What the first PLL-based digitally synthesized HF receiver?

[13dka]

Finding out how and when exactly DDS came into existence seems to be harder than I expected. It looks like patents for numerically controlled oscillators (NCO) emerged in the 70s but it seems it could be questioned whether or not the technology had a direct advantage for military or commercial usage yet. The "Speakeasy" SDR concept/project if I understood that right was a 1990s thing and commercial radio research had shifted largely towards cellphone technology I guess, so many "firsts" are to be found there, probably...

Poking around in ham radio service manuals instead, first clues were radios like the 1992 -DX update to the Kenwood TS-950S, the 1993 TS-50 or the 1996 AOR AR-5000 sporting the DDS acronym in their documentation.

It looks like JRC's JST-135 transceiver of 1988 (date of manual print) is the first radio I can find, and the NRD-535 seems to have inherited the dual DDS, dual PLL board from it in 1991, so that might be your candidate for a first SWL radio with DDS? JRC apparently produced their own DDS ICs for that board, so the internet claiming "mid 90s" for the first integrated DDS chips seems questionable.



However, comparing phase noise figures in order to identify DDS radios wasn't the bright idea I thought it was -- for example, the AOR 5000 has particularly high (103dBc/Hz @10kHz spacing) phase noise on Rob Sherwood's test bench, that's right among the noisiest of PLL synthesizers (like in the R600 - R2000, FRG7700... more than a decade earlier). I could not find phase noise measurements for the JST-135 or Kenwood radios, but the NRD-535 and -545 are only getting close to the level of their PLL predecessors and that's apparently 10dBc more to go to match the best PLL designs. What gives? Isn't DDS meant to have lower phase noise?

An answer seems to be that "analog output DDS synthesizers have both amplitude errors and phase errors on their outputs from basic sampling theory even if the clock input is perfect" (source). That, if still true with modern and probably much (!) higher clocked digital logic with increased bit depth, would explain why it looks like only SDRs end up in the phase noise ballpark of non-PLL radios around 140dBc/Hz @10kHz or even better. So it seems DDS wasn't exactly a revolution (or at least a real upgrade) for the amateur in the beginning, just an important new lego brick needed for the truly revolutionary things to come.

 

[13dka]

What I missed is the IC-781, also 1988 and it was preceded by a DDS in the IC-275 VHF allmode transceiver in 1986, but that one is pretty out of scope (pun intended). I also checked a magazine and it looks like the IC-R72 was introduced in 1990, not 1992 as Google claimed so Icom may have had the first SWL tabletop with DDS.

 

[K0WWX]

The synthesizer in e.g. the 8054 receiver is described somewhere down that 100ft long .pdf like...

That sounds to me like a digitally controlled and (to me) mind-boggingly complicated but analog PLL synthesizer? Whatever it is (or what I missed), boy did they mean the "control" part and the concept is impressive in many ways, documentation is plenty and an interesting read!

Good research, thanks. The 1984 manual supplement for the Rockwell-Collins 8054A receiver notes that it differs from previous configurations because it contains a DDS.

 

[prcguy]

Good research, thanks. The 1984 manual supplement for the Rockwell-Collins 8054A receiver notes that it differs from previous configurations because it contains a DDS.

View attachment 191653

The first PLL digital synthesized receiver will be much older than 1984. There were full blown SDR receivers and transceivers running in the late 80s.

 

[K0WWX]

The first PLL digital synthesized receiver will be much older than 1984. There were full blown SDR receivers and transceivers running in the late 80s.

Thanks, but my post was in reference to what the commercially-available receiver with a DDS might have been. If you know of one before 1984, please list a model number.

 

[13dka]

The 1984 manual supplement for the Rockwell-Collins 8054A receiver notes that it differs from previous configurations because it contains a DDS.

I guess I didn't have the patience to find that upgrade down the list. Good find!

There were full blown SDR receivers and transceivers running in the late 80s.

I was always assuming the same but I could never find any evidence for that, nowadays the WP article for "SDR" has a quite comprehensive history section explaining why -- apparently "digital baseband" receiver prototypes emerged as early as 1984 indeed but it seems the term "SDR" was coined and mass-produced implementations were rolled out at the earliest in the 1990s?

 

[prcguy]

I guess I didn't have the patience to find that upgrade down the list. Good find!



I was always assuming the same but I could never find any evidence for that, nowadays the WP article for "SDR" has a quite comprehensive history section explaining why -- apparently "digital baseband" receiver prototypes emerged as early as 1984 indeed but it seems the term "SDR" was coined and mass-produced implementations were rolled out at the earliest in the 1990s?

