TestO Datalogger - what do you think?

Pingu

Zorg Guru (III)
3rd Party Trader
Joined
Dec 8, 2011
Points
145
I was pointed towards a datalogger by @Murray Wall

The software is called TestO and some info about it (including download links) is HERE

I'm not a C++ programmer, but I have dabbled in Visual Basic and I was able to produce some graphs that make it very easy to see what is going on with your car. Both the TestO settings and the graphs need some tweeking, but I hope that you can see the possibilities, because I can.

TestO uses the EDIBAS Server, so if you have INPA, you can use TestO. If you don't have it, you can try the standalone version.

TestO

The software was originally developed to allow you to have multiple gauges on a tablet, or the like, but I think the datalogger is more useful.

There are around 100 parameters that you can download (or log) on the S50 engine (there are many more on a more modern car).

I have chosen to download the following datasets.(Click to open PDF). You will notice that the database is in German, but you can amend the text using a translator that is in the program directories.


Graphs

Exhaust VANOS

This graph (and the Inlet VANOS graph) can be used to see if your VANOS is working - and if so, how well it is working. If the Target VANOS graph is moving, this means that the ECU is sending a signal. If the Actual VANOS is moving, this means that the VANOS is working. How closely the Actual follows the Target tells you how well the VANOS is working. You want the lag to be as small as possible, but a small lag is acheived by having loose diaphragm springs. If the lag is excessive (as seen on my Inlet VANOS graph below, the diaphragm springs are possibly too tight - but I do have a much quieter engine. It's a compromise!!

Inlet VANOS


Ignition Cut-Out

My engine shows no signs of ignition problems, but if a spark plug or a coil pack were to break down, I would expect to see a step in the square graph. There are other data fields that could be downloaded to further investigate the problem if required.

Temperatures

This is very useful if you want to see the effect of modifying, such as fan delete or ceramic coating the exhaust manifold.

I've always cooled the car down after a hard drive by taking it easy for the last few miles. The evidence on the graph shows that I am actually heating the car up!! It would seem that in the cool-down phase, the oil temperature drops a little, whereas the coolant temperature and the air intake temperature increase significantly. I want to do some more experimenting to see what effect different actions have on the temperatures.

Throttle v RPM

This is another graph that can be used to measure how successful a modification is. Ideally, the Throttle %, the Air Flow Rate and the RPM should be overlaid exactly. This would never happen as it takes time for the air to move from plenum to the cylinder. A faster throttle response would squeeze the graph.

RPM v Speed

This graph would show how good you are at changing gear. The sample rate in the example shown is too low, but if I could download at 100Hz, I would be able to tell how much my acceleration slowed down when I changed gear (for the anoraks amongst us, not only should acceleration (a) be high, but ideally da/dt should also be as smooth as possible). Varying da/dt is usually shown by your passenger being flung forward in their seat when you change up. Ideally, they should be pinned to the seat during the gear change as well - this is impossible in a single clutch car.
 
Last edited:

billz

Zorg Expert (II)
Supporter
British Zeds
Joined
Dec 4, 2011
Points
226
Location
Langley Mill
Model of Z
Z4 3.0si Individual Ruby Black Last one produced in this colour combo
I was lost on the first graph =))
 

andyglym

Shiny Dust Caps Make Your Zed Go Faster.
Supporter
British Zeds
Joined
Feb 20, 2015
Points
231
Location
Moresby, West Cumbria, England
Model of Z
2.8 Roadster
Looks very intetesting stuff. Just read your link, I just don't understand how it works, not cut out for it :arghh: Thanks for the heads up though :thumbsup:
 

Redline

Zorg Expert (I)
British Zeds
Joined
Mar 10, 2015
Points
208
Location
Nuneaton
Model of Z
E89 20i msport
Fascinating read. This makes reading latched codes to find out what's going on a rather archaic way of diagnosing problems. How many parameters can you read out simultaneously? I guess the more you do the slower the sample rate you achieve.
Interesting that oil temp takes a lot longer than coolant to reach its max and notable that coolant blips upwards after you ease off.
 

Pingu

Zorg Guru (III)
3rd Party Trader
Joined
Dec 8, 2011
Points
145
Fascinating read. This makes reading latched codes to find out what's going on a rather archaic way of diagnosing problems. How many parameters can you read out simultaneously? I guess the more you do the slower the sample rate you achieve.
Interesting that oil temp takes a lot longer than coolant to reach its max and notable that coolant blips upwards after you ease off.
I was reading 55 parameters every 0.25 second, but it looks like the parameters were constant for 7 seconds. It looks like a different dataset was updated every 0.25 seconds until all the datasets were updated and the update process started again.

