serious (?) problem with Sony HXR-NX5U
 

Hi All

I'm one of the first normal mortals to have bought a Sony NXCAM. However, within a few days of getting the unit and shooting random stuff with it to get a feel for it, I experienced a serious problem which I have described below.

I would be very interested to know if any other NXCAM early-adopters have had the same or similar problems, and if anyone knows what might be going on here.

*** ‘Buffer overflow & crash’ problem with new Sony HXR-NX5U:

Problem noted twice while shooting random test footage in first week
of use (between 6 and 13 Feb) and replicated deliberately once on 16
Feb 2010. Not possible to replicate problem at will, i.e. is
intermittent.

Problem noted while shooting in 1080/30p FX mode (high-def) on to SD card and
FMU128 unit simultaneously, PCM audio mode. SDHC card is new SanDisk
Extreme 16GB class 10. Both SD card and FMU128 formatted on camera
prior to use, both have lots of space available. Camera serial 110136,
obtained from B&H Photo 5 Feb 2010.

Problem description:

A) Between one and two minutes into shot, footage freezes without
warning and grey overlay screen with words ‘Buffer overflow’ appears
for a few seconds on flip-out LCD screen. Screen then goes black with
words ‘Recovering data.’ in middle. Camera is now completely
unresponsive – even closing the screen does not cause it to turn off,
no buttons or dials respond, and even turning power off elicits no
response – I have left the camera in this state for up to twenty
minutes with no change.

The only way I found of getting the camera out of this deep crash was to remove
the battery.

B) On reinserting the battery and powering on, a screen appears with
the words ‘Writing to the memory card was not completed correctly. Do
you want to recover the data?’ A ‘yes’ and a ‘no’ option are presented
on the touchscreen and a small SD card icon is shown. On pressing
‘yes’, another screen appears with the words ‘Execute? Recovery may
damage data that was added or edited on another device’, and ‘yes’ and
‘no’ options are again presented with an SD card icon. On pressing
‘yes’ again the screen goes black for a few seconds while data is
being recovered. Then a similar message ‘Writing to the EXT media
device was not completed correctly…’ also with ‘yes’ and ‘no’ options
but an EXT media icon [referring to the FMU128] appears instead of an
SD card icon. On pushing ‘yes’ I get taken to a similar ‘Execute?’
screen and on pushing ‘yes’ there, I have a few seconds of black for
the data to be recovered and the camera then moves on into shooting
mode.

Twice out of the three crashes the clip that I was shooting as the
camera crashed was in fact recovered. Once it was nowhere to be seen.

COMMENT: Part A of the problem is really scary. Which buffer is being
overrun and why? Surely there should be no possibility of overrunning
any of the camera’s buffers while shooting? But this seems to be
happening. Also, if a buffer is overrun, why does this trigger a deep,
irrecoverable crash of the camera? This is really worrying!

Part B is the normal recovery process that the camera triggers after
it has run out of battery power in the middle of shooting a clip
(according to Richard Keys, technical supervisor at Sony Professional
Services in New Jersey). Interesting to note that writing data seems
to have failed to both the SD card and the FMU128. Recovery worked
only twice in three crashes. This is a little less scary than the
buffer overflow crash, but not much – I want my recovery to work every
time!

The time where I managed to replicate the problem I deliberately
stressed the camera’s processor by shooting a rapidly-changing,
high-contrast scene. I was not able to do this again – when I tried
again while recording to only the SD card or only the FMU128 (not both
simultaneously) I was not able to crash the camera during shots up to
7 minutes long, although I had only half an hour in which to try to replicate the problem so not too much should be inferred from this.

The camera is with Sony Professional Services in New Jersey, where I dropped it off on 16 Feb. As of Friday 19 Feb there was no concrete diagnosis for the problem, and Sony has promised me a new unit tomorrow.

