SSD Use During YouTube or Streaming

[bunnspecial]

That's awesome!!! LOL!

I have an old SCSI ZIP and an parallel ZIP drive here in my parts pile somewhere. Also had a JAZ drive, but haven't seen it in a while. That's quite ingenious!

Did you ever play around with the higher density 3.5" floppy drives? I had one of the 2.88MB ones but it was pretty shady.

No, never did. It sounded interesting but also scares me a bit.

What I did use, and actually back in the day(still using PCs) were my choice over ZIP, were LS-120 drives. The drives are convenient since they handle 1.44mb floppies also.

I'd have been interested to see what happened on Macs if Woz had still been active in the company when 1.44s became common. You may well know this, but a "normal" PC floppy is 360kb(LD single sided) or 720k(LD double sided). Woz realized that if you vary the rotational speed according to head position, you can squeeze 400kb on each side, or 800kb on a two-sided disk.

The Apple ][ used that technology, although I forget what capacities they managed on 5 1/4" disks.

The early Macs had a chip called the IWM, or "Integrated Woz Machine" which replaced the floppy card on the ][e to drive the disk drives.

Unfortunately only Macs with built-in floppies could actually handle Macintosh formatted low density disks, and of course you need a Mac specific drive. The beige G3s and Powerbook Wallstreet/PDQ were the last to ship with floppies. USB floppies like were popular for the iMacs can't do Macintosh low density disks. As a side note to that, though, the first gen iMac G3s have a SWIM chip(successor to the IWM) and IIRC actually have a floppy drive header on the board, or at least the solder pads to install a header. I know I've seen people hack together floppies on iMac G3s...

 

[OVERKILL]

No, never did. It sounded interesting but also scares me a bit.

What I did use, and actually back in the day(still using PCs) were my choice over ZIP, were LS-120 drives. The drives are convenient since they handle 1.44mb floppies also.

I'd have been interested to see what happened on Macs if Woz had still been active in the company when 1.44s became common. You may well know this, but a "normal" PC floppy is 360kb(LD single sided) or 720k(LD double sided). Woz realized that if you vary the rotational speed according to head position, you can squeeze 400kb on each side, or 800kb on a two-sided disk.

The Apple ][ used that technology, although I forget what capacities they managed on 5 1/4" disks.

The early Macs had a chip called the IWM, or "Integrated Woz Machine" which replaced the floppy card on the ][e to drive the disk drives.

Unfortunately only Macs with built-in floppies could actually handle Macintosh formatted low density disks, and of course you need a Mac specific drive. The beige G3s and Powerbook Wallstreet/PDQ were the last to ship with floppies. USB floppies like were popular for the iMacs can't do Macintosh low density disks. As a side note to that, though, the first gen iMac G3s have a SWIM chip(successor to the IWM) and IIRC actually have a floppy drive header on the board, or at least the solder pads to install a header. I know I've seen people hack together floppies on iMac G3s...

Oh yes, I have an LS-120 here somewhere too!

I think the highest capacity from the 5.25" drives was 1.2MB IIRC. My old Powerbook has a 3.5" drive in it.

 

[javacontour]

All this talk about SSD lifecycles...has anyone had one fail?

I replace enterprise grade SSD and NVME devices all the time. They fail. Everything has a failure rate.

Heck, at home, this past week I replaced my second WD Black HDD in three years. Can't RMA this one because it was the replacement for the last one, LOL.

 

[y_p_w]

SSD's don't need shock protection, that's one of the big plusses of using them in a laptop. Being shipped in a bubble mailer shouldn't cause any issues at all.

Like any kind of electronics they can be subject to physical damage. It wouldn't be like shock that causes a head to scratch a platter, but it's certainly possible to crack a solder ball connecting the controller or NAND flash to the board. But wrapped up securely, I wouldn't worry about it.

 

[y_p_w]

SSDs are far more resilient than HDDs for sure.
But are not fool proof as you suggest. Nobody would throw SSD drives across the room and expect them to survive.
Come on.........

