HR20 Issues -- Hardware vs. Software?

1743 Views 30 Replies 11 Participants Last post by Earl Bonovich, Jan 4, 2007

fancydancy

Discussion Starter · Jan 4, 2007

So clearly there is a broad spectrum of observed HR20 performance ... some people do not seem to have many problems and other people are having nothing but problems.

If this were purely a software issue shouldn't we be able to correlate these issues to the Series Links (specific shows) and user habits (how, how often a HR20 is being used)? When people talk about "getting lucky" with their HR20 after having many failed boxes beforehand that suggests the issues have to be at least a little bit hardware related, right?

How funny is it that I'm keeping my R15 as the stable, reliable box when the HR20 goes in?!

Status: Not open for further replies.

1 - 20 of 31 Posts

hasan

· #2 · Jan 4, 2007

fancydancy said:
So clearly there is a broad spectrum of observed HR20 performance ... some people do not seem to have many problems and other people are having nothing but problems.

If this were purely a software issue shouldn't we be able to correlate these issues to the Series Links (specific shows) and user habits (how, how often a HR20 is being used)? When people talk about "getting lucky" with their HR20 after having many failed boxes beforehand that suggests the issues have to be at least a little bit hardware related, right?

How funny is it that I'm keeping my R15 as the stable, reliable box when the HR20 goes in?!

No...wrong. How the box is being used plays a large part in the various interactions that take place in the software. There are hundreds and hundreds of variables involved in how the box is being used. What you are calling "user habits" is a very, very large universe, as well as what I/O devices they have connected: HDMI/TV and Ethernet. None of those necessitate "hardware" issues.

When a person gets a new box, they start with clean stacks, clean buffers and very few interactions. They build slowly over time. How soon they hit one of the firmware bugs is anyone's guess. Some of them have been correlated, but it is difficult to get anything meaningful out of this, as correlation is NOT cause and effect.

There could be some hardware issues...but we are CERTAIN there are quite a few software issues. Until the software issues are ironed out (or someone at D* stumbles on a clearly identified hardware failure), we won't know much about hardware issues. We have to work with what we have, and what we can put our "fingers" on. These are (to this date) software issues. I haven't seen one verified post of a hardware problem, other than failed hard drives.

Look what happend in the last release: black screen problems hit the roof. There was no hardware change.

Not that it matters (hardware vs. software)....we have to go through EXACTLY the same troubleshooting processes in either case. Of course, software issues leave us some hope, while hardware issues would require a massive recall. So, I sure hope we don't find a real hardware problem.

See less See more

brewer4

· #3 · Jan 4, 2007

I would definately say software and very correctable. Mine was fantastic before 10b and it hit the roof. Tons of locks and constant reboots. Before 10b, did everything I wanted and more. Great DVR. So I agree that its a combination of user configuration/variables and software. As much as I wish everything was flawless right now, there is hope that things can be fixed and greatly improved. Now if only they would double their team and provide additional QA time. But then, that might slow releases which would be a hard pill to swallow for our "Gotta Have it Now" culture.

fancydancy

Discussion Starter · #4 · Jan 4, 2007

I guess what I was trying to say was that we should be able to figure out in general what kind of things cause trouble. Obviously there is a tremendous amount of different variables at play, but it seems like # of Series Links and # of ToDo items are a few general things that might help people survive until the software is more stable. It also sounds like staying away from padding is a good thing.

When I got my R15 people were crying bloody murder at dbstalk so I kept Links and ToDo items low, RBR'ed at the first sign of a problem and I have not had many problems at all. I'm wondering if the same habits with the HR20 would serve me well. Are there other things along the same line that people recommend?

brewer4 said:
I would definately say software and very correctable. Mine was fantastic before 10b and it hit the roof. Tons of locks and constant reboots. Before 10b, did everything I wanted and more. Great DVR. So I agree that its a combination of user configuration/variables and software. As much as I wish everything was flawless right now, there is hope that things can be fixed and greatly improved. Now if only they would double their team and provide additional QA time. But then, that might slow releases which would be a hard pill to swallow for our "Gotta Have it Now" culture.

Stuart Sweet

· #5 · Jan 4, 2007

It would be very easy for me to say "hardware problem" but --

I'll admit that I'm having a lot more problems with HR20#2 than #1 but there are a lot of other factors, including my having changed my behavior, changed year to 2007, new software version, on and on.

