Article by Mike Barrett

Well they said we need to evaluate cards. Right I said and booked a flight to Las Vegas. "No, not that sort of card, you know chips". So I still kept my flight tickets. "No, not that sort of chip. Storage cards, memory-chips!". "Damn" I thought. I still went to Vegas though!

Over the last year or so there has been an explosion in the number, size and speed of storage cards in production. There is a lot of myth and legend about speed and compatibility issues with various devices. To try and cut through some of the misconceptions PocketGPSWorld.com has gathered together a large number and type of storage card and has created a compatibility/speed comparison test.

Thanks to the generosity of a number of the Memory Card manufacturers we have collected a number of different cards with varying storage capacities to help you determine which cards you should be considering for your navigation equipment. Our thanks go out specifically to ATP, Kingston and Sandisk for providing test samples, and also to Rob at ValueFlash.co.uk.

We did contact other manufacturers for test samples but received no response. This does devalue the test a little, but on the other hand if we have successfully tested the cards then you know that they should work for you as well.

The concepts of these tests are not to prove scientifically that one card is better than another in laboratory conditions. I have taken the view that benchmark testing although good to give an index to compare cards does not offer a real life test for the applications to which we put them.

We have therefore devised a number of tests which represent everyday usage of the storage medium in the navigation application usage model.

We have not made comparisons based on any manufacturing attribute or physical superiority of each different medium or type of medium. This is an objective and measured test in a controlled environment.

We tested 4 different types of card. Mainly SD and CF, with a couple of MMC cards as well. During the tests a number of these cards were used, but the ones we actually performed the speed and compatibility tests on are identified below.

Devices Tested

As the memory cards are designed to be used in a number of different devices we have gathered together as many devices as we could find and used them both for the compatibility and speed tests. We will add to this list as we get our hands on new devices.

The testing will take forms Compatibility and Speed. For compatibility we are just ensuring that the card works correctly over a period of time running both a GPS application and (where applicable) a media player.

Memory Cards
The raw performance of the memory card will be tested using a PC connected to a USB 2.0 Card reader. This will allow the cards to be tested under optimum conditions and not be restricted by the hardware implementation on the target (PDA) device.

This will test the usage scenario where the user uses the PC to copy various maps directly onto the memory card for use on the mobile device. This test will use TomTom Navigator UK maps, Co-Pilot UK maps, and a selection of Raster images for use with Memory-Map.

Application Speed Testing
Application testing will use the memory cards in the devices and where applicable will test the following:

• Creating a route from Lands End to John O’Groats.
• Creating the route with another background task running and accessing the storage card.

PDA Compatibility Tests
Prior to the running of the memory tests the PDAs will be hard reset and have the test applications installed on them.

The memory cards will be used in each of the available PDAs and the benchmark tests will be performed for the hardware to determine the underlying speed of the hardware.

For each PDA tested the stability of the memory card will be recorded. As it is speculated that some cards have problems when 2 applications are trying to use them concurrently we will test using the Media Player and the GPS application.

Dedicated GPS
Currently the only dedicated GPS devices we will be testing are the TomTom GO and Mitac Mio 268. This will have the navigation tests performed as per the PDA application testing.

There is currently no additional raster based navigation, or benchmark tests available for this device.

Digital Cameras
Using the Nikon D70 the compact flash cards were tested in the following scenarios:

Auto with manual shutter mode, this should take a number of images as quickly as possible and determine how quickly the camera/ memory card can process them. The images should be captured in both raw and jpg formats.

Compatibility Results

1) In one of the 3 tests performed the Integral card failed on the Palm T3 requiring a reset of the PDA.

Speed Results

Raw Speed Tests

Using an internal USB 2.0 card reader we first cleared the memory card then copied the data directly onto the memory card. We then did a reverse test reading the data and writing to a local hard disk. For memory cards in excess of 512MB we used 482Mb of data for cards of 256MB we used 237Mb. As we are measuring different data we present both the time taken and the MB per second transfer rate.

We had originally intended to take 5 times and then average the results. When executing the tests I found that we only needed 3 readings to get an accurate result.

Application Speed Tests

Using an iPAQ 4700 and a TomTom GO I took a series of timings, the 4700 was selected as it has both CF and SD slots, and the MMC cards can be read in there as well. First I took timings with just the TomTom application calculating a route by itself (the Route column). Then timings were taken with the Media Player playing the same track from the same storage card whilst the same route was planned (the Route+ column). The same test was then carried out on the GO using the SD cards obviously no media player is available on the GO (the Route column).

Digital Camera Tests

With a Nikon D70 set to manual 1/500s at f5.6 with manual focus the camera was set up to take continuous multiple shots. As all the factors affecting the timing of the capture were set to manual the only two things that could influence the speed of taking pictures and saving them to the memory are: the Nikon D70 internal buffer and the speed that the card can be written to.

The test involved taking as many photos in a 30 second period as possible then measuring the time taken to save those photos to the store card. As the number of images varies an average per image is used as the result for this test rather than the total elapsed time.

Conclusions

Although all the cards worked in the tested devices there are varying performance levels that affect the way the applications work. If my only experience with TomTom GO had been with the Integral 256Mb card I would have been very disappointed with the setup. Transferring data to the cards using the Card reader put in impressive timings for the SanDisk cards, which were not carried through to performance in the application testing. The 512 SD cards were a little slow when used for transferring data to the cards, but once it was there they performed extremely well.

The initial results of the raw speed tests suggested that the SanDisk would be the best and fastest memory-cards. Moving on to the application testing changed my views completely. Whilst the SanDisk cards were by far the fastest for reading serially that was not the case when reading in a random manner.

The application testing showed that in general the larger capacity CF cards all performed well within an tolerance of about 4 seconds to create a route with the exception of the SanDisk which took almost twice as long.

The Digital camera tests showed the expected results mirroring the raw speed tests with Sandisk top for writing, followed by the ATP card. The slowest one in this case was the Lexar.

The SD cards also showed varying results. Interestingly the TomTom GO tests all performed within about 6.5 seconds of each other with the exception of the Integral card which was the worst in all tests of the SD cards. Whilst there was a similar trend different cards performed differently in the different scenarios. Though there was little difference between the fastest cards here. Interestingly there were less differences with the SanDisk cards in this form factor.

The MMC cards matched each other and it really was a coin toss which would be the better of the two.

Compatibility with the different devices was definitely not a problem with any of the cards or devices tested. By this I mean that none of the devices refused to recognise the card or failed with the exception of the Integral card in the Palm T3.

Long Term Testing

These cards have all been used in a variety of devices on a daily basis since August 2004 so they have had a good test in a number of conditions from the dry deserts of Death Valley, Nevada and Arizona, the humid and wet atmosphere of Florida and the Keys, to the cold and frosty delights of the UK in winter. During this time whilst in actual use I never had a single card fail.

How can you help?

Well as I only have a limited number of resources I am unable to confirm the compatability of each and every card/PDA/application combination. This is where you can help us.

If you have a problem with a storage card then please post the details into this topic. Quoting the Make, Type, Size, Serial number, your PDA details, the application you are using.and the problems you experienced.

If you have a card that works well then please post the details into this topic. Quoting the Make, Type, Size, Serial number, your PDA details, the application you are using.

Resources

Mike Barrett

http://www.1st-ventures.com