Back when CDs were fairly new, I read about a retired politician (I won't name him) who was enamored with CDs. He'd buy them and throw them around the room like frisbees, and tell everyone how durable they were and show them full of scratches but still working. The only problem that most people don't realize is that it's only polycarbonate on one side. The read layer is actually not that thick, and the protective layer is actually sprayed on. One can press a fingernail into the label side and see it move from the other side. It's also fairly easy to just scratch it off. I did that with a few CD-Rs that failed the write and once it's scratched it just flakes off easily just like a car with bad paint.

The big worry about SSDs would be damage to the connector or board. The ICs themselves are probably extremely durable, but the connections can fail if poorly handled.

 

[y_p_w]

Yeah, the ESD emphasis I think stems from the earlier years of computing where components weren't so durable. When I first started working on computers, which was when I was about 8 years old, I had no idea about it of course. By the time I was 10 (1990) I had read about how it could damage semiconductor components and how you were supposed to wear an anti-static wrist strap or at least keep yourself grounded. Of course SIMM's or DIMM's weren't a thing then, even when my dad bought a TI 486SX/25 laptop, it took RAM that had pins that you had to push into the board.

That concern and cautionary principle carried well forward into today where it is still advised, despite increased durability and resiliency that has made its way into these components.

I used to work a company that handled a lot of electronics, even though I rarely need to go into the lab, I look a class to get ESD certified, and as a result I got a blue ESD smock with my name on a patch. I think the regular lab guys did as much as they could, but I've never really heard of anything really getting destroyed by ESD. I've done some lab work before - simple probing and some handling of the boards, but I don't recall too many cases where anyone bothered to use an ESD strap. I certainly wouldn't worry about a packaged SSD or board, although they're really paranoid about it and package everything in metaled bags.

 

[y_p_w]

I had a 250GB SSD fail in a 2012 Macbook pro after about a year earlier this year. The macbook got extremely slow, and would take like 4 minutes to boot, and 30 secs to open the browser, while i did not lose ANY information, i now make sure i back up often.

What happens is that as these things age due to more erase cycles (which cause a little bit of damage each time), it gets to a point where the error correction kicks in. A new drive will almost never use error correction. It's kind of technical, but it's not really reading data. It's testing the voltage threshold that turns on a single MOSFET. But the ability to do this consistently gets fuzzy the more erasures it gets. There are a lot of little things that are done including error correction codes and retries with different threshold voltages, but eventually it starts taking more time to read. If it's really bad, it's going to take several tries. I haven't seen it in an SSD, but I've got a few older USB flash drives that are painfully slow to mount and to read files. But the data is still 100% intact.

I remember way back when single-level NAND was the thing, and I thought that their 1 million rated erase cycles didn't sound that great since I thought that drives get rewritten all the time. Then it went to MLC (2 bits per cell) and down to 100,000. Then TCL and now it's 10,000. 3D NAND stacked on top of each other seems to help. Now it's quad-level and down to less than 1,000 cycles. But regardless of what most people thing, that should generally last the reasonable life of a product for nearly anyone. My 1 TB WD Blue SATA SSD has been in my computer for more than 2 years and one monitoring tool is still saying 100% wear level.

The big thing that a lot of these do is use a small (but expensive) SLC cache at the front end and hopefully that reduces the amount of writes to the cheaper NAND.

 

[OVERKILL]

Like any kind of electronics they can be subject to physical damage. It wouldn't be like shock that causes a head to scratch a platter, but it's certainly possible to crack a solder ball connecting the controller or NAND flash to the board. But wrapped up securely, I wouldn't worry about it.

Yes, hence my qualifiers in subsequent posts. As long as you don't crack a solder joint, damage the PCB or wreck a component they are incredibly resilient.

 

[y_p_w]

Yes, hence my qualifiers in subsequent posts. As long as you don't crack a solder joint, damage the PCB or wreck a component they are incredibly resilient.

Sure. The biggest danger with hard drives is scratching the surface where data is recorded while it's spinning. The systems that have a sudden motion detection mechanism usually just park the heads quickly where there should be no data, before there's any physical shock that might cause the head to crash.