I think that unless someone figures out how to enable some sort of "debug mode" on the HR20 we won't know.

Which brings up the question, what if you hook up a USB keyboard to the front USB port and start hitting common key combinations (ctrl-alt-delete, ctrl-ctrl, ctrl-esc)? Would you stumble upon such a mode? Is such talk considered hacking, and therefore verboten?

mtnagel

· #6 · Jan 4, 2007

fancydancy said:
Obviously there is a tremendous amount of different variables at play, but it seems like # of Series Links and # of ToDo items are a few general things that might help people survive until the software is more stable.

When I got my R15 people were crying bloody murder at dbstalk so I kept Links and ToDo items low, RBR'ed at the first sign of a problem and I have not had many problems at all. I'm wondering if the same habits with the HR20 would serve me well. Are there other things along the same line that people recommend?

Your wish is my command. Usage patterns were talked about in this thread. Don't know if anything concrete came out of it, but there you go.

And I'd probably say I'm the case that shots your theory down (about usage affecting the stabilty). I have 40-something SL's (I'd have more, but I just cleaned some out because I was getting close to the magical 50 SL limit). Close to 100 items in my To Do List and I think I saw almost 1000 items in history. Until vacation, the unit hovered around 30% free space and sometimes dipped into the teens. It's now around 50%, but I'm sure that will go down with the new shows coming back on and being away for 3.5 days this weekend. And, until 10b, my box was stable. Maybe 3-4 RBR's in close to 2.5 months (I've had that many in the last 1.5 weeks with 10b though). Never missed a recording. No major issues.

All I have to say is good luck to the programmers trying to figure this out and I'm glad it's not me.

mtnagel

· #7 · Jan 4, 2007

lamontcranston said:
Which brings up the question, what if you hook up a USB keyboard to the front USB port and start hitting common key combinations (ctrl-alt-delete, ctrl-ctrl, ctrl-esc)? Would you stumble upon such a mode? Is such talk considered hacking, and therefore verboten?

You first

veryoldschool

· #8 · Jan 4, 2007

hasan said:
No...wrong. How the box is being used plays a large part in the various interactions that take place in the software. There are hundreds and hundreds of variables involved in how the box is being used. What you are calling "user habits" is a very, very large universe, as well as what I/O devices they have connected: HDMI/TV and Ethernet. None of those necessitate "hardware" issues. So, I sure hope we don't find a real hardware problem.

I wouldn't say anybody is "wrong" yet. We would be limiting our scope, before it's fixed. Both sides are valid [software/hardware] and can be supported. Why so many problems after 010b? One side- bad software. The other side- the two new tuners being activated are showing there is poor power supply's output.
Do we know? Nah
Should we not rule out either? Yes.
We all have our experience to guide us in our quest, but let's not rule anything out until it works, since it could be a combination [hardware/software].

hasan

· #9 · Jan 4, 2007

My "no-wrong" was aimed at the idea that hardware issues are playing a significant role...there just isn't any evidence of it. Could there be hardware problems....sure. Are there software problems, NO DOUBT ABOUT IT. Until the latter are largely ironed out, short of a miracle discovery, there won't be any chance of progress on the hardware side (at least by us).

It's not wrong to consider the possibilty of hardware....but when we have the absolute certainty that we have software problems, there is little point focusing on an "as yet" established hardware cause, much less fix.

It's all a matter of priorities (and capabilities, as well)

BubblePuppy

· #10 · Jan 4, 2007

I have been reading all the reports of bugs in the HR20 and I thought I would add my two cents in.
When I had my D* installed the installer said that I was very lucky in that they just recieved a new shipment of HR20s to replace their inventory of defective HR20s that had to be sent back.
Now I have never had the serious problems with mine that others seem to be having. Mine has had only three updates with 10B being the latest. Every thing has on my TodDo list has recorded and played back just fine. I haven't had the BSOD and I've only had to to a RBR just once. I have had a few quirky things happen and CC was screwed up until the latest upgrade and now works fine.

I guess what I'm saying is that the problems could also be from defective boxes that can't handle the software properly. In that end I suggest to also include the serial # from your HR20 in your posts and when talking to D* and that way D* can also track any hardware issues and that a recall of certain serial number sequences might be what is really needed.

