Picture Problem

[/df]

The BCM7405 IC has HDMI outputs that are connected directly(no, see posts below) to the HDMI socket, with electrostatic protection diodes on each line, if the device's application test circuit is any guide:

IC pin(s)	signal(s)	HDMI pin(s)
A19/B19	DIFF_HDMI_0 P/N	TMDS-D0-P/N
A18/B18	DIFF_HDMI_1 P/N	TMDS-D1-P/N
B17/C17	DIFF_HDMI_2 P/N	TMDS-D2-P/N
C20/ B20	HDMI_CLK P/N	TMDS-CLK-P/N
D20	HDMI_CEC_1	HDMI_CEC
B16/D17	BSC_M1_SCL/SDA	TMDS-SCL/SDA

The following layout guidance is provided:

HDMI - Layout Guidelines:
- Differential pairs should be routed on TOP or BOTTOM
layers only.
- Trace impedance: 100 ohm differential impedance to the
ground plane.

• 5 mils trace width with 7.5 mil air gap on P/N pair.
• Match trace length of differential pairs, 20 mils max within

a pair, and 100 mils max between pairs.
- Adjacent TX/RX differential pairs should be separated by
more than 50 mils to each other.

So if any of those guidelines have been compromised by muck on the PCB or similar, that could be the source of the problem.

And the good news is that the traces should be visible to check.

But before opening the box, consider whether the routing of the HDMI cable may be to blame. Other users have found that moving the cable can affect the picture quality, despite shielding. Conceivably some other device nearby has started to play up and generate enough interference to break up the picture.

A "mil" is presumably a milli-inch or "thou", equal to 25.4 μm, rather than some weird alternate abbreviation for mm.

 

[rollsjohn]

That's a shame. With a now duff HDR Fox it's a difficult decision as to what to replace it with.Checked all solder joints which were sound so as BH says it a junk!

 

[/df]

Not junk but as I suggested before, a spares donor.

Get a used HDR; change the key to match the u/s one if you have precious recordings that haven't been decrypted; copy precious recordings to the newly acquired machine; retain power supply, disk(?), front panel (perhaps case too) of old machine for spares.

As no mainstream Freeview PVR product offers the facilities of the HD/HDR-Fox T2 there is a forum dedicated to home-brew replacements. No suggestion has yet beaten the price/performance of the HD/HDR-Fox T2 with CF, and certainly not at second-hand prices.

 

[xyz321]

I think there is more to it than just a direct connection between the SoC and the HDMI connector.

HDR FOX T2 Board.png

wiki.hummy.tv

I would suggest using the V-FORMAT button to see if it has any effect.

 

[Black Hole]

A "mil" is presumably a milli-inch or "thou", equal to 25.4 μm, rather than some weird alternate abbreviation for mm.

I agree with your analysis of "mil" (but that's the actual name, not an abbreviation), but 5 mil track width? That's about the width of a hair.

 

[Black Hole]

On a different note why are there no later models from Humax that have a dedicated bunch of tallented private individuals producing custom firmware?

Somewhere along the line Humax changed the format of the update files which, I believe, makes them more difficult to work out the hack.

The HDR-FOX update file is hash-signed to authenticate it, but not encrypted. It was therefore possible to look at it, but not create one until the signature algorithm was worked out (a simple industry-standard hash). With a known signature and an unencrypted payload, that made it possible to tweak the firmware and insert "our" own code, thereby making it extensible.

These loopholes have been closed for more recent models, no doubt having realised the security wasn't good enough. Ditto media decryption.

If I have to replace the Fox I'll certainly miss these facilities.

So get another one. Good second-hand goes for about £50 on a well-known auction site.

 

[/df]

I think there is more to it than just a direct connection between the SoC and the HDMI connector.
...

Clearly so, and a good point. But why?

We have a TDA19977A HDMI Receiver, but there's nothing that receives HDMI streams in the HD/HDR-Fox T2. I wondered if it was originally intended to have some function for recording from a separate HDMI input.

Also there is an IC marked HDMI Xmit. In some old NXP Application guide - Set-top boxes, we read about the TDA9984 "HDMI 1.4a Transmitter with upscaler":

Optimized for use in HD consumer applications, this device makes it easy to add an HD output to any legacy STB platform. ... For STBs based on source decoders allowing only 720p or 1080i HDMI output, using the combination of a TDA19977 (HDMI receiver) and a TDA9984 (transmitter with integrated upscaler) makes it possible to upgrade the STB to 1080p output capability with minimal hardware and software development

So this pair of devices is likely to be present because the bare BCM7405 SoC only offers at most 1080i, though perhaps the transmitter isn't actually TDA9984 as that claims HDMI 1.4a conformance while I believe the HD/HDR only have 1.3a. Actually, subsequent inspection of the HD-Fox T2 PCB identified U1650, NXP TDA9984A, and that variant is, or was at the time of production, HDMI 1.3.

This post on another forum came to a similar conclusion and the hypothesis above was confirmed by BobHumax.

 

[Black Hole]

but there's nothing that receives HDMI streams in the HD/HDR-Fox T2. I wondered if it was originally intended to have some function for recording from a separate HDMI input.

Clearly it is an unnecessary BOM cost to fit it if it has no function. I thought I had an idea about that, but on second thoughts the idea can't be right.

 

[/df]

You can see my suggested rationale if you read on.

Or, less likely, Humax had a large surplus stock of TDA19977 and TDA9984 and obtained a bargain contract for BCM7405 ICs whose HDMI outputs failed testing.

