Internal PAL tuner gets a sound upgrade!

All.

Some of you might be aware that I have been hard at work doing more hardware modifications for Series 1 TiVo's and have said that shortly I would be making public some news.

That day has come - if not a little later in the day than expected as I wanted the parts in my hand before going public and the couriers took their time....sigh :(

Anyway, with that little hiccup aside I am pleased to announce that after almost three months of research and co-operation with both "Thompson TiVo" in New Zealand and Keith Wilkinson in Australia, and finding stocks of the new parts at a reasonable price, that the method of obtaining stereo sound from an internally fitted PAL tuner module with an add-on stereo decoder board sitting on top of the tuner module as you have known it for the last 18 months is officially superseeded.

From this point on, any PAL tuner installs that I do (and I'm still doing on average three a week) will have what I am now calling the "Mark 2" stereo decoder fitted. This new way of obtaining stereo sound involves NO add-on stereo decoder board, but instead a complete replacement of the 3430G Micronas sound processor chip on the motherboard for a 3410D version. This is the same chip used in the UK Thomson model TiVo. Now that I am using both the same tuner as fitted to the UK TiVo and now the same sound processor that means there is not a lot of difference between a stock UK TiVo and a modified USA TiVo apart from the SCART connectors.

The benefits of a "Mark 2" decoder over the older (and what I am now calling the "Mark 1") decoder are as follows:

* Mono transmissions get detected and switch to a true mono output in under 2 seconds
* Poor signal areas now hold stereo better, and fall back to a true mono if necessary
* No interference from an internal “Airnet” wireless card as some people experienced
* It is more stable than the “Mark 1” in stereo separation and performance
* It has the ability to decode more stereo sound standards in other parts of the world
* NICAM stereo for New Zealand will (hopefully) one day soon be available^
* It looks more elegant. In fact it is hard to see that the board has even been modded!
* The "2nd Audio input" modification for two independent A/V inputs is still possible with the 3410D chip just like before

^ The 3410D does support NICAM stereo but there is a software conflict issue preventing this from being supported at present. For the time being the sound output is being decoded as mono from the 3410D and software testing is being carried out. It is recommended for anyone in NZ contemplating their options to consider a “Mark 2” decoder modification and when the issue is resolved a simple software upgrade will enable stereo decoding of NICAM transmissions at a later date. This applies only to New Zealand TiVo owners as Australian stereo sound decoding is NOT affected.

To enable this modification to function correctly a modified version of "palmod_config" is required. At present this is NOT INCLUDED on any OzTiVo image release but will have to be manually installed otherwise some weird sound effects will be heard. This modified palmod is Thompson's handy work and he has done a fine job incorporating the 3410D functions and also written into the module a logging function so that you can actually see in the kernel log what the decoder is doing. In fact his modified palmod can also be used on stock unmodified motherboards and will auto detect the type of Micronas chip fitted and switch accordingly. The software upgrade procedure along with the updated "how-to" document that encompasses all methods of obtaining sound from the internal tuner (ie mono, "Mark 1" stereo and "Mark 2" stereo) and a section on upgrading from either mono or "Mark 1" stereo to "Mark 2" stereo is available on Minnie here:
http://minnie.tuhs.org/twiki/bin/vie...erModALPSHowTo

While being able to deliver a more elegant and technically superior way of handling audio from the internal PAL tuner I've also been able to keep the cost of conversion to the same price as the older method. I also acknowledge that the new "Mark 2" may not be attractive to the D.I.Y. hobbyist due to the high level of soldering required which is why I am still offering "Mark 1" stereo decoders if you so desire to fit one of those yourself. All prices and an actual photo of a "Mark 2" modified board is available on my website here:
http://www.members.westnet.com.au/ki...tivo/index.htm

I would like to take this opportunity to thank Keith and Thompson for their software support as without them this would not be possible. I'd also like to thank all of the beta testers and especially Ron Davis who can confirm that proper mono resolution of TEN-Capital in Canberra is indeed a reality as this fact alone haunted me from day one of the original decoder. I'd also like to thank the half a dozen or so people who had faith in me and took me up on the offer for getting their TiVo's fitted with the new "Mark 2" decoder while I had their TiVo's in for PAL tuner fitting - all of which I am sure are happy with the result, or at least I presume so as I have not heard anything otherwise!

I always promised in the original stereo decoder document that a better solution would one day be available and anyone who knows me well enough knows I like to keep promises. Enjoy :)

Kind regards
Darren King

Outstanding job as always Darren.

Has there been any progress on the provision of NICAM decoding for NZ Tivo users?

Not being in a NICAM country myself the answer is "no" from this end, but I'm not the software guy anyway so I cannot 100% answer this question, sorry.

Quote:

---

Originally Posted by Islander

Has there been any progress on the provision of NICAM decoding for NZ Tivo users?

---

The (Mark 2) stereo modification did look promising for a while, however getting NICAM to work via the US mainboard does not look like it will be solved in the short term. There is a brief wiki on the subject of Stereo support in New Zealand.

