history of protocol decoding in scopes

[JPortici]

Inspired by some of the recent threads,
many of us (expecially the greybeards) used to decode digital communication by hand because one time (or even now sometimes) there was no tool available to do it for you

i was wondering.. when did al the protocol decoding start? an esoteric tektronix plugin? some fancy lecroy monster cube? whatever?

[ebastler]

I would assume it was first available in logic analysers, then MSOs, then came to "regular" DSOs?
Just guessing. I'm sure someone in the know will be able to provide the facts...

[irakandjii]

I have been decoding manually for the past 18 months.  It .. as a young'un would say SUX! 

I have no idea when it became a thing in scopes, but in many cases it seems a marketing thing because many implementations have limiting use cases.  The USB protocol analysers at $600 or so US have a pretty UI, but depend on having a readable signal.  Heck if I have a readable signal. I'll just get a serial output from a second device (in my case a working 180 Mhz teensy).  So a scope makes sense if you design for a range of use cases.

I suspect it is becoming a thing because the serial protocols are now being used for inter-processor communication in the Internet of Things (IoT) and between layered microprocessors in robotics etc.  A long way from the original intent of chip to chip communication.

Just my 2 cents

[JPortici]

I would assume it was first available in logic analysers, then MSOs, then came to "regular" DSOs?
Just guessing. I'm sure someone in the know will be able to provide the facts...

probably, too bad lots is available about scopes, not so much about logic..

[alm]

The old HP catalogs are available for free as PDF. If someone was so inclined, they could go through them and note when the first products with protocol decoding were available.

Since logic analyzers supported protocol decoding well before digital scopes were common, and since a general purpose protocol decoder in an analog scope would either require a full-fledged logic analyzer or digitizing scope, I am pretty sure that it first appeared as either dedicated devices or as part of logic analyzers. I believe HP made a dedicated HPIB/GPIB protocol decoder with a bunch of lights and toggle switches. Not sure if this was before or after protocol decoding in logic analyzers. Since companies like HP and Tek made both scopes and logic analyzers, I imagine the technology migrated from the logic analyzers to the scopes. Just like how modern logic analyzer probes were derived from scope probes.

[MarkL]

I don't know if this is the earliest, but the HP 1631D logic analyzer could decode RS-232, RS-449, and GPIB.  It had 2x 50MHz analog channels (200MSa/s) and 43x digital channels.

It was definitely more of a logic analyzer than a scope.  It couldn't decode from the analog channels, but everything could be displayed correlated in time.

It's in the 1987 catalog.

[joeqsmith]

Right, I was going to say the real question is when LAs started having oscilloscope features.

Just to the right of my 141T you can see my 1631.  A modern home made pocket scope may do a better job.  Calling these a basic scope would be an understatement. 

[tggzzz]

Inspired by some of the recent threads,
many of us (expecially the greybeards) used to decode digital communication by hand because one time (or even now sometimes) there was no tool available to do it for you

i was wondering.. when did al the protocol decoding start? an esoteric tektronix plugin? some fancy lecroy monster cube? whatever?

Greybeards like me used a scope to check the voltages were correct, the setup/hold times were met, and there wasn't too much noise or poor transitions. At that point we knew we had valid digital signals, so we flipped to using the appropriate tools for digital signals: logic analysers, printf statements, change i/o pins, and in-circuit emulators etc.

And that what we still do for most problems. More thought, imagination, and a good design/implementation/test strategy is required for mixed-signal systems such as mechatronics and robotics. It is usually possible, but hackers will find it more difficult.

[JPortici]

I'm just curious about history. Maybe your beard is grey now but when it was brown there were already logic analyzers with protocol decoding.
I'm looking for the long long time ago

[mikeselectricstuff]

The Agilent MSO6034 had it fairly early on, but I don't think quite from launch - that would put it about 2005-6.

