CompactFlash memory card

[RogerChristie]

I have a nominally 32GB Flash card. Can anyone tell me by how much I might expect the actual memory to vary from this capacity?

I have looked for the answer. I can't find it and I can't justify laying out $100 for the CompactFlash Association's specification 4.1 document - supposing that it actually contains what I'm looking for.

(I'm talking about total memory here, so I don't need another lecture about overheads and memory reserved for control purposes. I've had that already from the manufacturer, the supplier, and the Trading Standards office. Sorry for being rather blunt, but I've been trying and failing to get a sensible answer for some frustrating time now! )

[Slartibartfast]

Can I ask why you're asking as it might help to answer the question.

Some companies record a gb differently. So some will say it's 1000mb and others will say it's 1024mb. If the card manufacture say's 1000mb then the card will should have 32000mb on it but your PC might read that as 31.25gb meaning you've 2.4% 'capacity' before you even start thinking about anything else!

[RogerChristie]

Thank you for your prompt response Slarti. You rather caught me by surprise there!

It's a valid thought but I've already been there. 32GB = 29.8GiB = 32,000,000,000 bytes.

In fact I've been pursuing this on and off for the last three months now. The story is that I bought a CompactFlash card. It was advertised and clearly marked 32GB but only had 30.5GB (total) available (actually 30,515,249,152 bytes). (And, yes, I checked and the card is genuine.) The manufacturer - Kingston Technology, the supplier - 7dayshop.com, and now today (whoa, look at the time ... actually yesterday) the Guernsey Trading Standards Service have all advised me that 30.5GB on a 32GB memory card should be considered acceptable. I still don't think so! (I believe in the KISS principle (that is Keep It Simple, Stupid!). My simple view is if I bought a bag of flour clearly marked 1*Kg and it only had 950 grams of flour in it then I would have good reason to complain. So what's the difference?)

Now I'm well aware that flash memory may have a few faulty sectors disabled. I had hoped to support my argument by finding the acceptable shortfall in the specification. However the CompactFlash Association want $100 for a copy of the spec. I can think of more important things to spend my money on, and anyway I don't even know whether the information I want is included. Over the last three months I've read a lot about memory cards but I can't find what I want to know anywhere else (so far). (Incidentally, I see that several memory card suppliers are now including a "disclaimer" in their advertising to the effect that some cards may have less memory than that indicated, but they don't mention any numbers.)

I don't give up easily though - especially when I'm sure that I'm right. Chris Cheeseman should have my letter by now, and I've another ready to post to the Consumers Association. If you or anyone else here can help me gather more information I would appreciate it.

The bottom line is check what you get and complain to whoever will listen if you don't get what you ordered. If (as the GTSS says) "...30.5GB on a 32GB card is considered an acceptable amount throughout the removable memory industry", then it appears to me that an awful lot of memory card users are being cheated. The more of us that complain the more likely it will be that the cheating is stopped.

Sorry for the long answer, but you did ask!

Regards,
Roger Christie

[GeoffR]

If it were a hard disc you would expect 30GB on a 32GB drive, it has been that way for years. I never actually check my cards but then the biggest I use are only 8GB.

[PeteRob]

Prefixes kilo, mega, giga etc are all decimal. kilo got "misused" as an abbreviation for 1024 which is close to 1000 but rounding errors of course accrue from thereon. Maybe a binary notation would better have been adopted but who would have thought that such a thing as a 32 GB memory card could ever exist in the early days when public communication of memory amounts began.

I can remember everyone crowding the university lab when the first "Commodore Pet" was delivered - I think that had 4k of memory. The storage medium was cassette tape. The built in monitor was tiny - probably 8" - green on black. It gave rounding errors if you asked it (in basic) to add up to a (decimal) 1000 in steps of 0.01 - but it was a true marvel.

[Stevet]

If it were a hard disc you would expect 30GB on a 32GB drive,

Not quite - and I think the OP may have a valid point.