I need to do a lot more investigation about reducing the parameters to those required and trying to increase the update rate. I think the sample rate is appropriate for Temperatures and Misfire, but I'd want to get it down to 10 per second on the others. The downside is that would be 18,000 samples in 30 minutes!!!
 

Redline

Zorg Expert (I)
British Zeds
Joined
Mar 10, 2015
Points
208
Location
Nuneaton
Model of Z
E89 20i msport
I was reading 55 parameters every 0.25 second, but it looks like the parameters were constant for 7 seconds. It looks like a different dataset was updated every 0.25 seconds until all the datasets were updated and the update process started again.

I need to do a lot more investigation about reducing the parameters to those required and trying to increase the update rate. I think the sample rate is appropriate for Temperatures and Misfire, but I'd want to get it down to 10 per second on the others. The downside is that would be 18,000 samples in 30 minutes!!!
Data is data - it's the processor and interface bandwidth that constrains you. I guess from what you say you can set the sample rate per parameter. I'd have thought once per sec for some slow changing items like temperatures is more than adequate. Tying it together in the time domain is the trick.
Does the ecu push the data or it it pulled by the applications?
Are you post processing and graphing the data captured in flat files?

I wonder how long it will be before cars have all its subsystems communicating on an IP bus and you connect via a router and collect data via SNMP traps.
 

Brian H

Zorg Expert (I)
Supporter
British Zeds
Scottish Zeds
Joined
Dec 5, 2011
Points
205
Location
Killin
Model of Z
E36/7 3.0i - E85 Z4///M
I wonder how long it will be before cars have all its subsystems communicating on an IP bus and you connect via a router and collect data via SNMP traps.
We are almos at this point, the latest M2 collects all the car data and transfers it to an app on your phone, this is standard! If you track the car it will overlay the data on a track map :)
 

NZ00Z3

Zorg Guru (IV)
Supporter
New Zealand Zeds
Joined
Dec 9, 2014
Points
158
Location
Timaru, New Zealand
Model of Z
3.0L, 2.8L, 2.0L Z3 Roadsters
Great work @Pingu . You have done a lot of work to get this up and running.

I use testo for some simple checks of Vanos operation and do the data crunch manually. It's another tool in the toolbox that gets used for the complex issues.

Testo is a powerful tool and one that computer savvy people can put to use in all sorts of ways.

Edit:

To check the performance changes of your engine modifications:

- rev verses engine load is a surrogate for torque

- revs verses mass air flow is a surrogate for horse power.

See post 27 in the below link

https://www.bimmerforums.com/forum/...with-Ebay-headers/page2&highlight=Engine+load
 
Last edited:

Pingu

Zorg Guru (III)
3rd Party Trader
Joined
Dec 8, 2011
Points
145
Data is data - it's the processor and interface bandwidth that constrains you. I guess from what you say you can set the sample rate per parameter. I'd have thought once per sec for some slow changing items like temperatures is more than adequate. Tying it together in the time domain is the trick.
Does the ecu push the data or it it pulled by the applications?

I wonder how long it will be before cars have all its subsystems communicating on an IP bus and you connect via a router and collect data via SNMP traps.
I don't know whether the data is pushed or pulled. I imagine that it is pushed through an open door, as TestO doesn't work unless the car is powered (i.e. ignition ON).

Here's a PDF showing an extract from the dataset. It shows that the data is updated in 7 second "chunks", but is loaded at different times. At the moment, I don't know whether the upload is from a buffer or if it is instantaneous.

I'll have to do some analysis to see whether things that I know to happen at the same time are staggered or not.

Are you post processing and graphing the data captured in flat files?
The data is saved in a csv file by TestO. I load it into an Excel spreadsheet and manipulate it to present it in graphical form. The only clever bit is that I have written macros to make it easy to do. Rather than spending 10 minutes every time I want a different graph, I run a macro and the job takes seconds.
 

