Scanning TheBard's car: Input would be appreciated! :)

[Aurora40]

If anyone would be interested in offering their knowledge and experience up, that would be great! I went out with TheBard yesterday and we got a bunch of scans of his car running. It does seem his car is running nicely with no knock, nice boost control, and good power.

However there are a few areas that seem like they could be improved, and it would help to have input from people much more familiar with what kind of numbers these cars are normally making.

First, here is what we did. We took my laptop and my GM OBD-I to USB connector, and created two bent wires to reroute the hookups to match the 10-pin Opel connector on the Elan. I had TTS Datamaster (I am familiar with this software as I use their ZR-1 version) for the Elan M100 installed, as well as ElanScan. Then we went for a drive. Ian made runs by accelerating briefly in 1st gear to get going, then shifting to 3rd and keeping the pedal to the floor through as much of the rpm range as possible. We had a nice flat straight long section of road with no intersections to do this in.

The TTS Datamaster scans came out great. However, the ElanScan would output data, but appeared to update at a very slow rate, over a second. It also had odd things like the promid changing with each update. Looking at the data we got over the runs when it was running, it appears to be just a bunch of junk. I don't think it was communicating properly with the vehicle. We'll have to try this again sometime to see what we need to do to get it going. So the good data is from the TTS runs.

The good:
- The car had no knock the entire time. It did have knock counts, but the count never increased during our scanning. I suspect the initial counts were from starting the car up. In fact, to me, it had eerily no knock. I would expect occassionally to have slight knock retard or at least an event under WOT at peak torque, or to at least have knock initially when those secondary throttle butterflies open up and the charge momentarily goes lean. But maybe this is perfectly normal. It's certainly perfectly safe that there's no knock.

- The car seems to have great boost control. From ~1,000 rpm in 3rd gear, the car is at atmospheric around 1,000 rpm, it hits 1.1 bar by 1,600 rpm, 1.2 bar by 2,100, 1.4 @ 2,850 and 1.55-1.6 @ 3,250 which it held nicely all the way to ~7,000 rpm.

The bad, though not really that bad:
- The car never set the secondary throttle bit. This could be because TTS doesn't read it right, it could be that the ECM doesn't actually set it, or it could be that the car is not opening the secondary intake runners. This should probably be confirmed manually just to be sure they are actually opening up. It's too bad you can't override them. On my car I have a switch on the dash to keep the secondary injection/induction system off. It's easy to tell if it doesn't work by seeing if it feels the same with the 'valet' key in the on and off positions.

- The intake air temp seems to indicate the car would benefit from intercooling improvements. At least that what it seems like to me. On every pull, the initial IAT was fairly high. It was a 90F day, and the temps would be anywhere from high 120F to 140F. No biggie. They would then drop down to the mid-low 120F by about 3,500 rpm or so. But then on every run the temps begin to climb back up from that temperature, slowly at first through the 5,000's then somewhat more steeply in the 6,000's ultimately rising to the high 130F's each time. It seems to me like the intercooler hits some kind of saturation point or something, but I've never owned a turbo car. So perhaps that would happen no matter what size/efficiency the cooler was?

[Aurora40]

Now here's the weird stuff:

- The O2's seem lazy to me. The more I look at the scans the more they seem lazy. On many of the idle sections it has a frequency of over 1 second to go from peak high to peak low and back. This seems pretty slow to me. I expected it to jump back and forth much more quickly. Unfortunately the O2 crosscounts value was 0 the whole time, which I guess means TTS isn't reading it right, or that the car doesn't set crosscounts.

- The fuel trims while just watching them seemed fine. In the 120-130's pretty much all the time. However, when playing the whole driving event back and compiling a histogram, there are heavily rich blocklearns and heavily lean ones. I will attach the long-term fuel trim/blocklearn histogram as an image. It could be that the lazy O2's exaggerate this. I thought it might be the accel enrich and decel enlean functionality, except when those are on the blocklearns all seem to be 128 (which means no adjustment):



- The injector duty cycle gets up to 84%. Not concerning at that power level, but it seems like if you wanted to kick the power level up, the injectors would really be getting close to their physical limit.

- The O2's on WOT are RICH! I know they will be rich, but this is really rich. They are mostly around 980mV through the whole RPM band. On the lower end like 1-2,000 rpm they are closer to 950mV. However, to me that seems really on the high side. I'm used to them being more in the mid to high 800's, but then, as I said, I don't know what's normal for an Elan. If that's rich, it could explain why the car never has knock...


