Too bad we can not go back and edit the thread title. Should be FD3S!

In the past I have posted about how recording more info with the DL slows down the frames/lines/samples per second count. Thus for higher sampling rates, only record what you need. Examples of this are in my tuning notes.

* * *

Today I compared the last FD3S software to the current UNIV software with my beige box. Here are my test results on an old 700 MHz XP laptop.

What Was Recorded----------FD3S Rate------UNIV Rate
--------------------------------------------------------------------
AUX A/D with MapRef---------51cps------------82cps
ALL except sensors-----------17cps-----------19cps

 

Ok. Excellent, so they improved speed in newer universal software.

I will try to explain (as software developer) what happens. When you talk to hardware - you ask about data. In this case, you ask datalogit unit for data.

Datalogit has built in clock or it works on PFC clock (not sure) but it sends frames in specific periods of time (depending on what is clock speed)

What happens on your laptop? You get stream of data from com/usb and need to parse it and save it. This is your log. When packet of data comes in- you have to handle it. While you handling it there is possibility that other packed comes in and you still dealing with previous one. In this case you skipping new packet.


Here is example. You staying at conveyer line and get part to stamp manually every second. You can do it, but then you need to get away for a second to replace tool. You miss so many parts. Same thing with frames of data. Disk fragmentation, memory swapping to disk on laptop, etc - those cause missed frames.

You can do experiment. Try logging as you do right now and then execute resourse-intensive process on laptop (ie virus scan) and try again. See difference in frame counts?

So, you correct, ammount of data definitely will affect performance. But it won't affect PFC work. Better laptop should give much better results.

 

FastHatch should tell more about all that. But here is sample from my log.
I took time deltas from frames and built char. See spikes? It's when frames were missing probably due to disk writes or something else.

 

I have asked this forum many times to do some logging to compare logging rates but to no avail as they are lazy.

I know a faster laptop should log faster up to the point that the DL is the bottleneck. Nobody has listed any logging speeds with faster or slower computers than mine!

I would also like to see a comparisom between serial and USB ports if the computer has both.

 

I will post my results later. I just got all wiring setup and will start logging soon. From what I see - FC Edit written with Visual Basic and uses standard ways to read serial port. My chart shows me that it's either grabs every frame consistently or it misses every other or so frame because time deltas very good.

It might be true that datalogit/PFC can not send full packet on each clock tick and they skip some. I will post more findings later. But those spikes on chart I posted is obvious misssing frames.

 

Here are logs of "record lines per second" made with; v2.0 Black box and FC-Ecit v2.125 Universal, running on a 2 GHz, Pentium Duo, 2Gb RAM, USB port #4

Individual, one at a time

Basic 54
Sensors 59
Aux A/D 146
Map Ref 124
Advanced 42

Combinations * = on

*Basic 32
Sensors
*Aux A/D
*Map Ref
Advanced

Basic 71
Sensors
*Aux A/D
*Map Ref
Advanced

*Basic 22
*Sensors
*Aux A/D
*Map Ref
Advanced

*Basic 15
*Sensors
*Aux A/D
*Map Ref
*Advanced

 

Comparing that to my 2 examples, your faster computer on USB does not log faster than mine on serial.

The difference is that the UNIV software is faster than the FD3S software only when less is being recorded.

The bottle neck id either the DL hardware or PFC.

 

Well chuck i just checked and I think I have everything logging.

But this log is the one that i said I was running 80mph @ 18ish AFR Crazy gas mileage

Unzip it to open the *.txt file

I counted and it was 14 per sec with everything logging

Windows Vista
AMD 64 2000mhz
1gig ram
USB-Serial Convertor

 

Just did the Basic + AUX A/D + MAP REF with my 700MHz 128MB ram laptop and it got 33 readings per second. Just as good as the above super laptop.

The PFC/DL seem to be the bottle neck with the DL software neck.

 

Definitely not a super laptop

My Dell ((Was)) a super laptop, 3.2Ghz P4 w/Hyperthread 2gig ram, 160gig HD... It ate power supplys for lunch.

 

for what it's worth, my late 90's laptop running windows ME (originally ran '98. It's so old that I don't even remember what the specs are on it) gets 10 samples per second with absolutely everything being monitored. The reason why I monitored everything was because I didn't realize it made such a difference on the rate, but rather I just figured my computer was slow. I'm guessing this computer is slower than Chuck's, but I can probably check in the system properties section of the control panel when I get a chance. This is running the universal software.

When I get a chance to do some more datalogging in the next week I will see what my sample rate is when logging only Advanced, Map Ref, and auxilary inputs.

 

My old laptop is used strictly for tuning and nothing more. It was a rebuilt that I bought for this purpose. All unneeded services are disabled, all uneeded software is removed, and it has no anti-virus to slow it down further.

It is a stripped down "racing" version Compaq Armada M700. HaHa!

 

Whats that smell...............O must be a Computer Dyno........ Thought i smelt burning

 

When I went from a slower laptop to a faster one, 2 GHz Duo, What I noticed was how fast/easy it was to scroll thru the charts with map watch on, my older PC was choppy and hard to use. Has any one noticed how a faster PC makes it easer to review logs in chart + map ref mode?