In my previous post I looked at the Arcam Delta 150 NICAM decoder, said to be the best receiver ever produced. Spoiler: It isn’t.
For all of the Delta 150’s fanfare – in retrospect we have to accept that it was little more than a boilerplate Terrestrial NICAM receiver, using the same design and components as practically any other name-brand TV or VCR of the time.
For a Broadcaster, some kind of “monitor” type receiver would have been required to test and validate their infrastructure. A/V equipment designers also would have required an ultimate benchmark of the NICAM standard to test their own products against. Today we’re going to have a look at how the big guys designed such a receiver.
Like the Philips equipment I’ve looked at previously, there is no manual or specification available online, nor are there even any mentions of it other than a few having come up for sale. Unlike the other units, this is older, dated 1989, in less-than-pleasing 1980s “brown” livery. There was a newer white one for sale while I was writing my previous pieces but some bastard beat me to it. What the hey? Seriously? Someone other than me is buying this stuff in 2022?
Without having to open this thing, I know that like the equipment I’ve already looked at, it is going to be another masterpiece of electronic engineering which deserves to be seen and admired. That is the reason I have it here today.
The PM5688 features the same overkill mechanical construction seen on the PM5685 and PM5687, its front panel folding down, once again revealing a set of four different modules, but the modules contained within are all completely different, performing the opposite function as the transmitting end.
For this unit I’ll be starting at the analogue audio output, then working my way back to the RF input.
Unit 3 – Audio Output
This boards job it to take the digital audio data from the decoder (up next) and transform it back into analogue audio signals. The most prominent feature of this board is the TDA1541A DAC. Did you hear me complaining about the use of the TDA1543 on my previous post? Consider me silenced. The TDA1541 rose to fame as the DAC of choice in high-end CD players in the 1980s and 1990s. Even today much myth and hysteria surrounds this now discontinued chip, with audiophiles paying huge sums of money for higher grade samples for use in pre-made “DAC boxes“.
In the world of the audiophile, CDs are considered to be handed down directly by the gods. Perfect renditions of the music they contain. The DAC is one of the key conduits between the listener and true enlightenment, thus much emotion surrounds its selection.
In the world of NICAM, digitised data streams were thought to be read from golden plates found in hilly areas – with the NICAM system being only 14-bit, and with many opportunities for signals to be distorted at the source – depending on the individual in question the same level of faith couldn’t be placed in the received audio, however DAC selection was still critical to a small number of devotees.
In this application it is probably overkill, certainly there would not have been any justification for a “two crown” version of the TDA1541A. The TDA1540 would have matched the TDA1534 ADC at the transmitting end nicely however may have been “too old” or out of production when this was designed.
This assembly also includes a by-passable de-emphasis circuit, electronically controlled variable gain stage and a very nice looking low-pass filter. Down at the connector end are two massive and rather heavy audio isolation transformers.
Unit 5 – NICAM Decoder
This module is based on a mysterious proto-NICAM decoder, an SAA7272. Unlike the well-documented consumer focused SAA7280 which gets the job of decoding done in a single chip, this is part of a 5 chip arrangement including two of another mystery chip – SAA7273, as well as two SRAMs. My guess is this thing is probably more CPU than ASIC.
Also present on this board is the familiar SAA7220 digital filter seen previously in the Arcam Delta 150. Arcam may consider their decision to include this chip validated.
Unit 6 – QPSK Demodulator
This module’s job is to take an RF input signal and recover the digital NICAM data from it, feeding it to the decoder looked at previously. Like the DQPSK modulator module in the PM5687 this board is standard specific. Two versions exist – CCIR System I (PAL-I), and CCIR Systems B/G/D/K (PAL/SECAM). I have both.
This board mostly consists of bandpass filters of varying sorts. Unlike the inexpensive “Toko” filters found in consumer receivers, this board is packed with high quality tuneable inductors and custom value ceramic capacitors.
Here’s this one annotated:
Performing the DQPSK demodulation we have an interloper: A Toshiba TA8662N. Philips had their own DQPSK demodulator, the TDA8732, however I’ve not seen one dated prior to 1991. This product went to market in 1988.
