in the "custom" entry so the MS and the digital side readings match at the up[per and lower limits. Takes a minute or so.

2)If you are analog only, its not much harder--- Your WB will only output analog over a given range, for most folks 10-20 AFR.
Fuel cut is the upper limit (span), so it should ALWAYS read 20:1 on fuel cut, adjust the bottom line so this is so.
Upper line (zero) is a little dicier, as you have to feed enough fuel to get to 10:1, but sneak up on it so you don't flood it.

If you have a wideband that reads (really reads to free air, not free air cal) to free air like a JAW or SLC-pure or such, remember the upper limit of megasquirt is 25.4:1, so use that as the upper limit when doing your cal#, or use 20:1 if your engine can run there.
(for most folks 20:1 is a reasonable upper limit, although i cruised nice and cool at 22:1 for ages when I had my JAW installed)
BTW, the MS will not allow target a AFR higher than the upper limit of the WB you have set.

Those units have digital (serial) outs so its a bit of cake as long as you have a 2 channel serial adapter, i use a Microconnectors dual unit, but two single channel USB adapters will works as well.(but will probably cost more)

[jhoefer]

Couple weeks ago I pulled out my Megasquirt setup from boxes. I sold my DBL cab, but kept all my MSQ stuff. Sitting for packed for 3 years. I put about 10,000 miles on this setup to prove to myself Megasquirt was solid. It ran great very low oil temps and smooth and powerful.

What I have is a voltage mismatch between TS and the Innovate LM-1. All the way to the AD5-1(pin 28) I have Higher voltage in to the CPU than what TS is displaying on front bank voltage 1. My AFR reads richer than my LM-1 also.

This is the data,
The LM-1 is set for 2.34v = 14.70:1 AFR crossing.
The Jimstim set for 2.34v = 14.71:1 AFR
Both yield the same voltage at 14.7 to the CPU (AD5-1) 2.34v EXACTLY

Tuner Studio shows 1.50v = 14.7:1 AFR
I see this difference with the engine running. Before measuring everything; I predicted .5 to .8 volt differential. I guess I was correct.

Why? Any ideas?? It can't be resistance mismatch this is at the CPU pins and two separate source devices. Has to be software??
Can I custom make my slope to make sure my 14.7 crossing is 2.34v in TS?

Thanks,
Mark

What I have is:
TS MS Lite v2.6.19
v3.0 board MS I (extra code 029y4)
Gen wheel direct drive dual coil ; set for 0-5v Innovate LSU4.2 wide band
German plenum Vanagon 2.1 injectors

Tuner Studio scales the values according to the wideband calibration table. Sounds like you don't have the correct calibration to match what your LM-1 is actually doing. Because the LM-1 has programmable outputs, you haven't included enough information here to determine what the TS calibration values should be. Post the settings out of your LM Programmer software.

[markvwguy]

Thanks for the replies,
I'm just getting back to look at the post. Really good info this does sound like scaling.

First thing I'm not seeing a "custom" con fig box for doing this.

For setting the Custom linear wideband, file path is File_ Project_ Project Properties_ Settings_ ( First selection EGO sensor selection) Have selected Custom linear wideband.

From there I can't find any where to set scaling.

[Piledriver]

Overlooked the MS1 part, MS2E/MS3 I know have the scaling.
If MS1 had it it would be in the pulldown, a place to put afr vs voltage on two lines.

MS1e is comparatively old school and may use a different method, like directly editing the .ini.
On MS2E/MS3 its ~interactive so you can save the changes and see the results.

I have only run MS1 long enough to verify the mainboard worked before plugging in an MS3.
I'll see if I can still find the MS1 CPU to be of some help.

[markvwguy]

I found it. Went to TOOLS than to "calibrate AFR". Than "custom linear wide band" selections are present.

Haven't changed anything yet. It's 5am and have to go to work. See when I get home.

[markvwguy]

I found this on 14point7.com blog It's a very good read.

http://www.14point7.com/blogs/news/1653 ... megasquirt

[Piledriver]

NICE!

THAT needs to go in a sticky at the top of the Engine Management or Megasquirt forum.

You can do essentially the same thing if you know your min and max values of your controller, and drive it to its limits.
(Fuel cut works for the upper)

You can also do similar if you have a WBO2 with digital out: the digital data is going to be correct.