Computing – by Fred Howett

My Early Days of Computing – by Fred Howett

1. Stantec Zebra

My earliest encounter with a real computer was in 1969 when I was a young design engineer working in the Mobile Radio Division at STC New Southgate. We encountered an occasional problem with a UHF radio transceiver. In those unregulated days, the transmit and receive channel frequencies for a specific application were provided by a government agency (usually at the last minute). In a few instances we had serious problems with a spurious response of the receiver rendering the system useless. What was needed was some way of predicting the numerous spurious responses associated with the different channel allocations across the band. This was in the days when such calculations were carried out with an electro-mechanical (Friden) calculator – a process which was exceedingly lengthy, tedious and error prone. I learned that STC had its own computer a STANTEC ZEBRA on site, and best of all it was sometimes available in the evenings. With my colleague Geoff Lowe we wrote/evolved a programme which was able to predict the potentially problematic channel pairings, a feat which we were quite proud of as we were not programmers.

The computer was housed in a separate building at the far end of the north. playing field at the Southgate site. Programming was done at an assembly language level using either “Normal” Code or “Simple” Code. A high level language did not exist for the Stantec Zebra at that time. The programme was entered via a reel of punched tape and output to a printer. I remember that the computer memory was provided by a nickel plated drum -like a large dustbin spinning at ten thousand rpm. While this may seem slow by today’s standards, by positioning several sets of read/write heads around the drum it was possible to perform several operations per cycle. Unusually for computers at that time it had a large 32 bit word length which made it quite a powerful machine. Although ITT decided to kill off the Stantec Zebra programme, a number remained in use for many years, I believe the famous yachting sail making company Banks finally switched their Zebra off in 1981 after many years of faithful service.

2. Rockwell PPS4 Microprocessor

After STC sold off its Mobile Radio Division in 1972, I joined the Audio Division at New Southgate who were keen to add some electronic design skills to their established acoustic design expertise. An early development made use of the newly available microprocessors. In our case it was the Rockwell PPS4. The product was a 48 number repertory dialler with display and was aimed at British Telecom.

Early microprocessor based products were designed under very difficult conditions as the support tools we now all take for granted had not yet been developed. The programme was written in assembler language and was then commited to paper tape. This was used to send the programme via telephone line to Rockwell in California for compiling. Due to the time difference between UK and USA, we would, if there were no programming errors, receive a compiled programme (in hex code format )24 hours later.

Although there was an evaluation board, we had no PROM programmer so we had to build our own in the lab. The programmable memory chips were uv-erasable, and we purchased an ultraviolet sunray lamp from Boots the Chemist to erase our memory chips. The programme was written to the memory chip line by line. The data was entered in binary form so had to be mentally converted from hexadecimal as each line of code was entered. One mistake and the chip had to spend 20 minutes under the sunray lamp to erase it before starting again.

Despite the difficulties, the software was developed in under 6 months and working prototypes demonstrated to British Telecom. My recollection is that they were very impressed and said it would make an ideal 2nd generation repertory dialler, but at that time they were looking for a 1st generation product.

3. Rockwell AIM65