I was really annoyed when years later I found out that the 30 day trial could be reset by just resetting the  date 
 

[David Hess]

Inspired by some of the recent threads,
many of us (expecially the greybeards) used to decode digital communication by hand because one time (or even now sometimes) there was no tool available to do it for you

i was wondering.. when did al the protocol decoding start? an esoteric tektronix plugin? some fancy lecroy monster cube? whatever?

Tektronix made logic analyzer plugins for their oscilloscopes which did decoding but not of serial protocols.  They also made word recognizer units to provide oscilloscope triggering but again, that is not the same thing although I have applied them to protocol decoding to get a stable trigger.

Looking through the Tektronix catalogs, I get the feeling that decoding of serial protocols was done with dedicated instruments (asynchronous, HDLC, V.25, whatever) until serial protocols became more commonly used at the board level.  I see some parallel decoding capability going back a long time but no serial decoding capability.

[alm]

As usual, it depends on what you consider 'the first real ...'. If you limit it to serial protocols, then you might end up quite late ('90s or '00s). Parallel protocols were much earlier. Something like a word recognizer might not qualify since its decoding is extremely limited, but something like a Tektronix DF2 (could decode GPIB) or DF1 (format a parallel bus as a list of hex/binary states) could arguably qualify, through you could argue if this was for a logic analyzer (7D01) that happened to use a scope as display, rather than a scope. This stuff would have been mid-70s.

A HP 59401A Bus System Analyzer (would display the current status of an HPIB/GPIB bus with lights and allowed you to send messages using toggle switches), also around mid-70s, could be argued to be a stand-alone protocol decoder (similar to the TotalPhase products available these days). The HP 1601 logic analyzer (mid-70s) could do basic state display, and the later 1611 (late 70s or early 80s) could do actual inverse assembly, which I would consider a fairly sophisticated kind of protocol decoding.

So based on a quick look in a few Tek/HP Catalogs, I would place the first products around mid-70s. Basically when digital design and computer design started to take off.

[David Hess]

As usual, it depends on what you consider 'the first real ...'. If you limit it to serial protocols, then you might end up quite late ('90s or '00s).

I took it to mean serial protocols but even up through 1994 when Tektronix introduce their TLS216 "Logic Scope" which is 16 channel DSO, I did not find anything except dedicated protocol analyzers.  It would not surprise me if HP implemented this first but someone who knows more about their old products lines can answer.

[artag]

The first time I recall getting serial analysis on a LA was with the HP16505A, an external processor for the 16500C. This came out in about '96.
It used the 16500C to capture data, transferred to the the 16505 using SCSI (took a long time for large captures) and then did protocol analysis. The 16500 series could also have oscilloscope plugins so I guess you could consider that a scope with serial analysis - though I don't think it could use the analog data.

There were also disassembler add-ons for the 1630 range. Doubtless this could have allowed a serial analyser but I don't recall one. At the time, I probably used an HP4951A/b/c which could only do async, hdlc etc. : not on-board protocols like I2C.

[gslick]

If you wanted to do RS-232 type serial bus decoding with an HP logic analyzer you could have used a 10342B Bus Preprocessor. That was one of those interfaces that attached to the logic analyzer pods through a 10269 interface.

The 10342B used a Zilog SIO/1 controlled by a Z80 CPU to convert the serial bus data to the parallel logic analyzer interface in hardware instead of doing the conversion in software on the analyzer. Baud rates and formats were limited to those directly supported by the 10342B hardware.

The first version of the manual describes using the 10342B with the 10269A or 10269B interfaces with 1630 series analyzers in 1985.

The second version of the manual describes using the 10342B with the 10269C interface with 1650 or 16500 series analyzers 1987.

http://www.bitsavers.org/pdf/hp/te/logic_analyzer_preprocessors/10342-90901_10342B_RS232_RS449_HPIB_Operating_Manual_Aug85.pdf

http://www.bitsavers.org/pdf/hp/te/logic_analyzer_preprocessors/10342-90912_10342B_RS232_RS449_HPIB_Operating_Manual_Nov87.pdf