As stated already, computer memory uses multiples of 1024 when describing capacity, so for example 4MB = 4 x 1024 x 1024 = 4,194,304bytes

Hard disc capacity, on the other hand, uses multiples of 1000, so a 4MB disc would hold 4 x 1000 x 1000 = 4,000,000 bytes

Even if the CF manufacturer is using multiples of 1000 to describe capacity, a 32 Gb card should have 32 x 1000 x 1000 x 1000 = 32,000,000,000 Bytes, so the OPs card has a 'shortfall' of 1,484,750,848 Bytes. Then again, I don't know how the OP has measured the 'available' memory, maybe the card has 32GB but 1.5 GB is being used on 'overheads', though I would say this looks excessive.

[P_Stoddart]

In reality is does have 32GB. Computer electronic memory has to be built in multiples of 8 at the minimum. So in the physical sense you have 32GB (32x1024x1024x1024).

You say you have already had the overhead etc info. I presum they have explained that as memory goes up the structure to organise that memory has to increase as well. Example FAT to FAT32 then NTFS. In real term is just 6% of the total.

Actually this does not just occur on memory. It also happens with pixel counts. I've seen some cameras claim 10MP when the actual photo taken is just over 9MP. Rounding up for sales.

Bottom line every single 32GB card will do the same in the market place. Whether SD, XD, CF or any other format. Therefore where do you go? Don't buy memory? All memory is treated like hard drives so the formats are the same whatever make or specs.

Hence the trading standard's position.

[LargeFormat]

Formatted sizes are considerably less than nominal sizes. Just looking at my laptop in front of me the 256GB SSD shows a net size of 231GB while the 1.5TB shows 1.36TB.

[Benchista]

I have three of these cards (well, I'm assuming they're the same, as that's all 7Dayshop appear to sell). I've just had a look, and my computer reports the size of all of them as 29.8GB, or in each case, 32,010,731,520 bytes. However, my computer is actually wrong - by the official definition of the prefix "giga", that's 32GB.

The official definition can be found in this document published by the International Committee on Weights and Measures, and this is what they say on the subject:

"These SI prefixes refer
strictly to powers of 10.
They should not be used to
indicate powers of 2 (for
example, one kilobit
represents 1000 bits and not
1024 bits). The IEC has
adopted prefixes for binary
powers in the international
standard IEC 60027-2:
2005, third edition,
Letter symbols to be
used in electrical
technology – Part 2:
Telecommunications and
electronics. The names and
symbols for the prefixes
corresponding to 210, 220,
230, 240, 250, and 260 are,
respectively: kibi, Ki; mebi,
Mi; gibi, Gi; tebi, Ti; pebi,
Pi; and exbi, Ei. Thus, for
example, one kibibyte
would be written:
1 KiB = 210 B = 1024 B,
where B denotes a byte.
Although these prefixes are
not part of the SI, they
should be used in the field
of information technology
to avoid the incorrect usage
of the SI prefixes."

Which you clearly know, Roger, but I thought that bit might be instructive for others... and I think it might be at the heart of the statement from Guernsey Trading Standards - do they understand that you understand the difference between GB and GiB? If they do, I find there position slightly odd.

So, my 3 cards all meet the spec, which proves that it can be done by Kingston. Yours doesn't. I think you're right - is 95% of what you paid for acceptable?

[Steve52]

So how many 'less pictures' will the card hold? I only use 4MB cards and can take about 950 pictures on my A350 per card. If you are taking in the region of 7500+ pictures on your 32MB card, does it matter whether its 7500 or 7450?

I can see your point of 'doing what it says on the tin', but would you still complain if in the small print it stated 'actual memory 30.5MB'?

[Benchista]

I get around 1100 pictures on my (actually) 32GB cards - so if I was using Roger's card, I would get around 1045. Does it actually matter, especially given the ridiculously cheap price of these cards? Well that's a question only Roger can answer.

[P_Stoddart]

In real terms that 32GB card will have

343597338368 or in hex 800000000

Computer memory flash or otherwise is binary so it has to come out at number that divides by 8 alot.

so your format is 31.25GB therefore is 33554432000 of bytes.

