First Look: Dish Network's DTVPal Digital Converter

[rperlberg]

I think DISH should hire me to debug their stuff. I could actually use the work right now.

I hear you, brother. It's a constant source of annoyance to me when my bosses are so hard on me about minor problems in in-house systems that are only used by a handful of people and then I see far worse errors in major products. It sounds like they couldn't have spent more than 10 minutes on testing.

This array indexing implementation really does make it sound like this box (or at least the timer function) was designed by inexperienced programmers. This is first-year programming stuff. A more "hardened" approach would be to store the channel name (e.g., "2.1") in the timer, even if it meant storing it as a character string, then doing a lookup and mapping the channel number to the stored channel data at the time the event fires. I often get into arguments with inexperienced programmers about this, because their training has beaten it into them that they have to use more efficient techniques, but there are times when functionality takes priority over efficiency. In other words, who cares how fast you get the answer if the answer is wrong?

I'm curious as to whether this problem occurs without ever going into TVGOS mode. I'm not surprised that Dish didn't catch this bug since they probably didn't think people would be switching in and out of that mode. It's really more of a set-once configuration thing.

 

[Whidbey]

I have an idea for how software updates could be done without having to rely on the availability of an OTA broadcast. The vendor could distribute a DVD containing a video signal with the embedded software update data. Then, if the box was designed to accept OTA updates, you could just play the DVD into the box. One problem with this is that most DVD players don't have RF converters and it would be a bit much to have to buy an RF converter just to perform this operation, so another possibility would be if there were a mode you could put the box into that would cause it to temporarily go into a "read" mode on one of the A/V output jacks and you would connect the video output of your DVD player to that. You could monitor the operation from the RF output, or if they used an audio jack for the temporary video input you could monitor from the video output as well.

Another far out method would be through the IR input using an IR blaster hooked up to your computer.

I believe Dish has stated the only way to get a firmware update is to send in the unit for exchange, if they even decide to offer such an exchange program.

 

[A guy who read the manual]

I'm curious as to whether this problem occurs without ever going into TVGOS mode.

Yes.

I first discovered timer bugs happening most often when I'd set a timer about 9-12 hours ahead to trigger early in the morning. I was recording the result on a VCR.

Then when the CBS affiliate stopped sending PSIP for 3 days, every time I turned on the DTVPal (unless only turned off very briefly), the Guide data download (with progress bar) happened, and, when I checked them, the timers had been altered.

When CBS resumed sending PSIP, I thought there would be no predictable way to keep investigating the bug. But when I noticed the time setting on my DTVPal was 6 minutes slow, I started looking for a way to force the clock to reset, other than unplugging the DTVPal. When I tried going in and out of TV Guide mode, it was successful in resetting the clock, but it also had the side effect of shifting the timers.

So I was back in business for isolating the timer bug behavior.

 

[Anole]

what's inside

As for what's going on inside another such box, here's an interesting peek by Smith,P who always provides fascinating insights.

rperlberg, I loved some of your imaginative firmware-load ideas! - :up
Such ideas give the Echostar designers and programmers too much credit.
A close examination of the PCB will probably reveal a row of JTAG pads on the underside, where a bed-of-nails tester can reprogram the board.
( . . . suggesting we could do the same thing in the future . . . )

However, don't let me stifle anyones creativity.
I do love a good brainstorming session! -

 

[A guy who read the manual]

Catch the Brightest Stars on SEE B S

I believe Dish has stated the only way to get a firmware update is to send in the unit for exchange, if they even decide to offer such an exchange program.

Which could mean people in some areas will need to send in the unit for exchange, because the update can be made available by OTA only, and some areas will not get the OTA updates, and updates may not even work properly yet, even though they are designed into the unit.

When you read that statement, do you see a ring of truth, or do you SEE. B. S. ? *cough*

But I like the DVD idea and the IR idea. I'll bring you guys with me when DISH offers me a cushy job.

 

[Anole]

AGWRTM -

Well, considering that things said in public by the top leaders of the company have turned out to be wrong, or lies . . .
I would not take the word of any Echostar employee on the phone, nor anyone reporting such a conversation, nor anyone reading about such a conversation on another forum! -
It's all just conjecture.

And then there is the possibility of confusing good old speculation and wishful thinking, with "facts". . .
Again, it's all just conjecture.
Until Echostar actually -does- something, we're all just walking on quicksand.

 

[Whidbey]

Which could mean people in some areas will need to send in the unit for exchange, because the update can be made available by OTA only, and some areas will not get the OTA updates, and updates may not even work properly yet, even though they are designed into the unit.

When you read that statement, do you see a ring of truth, or do you SEE. B. S. ? *cough*

No, I just see a company taking the path of least resistance. OTA update would be a PITA for Dish, having to make arrangemnts with thousands of stations. Having people send in their DTVPal at their own expense, that seems more like Dish's style.

 

[TalkingRat]

When they tested all these CECB boxes, did they use real air waves or simulation? If it was simulation, maybe they could simulate again from a Dish store's pc.

 

[Smith P.]

When they tested all these CECB boxes, did they use real air waves or simulation? If it was simulation, maybe they could simulate again from a Dish store's pc.

From experience in development of smartfones - yes, it happens on EVT/DVT stages, but soon when you'll close to PVT - DTV tests will be expanded to outside RF lab, then FCC tests will be conducted and it should go into real environment tests.
And according Dish announcement - the box is Sling Media development.

 

[pabeader]

I hear you, brother. It's a constant source of annoyance to me when my bosses are so hard on me about minor problems in in-house systems that are only used by a handful of people and then I see far worse errors in major products. It sounds like they couldn't have spent more than 10 minutes on testing.