Unit 4 – EYEH & BER Detector
Lightly seasoned with mostly 74xx logic, this module prepares the signal quality and bit error-rate counts for the display on the front panel. I have no idea how it works. It probably connects to ancillary outputs on the TA8662N demodulator chip on Unit 6.
Front panel assembly
The front panel board is quite similar to that of the PM5687. Philips once again have reached for the trusty MCS-51 and built another complicated product with it. I dumped out the EPROM and found some rather interesting text in it:
Welcome to Philips at IBC-88. On display you will find a wide selection of TV test and distribution equipment. This year Philips presents various items for NICAM Sound Broadcast and testing – PM5685 NICAM Encoder – PM5686 NICAM Sound Modulator – PM5687 TV Digital Sound Modulator – PM5688 NICAM Sound Demodulator. Also NEW is PM5640 Video Test Signal Generator and a new sync change over unit. The PTV-staff is here to assist you – please enquire within. Did you see our new LOGO generator? [possibly referring to the PM5644]
It implies that all of this equipment was first introduced in 1988. Not the best English but we can forgive. Given that this likely came from The Philips TV Laboratory Copenhagen the Engineers were certainly Danish. They must have been expecting a large Danish contingent as the text also appears in the ROM in their language:
Velkommen til Philips på IBC-88. På display vil du finde et bredt udvalg af TV test og distributions udstyr. I år presenterer Philips forskellige enheder for NICAM Lyd Udsendelse og afprøvning – PM5685 NICAM Encoder – PM5686 NICAM Sound Modulator – PM5687 TV Digital Sound Modulator – PM5688 NICAM Sound Demodulator. En NYHED til PM5640 Video Test Signal Generator og en ny sync change over unit. PTV`s folk er her for at hjælpe dig – henvend dig blot til en. Så du vores nye LOGO generator?
(In case you were wondering, the non-latin characters were stored in the EPROM using non-standard encoding. It took a bit of google-translate-fu to reconstruct the above text).
The message was probably marquee’d on the PM5688’s front panel in a special demo mode. With just 8 characters per display I’d imagine it didn’t quite have the impact the engineers were hoping for. Hopefully someone saw their new logo generator. It would be nice if the “PTV-staff” had had left their names in the code.
IBC-88 was a Broadcast trade show held in Brighton in the year 1988. Philips must have had one of these on display. Unfortunately reporters were too fixated with the BBC’s HDTV display to have visited Philips.
Front panel controls
- DQPSK: Modulator is in sync with the source (Not illuminated when using digital connection)
- FRAME: Decoder is in sync with the source
NICAM MODE (Display)
Displays various operation states:
- Stereo/Dual/Mono/Data =1 / ≠1: Receiving valid signal. Audio mode of source and state of sound flag. “=1” if the source sound flag is set to S1=S2. “≠1” if the source sound flag is set to S1≠S2.
- No IF ?: No signal from source
- Wrong IF: Incorrect IF input detected
- IC In ??: Expecting Intercarrier input but none detected
- A bunch of other obscure states I can scarcely imagine how to activate
IF MODE (Intermediate frequency)
- INTERC: Tells the receiver that the IF includes both a vision and NICAM carrier. In this mode it can lock onto the vision carrier tone and enable automatic frequency correction.
- SPLIT: Tells the receiver that the IF contains only a NICAM carrier. In this mode there is the problem of carrier frequency “drift” potentially explaining why my unit powers up at 100% signal quality, then degrades down to 93% after getting up to operating temperature.
Switches the input between one of three different sources:
- IF: 33.05 MHz (or 32.348 MHz for the “System I” model) spectrally inverted DQPSK signal (IF connector on rear)
- INTERC (Intercarrier): 5.85 MHz (or 6.552 MHz for the “System I” model) DQPSK signal (INTER. C IN connector on rear)
- NICAM: Baseband digital input (NICAM connector on rear)
PARAMETER / “TRANSMISSION QUALITY” Display
Toggles the measurement on the second display:
- EYE: Signal quality as a percentage
- BER: Bit-error rate
Enable/disable de-emphasis circuitry on the analogue board. This function is the exact opposite of the pre-emphasis feature on the transmitting end.