KB has always since the 70s been 1024 bytes and 1MB has always been 1048576 (you see this in PC memory test on boot up)

Memory has to double to increase because of electronics.
So you always get 2 4 8 16 32 64 of memory. It is to either double the number of chips or double the electronics on the IC. So in real terms you are getting more than 31.25GB it is actually rounded down if you was using the 31250000000.

So you are not being cheated at all.

Take a PC that has 512MB of ram you might think you are paying for 512000000 bytes of ram. But you actually get 536870912. Which you would see if you change the BIOS to slow boot.

Hopefully that makes sense.

[RogerChristie]

Well that stirred up some interest! I’m sure you don’t want to read all the details. (As I recall, my submission to the GTSS ran to 14 pages!) However there are a few points I’d like to address.

Thank you for your observation, P_Stoddart. The architecture of the memory chips means that every “32GB” memory card will have a little over 32 Gigabytes (something like 32,212,254,720 bytes, which is 30x1024x1024x1024), but it is normal that a few blocks are found to be faulty and are disabled, so the indicated available memory will typically be a little less than maximum.

Yes, PeteRob, there is a binary notation. Benchista has it right. (But, sorry P_Stoddart, what you say in your last post doesn’t make sense.)
1 GB (gigabyte) is 10^9 bytes (so 32 GB = 32,000,000,000 bytes).
1 GiB (gibibyte) is 2^30 bytes (so 32 GB is about 29.8 GiB).
However Microsoft’s software engineers have been careless, so when you check the capacity of your memory hardware (hard drive, flash cards, or whatever) you will see GB where it should say GiB. You will see though that the dialog reports the figure in bytes as well. (So, for example, where Windows Explorer Properties dialog shows 7.49 GB (when they actually mean 7.49 GiB) the dialog also shows 8,049,807,360 bytes.) That also means, LargeFormat, that your 231GB should actually read 231GiB (which is 248GB), and the 1.36TB should say 1.36TiB (being 1.495TB).

I was very careful, Stevet, to say that the 30.5 GB on my card is total available memory, including any overhead. The overhead on the “empty” card was in fact 16 kilobytes (actually 16,384 bytes), so way way short of 1.5 GB!

I did in fact wonder whether I had made an error or misunderstood something. The clincher was when I checked my other memory cards (all using FAT32, P_Stoddart). The data were as follows:
Nominal memory (Gigabytes) …. Actual total memory (bytes) …. Overhead (bytes)
……………… 8 …………………………………… 8,016,560,128 …………………………… *
……………… 8 …………………………………… 8,049,807,360 …………………………… *
……………… 8 …………………………………… 8,049,807,360 …………………………… *
……………… 8 …………………………………… 8,049,807,360 …………………………… *
……………… 8 …………………………………… 8,016,560,128 ………………………… 32,384
…………… 16 ………………………………… 16,017,719,296 …………………………… *
…………… 32 ………………………………… 30,515,249,152 ………………………… 16,384

All of these are Kingston Elite Pro CompactFlash (CF) Memory Cards purchased from 7dayshop. The 16 GB card and two of the 8 GB cards had been used in the same tablet computer. Two of the 8 GB cards were in use in cameras, and the remaining 8 GB card is new and unused.
[* These cards all have data files added to them so I cannot quote the actual overhead on the “empty” cards without deleting all the files.]
None of these cards have been partitioned and all are using the FAT32 file system.
I did, in fact, check the same cards on two different computers with exactly the same results.

To those who imply that I’m just being mean: well, yes, Steve52 and Benchista, the odd 50 photos in a thousand maybe doesn’t really matter (although I actually got the card to supplement the little 40 GB HDD in my tablet computer). In fact, as the card ages, more blocks will fail and be disabled so the capacity will gradually decrease anyway. (And, unless you are looking for it you would probably never notice the deterioration.) It is covered in the specification and is quite normal.