Technically it's a Catch-22 situation... the MSP3410D is capable of decoding NICAM only when it is run as a I2S master, but we are having to run it as an I2S slave in order to get it working on the US mainboard. Flicking the MSP3410D between modes (slave/master) can be done via software, but it is not clear if the chip on the other side of the I2S bus can be configured via software.

The fact that the drivers running on the v3.0 software will recognise the MSP3410D and set it up as the I2S master suggests that the other component is not software adjustable. I guess what I am saying is that the software versions after v2.5.5 recognise the UK hardware and setup the MSP as an I2S master (which is why the v3.0 software will run on the UK hardware), otherwise it sets it up as a I2S slave - which implies that the chip on the other side of the bus is not configured via software (ie on the US board it is set, in firmware, as a master).

If you are technically minded and wish to dig around in datasheets and software internals and verify this then you are likely to need a FakeI2C package.

[I wonder who actually read this far :)]

Quote:

---

[I wonder who actually read this far ]

---

I do. Thanks mate for the detailed response.

All I can suggest to those that *may* want to see stereo eventuate one day is to get it done and just be happy with mono for now with just a simple software upgrade later for stereo if it happens. A lot easier (and cheaper shipping wise) to do it all the first time than to get the board returned for secondary modifications later..... but that's just my opinion of course.

I read the entire post thomson :) I'm a geek :)

Hrm.. sounds tricky.

Thanks for the reply(s).
I assume that the I2S device the MSP chip is connected to is the Tivo ASIC and the US ASIC is a different version to the UK one.
Assuming this to be the case I also assume that any detailed info on the functioning of the ASIC is not available hence the problem.
I am not very conversant with the operation of the I2S bus but wonder if it would be possible to make a slave to slave converter to interface between the two master devices ie. a pic or similar processor?

Quote:

---

Originally Posted by Islander

Thanks for the reply(s).
I assume that the I2S device the MSP chip is connected to is the Tivo ASIC and the US ASIC is a different version to the UK one. Assuming this to be the case I also assume that any detailed info on the functioning of the ASIC is not available hence the problem.

---

I have not had a look, but it would be a reasonable assumption that it is the ASIC chip on the other side of the bus... the only problem is that I think the markings on the ASIC chip seems very similar between the UK and US models... so it might not be the case...
eg:
UK: TMS1 REV A ASIC-00001-000 0036WX016
US: TMS1 REV A ASIC-00001-000 9919KX001

Your slave-to-slave idea might be possible... although I am not sure how the timing channel would be affected by this. The following is a short run-down on what I2S is, and where to find the datasheet:

I2S = Inter-IC Sound
Source: Audiotools.com
A high speed serial bus standard for use with digital audio and/or video communications developed by Philips in 1986 and based around their I2C chip-to-chip communication standard (aka 12S), were it differs from most such serial systems is that high resolution clocking signal is sent alongside the data itself to ensure synchronisation and reduce or eliminate problems such as Jitter. It was originally designed for on board communications between electronic parts as can be seen from the name and is used as such on everything from high end digital televisions to mobile phones, but has more recently been seen used on high end audio as an external interconnect standard in the form of I2S Enhanced. Basically an IS2 bus consists of 3 signal paths, one has 2 channels of audio or video data multiplexed onto it using TDM, the second channel has clock info and the third is a word select line. Get more info from Philips here (PDF file).

If you wish to go down a completely different track you might want to try a F15a board (Tomek found this)... you are likely to need the 38.9MHz version - price is somewhere around the NZ$40 mark. It's Polish, so contact, payment and shipping could be a hassle... I was in Poland a year back and still have some money was going to look into this before the whole MSP3410D swapout became viable. (FYI: banks here will neither provide nor purchase Zloty so I need to use it somewhere :))

This may be a wild idea.....but
Why not install a MSP 3410D on a daughter board, connect to the same conns as the mk1 stereo mod and command the new chip form the Tivo I2C bus. This would get around the I2S problem but present decoded NICAM to the inputs of the Tivo MSP chip.
I realise that this is not exactly a perfect solution but it would give the required result.
Any thoughts?

That's not as silly as it sounds, but not as elegant as everything all in one chip on the motherboard. Still if the end result is stereo then by all means!!!

A couple of points from the above posts:

1. There is no difference between the ASIC chips on a USA or UK motherboard as I have swapped chips on a board that although was not 100% working and it wasn't the ASIC at fault like I thought. I have brand new ASIC chips that were destined for UK TiVo's on hand. I'll repeat the procedure soon on a totally working USA board just to rule it out soon*

2. I was tracing the I2S (yes that's right it is I2S which is the sound version of the data I2C bus just in case other's were wondering if we have all been doing typos). From what I have gathered the main players involved here are the Micronas, the TiVo ASIC and the Analogue Devices audio encoder which is a microcontroller in itself and is quite possibly where our master/slave culprit is. I'll do a full trace on both a UK and USA board soon that are essentially bare so I do not miss any diversions on the bus to other weird places*

*both points have the asterisk (*) as time at present is a little limited. For anyone wondering why you are quite welcome to visit my website (link in signature) and see why! :)

Cheers

... And from this day forward, he shall be known as 'Daddy King' :)

To get a US TiVo with the new Micronas chip to run the UK image?

