Tivo is blaming V53 error on SNR ratio?

Advertisements

My 4-tuner Roamio (OTA only) had an interesting event last night, which I might blame on a tuning adapter if I had such a thing. I noticed that the main HD channel of a local station with serious multipath problems was showing an error (V53, I think), as it does sometimes, but oddly the .2 subchannel was being received just fine on a different tuner. So I checked my 2-tuner Premiere, and it was receiving the main channel just fine. After checking back and forth a couple of times, I used channel-down to retune the Roamio's tuner lower, then channel-up to retune the problem channel, and it was immediately received perfectly. I realized afterward that I should have checked the Diagnostics screen to see if this looked similar to the odd error-counting situation I saw on March 1st, but in that case I think the signal was actually glitchy when the strangeness occurred.

I'm thinking that under some circumstances the Roamio tuner can see a glitch that it treats like an infinite loop in the data stream, causing it to seize up and stop accepting data. And I wonder if a signal strong enough to overdrive something could cause such a problem. If that sort of lockup isn't impossible by definition in the design of the modulation method, then the chipset should immediately detect when it is being overrun by data and clear its buffer. And if the chipset doesn't do that, then some task in the supporting software needs to monitor the chipset closely to detect the seized-up data-overrun condition promptly and tell the chipset to clear or retune or reset or something to clear the lockup. TiVo's programmers need to review the functional characteristics manual for the chipset and consult with the chipset engineers as necessary to get this fixed.

 

I've experienced issues like this as well. I have no idea what a "multipath problems" are. I posted a longer reply in the thread you referenced.

 

I'm sorry to see that this thread has more or less died, although the problem continues.

The 10db pad I added a few months ago actually seemed to cure the V53 error problem, but at the expense of adding a new problem: severe pixellation, primarily on Comcast channel 621 (USA Network). In the TIVO DVR diagnostic screen I found that the signal level for ch. 621 was rapidly fluctuating all over the place from 80% down to 50% and constantly changing by dramatic value differences with each successive sample. Oddly enough other channels were solid (no fluctuation in levels) WTF/

I changed the pad to 3db which mostly cured the pixellation on 621, but levels (only on 621) continue to jump around a bit, but not as severely. USA network is again watchable, but on rare occasions will still take spells of pixellating, but only on this one ROAMIO. Our 2nd Roamio and our Premier have never had this problem with pixellation, tho the 2nd Roamio will on occasion throw a V53 error and appear to lose the channel for some period of time.

Has there been anything official from TIVO on this? I would sooner take a whipping than have to go through dealing with Comcast support again. Thanks!

 

I don't know if this helps or adds to the general confusion, but I got so sick of tuner errors that I decided to see if lowering the number of tuners would solve my problems. Amazingly enough it worked! I've got a Roamio Pro and for several months I was getting channels that would lose their authorization almost every day. My soluition was to reboot the Tivo every evening. Tivo support told me that the signal was too strong, but when I added attenuators half my stations were plagued with pixelation, so that wasn't a solution. Comcast support came over 5 times and said everything was well within spec.

3 months ago I lowered the number of tuners from 6 to 4 and I've yet to loose a channels authorization. I tried 5 tuners, but it still had the problem. So I'm wondering if having hot signals on too many tuners overloads something in the cable card and channels just get blocked.

I'm not thrilled to have lost 2 tuners, but am otherwise fairly happy to have found a solution. I use a 2-tuner premiere for OTA channels and the Roamio for cable channels, and so far have had no need for 6 tuners on the Roamio.

 

Advertisements

Thanks Hyrax.

My Roamio is only a 'basic' but (conceptually at least), what you may have accomplished by dropping the tuner count from 6 to 4 has (again conceptually) resulted in a 3db increase (doubling) of the signal strength available to the remaining 4 tuners.

