HDR-Fox T2 Freeview box

[Ezra Pound]

We are probably further along that curve than the flurry would suggest, you would only get the real picture if users were forced to keep their Humax box until it actually failed, but they also have the option of switching to get new features.
I doubt if any PVR manufacturer expects (or wants) users to keep their old tech. for 10+ years, unfortunately for them there is very little extra that I (and the others on this forum) actually want to add, more tuners maybe ?, 4K ?, no they won't be transmitting that any time soon, working (or better) Internet services ?, Yes I suppose so but add on boxes are quite cheap and all new TVs have this anyway

 

[Black Hole]

It may be my imagination, but there seems to have been a flurry of people recently with dead or sick FOXs. Are we finally hitting the far end of the bathtub?

I've been expecting it. If they are mostly dry capacitors we have the means to repair, but I anticipate Alzheimer's in the "non-volatile" store!

 

[Owen Smith]

I've been expecting it. If they are mostly dry capacitors we have the means to repair, but I anticipate Alzheimer's in the "non-volatile" store!

Flash life is determined by how many times you write and erase it, which is typically very infrequent on this sort of device but I don't actually know for the HDR Fox T2 with or without custom firmware. Given there is a hard disc, why would anything be written to Flash except when updating the firmware?

 

[aw1]

My main HDR freezes at least once a week. I do not use it much to record (nowadays watching mainly internet streams), but have 2 TB of recordings back to 2011 which I sometimes watch.
This I bought new and later upgraded to a 4 TB disk.

My remote HDR bought second hand is seldom used, and my spare (also second hand) I only used once to check that it worked and install the CF.

I have the main one backed up to a NAS, and am backing up the remote one as I type, with the intention of decommissioning all of them, viewing the recordings via a PC connected to my TV.

So. Anybody interested in 3 free HDR Fox T2s, for collection from near Crewe or Bridgwater in a month or two, or even relatively local delivery? I don't want to faff around with shipping or EBay selling.

 

[Black Hole]

Flash life is determined by how many times you write and erase it

Wear out yes, data retention no.

 

[Black Hole]

Bridgwater

I take it you don't mean Somerset?

 

[prpr]

Anybody interested in 3 free HDR Fox T2s, for collection from near Crewe or Bridgwater in a month or two

Yes, if you mean in Somerset.

 

[prpr]

Given there is a hard disc, why would anything be written to Flash except when updating the firmware?

Normal operation of the box writes to the flash - retuning, changing the recording list, even changing channels and adjusting the volume.

 

[aw1]

Yes, if you mean in Somerset.

I do indeed. I will bring my other two HDRs the next time I come down to visit Mum, probably towards the end of January.
We can then discuss delivery, collection or handover. How far are you from Bridgwater?
I am glad that they are going to someone who wants them rather than joining my Thomson TiVo in the recycle pile. You can have that too if you want. I am amazed to see that one was sold recently on eBay for £80.

 

[Owen Smith]

Normal operation of the box writes to the flash - retuning, changing the recording list, even changing channels and adjusting the volume.

OK so that will gradually wear the flash. However, you get 100,000 erase cycles per block on average, and you get a lot more writes provided it gradually journals the writes across each block. And then how many blocks the writes are spread across over time makes a large difference.

* 100,000 cycles is what most flash datasheets say, but I don't know the figure for the one in the Humaxes.

 

[Owen Smith]

Wear out yes, data retention no.

I work with flash devices in my day job. Retention when simply reading them has no quoted lifetimes in the manufacturer datasheets, if the data was written correctly in the first place the retention life is deemed to be infinite. I suspect several decades would have to pass before problems were seen.

 

[MymsMan]

While the firmware part of the flash memory is largely static the EPG and Recording schedule databases are also held in flash memory and these are updated multiple times daily so are we not liable to be reaching the limit for writes to flash in those critical areas?

 

[Black Hole]

has no quoted lifetimes in the manufacturer datasheets, if the data was written correctly in the first place the retention life is deemed to be infinite.

So I take it your work doesn't have to pass formal verification, or the MD signs off a waiver.

 

[Ian Manning]

I was interested to read of the Manhattan T3-R on here - I'd never heard of it before. I've been looking at alternatives for my ageing HDR but the one feature that I can't easily find elsewhere is the ability to decrypt and copy HD recordings onto other devices (either for watching on portable devices when away from home, or for long-term storage - e.g. music shows that I might want to watch again).
I was thinking that the HDHomerun might provide a solution to this. Presumably the Manhattan box doesn't?

 

[Owen Smith]

So I take it your work doesn't have to pass formal verification, or the MD signs off a waiver.

The flash manufacturers quote no retention lifetime in their datasheets. What do you expect my employer or any other company making products that contain flash to do about it? We take the datasheets at face value. Developing products is time consuming enough as it is without also assuming you have to distrust every line of every datasheet for parts you buy in.