This array indexing implementation really does make it sound like this box (or at least the timer function) was designed by inexperienced programmers. This is first-year programming stuff. A more "hardened" approach would be to store the channel name (e.g., "2.1") in the timer, even if it meant storing it as a character string, then doing a lookup and mapping the channel number to the stored channel data at the time the event fires. I often get into arguments with inexperienced programmers about this, because their training has beaten it into them that they have to use more efficient techniques, but there are times when functionality takes priority over efficiency. In other words, who cares how fast you get the answer if the answer is wrong?

I'm curious as to whether this problem occurs without ever going into TVGOS mode. I'm not surprised that Dish didn't catch this bug since they probably didn't think people would be switching in and out of that mode. It's really more of a set-once configuration thing.

you are, of course, assuming that there are no ways that a program glitch can be introduced after the program is written and compiled. I can think of three possibily more.

 

[TalkingRat]

Smith, P -- I was just wondering if they could use simulation at the local level to fix the box if OTA signals wouldn't do it because of something like bad timekeeping.

 

[Malouff]

you are, of course, assuming that there are no ways that a program glitch can be introduced after the program is written and compiled. I can think of three possibily more.

I know that Dish Network would have had to have tested the compiled program and see if it was functioning properly.
I don't accept the compiler as the problem for the current glitches.

I know that you can't share but hopefully we can access a debug screen to see what is going on.
Maybe that is the difference between the 101 software and your beta tester software (debug was removed - if it was ever there)

Edit: pabeader, I know some of what I am saying sounds personal but it isn't.
I have great respect for you and consider you a friend.
It's just that I like others are frustrated with Dish Network and we have no where to turn for possible answers.

 

[Smith P.]

Smith, P -- I was just wondering if they could use simulation at the local level to fix the box if OTA signals wouldn't do it because of something like bad timekeeping.

Hard to say; if we will have access to debug console and will see some info there while the box getting all problems from the air: multipath, low signal, high BER, wrong/garbage PSIP packets/tables, etc.

 

[pabeader]

I know that Dish Network would have had to have tested the compiled program and see if it was functioning properly.
I don't accept the compiler as the problem for the current glitches.

I know that you can't share but hopefully we can access a debug screen to see what is going on.
Maybe that is the difference between the 101 software and your beta tester software (debug was removed - if it was ever there)

what i meant was, after the compiled program is sent off to the chip burner. no possible inhouse testing at that point.

 

[pabeader]

The thing is, it's repeatable for me, but not for pabeader. The difference is in the broadcasts we're receiving. DTVPals apparently work fine with no bugs (at least no timer bugs) if you are getting strong signals and good PSIP on all channels. The CBS affiliate in each market seems to be especially important, but that came to my attention because mine sent no PSIP at all for a 3-day stretch.

i just tried again i even set 5 different timers, just to be sure. I still can not get my unit to shift the channel index. I tried going into and out of TVGOS twice, just to be sure. I did see the EPG get updated.

one thing i just thought about. you have renamed the channels haven't you? i have not. maybe i should try that and then try to duplicate the "channel index" shift error,

 

[TalkingRat]

Hard to say; if we will have access to debug console and will see some info there while the box getting all problems from the air: multipath, low signal, high BER, wrong/garbage PSIP packets/tables, etc.

I wasn't thinking of diagnosis, just upgrade solution. I was thinking if upgrades come OTA through one network and some regions can't get reception to get the upgrade, a local dish dealer could have a simulated broadcast (in a room) to upgrade boxes, similar to how they seem to have tested with a standard set of channel signals for FCC certification. It would save boxes having to be sent back to the factory if a fake OTA broadcast could upgrade the software. OK, I have ideas, sometimes they are the kernel of great solutions, sometimes just way too weird.

 

[noremac]

I have an idea for how software updates could be done without having to rely on the availability of an OTA broadcast. The vendor could distribute a DVD containing a video signal with the embedded software update data. Then, if the box was designed to accept OTA updates, you could just play the DVD into the box.

I'm not aware of any DVD player or VCR or, for that matter, any other consumer playback device that has an ATSC modulated output capability. And the DTVPal does not seem to have any capability for NTSC reception (this might even violate the CECB rules).

I'm guessing that Dish has zero plans for updating firmware. Remember, it's only been very recently when we've all come to expect our electronics to be upgradeable! Even now, how many TVs, microwave ovens, automobiles, etc. are actually upgradeable by the consumer?

Also, the DTVPal is essentially a throw-away disposable piece of plastic, whose manufacturing cost is probably lower than the cost of two-way shipping and technician time to bother with upgrading. I also don't see the market for converter boxes extending beyond another 18 months or so. After that, it's a dead-end product.

Some things to keep in mind...

 

[pabeader]

the Pal receives and displays both NTSC and ATSC that is it's reason for existance after all. it just can't tune NTSC but it does pass it through.

18 months? that's only true if the average price on an ATSC set drops dramatically. You know something the rest of us don't?

 

[TalkingRat]

There won't be 600,000 boxes a week sold after the coupons expire, but I would expect places like Radio Shack, Frys, and Amazon to do business in boxes still. Canada's later transition means there they will still be produced in quantity next year.

 

[Malouff]

Speaking of a possible debug screen I wonder if there is one.
Any hints pabeader?

Maybe it was jut enabled on the beta software and that is the real difference between the 101 software.

pabeader do you also know what the difference between the 100 and 101 is?
I still don't understand why they decided to distribute both of them.

I guess to show that the software can't be upgraded or they would have been the same.