Redline

Zorg Expert (I)
British Zeds
Joined
Mar 10, 2015
Points
208
Location
Nuneaton
Model of Z
E89 20i msport
Think I'll stick to things I can hit with a hammer.=))
Best stick with a metaphorical hammer although the temptation to use a real one is an ever present threat :whistle::bag:
 

the Nefyn cat

Zorg Guru (IV)
Supporter
Joined
Oct 19, 2014
Points
174
Location
Actually in Nefyn. My, that took a while.
Model of Z
2.2i Sport
Got several hammers, none of them metaphorical. Just to win a bet many years ago I changed a lightbulb using a hammer and screwdriver, army Landrover with lenses that unscrewed. Damn, strayed off-topic there, hope no-one noticed:whistle:
 

Redline

Zorg Expert (I)
British Zeds
Joined
Mar 10, 2015
Points
208
Location
Nuneaton
Model of Z
E89 20i msport
Got several hammers, none of them metaphorical. Just to win a bet many years ago I changed a lightbulb using a hammer and screwdriver, army Landrover with lenses that unscrewed. Damn, strayed off-topic there, hope no-one noticed:whistle:
Who held the lens while you unscrewed the Land Rover? =))
 

Pingu

Zorg Guru (III)
3rd Party Trader
Joined
Dec 8, 2011
Points
145
Got several hammers, none of them metaphorical. Just to win a bet many years ago I changed a lightbulb using a hammer and screwdriver, army Landrover with lenses that unscrewed. Damn, strayed off-topic there, hope no-one noticed:whistle:
Sputniks or those new fangled ones that need a plumber's wrench ;) ?

I've done a few oil filter changes with a hammer and screwdriver.
 

Brian H

Zorg Expert (I)
Supporter
British Zeds
Scottish Zeds
Joined
Dec 5, 2011
Points
205
Location
Killin
Model of Z
E36/7 3.0i - E85 Z4///M
I don't know whether the data is pushed or pulled. I imagine that it is pushed through an open door, as TestO doesn't work unless the car is powered (i.e. ignition ON).
The data will be present at all times (pushing) to the cars onboard controllers, you have configured TestO to sample the data therefore tapping in to an existing resource.
 

the Nefyn cat

Zorg Guru (IV)
Supporter
Joined
Oct 19, 2014
Points
174
Location
Actually in Nefyn. My, that took a while.
Model of Z
2.2i Sport
Sputniks or those new fangled ones that need a plumber's wrench ;) ?

I've done a few oil filter changes with a hammer and screwdriver.
Guess they'll be Sputniks, fins all over them, seemed to be fitted to everything Army back in the seventies. Have indeed changed a few filters with hammer and screwdriver, teaches them a lesson they won't forget. And still got the chain-wrench I made out of my old Morris 1000 timing chain when the rattle got too loud to ignore, a bit more subtle than the hammer way.
 

pheno

Newbie
Joined
Jun 15, 2017
Points
3
Hi guys!

Gary pointed me here :) Looks like a nice forum you have here.

I wrote something about the data rate to the other discussion. But I assume you're using MSS50 ? if so then I suggest trying the 'Serial' version of the ecu, named: MSS50 (Serial). It is different from Ediabas jobs in a way that it interprets all data points in the reply message from ecu. Ediabas job usually extracts only one value. It is simpler but much more slower. For 'Serial' version you'll have to change serial port in testo.ini .

This 'Serial' version of ecu is available for all MS4x ecus. Diesel owners (DDE) and newer gasoline ecus (like MEVD, MSS60, etc) are much happier because they can use 'custom jobs' where you can build a custom set of parameters.

Thanks!
 

Ianmc

Zorg Guru (IV)
British Zeds
The M44 Massive
Joined
Oct 12, 2014
Points
165
Location
New Forest
Model of Z
Z3 (M44)
Hi guys!

Gary pointed me here :) Looks like a nice forum you have here.

I wrote something about the data rate to the other discussion. But I assume you're using MSS50 ? if so then I suggest trying the 'Serial' version of the ecu, named: MSS50 (Serial). It is different from Ediabas jobs in a way that it interprets all data points in the reply message from ecu. Ediabas job usually extracts only one value. It is simpler but much more slower. For 'Serial' version you'll have to change serial port in testo.ini .

This 'Serial' version of ecu is available for all MS4x ecus. Diesel owners (DDE) and newer gasoline ecus (like MEVD, MSS60, etc) are much happier because they can use 'custom jobs' where you can build a custom set of parameters.

Thanks!
I wish I wasn't so dim, haven't a clue what this meanso_O
 
Top