--- In TekScopes@..., "Tim Phillips" wrote:
> from Tim P (UK)
> just had a filing-cabinet 'purge' and turned up a copy of something called
> 'Handshake - Newsletter of the Signal Processing Systems Users Group'
> It's dated Fall / Winter 1979, presumably a quarterly.
> It has examples in TEK SPS BASIC with code-segments for (in my copy)
> driving a 7912AD with a 4050 controller.
> Anyone familiar with this publication ? Are there any more issues archived anywhere?
When I started at Tek as a design engineer in TM5000, we were in a division that included a sister product line – the transient digitizers. This included the engineers who wrote SPS Basic and the transient digitizer (hardware) group that the software supported.
SPS stood for "Signal Processing Software". This group pioneered the algorithms that went from mathematics text books to the practical application with digitized waveforms. Essentially, they did the ground work of what became ALL of the classical measurements in all DSOs today – including the first practical application of the Fast Fourier Transform (FFT) that generated frequency domain representation of time domain data.
The original code ran on a PDP 11 series minicomputer. We had several PDP11-34s for test beds in the department. It ran very fast and was very efficient – better than many of the early implementations in DSOs.
The hardware that acquired the signal that fed the code were a series of transient digitizers. I worked on the hardware of one of these – the 7612D. These used scan conversion technology to acquire a very short record (512 – 1024 samples) at a very high sampling rate for the day. The short capture time and relative long rearm time essentially limited them to fast single shot capture – which greatly limited the applications.
To overcome this limitation, a new product was proposed. The hardware group worked with the 7000 series design team to develop what was really the first DSO. The 7000 series team was starting a project to upgrade the 7704 into an "A" model, and with some political pull, the digitizer team was able to get the 7K team to add a feature to the new design – make the display section in a unit that could be separated from the "acquisition unit", similar to the topology used in 5000 series scopes. It was really a simple process, which is probably why it was approved. Essentially, the Acquisition unit was the lower half of the scope and contained the plug in slots, channel switching and trigger selection, and power supply. The Display section was the upper half and contains the CRT, Z axis circuit, and readout logic board. A single connector that separated the low frequency and differential X, Y, and Z signals from one side to the other was included – probably the most expensive part of this addition. Finally, the two units could be mechanically separated by removing a connection bar that latched them together. All 7704A units have this feature, and if you own one, you can separate them easily.
With this modification in hand, the digitizer group designed the "W7001". It was a module that would mounted between the 7704A Acquisition and Display sections. It contained a digitizer, and a waveform processor, based on the TI99000 16 bit microprocessor. (I believe it was the only application of that processor at Tek). The front panel included a key pad that had many pre-stored math waveforms, along with some user keys that allowed the user to program custom waveform analysis subroutines.
The waveform analysis was all based on SPS basic routines, and the same software team coded it. While the user interface was crude by today's standards (Still infinitely better than 11000 series which followed nearly a decade later), it had the basic concepts of waveform storage and recall worked out. I don't know the sampling rate, but it was not the full bandwidth of the 7704A. Unlike the transient digitizers, that had rather long rearm times, this acted more like a modern DSO, and allowed reasonable up date rate.
The product only really sold in the research and academia segments, and not to general scope users. But that was the way it was marketed. The marketing was done by the SPS group, not 7000 series. But the product paved the way for the 7854, which was grew on many of the W7001 basic concepts. You can almost equate it to how the Apple Lisa did the marketing groundwork for the Mac. Few people know of the Lisa or had ever seen one (I used one quite a bit at Tek), and by itself, it was a financial loser for Apple. But that was not the point. Rather it was a test platform to work out the concepts that ended up in the Mac. Same with the 7854.
Getting back to the original question – "Handshake" was a quarterly publication of the marketing group from SPS. It featured applications learned from customer interaction of the SPS software, as well as discussed the new product offerings. As the marketing group was in the same division as TM5000, when it came out, Handshake took on an additional role covering applications and technical notes from this product line. The publication morphed into a definition of applications news on any programmable instrumentation. At that time, it was pretty much limited to the digitizers and TM5000. Wilsonville was putting GPIB in their computing products (starting with the 4051 graphics controller), and because they chose an incredible intelligent IO addressing scheme in their architecture, these quickly became the controller of choice for automating data collection in engineering characterization applications.
(IMO – to this day, no one who has integrated GPIB into a PC platform (such as National Instruments) has come up with such an elegantly simple and easy to use method to pass data to and from the instruments to the program than the 4050 series does.)