Juan Martinez, Sony product manager, suggested in a phone call on Friday that there could possibly be some sort of incompatibility between the controller chip in the SanDisk SDHC cards and the camera causing slow writing of data to the cards -- but he stressed this was only an idea and their testing had not confirmed this. He doubted it was any problem with the FMU128, as that unit's interface with the camera is extremely fast.

It strikes me as odd that there should be any issue with one of the fastest SDHC cards made by one of the best-known brands in the business, but who knows?

I look forward to the input of wiser minds. Hopefully my cam or cards are just lemons and it's not a design fault of the HXR-NX5's -- I am really looking forward to using this cam as the feature set, layout etc. is really good for the sort of shooting I need to do, and the few random clips I have shot with it look just like I want them to -- a little rougher than XDCAM-EX, but definitely finer and less noisy than the HDV I used to shoot on the Sony HVR-V1...

Cheers

Adam Welz

just Guessing (as usual)
buffer overruns, are when the buffer fills up. With contraptions like these you got buffer everywhere :-) The SSD hard drive has buffers , the Flash Card has Buffers, both of them have holding areas designed to speed-up the solid state memory by grabbing the incomming data quick, then sending it out to its places in the memory.

The device writing to the Solid state memeory devices, has ITS buffers, and it is most likly that these are the ones it is talking about. Generally this would indicate that the writing to the media was not going fast enough, and the camera post encoding write out buffer was filling. Usually that would mean a media issue, WHICH could be caused by a media error, because remember too that these medias are "self-correcting", when they cant write to an area of memory, they balk and try to work around the problem.

Once you had the problem the first time, it could have become compounded, by various internal bad block sectioning and re-writing and any corrupt stuff on either solid state units. So if I myself had gotten into that situation, i would want to attempt a Fuller format of the solid state items, completly clear them out instead of these quickee format things (full format in the PC). then return to the camera and reset the formatting there in the camera, for the camera use. THEN hope that the first one was a fluke.

once that didnt work , just like you it would be off to the shop, and at that point It would take weeks of the camera working perfectally before i would trust it again.

but of course there is more , lots more, everything has buffers and caches or whatever you want to call it from what i am reading. Scan the Imager chips into WHAT? of course another buffer, process the image, from another buffer. then compression has to be done by comparing multiple frames , and compressing based on the previous frames, and where in digital land would those frames be held while processing them? another buffer :-) Then passing from the imager buffer to the compression processing , mabey another one again.
makes you wonder by the time the chips see the data just how many frames has already gone by huh :-)

and just like your computer, it IS a computer now you know, not a camera, you gotta be really careful about crashing it, turning it off when it isnt finished writing, power surges, brownouts, pulling media that isnt done writing, and all that stuff. If it is corrupt, when it comes back up it will either scan the blocks and TRY to fix everything that it presentally sees as wrong, or your just going to have a mess. So some extra USER care has to go into the new computer (freaking) camera , all the other camera care still applies, plus the new computer care too. Oh joy its digital :-)

card source
 

I bought the card from B&H at the same time as the cam. I've generally found them to be reliable.

Adam

You didn't mention the source for the card - any possibility that it's a knockoff? Memory cards are one of the most counterfeited tech items around, and they often "look" legit (right down to the name brand logos), but may have compatibility problems. That's my first guess, as I've seen buffer over-run on consumer Sonys with fake MS Duos... usually with data loss.

I seem to recall hearing of compatibility problems even with legit cards, although I have to wonder when I hear those stories whether the cards are really legit or not...

This is the biggest "danger" of any device using a memory card/stick - 90%+ of the "Sony" memory sticks on eBay are knockoffs, and I got some fake "Sandisk" SD ones at one time... it's a HUGE problem. There are a few guys on eBay that have details on how to spot the fakes, but I can tell you from experience that the counterfeit sources alter their "product" to look legit almost as quickly as the "tells" are exposed.

Bottom line, "saving" a few $ on memory cards/sticks may not be a wise plan... Remember that the knockoff sources will typically revise/relabel or repackage so it looks like you're buying the latest/fastest (and most expensive) legit product... but what's inside is just junk.