I don't think normal vibrations are likely to damage an SSD. But it's like any other piece of board level electronics. Heat cycles can be bad, and possibly the worst thing for longevity is electromigration. That's where the current passing through a conductor slowly causes the material to migrate - both in silicon and on printed circuit boards. I remember hearing about ways to mitigate the effects, like using specific metal alloys as well as making sure the traces were a minimum width/thickness.

 

[OVERKILL]

Sure. The biggest danger with hard drives is scratching the surface where data is recorded while it's spinning. The systems that have a sudden motion detection mechanism usually just park the heads quickly where there should be no data, before there's any physical shock that might cause the head to crash.

I don't think normal vibrations are likely to damage an SSD. But it's like any other piece of board level electronics. Heat cycles can be bad, and possibly the worst thing for longevity is electromigration. That's where the current passing through a conductor slowly causes the material to migrate - both in silicon and on printed circuit boards. I remember hearing about ways to mitigate the effects, like using specific metal alloys as well as making sure the traces were a minimum width/thickness.

Yes, a head crash can be devastating and the motion/drop detectors aren't always fast enough to catch it, seen enough of those. The Seagate mobile drives were the worst I've seen for that with the head crashing and taking the coating right off the glass platters.

Even abnormal vibrations aren't going to damage an SSD, they are very much like a stick of RAM in that regard in that they aren't dampened from vibration or impact because it's not necessary, even in ruggedized units like toughbooks. Think of automotive ECM's that aren't dampened and take wicked shock loads, many of them (like the old Ford EEC-IV ones) weren't even mounted with isolators, just straight to the body. As long as the PCB is rigid and not being manipulated the vibration/shock they can take is quite impressive, and this is of course aided by mounting them in a housing (like the Kingston one I posted earlier).

Now, of course there are limits/thresholds, but for an SSD versus a spinning disk, that is much higher since there are no moving parts.

That said, short detour:
A good friend of mine was in a rather severe car accident many years ago and had an ASUS 17.3" notebook in the car with him, just in a plain laptop bag, nothing fancy. He was nailed from the side when he was doing about 90Km/h our thereabouts and careened into the ditch where the laptop exited the vehicle and landed in a field some 100ft away (still in the bag).

I was able to take that drive out of that unit and copy all of his stuff off it no problem, and it was only mounted with those tiny rubber isolators. In fact, the entire unit continued to work (I gave it to one of my kids) but wasn't fit to be used for work anymore given what it had experienced.

 

[y_p_w]

Yes, a head crash can be devastating and the motion/drop detectors aren't always fast enough to catch it, seen enough of those. The Seagate mobile drives were the worst I've seen for that with the head crashing and taking the coating right off the glass platters.

Even abnormal vibrations aren't going to damage an SSD, they are very much like a stick of RAM in that regard in that they aren't dampened from vibration or impact because it's not necessary, even in ruggedized units like toughbooks. Think of automotive ECM's that aren't dampened and take wicked shock loads, many of them (like the old Ford EEC-IV ones) weren't even mounted with isolators, just straight to the body. As long as the PCB is rigid and not being manipulated the vibration/shock they can take is quite impressive, and this is of course aided by mounting them in a housing (like the Kingston one I posted earlier).

I'm not really that well versed on PCBs, but I do remember where electromigration was discussed in a digital design class. We were actually doing layout projects, and the talk was about how wide the traces had to be to reduce the chances of electromigration. Doubling the width would often reduce it to almost nothing, but that of course had an area penalty.

But PCBs can be really interesting how they can fail. The most obvious would be poor quality solder joints that break off, but this paper talks about ionic migration, electromigration, stress migration, and thermal migration.

http://www.narrowbandimaging.com/incoming/about_AMI__Read-Only__01.pdf

I guess they call some forms of ionic migration "dendrites" since they look like the dendrites in neurons. They increase the resistance of the trace before they become open, but worst case is that they short to the next trace.