So to sum up, the things I would look at if it were my car is:
- Is the physical timing too conservative?
- Are the secondary throttle blades opening?
- Is the O2 sensor shot? If this is cheap on your cars, I'd just replace and be done with it.
- Can the intercooler be improved, maybe it's something simple like leaves and road debris lowering it's effectiveness, or maybe that's just normal.

Other than that, though, I would certainly drive it and not worry as there really doesn't seem to be anything concerning in the scans.

I'd be happy to post the TTS files. TheBard also has them. TTS Datamaster is a free download, and free to use to review saved scans, but you only get 20 free recording sessions before you have to buy it to record with. I can also post in comma separated value files.

[TheBard]

Thanks to Bob for a fun afternoon, and for posing the questions I'm not auto-minded enough to ask.

Here is the ElanScan data, as Bob said, we reckon we hadn't got it hooked up right, but any advice/hints/tips would be great.
http://www.bardsworld.netfirms.com/lec/elanscan1-20060715.ecu

and here is the TTS Datamaster log
http://www.bardsworld.netfirms.com/lec/TTSLog1-20060715.uni

You can download TTS Datamaster via there website http://www.ttspowersystems.com/diagnostic.htm or just grab the M100 version of the software from here...
http://www.ttspowersystems.com/software/dm350/L1/setup-L1.exe

Now does anyone have an easy way to check if the secondary throttle is opening?

Cheers,
Ian

[ESM]

On ElanScan, the screen updates are much slower than the data capture (both are adjustable) this was specifically to maximise the rate of data capture as much as possible.

As to the garbage data - you didn't tick a menu item called 'Test mode' did you? This is a feature I put in to let me tests software features without connecting to the car - it literally outputs random numbers if its enabled.

By the way, TTS only know the format of the ElanScan data stream because they asked me...

[TheBard]

Ah, looks like I'll have to pop over to Bob's for another set of testruns then on the Elanscan software... Might even do some forum and website reading first so we know what we're meant to be doing.

Cheers for the advice and good to know TTS were working from reliable information.

Ian

[Aurora40]

On ElanScan, the screen updates are much slower than the data capture (both are adjustable) this was specifically to maximise the rate of data capture as much as possible.

As to the garbage data - you didn't tick a menu item called 'Test mode' did you? This is a feature I put in to let me tests software features without connecting to the car - it literally outputs random numbers if its enabled.

By the way, TTS only know the format of the ElanScan data stream because they asked me...

Ah, that's quite possible. It kept having errors connecting, and I was clicking on anything I could try to see if it would get working. "Test mode" sounds like it might test the connection or something, so that could be why it suddenly started "working".

Any suggestions on troubleshooting the connection then? I can read with TTS fine. What I'm using is a USB adapter from Moates, and then their GM OBD-I cable with two copper wires mating it up appropriately to the Elan interface:

USB: http://www.moates.net/aldu1-usbtoaldl-c ... -p-49.html
Cable: http://www.moates.net/cabl1-obd1style-a ... -p-50.html

As an aside, I hope you don't take offense at the mention or use of TTS. I realize this is the "ElanScan" section, and I think it's great you wrote it up and made it freely available. More importantly, you offer up this, excellent support of the product and consultation on the results, something one wouldn't get from TTS even with the ~$95 price. But as a consumer of the products, both are viable options, are free (tts at least for a few runs) and offer a different style of interface that may appeal to different people. I'm familiar with the TTS style because I use it on my 'vette.

Also, if you are familiar with it, as far as you know, does the display of the secondary throttle bit work properly? I.e. if we didn't see it, is it probably not working? It also doesn't display the PROM ID, but I suspect it could be retrieved using their custom data definition functionality if one knew what bit of data it was in.

Thanks again for your help Alan, and helping me in the first place with the pinout mapping!

[Aurora40]

So to sum up, the things I would look at if it were my car is:
- Is the physical timing too conservative?
- Are the secondary throttle blades opening?
- Is the O2 sensor shot? If this is cheap on your cars, I'd just replace and be done with it.
- Can the intercooler be improved, maybe it's something simple like leaves and road debris lowering it's effectiveness, or maybe that's just normal.

I was looking back at my posts, and they are quite lengthy... Sorry about that! So perhaps I should summarize the questions I have too, in addition to the previous summary of things we could use suggestions for troubleshooting, for folks that don't want to read a novel about it.

- Do most people have at least some slight knock retard occasssionally occur?
- Do most people have O2 readings in the ~980mV range on wide-open throttle?
- Do most people have intake air temps that start to climb again on the second half of the rpm range?

[ESM]

Ah, that's quite possible. It kept having errors connecting, and I was clicking on anything I could try to see if it would get working. "Test mode" sounds like it might test the connection or something, so that could be why it suddenly started "working".