Yup B&H is reliable. and extreeme is a great card that should have no trouble keeping up with the data rate at all, or even 2time that.
everytime i get in any storage media, from hard drives to cell phones, i run one of them Read write tests, in the main PC, and check how the media is working. that way i can HOPE that i have one thing accounted for as being (part of the) problems i might have or not.
there is a bunch of Drive tests out there that are free, even if they dont all show the same numbers, or do the same thing the same way, a person can run various tests on each of thier storage items and see if one gives worse numbers over another.

you mean even AMASON , might have aquired a batch of Faux cards :-) naw it would never happen :-)
that and cheap batteries that are overrated, are becomming more plentifull than the real stuff. probably because nobody WANTS it .

Sorry to hear about your problems. Have you ever been able to reproduce the problem with just your SD cards w/o the FMU128?

If it is a problem with the SD cards then you should still get the error w/o the FMU128. I have the AX2000 and have never seen that error, even though I'm using cheaper class 6 cards (Delkin eFilm Pro 32 GB's).

It would be easy to imagine that there is some firmware bug related to a specific combination of recording devices with different data rates. Sony probably couldn't test every different combo, so you may be the unfortunately soul that discovered it. I could even imagine that the firmware expects the FMU128 to always be faster than the SD Cards, and that your class 10 cards are faster than anything they could test during development.

I expect the card or the FMU has a speed problem. Check both on the PC with speed test as has been suggested. Even the Single chip hard drive cameras don't like being interrupted during a write operation and it is very easy to switch off rather than stop then switch off.

Ron Evans

nx5 problem
 

Was the problem solved?
Did Sony give you a new unit?

Woah thaz a RED LIGHT buddy. Keep us updated, I will feedback the issue to Sony Singapore.

I'm suddenly pausing in my decision to buy one. A friend of mine pointed me to this thread and now I'm a little nervous about dropping $4,000 on a faulty gen 1 cam.

Keep us up-to-date! Any other NXCAM users having similar problems?!

Thanks,

Heath

though different camera, I have been sticking with the Delkin class 6 cards (3-32gig, 3-16gig) and I have the "free" 16 gig class 10 that came with the camera. All my stuff is from BH and I did have one card fail even before I got to shoot anything on it. It wouldn't format in the camera or on my windows or mac computers. BH replaced it with no issues.

When asking BH about the reliability of these cards brand to brand, they said all the brands have had a certain percentage of returns and there really wasn't a reason to switch brands or go to a more expensive card OR class 10 for AVCHD recording unless I felt like it. So far, I've put a bunch of hours on all my cards and no issues...yet.

Reality is that electronics will fail. Just a matter of when. Could be before you unpack it or 20 years later...that's just the way it is! All we can do is run redundant systems and when something happens, hope the manufacturers are cool about honoring the warrantees!

Always bummed to hear of anyone having issues and I was tempted to jump from Panny to Sony 'cause it looks like a great system. Instead I got lured to the DSLR side!
Hope things work out!

update on Sony HXR-NX5U problem
 

Hi All

Sony replaced the camera on Monday with camera #110060 (an even earlier serial than the one I had issues with). They did not tell me exactly what caused the issue (I'm not even sure that they know) but the tech who worked on it said he thought it was possibly one of the camera's chips. He said class 4 and above cards should be fine for HD recording.

I have reformatted all memory chips and the FMU128. I am currently travelling but will be shooting by the weekend. Will update this forum if any problems strike again, or if they don't.

Best to you all.

Adam

That's great news!

Sorry to hear of your NXCAM problems. The first mistake you made was not heading 1.1 miles south of Sony Pro Services and coming to visit me ;)

I'm pleased for you that Sony quickly diagnosed a hardware problem. I suppose these things can happen with any piece of equipment.

At least from the perspective of ergonomics and feature set, it is an impressive camera for the money. I'll hopefully have one here for testing and review soon.

Ned Soltz