What we do is accelerated life testing of the final assembled products, at high temperature and raised voltage rails and simulated usage. This is industry standard practice and simulates 5 years of life in about a month.

 

[Owen Smith]

While the firmware part of the flash memory is largely static the EPG and Recording schedule databases are also held in flash memory and these are updated multiple times daily so are we not liable to be reaching the limit for writes to flash in those critical areas?

It depends. If they are constantly written to the same place in the Flash then yes. But if they are written into a wear levelling flash filing system then each time it is written it will go somewhere else in flash. So then it is the number of times it is written multiplied by the number of different places it can go in the flash, which can be hundreds or thousands. The firmware is static and probably won't be moved as part of the wear levelling, so that reduces the space available for levelling. Some early Tesla cars are struggling with this because the firmware has got a lot larger over the years, and now so little flash is left for wear levelling of the dynamic parts it's been wearing out.

 

[Black Hole]

The flash manufacturers quote no retention lifetime in their datasheets. What do you expect my employer or any other company making products that contain flash to do about it? We take the datasheets at face value. Developing products is time consuming enough as it is without also assuming you have to distrust every line of every datasheet for parts you buy in.

What we do is accelerated life testing of the final assembled products, at high temperature and raised voltage rails and simulated usage. This is industry standard practice and simulates 5 years of life in about a month.

I was in military design, where a service life of 20 years was required. If there wasn't a datasheet specification to verify some parameter on which a design depended, we couldn't use it (or if we did there would have to be justification and for that to be signed off by the customer). UV EPROM data retention life was a problem, but some manufacturers did accelerated lifetime testing for the military market.

Fundamentally: if it's not in the datasheet, the buck stops with you. What use is a NV storage device without at least some idea how long the data is guaranteed to be retained for? It certainly isn't "infinite", and I find it hard to believe device manufacturers give no indication. It's not a case of disbelieving the datasheet, this is a case where the datasheet gives you nothing to believe in.

But if they are written into a wear levelling flash filing system then each time it is written it will go somewhere else in flash.

You honestly think Humax has provided for this?????

 

[Owen Smith]

You honestly think Humax has provided for this?????

Yes because a wear levelling flash filing system is a completely standard feature on linux. All Humax had to do is not screw it up, which admittedly they have a mixed record on.

The fact remains that to make any predictions about flash life we'd need to know what wear levelling is being used across what section of the flash, and I believe we don't have that information.

 

[Owen Smith]

I was in military design, where a service life of 20 years was required. If there wasn't a datasheet specification to verify some parameter on which a design depended, we couldn't use it (or if we did there would have to be justification and for that to be signed off by the customer). UV EPROM data retention life was a problem, but some manufacturers did accelerated lifetime testing for the military market.

Fundamentally: if it's not in the datasheet, the buck stops with you. What use is a NV storage device without at least some idea how long the data is guaranteed to be retained for? It certainly isn't "infinite", and I find it hard to believe device manufacturers give no indication. It's not a case of disbelieving the datasheet, this is a case where the datasheet gives you nothing to believe in.

Military design is an entirely different level, with as you say a 20 year life and with generally harsher conditions to contend with too. Military users are issued equipment and they abuse it because it isn't theirs, or they're more concerned with avoiding being killed. Domestic equipment is made to typically a 5 year life (more than double the 2 year EU warranty), encounters less harsh conditions and is looked after better on average. If you want domestic gear designed to military standards be prepared to add a zero or two on the end of the price.

Yes UV EPROM data retention is not infinite, this was well known when they were current technology. But this is Flash not EPROM, and I've described the accelerated life testing that we do of final products. I cannot, on holiday away from work, remember if the datasheets say absolutely nothing or if they say assume it's infinite or 100 million reads or 20 years or whatever. What I know is no-one worries about it, because Flash has been in use for three decades and read/retention life has not become a problem.

 

[xyz321]

While the firmware part of the flash memory is largely static the EPG and Recording schedule databases are also held in flash memory and these are updated multiple times daily so are we not liable to be reaching the limit for writes to flash in those critical areas?

The EPG is stored on the hard disk but the settings and channel database are stored in flash.

It depends. If they are constantly written to the same place in the Flash then yes. But if they are written into a wear levelling flash filing system then each time it is written it will go somewhere else in flash. So then it is the number of times it is written multiplied by the number of different places it can go in the flash, which can be hundreds or thousands. The firmware is static and probably won't be moved as part of the wear levelling, so that reduces the space available for levelling. Some early Tesla cars are struggling with this because the firmware has got a lot larger over the years, and now so little flash is left for wear levelling of the dynamic parts it's been wearing out.

The flash is partitioned with the firmware rootfs and kernel vmlinux written to separate partitions in flash. The variable data is written to another partition humaxtv_db which has been formatted using jffs2 (journalling flash file system). This file system is designed for flash devices and includes wear levelling etc.

Internal Flash Layout

wiki.hummy.tv