The thing that really annoyed me was that Kingston is packaging and 7dayshop is advertising memory cards clearly stating 32GB when mine had no such thing. And they then tried to palm me off rather than being honest about it. If they had held their hands up and acknowledged the fault I probably wouldn’t have pursued the matter further. Strictly, as I see it, the manufacturer and the supplier are misrepresenting the product. I’m not prepared to let them get away with it. The comment from 7dayshop: “…I would suggest that you do not purchase another 32gb cf card as they will all be the same” just made it worse. Thank you, Benchista, for checking your three cards. The figures you report would be about what I would expect, and they indicate that the comment from 7dayshop is not true.

I believe the matter is simple (as in my “flour” analogy), so it seems very strange to me that GTSS have decided not to take any action, and even more worrying that they suggest that I should accept a 4½% shortfall on the grounds (and I paraphrase) that they all do that!

So, to get back to my original question, does anyone have access to the CFA specification issue 4.1 (or an equivalent document such as the spec for PCMCIA or SD cards), and is there any indication there as to what is an acceptable memory shortfall in a new unused card?

Regards,
Roger christie

[RogerChristie]

I do feel obliged to add that I don't want to put off anyone buying from 7dayshop. I have otherwise been delighted with their goods, services, and prices over several years now. That only makes it doubly disappointing that they have handled this particular business so badly.

In this case they did eventually offer a refund, but they are not prepared to acknowledge that the memory card is faulty. I see that the item is still listed on their website. How many other customers have unknowingly received a memory shortfall?

[Benchista]

Roger, I don't believe the product in general is faulty in any way. Mine all make 32GB, and they are indeed the same card. Rather I suspect there's either a faulty batch, or your particular card is faulty - not sure I see any need for them to withdraw the product if yours is the only complaint they've had, and as my experience shows, there's nothing intrinsically wrong with the base product. (Far from it; I think they're great and great value cards, well capable of recording video on the 5D II and 7D, and of exceeding the stated burst depth for JPEGs on the 7D at 8 fps - not bad for a 133x card.) So if they've offered you a refund, in all honesty, that's the best you're going to get. Personally, I would be suspicious of your card on the grounds that it's not up to spec, and if there are physical problems before you start, I wouldn't want to trust valuable data to it. And personally, I wouldn't accept a capacity more than 2% below the quoted level for a cheap card (which to be honest, this is), and would expect cutting-edge cards not to dip below at all - but that's just me. I suspect you won't actually find the figures you're looking for.

[RogerChristie]

What you say, Nick, is just what I expected three months ago. I don’t want to write another essay but this is getting to feel like pulling teeth so I’ll have to give you more details. To save the stress on my keyboard, fingers and brain (not necessarily in that order) I’ll try some stream-of-consciousness bullet points (hey! sounds good!) with the occasional comment inserted:
• I discovered “32GB” is really 30.5GB.
• I asked Kingston’s “help desk” for advice but just got a standard text without the blanks filled in!
• asked 7dayshop for advice. No response, so three weeks later asked again.
• advised to complete Fault Report and return the item.
• item came back with “can’t find any fault” comment.
Doh! It’s only got around 5% of its memory missing guys. How hard is it to see that? Especially when I told you! Reading between the lines has never been my strongest suit, but it’s starting to look suspicious to me.
• sod the e-mail/internet (can I say “sod” Mark?), I write a paper-and-stamp letter. In my customary thorough (some might say pedantic) manner I include everything so far: thorough testing; researched info (GB, GiB, overheads etc); assurance that I’m not confused, stupid or ignorant; threat to contact Trading Standards; copies of all other communications. It ran to nine pages. I also copied the letter to … well I couldn’t find any reference to the MD or any senior managers, but I did find the name of a guy listed as a Partner, so I sent him a copy.
• Message via website offering rebate. The key point was “…I would suggest that you do not purchase another 32gb cf card as they will all be the same”
• Sent all info to Guernsey Trading Standards Service.
• After six weeks of “research”, now GTSS tell me “…30.5GB on a 32GB card is considered acceptable throughout the removable memory industry”.