To explain - -

I have no idea how they're accomplishing the internal signal splitting in the Roamio, but at the risk of showing my ignorance about this, RF signal splitting normally requires ganging multiple 2-way splitters together to achieve the desired number of output ports. (even though it may be done inside a common housing) For example, a 4-way split requires (electronically) 3 two-way spliters, with 1 driving two to give you 4 outputs, with a net signal loss of typically 6.5 db per port. Even a 3-way splitter consists (electrically) of 3 two-ways wired in tandem with one port terminated (internally) so as to give you only 3 output ports, but with one of those ports at -3 db and the other two ports each at -6.5 db each. (buy one and look at how the ports are marked)

If such a thing as a 6-way splitter existed, it would (electrically) consist of 6 two-way splitters wired in tandem, thusly:
1 driving 2
each of those 2 driving 3, with the 2nd port of the 2nd tandem 2-way terminated
net result: 6 output ports, total
The (approximate) output levels at each of the 6 ports would be the sum of the losses of each splitter in tandem (3db each, cumulative) plus a slight additional fractional loss. say an additional ½ db overall at each transition. In other words, roughly -10db per output port as compared to the signal level entering the input.

By reducing the signal split from 6 to 4, you've effectively doubled your signal level (increased it by 3 db) at the 4 remaining tuners (here again, conceptually at least)

the beatings may now begin...

HTH

 

I can't speak for the Roamio (yet), but I did pull the tuner shields on a 2-tuner Premiere, and found the "splitting" to be nothing more than a PCB trace being sent to two different chips. I found the same with the TiVo HDs.

What stood-out, to me (on those platforms), was:

1. Why not use shielded coax inside and use true splitting w/isolation?

2. How is it acceptable to split this way inside the TiVo, when splitting this way outside it is completely unacceptable (I know why, outside)?

3. One PCB trace always had a shorter run, and 90-degree angles (also a no-no) were in the PCB traces. This seemed to be reflected in the signal strength and SNR being different, when the TiVo had both tuners set to the same channel.

As somebody who used to repair TVs for a living, I don't recall any CRTs ever using anything other than shielded coax between the RF processing components.

It would be nice if somebody could pull the tuner shield covers on the Roamios and verify the path of the RF, and how it gets divided-up. I would think that in this day/age of technology, there should/could be a single, shielded, path to a chip, and that chip could do the splitting (and, if needed, amplify/attenuate on-the-fly), and should be fully RF shielded, both top & bottom sides of the PCB.

If I ever have to pull a Roamio out of service again (base-models), I'll be sure to check if TiVo has improved their designs, and try to spec-out any chips involved. When that time comes, I'll share what I find (unless somebody beats me to it).

 

Frankly I would be shocked (horrified might be a better term) to find only a PCB trace tying all the tuner inputs together. Whatever happened to the concept of impedance matching and the myriad perils of not doing so, especially where multiple tuners are concerned? (although this would possibly explain the rapid level fluctations on ch. 621... it's called intermodulation distortion, aka "interference") But what do I know, I'm a lowly peon from the vacuum tube era.

What I was sorry to see in the Roamio design was the disappearance of the separate OTA signal input and the ability to leverage both CATV and OTA at the same time. Here in Comcast hell (Houston western suburbs) CATV signal interruptions are almost more frequent (and certainly much longer duration) than satellite rain fade. With our Premier we would routinely schedule recordings from ABC/CBS/FOX/NBC to utilize the OTA tuner rather than risk losing an episode of something because Comcast dropped out right in the middle of a "who dunnit scene". Alas, I digress.

I see nothing in any of the firmware release notes suggesting anything has been done to address the V53 error problem, yet the discussion thread seems to have died. What have I missed?

