In my previous post we took a look at some of the best equipment available for NICAM transmission. Now we’re going to take a look at the receiving end.
NICAM by its very nature delivers us to the world of the audiophile, where the prospect of a 1/1000th improvement in audio quality demands a £1000 device to capture it. Today I’ve got just such an item on the bench: The Arcam Delta 150:
Before we get into the test setup, let’s whip the lid off this thing and take a look:
The Delta 150 is essentially a terrestrial television receiver, including NICAM audio with some extra signal processing magic thrown in. It was designed in the UK and is hardwired for the PAL-I standard – limiting its use to the UK, Ireland, Hong Kong and South Africa. I doubt many migrated from these shores.
How it works
For the most part the design of this unit is quite typical of any terrestrial receiver at the time. The SAW IF filter after the tuner has two outputs, one covering the IF frequency of the vision carrier. This is fed into the TDA8341 whose job is to recover the composite video signal. We don’t really care about this part. Let’s move on.
The other output of the SAW IF filter heads over to the TDA2546A which contains the FM mono receiver. This chip also outputs a second IF from which the 6.552 MHz (PAL-I) NICAM carrier is selected. This signal then passes through the 4 pole Toko filter (specifically designed for NICAM) then to the TA8662N DQPSK demodulator.
The TA8662N outputs the baseband NICAM data and recovered clock from the transmitter (the very same signals outputted from the back of the PM5687’s “NICAM OUT” BNC connectors) then feeds it to the CF70123C, which is the NICAM codec. The signal should then head onto the TDA1543 DAC, but it doesn’t.
This unit has an extra chip which sits in between the codec and the DAC: An SAA7220 from Philips. The special sauce of this unit. Its datasheet describes it as “Digital filter for compact disc digital audio system”.
This filter is intended to reduce (not necessarily eliminate) the anti-aliasing filter at the output of the DAC to satisfy the Nyquist–Shannon sampling theorem. Instead of having a large and complicated analogue filter network rammed with obscure and expensive adjustable components, instead things are mostly done in the digital domain with a simple chip the manufacturer can whack on the board without requiring any manual adjustments on the production line. A simpler filter arrangement after the DAC may remain.
Philips’ second (and third) generation NICAM decoders (the SAA7282 & SAA7283) had a built-in DAC adopting a 1-bit architecture, rendering this approach and any of the original benefits of the Delta 150 redundant.
After that we’re of course at the TDA1543 DAC. A popular but austere DAC from 1989 often found in older CD players. It is interesting that we don’t see more money thrown at this component however given that the source ADC is only 14-bit this might have been hard to justify.
Getting it working (Intercarrier method)
So how do I bring the Arcam back to life, using the unfamiliar connections on the back of my broadcast NICAM equipment?
I reached out to Nathan Dane from Ireland who’s also recently been looking at this and was informed of a “hack” to make it work quick and easy:
Above, the Intercarrier output from the PM5686A is simply connected to the composite video input of a cheap CCTV modulator, which in turn is feeding the Arcam. This worked first go and required no special expertise or equipment.
Feeding what is effectively an audio signal into a video input is a little counter-intuitive. Let’s bust out the spectrum analyser and see what’s going on here:
Looking at the PM5686A’s intercarrier output we can see a modulated NICAM carrier at the absolute frequency of 6.552 MHz. The exact spacing from the vision carrier in a PAL-I system.
Now when we look at the CCTV modulator’s output with an empty video carrier at 471.25 MHz. We can also see that familiar NICAM carrier again to the right, after the empty vision/FM mono audio carriers. All the modulator does to the video signal is shift the frequency up by the carrier frequency, putting the NICAM carrier at 477.802 MHz. Exactly where it needs to be.
The beauty of this method is that the Arcam won’t tune to just a NICAM signal by its self. It requires the accompanying PAL-I transmission, which has been provided.
The drawback however is that the composite video input is occupied by the NICAM connection. To have both video and NICAM on this input would require some additional circuitry which I won’t be looking into at this time.
I doubt this method would work on a Broadcast grade, or even a higher quality vision modulator as such equipment is likely to filter out signals above the maximum allowed bandwidth for composite video.
In a large scale setup, the carriers are more likely to be combined on-channel as RF signals. Before I can get started on this, there is an annoying problem. The Labgear MOD111’s carrier frequency is quite unstable, drifting +/- 5 KHz on a good day. Given that it has a crystal reference, this is a bit odd. In its normal application this doesn’t matter as the AFC (automatic frequency control) of an analogue receiver is more than happy to chase wandering carriers around.
The problem is that my NICAM carrier is going to be generated by high quality equipment stable within a 30 Hz margin at this frequency. Because the receiver will be following the unstable vision carrier, that’ll mean it’ll fall out of tune of the NICAM carrier, and in my initial testing, it indeed does.
Before I can go any further I need to get this thing under control. I had grand visions of clocking the MOD111 from an ultra stable 10 MHz reference however found that the MOD111’s internal crystal is 4 MHz. Converting from a 10 MHz reference requires a phase locked oscillator or an “RF Grade” fractional divider (a “digital grade” divider would have too much jitter) – not something that amateurs like myself can easily build. The PM5687 contains two in its master oscillator module but I am yet to unravel its secrets.
To solve this I whipped out the supplied crystal and replaced it with a DIP-14 oscillator. Having done this conversion I still have a 2.5 KHz “error” on the carrier frequency however I’m now stable within a range of about 500 Hz which is good enough for this experiment as NICAM tolerates +/- 2 KHz of carrier frequency error.
Throwing it all together – I now have this stack up:
The NICAM signal is now fed into the antenna input of the MOD111 having just been upconverted by the mixer on top of the PM5686A.
The mixer is driven by my ESG signal generator. As overkill as this appears its ultra fine frequency granularity allows me to exactly match the error of the MOD111.
The signal generator is set to 510.1525 MHz – which is 471.25 MHz (UHF channel 21) + 32.348 MHz (PM5686A IF frequency) + 6.552 MHz (PAL-I NICAM offset) + 2.5 KHz (oscillator error in the MOD111).
And we’re in business. The “NICAM” LED on the Arcam receiver is illuminated once again. Did I go to all of this trouble just for the LED? Of course not. I’ll be listening to some music through this setup for the next few days.
After much fussing around I can now present a full PAL-I spectrum including NICAM. Something that hasn’t been seen for a long time 😉
You thought your Arcam Delta 150 was the be-it-all-end-all receiver? Think again.