I was wondering if this was possible and would it result in NICAM functionality.

I have made a few basic attempts but have had limited success.

Problems to date:

TiVo boots in NTSC mode and continues through GS in NTSC output as PALMOD not loaded to convert output back to PAL.
GS fails at import data stage (series fail).

I have tried to alter rc.sysinit to correct PAL / NTSC on GS but so far have been very good at creating reboot loops.

The reason for this tack is that I figure that the hardware after a Micronas and tuner change is getting closer to being a UK than a US and maybe that an insight into the I2S master / slave problem can be gained from this approach.

Any ideas??

Have you running the NZ Thomson image?

Quote:

---

Originally Posted by Islander

This may be a wild idea.....but
Why not install a MSP 3410D on a daughter board, connect to the same conns as the mk1 stereo mod and command the new chip form the Tivo I2C bus. This would get around the I2S problem but present decoded NICAM to the inputs of the Tivo MSP chip.
I realise that this is not exactly a perfect solution but it would give the required result.
Any thoughts?

---

Building on this idea, I'm sure there are other ICs which could be used without having to program them for to specifically work with NICAM - briefly looking at STV8216 looks like it might do the job with its default reset settings and automatic detection, and it comes in a PDIP package which would make building a small board like the "mark 1" board much easier.
http://www.st.com/stonline/products/...re/ds/9786.pdf

Looks like that could be an option, and Table 9 reset defaults show it is indeed a good candidate and in default setting can accept input from the SIF and output it on SCART1. From there you feed the audio into the Micronas as the Mark 1 stereo decoder or mono hookup does via pins 56 and 57 of the Micronas.

You know, it's a shame the STV8216 outputs it's I2S sound data in Master mode only (refer part 6 of the document) otherwise it would have been a good replacement for the Micronas and solve everyone's sound issues... If it could have been configured for Slave I2S and someone was willing to write a version of palmod for the different I2C control commands and the necessary adaptor board (it isn't a simple "drop in" pin-for pin Micronas compatible replacement) then it would be an excellent chip to use in place of the Micronas altogether as it has both SIF sound input for use from the tuner and analogue sound inputs for use from the rear A/V input(s). Close... but no cigar :(

I have been able to get through GS using the current NZ UK 2.5.5 image on a US Tivo but have now got two basic problems.

1. As the boot PROM is for a US TiVo it starts up as NTSC and remains in NTSC after GS has finished . If I try to change to PAL using TWP the Tivo goes into a reboot loop.

2. There is no live TV. I am not sure if this is related to 1 above or the channel settings . I have tried manually selecting individual channels but have had no luck. I have setup using 2121 , cable, Auckland FTA.

My questions are:

Is there a way to set the TiVo to PAL working at the start of sysinit?
Is the lack of live TV from the tuner related to NTSC or a tuner setup problem?

Any assistance greatly appreciated.

I'm not familiar with 2.5.5, but there is code in 3.0 rc.sysinit which detects which hardware it is being run on. If it detects US hardware then it runs palmod.

Check the code.

When you say there is no live TV is the sceen blank or do you have the no signal detected message. from past memory when a PAL signal is input to a US TiVo in NTSC mode, you get the message screen.

HTH

Peter

Quote:

---

Originally Posted by petestrash

I'm not familiar with 2.5.5, but there is code in 3.0 rc.sysinit which detects which hardware it is being run on. If it detects US hardware then it runs palmod.

Peter

---

Thanks for the reply.

I have tried a brief attempt at tweaking the code in rc.sysinit to con it into thinking that it should config for PAL early on in the script but this quickly failed into a reboot loop that could only be stopped by removing the drive and inserting into a PC to rectify the code.

This then led me off in an attempt to have a test for a reboot loop at the beginning of rc.sysinit and branch to a safe version of sysinit on the second reboot but my script writing is not great and has to date has been unsuccessful ( the reason for the above is to reduce the swapping of drive to pc etc as the areas of sysinit to do with PAL/NTSC are well before bash/telnet start. I also feel that this could also be useful as a prevention of reboot loops for other reasons).

I had reasoned that the no signal message was due to the TiVo being in NTSC mode and the tuner O/P being PAL.

The TiVo also starts / remains in Maintenance mode NTSC output even after successful completion of guided setup but I assume that this is also because the PAL/NTSC/GS settings are not being set correctly for this firmware/software/hardware configuration.

My next thought is to load Palmod early in the rc.sysinit script but I suspect that this will only uncover even more problemsresulting in reboot loops.

Presumably what is required is a versioin of Palmod that runs once early in the sysinit script that configures the tivo in PAL mode and the leaves the rest of the operation up to the UK 2.5.5 software. Unfortunatly this is beyond my coding abilities.

If anyone can assist with the problems in the above ramblings this woulld be much appreciated.

Thanks in advance.

No help on the software side ATM, but I have one TiVo with a cheap $5-10 removable drive caddy and a spare housing in the PC I use for TiVo work. It makes running the drive between a PC and the TiVo much simpler during dev work.

Peter.