Data Logger Drive recommendations ?

Cheese

Hello All:

I have a parallax data logger and seem to be having difficulty with different drives.
I am saving data in the format of ####,###,#,# at a rate of the 10+ samples per second.
I am storing a folder for each day, then within each folder many files of 1900 samples each for that day.


The problem seems to be in the USB drive that I use. I have tried several and only one is working consistently.
I need one that is readily available in quantities the only one that works good is one that was given free with advertising stored on it, no brand name on it.

Some will not initialize others will hang up after three files and others stop in the middle of a logging sequence.


Gene

Tracy Allen

Errant behavior such as that you describe can often be traced to the power supply. It has to be solid, well filtered with at least 300 mA of capacity. The drives take pulses of current, particularly at startup and when they write to a sector. Also, there is quite a bit of difference between drives in how much power they need.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

Unsoundcode

Hi Gene, I have had a problem similar to what you describe. I found·I··could cure the problem in the way I formatted the drives.

This applies to WinXP.

Control Panel--->Administrative Tools--->Computer Management--->Disk Management

Right click on the panel related to your drive and select format, select Fat 32 and Allocation unit size 512.

Yours may be a different problem but every drive I have formatted as above has worked fine for me.

Jeff T.

Cheese

Tracy:
I do not think the problem is the power supply one is a 800 mA 7.5V. and another @ 4 amps, both good quality industrial type(going through the regulator on the board of ed @ 4.95V.).
However I put a meter on the power pins and measured during data recording and the meter was steady @ 4.95V.

Jeff T:
This was a first for me, I have never used that tool Before Hence don't Know what it does, However I did as you suggested and no luck.

Tracy:
Noting your suggestion of power being at fault I checked the mA. draw and found the one that works is half the draw of the one that fails....
I will see if I can find another with the lesser draw and try that one.

Thanks Gene

Post Edited (Cheese) : 4/18/2008 11:52:25 AM GMT

Cheese

To All:

Thanks again
Problem solved!

The Data Logger is very sensitive to the type of drive used!!
Went to wallmart and bought a handfull of drives having verious results. The one I found to be best was a Lexar 2GB Hi-Speed Flash Drive SKU#50590 14184
Some worked sometimes others not at all, This one seemes very forgiving!! (OFF THE SHELF) recoveres well when misused without safe removal and hot swapped even when removed during read/write, yes abused. This device must be tested jackass proof. Have run over 1000 logs files with good results so far, will continue thru the weekend abusing the logger and drive and will post the results. Must say it looks GREAT!

Tracy:
I noticed you have a logger for sd cards! I would like more info on that device.

Thanks again!!!! [noparse]:)[/noparse] Gene

Tracy Allen

I had a chance on Thursday at the embedded systems conference to talk with engineers at the FTDI booth. They said they did not have available a report or matrix of different thumb drives they had tested. My question was more to the point of the current consumption than the bare operation. You'd think they would have done a pretty thorough vetting. But apart from the issue of formatting that Jeff mentioned, they saw no reason it wouldn't work with most drives, and agreed that physically smaller drives seem to draw more current. Your observation is very interesting with regard to the Lexar drive. Something to write in on the working papers. I've been using Sandisk products successfully but have not done much looking around.

My setup uses a switched voltage regulator to control the 5 volts to the vDrive, so a a Stamp pin can turn it on and off, powered down between readings. When it powers up, the Stamp can capture the monitor ready prompt. The SUM command (suspend monitor) is also useful to reduce power consumption to a couple of milliamps, and it wakes up fast when the RI input line is pulsed.

I've tried several SD card interfaces. By far the easiest to use from the hardware standpoint is the SDdatalogger from Hitt Consulting. It uses only one Stamp pin for communications, plus 5 volt power, and it has nice Stamp-friendly i/o formatting features. It is limited to 32 meg or less FAT16, so the media has to be preformatted to that.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

Unsoundcode

Hi, I don't profess to know about sectors and formatting in fact it's the opposite.

I have used various sizes of Lexar ,·PNY , PQI and Sandisk.

In the early days of testing I would write short strings of data without any problem. When I got familiar with the VDAP commands the files became longer and faster. Thats when I noticed the files would appear to be corrupted, with the exception of the smallest drive I had, a Lexar 64 Mb Jumpdrive.