ElanScan logs every coms error it encounters - don't assume because the occasional coms error pops up that there is a big problem. A few errors just means a few dropped points - it is pretty error tolerant (so long as you've got the Moates device set to a 1ms latency). I should hide the log really, a high proportion of people think logging isn't working because of the odd glitch recorded in the error log. Unless its firing out several errors a second don't worry about it.

[Aurora40]

ElanScan logs every coms error it encounters - don't assume because the occasional coms error pops up that there is a big problem. A few errors just means a few dropped points - it is pretty error tolerant (so long as you've got the Moates device set to a 1ms latency). I should hide the log really, a high proportion of people think logging isn't working because of the odd glitch recorded in the error log. Unless its firing out several errors a second don't worry about it.

Thanks Alan! When we try it again, I'll pay less attention to those then. It was pumping the errors out fairly quickly though. I had set the device driver for the Moates adapter to something, it was whatever it said in the instructions that came with it. I'll check it out again and see what it was set to, and set it to 1ms if not already there.

[Doug]

Wow, glad you popped in here, Bob, cause I never heard of TTS before ( Alan ) Too bad I don't have an Elan to play with this program on anymore . . .

I don't think there's too many people here who have finely analyzed their ES logs here to give a lot of feedback, Bob.

As an alternative, you could probably get people to send you their scans and you can cross check them as needed . . .

Or we can start a new area for detailed ElanScan discussions

[GeoffSmith]

It was pumping the errors out fairly quickly though.

If you've lashed up an interface to the ALDL socket, be careful how long your leads are. I've had problems maintaining a consistent connection when I was probing the interface with a 'scope - it was due to the increased capacitance of the extension lead.

BTW, if you need a proper ALDL plug, I'm just about to order some. If you want one sent over, give me a shout.

Geoff

[TheBard]

Hi Geoff,

Might be interested in a proper connector, as our two bits of bare copper wire and three elastic bands were a tad Heath Robinson. Let me know the cost before I stick my name down for certain.

Cheers
Ian

[GeoffSmith]

Ian,

P&P will probably exceed the cost of the connector. I've ordered a batch today - once I've got them, I'll find the cost of postage for you.

Geoff

PS Three elastic bands is a bit excessive

[Aurora40]

Wow, glad you popped in here, Bob, cause I never heard of TTS before ( Alan ) Too bad I don't have an Elan to play with this program on anymore . . .

I don't think there's too many people here who have finely analyzed their ES logs here to give a lot of feedback, Bob.

As an alternative, you could probably get people to send you their scans and you can cross check them as needed . . .

Or we can start a new area for detailed ElanScan discussions

Thanks! TTS and ElanScan are just different ways of doing the same thing. It's the same data coming out, they just have different interfaces. And ElanScan has some neat extras like dumping the bin and viewing the prom ID. ElanScan also seems to have more derived data like horsepower and such. TTS has calculators to do this and 1/4 mile estimates, but it requires more user input (though it means you can tailor the data to match difference things).

I just like the TTS interface. It has a short wide graph of the whole run, with 5 data elements on it (things like rpm, MAP, throttle, timing, retard). I find that easier to see what was going on. When the throttle shoots up, I know the pedal is down and things are going to start happening. Plus you can easily zoom in on any section of the graph by right-mousing a box over the section of graph you want to zoom.

No need for the separate section. I'm no expert on the data anyway, I just find it interesting and always have ideas in my head. Sending scans is a great idea, though. If anyone has one they want to shoot to me, my email address is in my profile. I thought there were a bunch attached in this forum, but there was actually only one and it had limited WOT activity. Really the more I think about it, I'm sure everything's fine. The high O2 on WOT is probably normal, as the target A/F displayed was something like 11:1, which I believe is a bit richer than a Corvette (what I'm used to) would shoot for, hence the slightly higher O2 mV readings. And I'm sure Ian's secondaries open up. I imagine the car would really fall off on the upper end if they weren't. The "dyno" feature of TTS would indicate that too, with power climbing steadily until the 6,375rpm shortcoming of which value TTS uses.

If you've lashed up an interface to the ALDL socket, be careful how long your leads are. I've had problems maintaining a consistent connection when I was probing the interface with a 'scope - it was due to the increased capacitance of the extension lead.

Hi Geoff. It is about 1.5" of copper wire we were using. The thing is, is TTS could read with it. I'm sure it's probably just some simple config that has to be changed with the device driver or some such thing to get ElanScan going. Whenever Ian wants to try it out again, we'll see if we can't get it working.

[GeoffSmith]

It is about 1.5" of copper wire we were using.

That shouldn't cause you a problem then. I had a couple of feet of three core cable and it slugged the data line to death. Good luck next time.

Geoff