When the transmitting side is in bilingual mode (dual sound) this switch toggles this units’ two (now mono) outputs between the ‘A’ input (primary language) or ‘C’ input (secondary language) of the transmitting end.
This is a four way switch which allows the three “test” BNC connections on the rear to output various signals of interest to Broadcast engineers. The four options are:
- 1: I/Q components from the demodulator. Allows these signals to be examined on external specialised test equipment
- 2: Audio A/B: Outputs single ended audio signals for connection to an oscilloscope
- 3: NICAM Data / Error flag. Baseband NICAM data output and error flag (uncertain what this refers to)
- 4: NICAM Data / NICAM Clock. Baseband NICAM data and recovered clock
First IF input for a signal from a 38.9 MHz centric IF system. NICAM carrier is expected at 33.05 MHz (or 32.348 MHz for the “System I” model) It must be spectrally inverted. This connection is most likely how a large broadcaster would provide the input to this unit. Prior to this unit would be another piece of equipment performing a single down-conversion from the on-air channel, inverting the spectrum in the process.
INTER C. IN
Second IF input. 5.85 MHz (or 6.552 MHz for the “System I” model). If the Broadcasters test down-conversion apparatus has both down-conversion stages, then this input would be used instead of the “IF IN” connection. It is not spectrally inverted because each down-conversion stage in the equipment before this unit would have inverted, then un-inverted the spectrum.
Baseband NICAM digital data input. Allows the DQPSK modulation/demodulation stages in both the transmitter and receiver to be bypassed. This connection may be used if this piece of equipment is testing a signal received over a leased line or microwave link, instead of on-air however much of its functionality would be wasted in this scenario.
NICAM clock signal, if one happened to to need it. Typically not required as the receiver can recover the clock from the NICAM data signal its self.
Both of the above connections appear twice to allow signals to be looped through this unit for convenience.
CH A/CH B/TRIG OUT
Three test signals configurable by rotary switch on front panel. These would typically be connected to an oscilloscope to allow Broadcast engineers to perform various testing on their infrastructure.
AUDIO 1/L & AUDIO 2/R
Balanced analogue audio outputs.
GPIB remote control port. Like the PM5685/PM5687 the settings on the front panel can be configured remotely over this interface.
Outputs the “spare” bits which can optionally be encoded into NICAM frames. I’m told Channel 4 (UK) used these for OB Talkback data. As far as I know Philips Modulators cannot splice these into the data stream, thus this would be for testing bespoke coders.
Testing the PM5688
Being possession of a PM5687 with a System B/G/D/K modulator, and a PM5686A with a System I modulator, it was quick and easy to throw together a test setup.
In the above setup I am reproducing the audio twice from a signal source, via two different broadcast standards.
Before I got the PM5688 I didn’t have a lot to say about this. I put together a test setup where the PM5687 is fed by my CD player, and the PM5688 feeds my headphone amplifier. I splashed out on a couple of balanced-to-unbalanced conversion boxes to assist. I had the ability to rapidly bypass the NICAM setup so I could compare to the “straight through” scenario bypassing this equipment.
The reproduction of the audio is exceptional. Being only a 14-bit system, limited to 16 KHz one expects degradation. I couldn’t hear any by my ears (but I can’t hear anything much above this frequency anyway). What really sets the PM5688 aside from the other receivers I tried (I have 3) is how accurately equalised the received audio is. It is reproduced exactly as transmitted. The other do not do this.
This experience echoes the sounds of the original 14-bit Philips CD100: It’s not how big your word is, it’s how you use it.
Back in the 1990s I thought it was amazing that we were receiving a digital audio Broadcast. DAB Radio, a better known digital audio broadcast – the original implementation was inferior to NICAM did not arrive for another decade or more. It has been a fascinating journey taking delivery of all of these items behind that transmission, taking them apart, and figuring out how they work.
The engineering, particularly in the transmission equipment is brilliant. The signals these generated were received and enjoyed by millions, if not tens of millions worldwide. For most engineers this is sufficient gratification but it’s always nice for your creations to be seen.
I still have plenty of mini-projects and investigation I want to do on these items, some of it may be written up in future.
I hope that some day people who worked with it will find this page and get in touch. I’d certainly like to learn more about this most mysterious, unknown world.