I find the Harris RF-5000 series of SDR HF transceivers was introduced in 1993 and the US military bought lots of them. Considering how complicated this radio is, and how beautiful its designed it had to be in the works for a couple of years. The first commercially sold handheld I find is the Racal MSHR in VHF with an early contract in 1995 and it had to have been in development for some time. I worked with a Hughes Aircraft HF/VHF/UHF SDR transceiver in the early 90s and it had several years of development before I got to it. That was not a mass produced item but they did sell some. It could receive AM broadcast and VHF airband and 800MHz cell phone while simultaneously transmitting on 40m SSB amateur, etc.

I'm sure there are others that predate what I've mentioned with development that goes back into the 1980s.

 

[13dka]

Harris RF-5000

Look, that's how we get an idea here - someone drops a name, someone else thinks that can't be right and every time we learn something.

I spent quite some time to find some kind of circuit description/block diagram revealing the receiver architecture for that radio and couldn't find anything that would show me how exactly this was an SDR. What I found is not my favorite kind of source but it'll do for now: Source

Timothy: Harris’ first SDR, the RF-5000, was released back in 1986 and we have continued to develop and deploy generations of SDRs since.

So that's that, we have preliminary milestones for PLL synthesizers and PLL synthesizers with analog decade switches, PLL synthesizers with digital control (local and remote) and DDS radios. Now we are getting a clue that the 1984 digital baseband receiver may have found its way into production pretty quickly.

The Harris website begs to differ about the roll-out date tho:

 

[prcguy]

Look, that's how we get an idea here - someone drops a name, someone else thinks that can't be right and every time we learn something.

I spent quite some time to find some kind of circuit description/block diagram revealing the receiver architecture for that radio and couldn't find anything that would show me how exactly this was an SDR. What I found is not my favorite kind of source but it'll do for now: Source



So that's that, we have preliminary milestones for PLL synthesizers and PLL synthesizers with analog decade switches, PLL synthesizers with digital control (local and remote) and DDS radios. Now we are getting a clue that the 1984 digital baseband receiver may have found its way into production pretty quickly.

The Harris website begs to differ about the roll-out date tho:

View attachment 191757

I've owned a large number of Harris RF-5000 series radios including the original RF-5010s and various RF-5020s and have the manuals with circuit descriptions. Its an SDR that digitizes the IF and uses digital signal processing for demodulation, AGC, IF BW and so on. The transmit path digitizes the mic audio then uses DSP to create AM, FM, USB, LSB and a couple of digital modes then mixes that computer created waveform with an LO to operating frequency and amplifies it to rated power. All these things are changeable or programmable via software or firmware downloads including new waveforms.

 

[13dka]

All these things are changeable or programmable via software or firmware downloads including new waveforms.

Just so you understand my angle: It's not that I didn't believe you the first time but I'm a wannabe leisure-time historian for some other tech topic and there I learned how important and helpful it is to (mutually) fact-check everything, even if heard from the most knowledgeable guys I know (and rely on). Our topic here brought up a good example why:

We have now established (with your help and sufficient indications, if not evidence) that there was a complete commercial SDR system available by (at the latest) 1989 if I read that Harris timeline graphic correctly. The Wikipedia article seems to contradict this, it reads like everyone was dragging their feet until practical implementation happened almost a decade later. Maybe I read it wrong but for now I consider that article confusing enough to be a bad source. OTOH, I learned that the guy (U.L. Rohde) who made that allegedly first digital baseband receiver in 1984 is the son of Mr. Rohde of Rohde & Schwarz fame.

@All -- To understand the (usually military and commercial) first implementations of all the technology leaps discussed here it seems crucial to keep their different motives and requirements in mind. For example, it's obvious how "frequency agile" and precise synthesizers are beneficial for all radio users including amateurs, but particularly military users obviously had very different spectrum usage and modulation schemes in mind (Frequency Hopping/FHSS etc., encryption and error correction, practically implementing Hedy Lamarr's famous wartime ideas post war) rather than convenience.

Networking radio coms was another early requirement that made voltage control and a digital control interface necessary and at that point radio tech could assimilate practical developments in the equally emerging computing sector. That brought us the "digital" PLL radios we maybe grew up with at the end, but convenient station memories as we know them were apparently not the top motivation that led to them.

I assume that not only Increasing bandwidth requirements for the ever-increasing amount of information to be exchanged brought (in their core still) analog synthesizers to their limits? That must have made the idea of a digital oscillator that doesn't even need an interface to the digital domain very compelling and it (as I picked up somewhere) multiplied the hopping performance for example. So that brought us DDS and further down the signal path, ever growing baseband (=IF) portions had to be processed in ever growing complexity, to summarize what ended up in the development of SDR -- letting you process a (nowadays gigantic) baseband without ever leaving the digital domain, in every way that can be defined in software.

We probably all assumed this all along but this thread seems to confirm that almost everything we appreciate in our stunningly capable, yet affordable little radio wonders has its origin in technological quantum leaps propelled by the signal intelligence war, which then trickled down into commercial, then our hobby stuff through ideas and chips that were more often than not initially developed with your valued tax dollars.