 

[OVERKILL]

I'm not really that well versed on PCBs, but I do remember where electromigration was discussed in a digital design class. We were actually doing layout projects, and the talk was about how wide the traces had to be to reduce the chances of electromigration. Doubling the width would often reduce it to almost nothing, but that of course had an area penalty.

But PCBs can be really interesting how they can fail. The most obvious would be poor quality solder joints that break off, but this paper talks about ionic migration, electromigration, stress migration, and thermal migration.

http://www.narrowbandimaging.com/incoming/about_AMI__Read-Only__01.pdf

I guess they call some forms of ionic migration "dendrites" since they look like the dendrites in neurons. They increase the resistance of the trace before they become open, but worst case is that they short to the next trace.

We never got into it at that level of depth when I took microelectronics, but it clearly can't be much of an issue with production electronics given how long this stuff lasts. I assume it's just factored into the design process (mitigation) so that it's avoided or reduced to a level that it doesn't matter. Cold solder joints on the other hand, well, those have been an issue with all manner of devices from Xbox's to NVidia chipsets in notebook computers (remember those?).

 

[y_p_w]

We never got into it at that level of depth when I took microelectronics, but it clearly can't be much of an issue with production electronics given how long this stuff lasts. I assume it's just factored into the design process (mitigation) so that it's avoided or reduced to a level that it doesn't matter. Cold solder joints on the other hand, well, those have been an issue with all manner of devices from Xbox's to NVidia chipsets in notebook computers (remember those?).

Electromigration is still an issue - especially with smaller and smaller process nodes. That's supposed to be mitigated with automated tools that perform checks on the layout. But it's still possible to ignore the checks. I've heard of some companies that were so cost conscious to reduce pin count that they didn't follow best practices for the number of power and ground pins. I guess it's possible for someone to just take a chance.

 

[OVERKILL]

Electromigration is still an issue - especially with smaller and smaller process nodes. That's supposed to be mitigated with automated tools that perform checks on the layout. But it's still possible to ignore the checks. I've heard of some companies that were so cost conscious to reduce pin count that they didn't follow best practices for the number of power and ground pins. I guess it's possible for someone to just take a chance.

Yep, I suspect you are right, but it sounds like for the most part, the checks and mitigation mechanisms should prevent it from being a problem.

Great discussion BTW!

 

[Accent Abuser]

How much SSD use is there when watching YouTube ... or for that matter streaming anything on-line? Since SSDs have a finite life, wondering if watching a lot of YouTube or doing a lot of on-line streaming would be cutting down SSD life.

I'd imagine very little if any unless you're pushing the limits and running something stupid like 2gb of RAM. Anything open that can be fit into the RAM is, if RAM starts running out, then the operating system will start prioritizing certain things it thinks need immediately and other things (like that tab you've had open for weeks but never went back to) will get written to the page file. If you have 8-16gb of ram it's highly unlikely the page file is even used at all.

Windows 10 also knows if you're using an SSD and will behave differently to preserve the lifespan (like not defragmenting the drive).

 

[y_p_w]

I'd imagine very little if any unless you're pushing the limits and running something stupid like 2gb of RAM. Anything open that can be fit into the RAM is, if RAM starts running out, then the operating system will start prioritizing certain things it thinks need immediately and other things (like that tab you've had open for weeks but never went back to) will get written to the page file. If you have 8-16gb of ram it's highly unlikely the page file is even used at all.

Windows 10 also knows if you're using an SSD and will behave differently to preserve the lifespan (like not defragmenting the drive).

I only have the original 4GB and an SSD. I figure that a lot of the problems I had before were memory related, but when I got an SSD, it sped up like crazy. I suspect that a lot of that was using the drive when I maxed out, and the speed of the SSD covered up some of that delay when substituting partially with a hard drive.

Even then, there's almost no wear on my SSD after a couple of years. Most people don't actually use as much as they think. What wears an SSD is erase cycles breaking down the floating gate, and there are ways to reduce that. I previously mentioned the use of a single-level cell cache. That helps some with speed, but also a lot with longevity since it's nearly impossible to wear that out.

There are some weird performance issues when the SLC cache fills up, but only for faster SSDs.

