Weight of a gigabyte

[koti]

1 minute ago, swansont said:

The charge is trapped within the transistor, and leaves when the bit is erased. 

Okay, but does the number of electrons stay the same within the transistor through out the process of writing/erasing or not ?

[swansont]

3 minutes ago, koti said:

Okay, but does the number of electrons stay the same within the transistor through out the process of writing/erasing or not ?

Within the transistor? No, probably not, and mentioned in the link. But that's not what was under discussion.

[koti]

1 minute ago, swansont said:

Within the transistor? No, probably not, and mentioned in the link. But that's not what was under discussion.

Okay, so let me see if I got this. The only "significant" mass change comes from the energy difference of the states like John Cuthber and you are pointing out. The electron mass contributes but it's far less significant of a contribution than the change in energy states? Does that sound right or am I still shooting in the dark here?
If this is right, I still don't get how the energy difference is achieved without moving electrons in and out of the flash memory device. 

[swansont]

1 hour ago, koti said:

Okay, so let me see if I got this. The only "significant" mass change comes from the energy difference of the states like John Cuthber and you are pointing out. The electron mass contributes but it's far less significant of a contribution than the change in energy states? Does that sound right or am I still shooting in the dark here?
If this is right, I still don't get how the energy difference is achieved without moving electrons in and out of the flash memory device. 

I'm saying the electron mass contributes nothing, since the electrons are in the chip the whole time. You move electrons from one part of the circuit to another.

[koti]

3 hours ago, swansont said:

I'm saying the electron mass contributes nothing, since the electrons are in the chip the whole time. You move electrons from one part of the circuit to another.

Well you could have saved us quite a few threads by stating that in the beginning. Assuming you didn't have to researching it like I had to while participating in this thread
Anyway, after reading up a little I agree with the above and my electron/mass calculation a few posts back is useless in this context. 

[swansont]

2 hours ago, koti said:

Well you could have saved us quite a few threads by stating that in the beginning. Assuming you didn't have to researching it like I had to while participating in this thread
Anyway, after reading up a little I agree with the above and my electron/mass calculation a few posts back is useless in this context. 

You mean like from my second post?

Quote

You trap some electrons in a transistor that makes up the bit. The electrons come from somewhere else in the circuit.

 

Yeah, I should have said that.

[John Cuthber]

Is this an entry for the competition to make  longest thread about a fake video?
It's an old joke
http://dilbert.com/strip/1995-01-30

Just to prolong the agony, I rather suspect that the chips use error correcting codes.

https://en.wikipedia.org/wiki/Error_detection_and_correction

the practical outcome of which is that the number of 1s ad 0s won't actually depend strongly on the data.

So the actual stored energy associated with the transistors is even smaller than you might first expect.

 

Now to really put the cat among the pigeons:

Does the T Delta S term for the entropy of the encoded data mean that (except at absolute zero) there is energy associated with the randomness of the bits?

 

Edited August 14, 2017 by John Cuthber

[koti]

20 hours ago, John Cuthber said:

 

 

Now to really put the cat among the pigeons:

Does the T Delta S term for the entropy of the encoded data mean that (except at absolute zero) there is energy associated with the randomness of the bits?

 

If yes, it would mean that the mass difference is even lower. So your original calculation might be wrong?

[Carrock]

On 14/08/2017 at 9:27 PM, John Cuthber said:

Now to really put the cat among the pigeons:

Does the T Delta S term for the entropy of the encoded data mean that (except at absolute zero) there is energy associated with the randomness of the bits?

 

No. But erasing low entropy data increases overall entropy. The necessity of discarding data gives a theoretical limit to the efficiency of computers.

The effect has been observed.

 

[John Cuthber]

On 15/08/2017 at 6:14 PM, koti said:

If yes, it would mean that the mass difference is even lower. So your original calculation might be wrong?

My calculation was based on their "pretend" mass change, rather than any model of how the energy was held.

 

[koti]

1 minute ago, John Cuthber said:

My calculation was based on their "pretend" mass change, rather than any model of how the energy was held.

 

Sure, I get that. Just an intuitive approach which is not a good one in most cases but in this case it's obvious that it's just impossible for such a small energy change be worth 0.186mg of mass. We would see every day life machines and devices change their mass significantly enough to notice it if that were true.

[John Cuthber]

On 8/16/2017 at 4:47 PM, Carrock said:

No. But erasing low entropy data increases overall entropy. The necessity of discarding data gives a theoretical limit to the efficiency of computers.

The effect has been observed.

 

I'm not sure about flash drives, but in many cases the file system doesn't actually erase the data, it just sets the pointers so that the data can be overwritten.
So, if you fill a hard drive with data - say a video file- then "erase" it , you hardly change the contents of the  disk at all.

[Carrock]

1 minute ago, John Cuthber said:

I'm not sure about flash drives, but in many cases the file system doesn't actually erase the data, it just sets the pointers so that the data can be overwritten.
So, if you fill a hard drive with data - say a video file- then "erase" it , you hardly change the contents of the  disk at all.

The data, or most of it, is not erased and can be recovered.

Only permanently erasing data necessitates increased entropy.

Overwriting and destroying data increases entropy; the increase in entropy has just been delayed until the original data is destroyed.

 

[Strange]

10 hours ago, Carrock said:

The data, or most of it, is not erased and can be recovered.

The data in flash memory can be quickly totally erased (in a flash, hence the name). Although file system software does not do this when a file is deleted, but when space has to be recovered.

[Uses_x]

Wouldn't the secretions from the sweat glands on his hands massively affect the results? At several points on the video, he touches the SD cards without even having gloves on; from the time behind the scenes, where he must individually touch every SD card, a lot of substances probably build up on the SD card.

The sweat glands on the hands (which are found on most parts of the skin) are the eccrine sweat glands, which secrete (from Wikipedia; it's the only thing I have right now):

Quote

The secretion of eccrine glands is a sterile, dilute electrolyte solution with primary components of bicarbonate, potassium, and sodium chloride (NaCl),^[5] and other minor components such as glucose, pyruvate, lactate, cytokines, immunoglobulins, antimicrobial peptides (e.g., dermcidin), and many others

 Many of those substances don't evaporate, and I suspect that the pores in the plastic would fill with those substances.

 

I guess it could vary depending on what order he did it; if he went from filling an SD card and recording its weight, to deleting everything from the SD card, the 2nd step would have more collective residue, and as the mass is lower, it would prove that the experiment is true (not too scientifically, however), but I can't tell from the video.

I'm very surprised at the large change, which what I explain above might be the reason.

Edited August 18, 2017 by Uses_x

[John Cuthber]

It's more likely that they just faked the video.