Hi all, an update on my Instant Record issue on my 5000T, which despite reporting that it’ll record until the end of the EPG time, actually only records what’s in the buffer plus approx 1 minute. The machine also refuses to reformat the hard drive, reporting the failure to do so as ‘due to an unknown error’.

Sadly, no amount of unplugging it from the mains or factory resetting made any difference to either issue.

I removed the hard drive and installed it in my W10 PC, as I suggested I would. The SMART data showed both the drive as whole and its individual parameters to be in good health. It had 4 healthy active partitions and a small unallocated area of around 34MB. I decided to delete all 4 so that the drive effectively became new/blank with just one large unallocated area.

I reinstalled it in the 5000T and the machine immediately formatted the drive, even before I then told it to! It now happily reformats when requested to, so did another factory reset and reformat. On removing the drive and installing it in the PC once more, it revealed that it was still healthy and had been formatted exactly back to the partitions as they originally were (but without the error which was stopping the 5000T from reformatting it I assume, whatever that was!).

OK, now the bad news! The machine works perfectly, including being able to reformat but the Instant Record function remains broken exactly as before.

To me, the narrowness of the fault still seems to point to a Jan 21 update bug which seems to affect some versions of hardware. Any minor flaw in its actual loading would surely cause far more catastrophic failures in the machine’s operation, as would a hardware failure?

I can sadly go no further but hope my investigations may help someone else to crack the problem!

Thank you Martin. The faulty behaviour occurs in all situations the Instant Record is used and 100% of the time. In the situation you describe, I would have ended up with approx a 46 min recording, 45 from buffer plus about another 1 min. I have uploaded two JPEGs, one of my reported SMART values and one of the partitioning on the disk (Disk 1). My ID values are slightly different to yours but the data you request is there I think. Humax also made me aware of the nature of the update, however I do not share their belief that it couldn't have effected other elements of the software. I accept that it may not have done but such coding changes often have unexpected consequences or accidentally inject errors. It's a 5000T.

[attachment=87495,1836] [attachment=87495,1837]

My problem sounds almost exactly the same as RichHumax describes, on pressing the instant record button it would record what was in the buffer plus a variable but short additional amount of 1-5 minutes (and I may be getting confused with how much EXTRA was recorded after pressing the button and how much had already been buffered, it may have been always only 1 minute extra). This was happening for me before the recent software update so I would discount the YouTube changes. Also my problem has now gone away after a simple power off/on. I know this has not happened for RichHumax but in my experience as a software developer this screams uninitialised variable. For example, for me it powered up in a low or zero value that allows it to work up until such time as it is left with a high value whereas for Rich its power up value in ram is already too high to allow the function to work properly. Its a simple fault in software, easily fixed once you find the right variable.

Also it seems to me to be more linked with the program guide (or whatever gives the program information) rather than the delay buffer because it always captures what has already been watched, just not what is yet to come for that program

RichHumax – 18 mins ago  » 

OK, all round joy, the Instant Record now keeps going, problem over, it must have been something to do with the NVRAM being slightly corrupted perhaps, although for the machine to perform as well as it did under such circumstances felt wrong.

However, long story short, my joy was short lived, I soon discovered that it actually only keeps on recording if the internet is not reconnected (via LAN or WLAN). Once the internet is re-established, the recording stops after approx 60 secs, even when its been happily recording for 30 or more mins up to that point. The behaviour is perfectly reproduceable 100% of the time and once connected to the internet, the original faulty behaviour remains until it is once again removed.

That is a very interesting discovery but I don’t understand what is happening. For what little it is worth my Internet is permanently connected.

I sadly don’t have any current experience of the protocols but assume once again that the Jan 21 update must have a bug but then again how do you explain that so few people seem to have it? Perhaps a NVRAM flush is needed or an earlier Software Version to see if the issue goes away?

I just don’t know. It isn’t unknown for one person to have a very reproducible problem with their Humax that nobody else can replicate; often the problem eventually goes away after a while.

Fyl – 1 hour ago  » 

Hi Rich,

I’ll have to try that.

I tried a long power-down – that didn’t fix it on my 4000T.

Then I did Graham’s suggestion from post #8 and left it paused for a while. It then created a recording similar in length (I didn’t check the exact length relative to the pause).

Now it seems to be creating a recording the length(-ish) of the buffer.

If I remember correctly a second push of the instant record button allows you to extend the recording time. So you actually get some of the next programme on the end of the recording.

At the moment using the Aura as the primary recorder.

There is definitely an issue (as I have mentioned before) because I have had two “instant record” failures today, one stopping after 4 minutes and the other after 27 minutes. However, I cannot reliably replicate the problem intentionally.

I did not have this problem prior to the “youtube” update.

SoundmanLedbury – 37 mins ago  » 

If anyone has a solution I’d love to know.

I don’t think anyone has got a solution; it works perfectly for me but a number of other people have made the same reports as you about failures and I have seen enough evidence to fully believe it is a genuine problem.

Hi all, I posted a few fairly long comments earlier this year. I won’t waste people’s time repeating them but I thought an update might be useful following the renewed interest in the subject. My machine continues to function as last described, everything works normally except; instant record only continues for 1min but adds anything in the buffer to the saved recording and the hard drive won’t format (unless removed from machine and even then it soon becomes unformattable pretty quickly after re-installation).

I’d be interested to know, if, like for me, the instant record function can be temporarily ‘fixed’ on people’s machines, although not the re-formatting issue, in the same way as I described earlier this year? This morning I repeated my ‘fix’ and it is still 100% reproducible! If I disconnect my machine from the Network, both wired and wifi of course, the instant record works perfectly again. In fact, to prove the point, if the Network connection is re-established during an instant record, the recording will even stop exactly 1min after its re-establishment. Of course, this is not a ‘fix’ and will cause other undesirable issues but I suspect demonstrates some bizarre connection between the instant record issue and WAN and the EPG. Also of note, is that in some multiple instant record activation scenarios, involving a channel change at some point, the instant recording labels the recording incorrectly as being of the channel and programme of the previous instant record and often also has a projected end time of the programme being recorded as being exactly 1hour later than the EPG states.

For what it’s worth, I remain bewildered why some lucky people’s machines remain unaffected, it appears to defy fault finding protocols! However, due to the very narrow and repeatable nature of the issue, the cause must be very obvious to whoever designed the machine’s hardware and or firmware.