 

The question is, is the RF (radio frequency) being down-converted to IF (intermediate frequency) in the demodulation process before or after this pcb board split? If it's IF and then split, depending on the IF frequency, it may not matter as much. (Similar to what DirecTV does with its sat signal on your coax after the LNB)

 

With Satellite, the LNB (as its name implies) performs a block conversion of an entire band of frequencies, converting them from either Ku or Ka band down to L band (2 Ghz) primarily to make it economically possible to send the signals down a reasonable length of inexpensive consumer-grade coaxial cable as opposed to extremely expensive, semi-rigid pressurized waveguide (or heliax). Downstairs the set top box still has to "tune" among the individual transponder frequencies within the L band, as well as select right-hand or left-hand circular polarity. Ergo there is still an LO involved here, at least on "Charlie" systems. I've been away from sat for 15 years & haven't had time (nor desire) to study how "Dave" systems accomplish single wire multiplex (SWM) while tracking up to 5 birds with a single dish.

 

The trace going it's separate ways (making a "split") on the HD and Premiere 2-tuner models happens with an unshielded trace leading from the coax-in, up to the shielded parts, without any component along the way to do anything to the signal.

While in another realm, I've noted that even some of the cheapest wireless (WiFi) devices, use shielded mini-coax, instead of PCB traces, in the path to their antennas, while some of the big brands use unshielded traces for the same thing, or use ordinary unshielded wires from the wireless module connections, to the antennas (just soldered right to the center pin of the connections designed for shielded wires).

I really gotta get around to taking a look at what the Roamios do. However, if I recall correctly, the only reason I never did, was due to my base-Roamios having shields w/out removable tops (one-piece shields, soldered to the board on all sides). It might require one of those optical probes to see what is inside, without a soldering iron (or hot-air soldering/de-soldering equipment). I'm also seeming to recall that there was no exposed PCB, or PCB traces, to view, as the coax-in went straight into the shielding (that's an improvement over the older models).

I had a base-roamio that worked fine, except it couldn't tune channels >900MHz. I did some inspection, which is what I am now recalling. It was impossible for me to see anything that I was looking for at the time, except noting that the soldering on the shielding on that one, hadn't solder-flowed anything except a minute bit of one corner. Comparisons to one that could tune the >900MHz channels, revealed the fully-working one to have solder-flow all the way around the shield.

I think I was hoping to find a bad solder-joint on the inside (to quickly repair), and found that unless I wanted to unsolder the shield, there was nothing I could see, or re-solder, due to the new one-piece shielding design, without a removable lid, like the older models had.

That doesn't bode well for trying to figure out how the Roamios "split", route, and process the signal, unless somebody is willing to risk killing a Roamio in the process.

 

Tivo support told me the same, to get SNR under 35 dB. For about $15 I bought attenuators at 3, 6, 10, and 12 dB and got it down. I'm having different problems now, but not that one.

But I did that just to appease Tivo support. As a former working communication engineer there is no such thing as too high SNR - it's Signal to Noise Ratio, are you telling me I have too little Noise? Really? Infinite SNR is perfect! If I'm wrong I'd love to hear a knowledgeable explanation.

I think CrispyCritter may be close when he said that Tivo has a different meaning for what they call SNR. It is true the signal can be too hot if they don't have limiting circuitry, and reducing overall signal level is exactly what the attenuators do, so signal level is really their issue and they've mislabeled it as SNR.

Anyway, for barrett14 I'd suggest going ahead and buy another attenuator or two and perhaps put them in series until the Tivo box measures under 35. And hope that really does solve it. Good luck on that.

Tivo just 2 days ago did agree to replace my 5 month old Roamio Plus because it keeps deleting recordings with under 63% disk usage. But not really related to SNR, so irrelevant to your problem.

I also struggled as apparently many did during March with the Roamio rebooting at random times and for no visible reason. I see that Tivo accepts that was a problem in their servers and they fixed it by early April, 2014. I've not had that problem since. But with 3 or 4 different concurrent problems, diagnosis was impossible. Now I'm down to just the one problem of premature deletions and they're going to fix that. I hope. Good luck to barrett14 and to all. It's a great system when it's working right.

 

Yes I know all that. I was making an analogy, not a comparison. Just pointing out how a signal can be down converted to lower freqs as an "intermediate" step, as in "intermediate" frequency, i.e. - IF.

PS - I just knew someone was going to comment like that and should've deleted to save confusion.