With experimentation I found that the usual windows format was formatting the larger drives differently to the Jumpdrive, or at least that is the conclusion I came to. Because when I formatted as described in my previous post every drive I have, that gave errors originally, worked without a problem afterward.

I think it's worth mentioning and something to try before discarding a drive.

glad you found a drive to suit your purpose.

Jeff T.

Cheese

Tracy:

The physical size of the lexar drive is small (very small) I did not do an amp draw on it because it "just worked"
also it is a 2 GB drive, someone told me 1GB would be the Max the logger would handle, so I don't know if the Jeff T thing plays into it? (I used it off the shelf untouched.)

Jeff T.

My father is a perfectionist, Contractor for 60 years.when he built something it was designed to stand the test of time like the pyramids or great wall of china. throw every thing imaginable at it and only when it doesn't even scar is it good enough. So I guess I come by it honestly when I test a device. While some of the drives I tested might work for some applications it was not 100%.
This programming thing for me is a KILLER. I like computers because they don't goof up. only the programmer that didn't figure something some moron might do with it. So I am the "#1 test REDNECK MORON!."

I am researching the lead you gave me because I think it is either there or in the file handling structure of the logger. Not being a guru the search for bullet-proofing the logger is difficult. It is Sunday morning and so far the logger with the lexar drive is proving positive. I will continue testing till I crash the drive with Data overflow. To this point the only file lost is the one that I unplugged the drive during a write sequence. but when I reinserted it, it woke up and the subsequent files were good. Next is to clean up the code and put this thing to rest.


Thanks again for your TIME!
Gene

Tracy Allen

Gene,

There might be a firmware issue. Note that you can download the latest Vinculum firmware from the FTDI site, give it a specific name on a flash drive, and then the Vinculum will update itself when it finds that file on the drive. The documentation says it can handle FAT12, FAT16 and FAT32 so long as the sector size is 512. And it has the FS command for disk sizes up to 4gb, and the FSE command for disks over 4gb (FS=Free Space).

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

jrisse

Does your USB Drive have any program software installed on the drive?· If so, please remove the software that came with the USB because it may prevent data from parsing to your data folder.

Cheese

Tracy:
Where did you find the info on the installation? Link would be great!

Much Better But it will still crash! That is the latest to date!
Large Files OK
Large quantity of small files, crash!
Still looking into the firmware update to go to Fat32 sector size 512k.
Have not worked out 100% of bugs, But am still working on it.

Jrisse:
The drives Have been re-formated to the Fat32 sector size 512k. and is clean. That was an issue we found early on.

Thanks to Jeff T.
<<Control Panel--->Administrative Tools--->Computer Management--->Disk Management
Right click on the panel related to your drive and select format, select Fat 32 and Allocation unit size 512.>>

Butt until I can get a handle on the firmware update from FTDI I still cant put it to rest.

Thanks again for your Time!
Gene

Unsoundcode

Hi Gene, this may not apply but I think it is something worth knowing. It is also something that may be dependent on your hardware and software configuration. Before Parallax sold the datalogger and Chris wrote the software with RTS/CTS flow control another forum member and I were using only 2 wire communication (RX/TX)
with a Vdrive 2 which is basically the same thing as the datalogger. We were having issues where data would get corrupted at almost predictable intervals. It turned out the Vdrive could not process the data as fast as was being sent, and if I remember right that was also at the rate of 10 samples a second at 9600 baud (GPS data). So what happened was that fresh data was sent before the Vdrive was ready and that data became corrupted. In that instance we overcame the problem by waiting for the drive to return a prompt signalling a succesful operation. The timing of the prompt return could vary and be over 250 mS. Since then that is the method I choose to use so I really can't say how that would effect RTS/CTS operation.

For a test you could try and half your sample rate and see if there are positive results.

Just for reference, hope it's useful information

Jeff T.

Tracy Allen

Gene,
The information is in the Vinculum Firmware User Manual version 2.3 Issue date 2007-12-03 or the most recent you find at the link.

In line with what Jeff said, I have been using the full monte of signal lines, rx, tx, rts, cts, for flow control. Do you have a simple program apart from your application that illustrates these failures you are seeing?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com