There is a lot of confusing nonsense about electronic data storage devices. Sadly (being blunt) some of it is being parroted above! (My cynical view is that it is generated by the manufacturers and promulgated down the supply line to confuse the ignorant and allow them to sell off-spec items that they would otherwise have to scrap.) Now it doesn’t really require my 30 years experience in a high-tech industry, but the technology doesn’t phase me, I am entirely confident that I know what I am talking about, and the “explanations” from the manufacturer, supplier, and Trading Standards are just plain wrong.

I’m afraid your last comments, Nick, just serve to confuse matters further. I never suggested that there is a general fault with the product. In fact I had no reason to suspect that this might be anything other than a one-off until 7dayshop’s comment “they are all the same”. I certainly never suggested that the product should be withdrawn. (Having said that, you have little to justify your claim that there is nothing intrinsically wrong with the base product based on a sample of three!) As to good value, yes they are; and if I’d bought it as a 30.5GB card I would have been quite happy. And I am satisfied that I got a refund (although I had to bully them for it). If you think this is the start and end of it then you are way off-track. My main complaint is that I received a faulty card and no-one will acknowledge the fact; and I’m damned if I can see why they would want to do that.

Let me try and cut through the eight-letter B word that Mark won’t let me use. Let’s forget memory cards for the moment. It really doesn’t matter whether it is gigabytes, apples or six-inch nails. Your supplier is offering a pack 32 widgets. You order and pay for a pack of 32 widgets. When it arrives you find the pack contains only 30 widgets. In general terms it’s called short measure. Certainly when Trading Standards was known as Weights and Measures it wasn’t tolerated. Where’s the problem?

Regards
Roger

[Scphoto]

Seriously there's more things wrong in the world than a few missing or misrepresented GB's.

I know the capacity of anything I buy isn't going to be what it says on the box, maybe the manufacturers could be more honest - problem is they've been working this way for many years. When HDD's were smaller I guess it wasn't an issue as you 'lost' less.

Old CRT screen sizes were the same, measured diagonally to a position inside the case!

Does it bother me, no cos I know what to expect.

[Roy5051]

This really is quite easy to explain - 30.5 x 1024 x 1024 = 31,981,568 which is near enough the same as 32 x 1000 x 1000 = 32,000,000; this means that the manufacturer is using 1000 bytes to equal 1MB and 1000MB to equal 1GB. Wrong, I know, but most manufacturer use the same equation. There is nothing you can do about it, so just accept it.

[RogerChristie]

So what you are saying, Stephen, is that it is only a little lie and doesn’t really matter. Have you any idea how profoundly disheartening it is to read what you’ve written?

It may seem over the top to cite Edmund Burke’s "Evil triumphs when good men do nothing". A simpler question may be: How many little lies are you prepared to tolerate?

It’s rather turgid, but let me quote Erik Naggum:
“If a lie is obvious, people will understand it right away, or as soon as they try to act on it, tell it to someone else, etc. This makes big lies harmless; they can even safely be used as humor. In contrast, small lies are capable of serious, long-term damage, and even if it is uncovered, people may have based too much on them to be able to correct it. In a sense, small lies may become truth because people act as if they are. This is why they must be stopped, and stopped in time. Stopping a small lie may require much effort, especially in alerting the complacent masses of its falsity, but stopping it now will require less effort than stopping it later, especially if it spreads.”

So I’ll ask again: How many little lies are you prepared to tolerate? You can be sure that the liars think they know. Once you have been conditioned to accept little lies, do you then become immured to bigger ones? And one day you wake up and find it is all out of control and beyond recovery.

And when the majority finally recognize Edmund Burke’s warning it is undoubtedly too late.

[Benchista]

the manufacturer is using 1000 bytes to equal 1MB and 1000MB to equal 1GB. Wrong, I know,

Actually, it's right - using multiples of 1024 is what is actually wrong. And you've misunderstood Roger's issue.