 

[bunnspecial]

I'd imagine very little if any unless you're pushing the limits and running something stupid like 2gb of RAM. Anything open that can be fit into the RAM is, if RAM starts running out, then the operating system will start prioritizing certain things it thinks need immediately and other things (like that tab you've had open for weeks but never went back to) will get written to the page file. If you have 8-16gb of ram it's highly unlikely the page file is even used at all.

Windows 10 also knows if you're using an SSD and will behave differently to preserve the lifespan (like not defragmenting the drive).

Tell that to my 8 month old Macbook Pro, which has been thrashing its drive since day 1. It's a known "defect" that is better than it once was, but still isn't good.

I've lost roughly 1% of useable life of the drive per month reliably...

 

[Leo99]

I've worked professionally in IT for over 20 years. I'll gladly throw that Kingston across the room right now, plug it into a a USB connector and it'll work. There's nothing that's going to be shock impact damaged inside an SSD, that's the 2nd biggest benefit of them, the primary of course being the massive speed increase over a spinning disk.

Now, if I go put a .308 round through it, that's going to kill it, because I'm destroying the PCB, but just like with a stick of RAM, as long as you aren't bending it, breaking the solder joints on the PCB or destroying it with ESD, they are incredibly durable. This goes doubly for those securely mounted inside a protective aluminum housing like the SATA ones.

I'm not saying they are "fool proof". They are vulnerable to certain conditions/exposures (like ESD) and of course they can suffer a component failure like any other piece of IT equipment. However, they are not vulnerable to shock impact unless you reach the point of physical damage to the PCB or its components, which requires considerable force given how light they are.

This post made me chuckle. I remember years ago when I was first building my first computer and people at my work were showing me a stick of RAM and told me the precautions. We were holding the RAM like baby Jesus fresh out of the womb with anti static wrist cables and whatnot. Then I went to a computer show and bought some RAM from the RAM purveyors there. They were not so careful. LOL. They threw them around like a bar of candy.

 

[bunnspecial]

I have a spare watchmaker's bench that somewhere along the way became my organized dumping ground for RAM and other computer parts. I had drawers for RAM type-i.e. SDRAM, DDR, DDR2, etc. There were certain sticks that I had bunches of, like somehow or another along the way I got probably 30 sticks of 128mb PC133 ECC, and I use to toss them in zip lock bags.

I have another box somewhere of some older RAM types-30 pin SIMMs, 72 pin SIMMs, and some obscure stuff like Mac VRAM SO-DIMMs. A lot of these weren't explicitly marked with capacity, so I'd sort identical sticks, then Google the chip IDs on them, find the capacity of each chip, and multiply it by the number of chips ion the stick to arrive at capacity. Once it was done, I'd take a Sharpie and write it on the PCB. I'd do this in the evening while sitting on the couch watching TV on a TV tray.

I made a bunch of folks panic with how I was treating the stuff, but I never killed a stick.

BTW, at one point maybe 5 or 6 years ago I'd bought some 256mb PC-133 SIMMs from a seller in China. I bought 20 of them, and they arrived in plastic trays of 10. The problem was I was buying them for a specific application where I needed low density 256mb sticks(16 chips), which is what a lot of G3-era Macs require. They will only read 256mb high density at half capacity. These were advertised as low density, and I received high density. The seller sent me another 20 low density sticks, refunded my money, and told me to keep the HD sticks.

I was left with a bunch of brand new obsolete RAM that I didn't have a ton of use for. I have a friend in London who I talk computers with and trade parts sometimes, and one of the things he was short on in his personal stash was PC133 SD-RAM to fill out some of his computers. I went to my RAM drawer and quite literally started just piling sticks in, including some of those 128mb ECCs I mentioned above and the HD 256s. I don't know how much RAM I sent him other than that it was a 4lb box of RAM. I still joke with him about how many people can say they've received RAM in quantities measured by weight

 

[Trav]

Samung 850 EVO in an old workshop laptop. Fox news streaming most of the day, it is about 4 years old, never any problem and has over 20 years left in it at this rate. I have 6 of these drives, all have been trouble free.