I've been datalogging with my new zietronix WB and I've run into a rather stupid problem. The datalogger runs at around 78 frames per second!!!!! So I get to see lots of detail. The problem is I have an excel program made up to take a datalog and put it into easy to read charts much like WINALDL does only using the WB. The problem is a 10 minute drive makes for a 50,000 entry datalog and my poor little 150 mhz laptop can't take it. It can barely handle 15,000 entries and still do the necesary math in less than a minute or two.

My question is do you guys know of an excel function that will delete every other data point in order to cut my data logs down to size. I only need 10 frames a second or less for what I'm wanting to do. I've messed with it for hours and can't seem to find any function that will allow me to do this. I'm gonna send an e-mail over to zietronix to see if they can add a feature to their sofware that will alow you to do this. They do have a setting that changes the display rate of the data just not the recorded rate.

 

You can do a lookup.
Type a column of 0, 10, 20, etc....
in the next column type =lookup(single entry from column of (0,10,20, etc) then comma, then entire column of data to look up from like WB A/F, then comma, then entire column of data to look up again).

SO, if column A has WB A/F, Column B has 0, 10, 20, etc, and Column C will be =lookup(B1,A:A,A:A) in C1, C2 will be lookup(B2,A:A,A:A). To make it fill to the bottom, you can double click the bottom right corner, like if you type 0 in B1, 10 in B2, click and drag over both, then double click the bottom right, it will fill to the bottom of data in column A. Then in column C, type the first lookup as shown above, then (with that cell single clicked upon), double click on the bottom right hand corner, and it'll fill until finished. You'll see that it'll fill 0, 10, 20, etc until the number is much bigger than you need (you only need to go to 50,000), so you may want to delete numbers above 50,000 (and you may want to do this before you do the lookup).

I fear that this will still take quite a while on your slow computer, and that you may have wanted to import the data in a more efficient manner, rather than manipulate the data, but this is my only idea.

 

Kinda a little early, but..

I noticed the same thing while logging.

I am almost done with a Pocket PC program that lets you view the data as it is captured, and also logs the data to a file.

It has an option that allows you to log every frame, every other frame, or what ever number you pick, up to just logging 1 frame per second. I did that because file size can be an issue on a PDA.

I wrote a standard laptop version of the capture program, as a test first. That was much easier.

Hopefully, another week or so, and I can release the PocketPC software, but if you do not have a PocketPC, that doesn't help you much does it?

The PocketPC software also has a option to re-scale the user1 input into GM type MAP reading. That is handy. Working on a routine that will measure RPM change at WOT. Kinda like a poor mans rolling dyno.

Who would have thought that too much data was a bad thing? Lol.

 

7point4, sounds like your sofware will have all the features I wish my zietronix software had.

Redngold, I'm not real sure how to implement the lookup funtion. What I really need is to systematically remove data from my large file in order to make a smaller one. My left most column is time so I guess you could some how have it look up just odd times, but tht doesn't do me much good since I have 73 entries for each second. I think my laptop might be able to take a 50,000 line program as long as that was the only math it had going on. My excel file I have for analyzing my data uses a lot of pivot tables and even has a feature that lets you input your old VE curve and then it spits you out a new one based on your actual AFR vs your desired AFR, so that takes alot of math with 15,000 entries.

I think my best bet might be to hound zietronix for options in the software, or just wait for the speadreader and run my WB through it that way I'll have all my data in one place.

 

I was certain that the Zeitronix software had a slider bar to set the sample rate.. maybe it just adjusts the display rate.. I will have a look at it when i get home... Do you have the latest software version, its not the one posted on their website. I have posted a link to it here before. The easy way to tell is if your version has auto-scaling of the tps and 'user input' inputs.
Ben.

 

Yeah I have the new software, it came on a cd that came with the WB, but it does have the auto scaling of the TPS and user input. It's the 1.04 version. I tried turning down the slider and it only slowed down the display rate, the recorded logs were still at the 70 somthin frames per second. The slider says "text display speed".

 

BBMonte,

Have you tried "splitting" the datalog into smaller chunks? The ZT software allows you to do that. On some of my runs, I'll log 160k+ frames (the slider is for display rate only) sometimes which is too big for .xls spreadsheet. So I'll split the sample up into multiple .xls sheets and then combine them for the rpm/kPa ranges I need for tuning. A friend of mine has a tuner program that will take ALDL datalog like this and crunch it down, analyze it and spit out a new VE table like WinALDL does. I have asked him to look at the ZT datalog and try to do the same with that data. The input for his program uses NotePad.

 

Yeah my excell program does the same thing only I can't feed it any more than 15,000 entries at a time. The problem with the smaller chunks is that it would only give me about 2 min of run time for each chunk. It would take me forever to tune my VE's that way. I guess I'm spoiled on WINALDL, I also have a excel file that will take the tables it spits out and modify my existing VE curves based on that. I just like to take really long drives and get some good average numbers, it keeps a hicup from skewing the results too much. If anyone wants to look at my files I'll send them to ya. The WINALD one is pretty much complete the WB one is a work in progress. I also have one for the standard moates data logger. Makes neat little graphs of where your blm, knock, and the likes are. Can you tell I'm a visual learner.

It wouldn't be so bad if I could just turn down the log rate or input my WB into my aldl stream. Ahh where's the speedreader when you need it. I sent an e-mail to zietronix. Just seems like there would be an easier way to just delet every other entry. If I did that a couple of times I would have a nice small datalog with lots of info. As a side note the super fast data logging is great for AE and other "momentary" issues.

 

I just got an e-mail back from zietronix. They said they would definately consider that for the next software release. He also suggested I some how filter the large file in excel some how. Go figure. He said they would probably keep the fast data logging rate and just have a setting that would average every 10 cells into one. I'm impressed with this company, they responded to my e-mail within 2 hours.

 

Upload a snipet of the file and I'll see if I can find something that will do it in excel or elsewhere.

 

BM,

Let me know when you have that one for the WB. I would love to just feed in the WB output and get a VE table. One thing I will say tho, is that I still find it necessary to give the data an "eyeball". With the 7747, I tune BOTH the VE and the SA tables simultaneously. I find that the tune is much smoother if you tune the VE in those 400rpm increments and then use the SA to tune the "in between" cells like 1400, 1800, 2200rpm.

 

Yeah, I always go over the changes to weed out oddball data and fill in the cells that didn't get any data. The way I have it set up there is a page to enter your old VE a page to name the AFR your aiming for then there is a page that gathers what AFR you have currently based on your datalog, then it spits out a new VE based on all that, if it works it saves a ton of time. The one I made for the WINALDL data logs saved me a ton of time, 2 or 3 half hour datalogs got me within 2 or 3 BLM.

I think I might be into something in excell, using the advanced filter feature there is a check box that lets you return only unique entries...which sounds perfect, but for some reason it DOES find something different about the 72 entries that have the same time. Might need to round those then try it. Bad news is that if I try that feature on anything larger than 1000 entries it takes 10 min. to do.

 

BM,

My friend's VETTtuner does something very similar to that. Programs like this only help and save tons of time. I still think you want to eyeball the results just to make sure. Its a short leap to imagine having something like this in a "closed loop" with a scantool such as WinALDL so that the tuning is taking place in real time modifying the VE table as you drive and as cells are populated with more and more samples. If anyone can do that its the talent on this forum.

 

Actually I think JPrevost is working on just that. Once you get into code the sky is the limit, I don't see any reason why you can't use the WB just like you would the NB, only with the ability to truly name your AFR at any load or rpm.