Just a thought.

Chuck

See less See more

veryoldschool

· #11 · Jan 4, 2007

hasan said:
It's not wrong to consider the possibility of hardware....but when we have the absolute certainty that we have software problems, there is little point focusing on an "as yet" established hardware cause, much less fix.
It's all a matter of priorities (and capabilities, as well)

I admire your dedication & zeal.
I worry about "absolute certainty", but then I'm a very old dog who's learned "never & always" are just concepts you teach children. If you live long enough "never" will happen & "always" won't.

veryoldschool

· #12 · Jan 4, 2007

BubblePuppy said:
I guess what I'm saying is that the problems could also be from defective boxes that can't handle the software properly. In that end I suggest to also include the serial # from your HR20 and that way D* can also track any hardware issues and that a recall of certain serial number sequences might be what is really needed.
Just a thought. Chuck

What your date of manufacture?

BubblePuppy

· #13 · Jan 4, 2007

veryoldschool said:
What your date of manufacture?

10/10/06 Made in Mexico

HarleyD

· #14 · Jan 4, 2007

As has been pointed out, event driven code is a very tricky animal.

To me that makes it the prime suspect, but I couldn't rule out something like a bad lot of some subcomponent actually used in the builld of the box.

Like a semiconductor of some variety where the manufacturer of the box received 50 different lots in their shipment and ONE of those lots has a problem that only shows up under SPECIFIC circumstances as it is being used in just this configuration/application.

Stranger things have happened.

Stuart Sweet

· #15 · Jan 4, 2007

HarleyD said:
As has been pointed out, event driven code is a very tricky animal.

To me that makes it the prime suspect, but I couldn't rule out something like a bad lot of some subcomponent actually used in the builld of the box.

Like a semiconductor of some variety where the manufacturer of the box received 50 different lots in their shipment and ONE of those lots has a problem that only shows up under SPECIFIC circumstances as it is being used in just this configuration/application.

Stranger things have happened.

I agree, but I think we need bulletproof code before looking at the hardware layer.

veryoldschool

· #16 · Jan 4, 2007

BubblePuppy said:
10/10/06 Made in Mexico

Thanks, they're all made in Mexico. I've had a couple of those & now have a reconditioned 9/?/06. D* is sending me a new one so I'll have a new one to test.
There is something we called "Friday night maintenance" back in the day, that would equate to the rush to meet shipment. Neither are a "good thing".

veryoldschool

· #17 · Jan 4, 2007

lamontcranston said:
I agree, but I think we need bulletproof code before looking at the hardware layer.

Or do they need bulletproof hardware to test their code?
The chicken or the egg?
Bad parts, bad process, bad code... The point is let's not rule out anything & limit our efforts. We can all "do what we do best" and collectively "work the problem".

BubblePuppy

· #18 · Jan 4, 2007

lamontcranston said:
I agree, but I think we need bulletproof code before looking at the hardware layer.

If it was a software issue wouldn't the problems be more consistant from one box to another? If it was a hardware issue with different suppliers of parts then I would think that would be a more logical explanation as to the wide range and various problems that some are having and why others aren't having the problems. If it is a hardware issue then no amount of code tinkering will be the cure.

Chuck

HarleyD

· #19 · Jan 4, 2007

I doubt if they have adequate manufacturing configuration data to track semiconductor lots used in different serial numbers anyway.

Anything less mission critical than products by say, Lockheed, Boeing or something military or aerospace like that where lives depend on it, they just doesn't normally track configuration to that kind of level.

So if something that obtuse actually is the problem (admittedly unlikely) attrition is probably the only thing that is ever going to eliminate it.

veryoldschool

· #20 · Jan 4, 2007

HarleyD said:
I doubt if they have adequate manufacturing configuration data to track semiconductor lots used in different serial numbers anyway.
Anything less mission critical than products by say, Lockheed, Boeing or something military or aerospace like that where lives depend on it, they just doesn't normally track configuration to that kind of level.So if something that obtuse actually is the problem (admittedly unlikely) attrition is probably the only thing that is ever going to eliminate it.

Can you say "a DVR for $10,000?" I've spent a long time with that type of traceability, & the commercial market can't afford it.

1 - 20 of 31 Posts

Status: Not open for further replies.