 

[Black Hole]

You mean this bit?:

For STBs based on source decoders allowing only 720p or 1080i HDMI output, using the combination of a TDA19977 (HDMI receiver) and a TDA9984 (transmitter with integrated upscaler) makes it possible to upgrade the STB to 1080p output capability with minimal hardware and software development

So your hypothesis is the SoC outputs HDMI to the receiver chip, which provides the conversion from i to p and then outputs through the transmitter chip. Seems reasonable.

 

[/df]

...
So your hypothesis is the SoC outputs HDMI to the receiver chip, which provides the conversion from i to p and then outputs through the transmitter chip. Seems reasonable.

I didn't study the details of the application suggested by NXP (which BCM7405 signals to which TDA19977/TDA9984 signals, eg) but the end result is capable of a much wider range of HDMI resolutions (the ones seen when cycling through the V.Format options) than the two that, I discovered, the SoC itself supports.

 

[rpb424]

Looking at the SoC datasheet from here on page 1-37, the decode/display formats matrix appears to be missing all 1080p display options from 50 frame (or field) decode formats, which supports your theory. 1080p is available in 24 and 30 frame though from some progressive sources.

 

[/df]

The document "PRELIMINARY HARDWARE DATA MODULE - BCM7405 - Pt 1 General Information" lists the supported HDMI modes for the SoC in Table 1-10 (DVI Supported Modes) as: 720x480i@60 Hz, 720x480p@60 Hz, 1280x720p@60 Hz, 1920x1080i@60 Hz, 720x576i@50 Hz, 720x576p@50 Hz, 1280x720p@50 Hz, 1920x1080i@50 Hz.

Whereas TDA9984 claims to convert between all standard EIA/CEA 861B modes, which I believe means the above plus 1080p@50/60Hz.

Apparently Rev.04 of the TDA9984 (2009) had HDMI 1.3, while the previous version was HDMI 1.2 and later ones HDMI 1.4+. So from functionality, era and package (4 x 20 pin, 12mm square), that looks like the one in the HDR as shown in the Wiki image above.

 

[lc200]

You mean this bit?:


So your hypothesis is the SoC outputs HDMI to the receiver chip, which provides the conversion from i to p and then outputs through the transmitter chip. Seems reasonable.

I seem to remember this from the launch and some discussion around it. The extra LSIs provide a 1080P@50Hz output, rather than limited to 1080i@50Hz on the Broadcom SoC. For broadcast TV in the UK this is pretty pointless, as Freeview is only 1080i @ 50Hz or 1080P@25Hz with 1080P@25Hz transmittable over 1080i@50Hz as progressive segmented frame.

I think the main reason for this small jump up to 1080P@50Hz is to allow any content that is 720P@50Hz being streamed or from file leave the Humax without any nastiness when the Humax is set to maximum resolution at 1080. Without the extra chips, a Humax box set to 1080i @ 50Hz would be fine for any broadcast TV, but when asked to play any 720P content at 50Hz would need to upscale it to 1080, then interlace it to 1080i @ 50Hz as that is the maximum it can output, this would make the content look pretty rough, forcing the user to change resolution to 720P to fix it or more than likely just complain about how poor the picture is with some content.

So by having the extra LSI, the Broadcom SoC can drop its HDMI output down to 720@50Hz when playing such files, feed it into the extra LSI and those extra LSIs then upscale it to 1080 and keep the 50 progressive frames as is.

 

[/df]

And just to continue, and perhaps complete, the massive hijack of this thread, the SoC just used some HDMI circuitry that Broadcom had on hand (BCM7501), which didn't support 1080p (same reference).

 

[rollsjohn]

Love the tangents that throw off in these threads - don't understand a lot of it but very informative and good to know that we are all learning!
Pulled out the motherboard from my Fox, found very little in the way of dust for a machine that's about 9 years old except around the fan blades. Carefully dusted with a paint brush top and bottom especially around the HDMI socket. Also cleaned connectors inside connector with pure alcohol. Result - only a couple of brief screen flashes in the last 2 days since clean. Guess it doesn't take much dust upset output. Thanks for your collective help.

 

[Black Hole]

Result - only a couple of brief screen flashes in the last 2 days since clean. Guess it doesn't take much dust upset output.

That's curious. Far more likely the act of disassembling and assembling again has disturbed a joint, which suggests it is repairable.

 

[Trev]

Work the plugs/sockets in and out a few times. That usually 'cleans them up' for another few years.

 

[Black Hole]

Love the tangents that throw off in these threads - don't understand a lot of it but very informative and good to know that we are all learning!

Agreed. It's fair enough to complain if a thread drifts completely away from technical discussion, but the forum would be a very sterile place if no discussion or debate went on. Sometimes the simplest thing can raise deeper question, and this is the way progress has been made.

 

[rpb424]

I'm quite prepared to believe that it is simply some influence from dust that causes it, either conductivity somehow or maybe acting as thermal insulation causing overheating. We have two HDR's and they have both had a seemingly permanently 'failed' HDMI output during their life , but the message on the TV screen was different for each ('No Video' in one case, and 'Video format not supported' for the other). In both cases I whipped off the lid and gave them a big dose of a compressed air can, blowing out several years of dust in the process, and they were fine after that. When in normal use I can't waggle any connectors on either of them to induce any kind of disturbance on screen, making me suspect it was purely the dust causing it.