12-08-2012, 09:09 AM
All my yesterdays!
This was the first fairly complex project that I built – a counter/timer designed by two radio amateurs, G Firth, G3MFJ, and David Pratt, G3KEP who later became Chief Examiner for the C&G Radio Amateurs Exam for many years until the RAE was made more elementary by Ofcom and dropped by C&G. The design featured in the RSGB’s Radcom magazine in March 1976. Back then, such a commercially made instrument was far outside the pocket of the average amateur, and as it was, it cost me £35.00 in components, which equates to £240 today. It had quite a high component count – 35 ICs, 33 caps, 36 resistors. It has a seven-way push button switch bank, and 5-pole 11 way rotary switch. The mains transformer had three windings of 9V, 6.3V and 250V to supply 250V HT for the nixies, +5V and -6V stabilsed for the ICs via the PSU.
I built it to satisfy the terms of the Amater Radio Licence at that time (then administered by the GPO). Your station was periodically inspected back then, and you had to demonstarate to an inspector that you had the means and the abilty to measure the frequency of any signals you transmitted. This counter satisfied that requirment. (There has been no such requirement or routine inspections by Ofcom or its predecessor, the Radiocommunications Agency for many years now).
Although the counter will measure up to 200 MHz, unusually, it was built on strip-board – not usually associated with high frequency applications. The secret was the input circuitry. The HF input circuitry was build ‘dead bug style’ with the components wired directly to each other, including the two input ICs - long since obsolete - (one - an NE529K - has a circular 10-pin base like a large transistor) with wiring as short as possible. One IC divided by ten, the other by four, so whatever the RF input, the output to the rest of the counter was divided by forty. Thus, if 200 MHz was input, by the time it was passed to the rest of the counter, it would be down to 5 MHz. A further precaution was the removal of all the unused copper tracks of the strip-board – a laborious delicate task, but it had to be done.
The counter used five small nixie tubes, but would display more than five digits of a particular frequency. For example, on one of the attached pics the counter can be seen displaying 100MHz from my Heathkit RF1-U signal generator, showing two decimal places, giving a readout of 100.00. However, if the display is switched to KHz, the counter will then display the KHz beyond the decimal point to five digits, so will read the frequency to more than just two decimal places, and would read to within 10Hz (Albeit within the limitations of accuracy of the 1 MHz crystal in the timebase).
The nixie tubes had short leads intended to be plugged into sockets rather like valve pins. Longer leads had to be soldered to 12 of the 14 pins on each tube, taking care not to fracture the glass seal by excessive heat. So, 12 soldered joints to each tube, and 12 soldered joints at the other end to each of the five boards which drive the display. That’s 120 soldered joints for the five tubes alone! Each of the five boards had 22 wire links, so that’s another 44 soldered joints per display board - 220 joints, and 46 IC pins for the three ICs on each display board - another 230 joints. Hence, just the five display boards had almost 600 soldered connections.
The ‘clock board’ has seven ICs which divide the 1MHz signal from the crystal oscillator to produce 100kHz, 10kHz, 1kHz, 100Hz, 10Hz and 1 Hz. There are another six ICs on the main control board. All in all, I estimate that there are well over 1,000 soldered connections, but as each board was built separately, it could be tested and any faults highlighted.
The underside of the chassis is quite spacious and uses several construction techniques – a tagboard and tagstrip for the power supplies as well as the ‘dead bug’ input sections in two little screened boxes made from tinplate. Above the chassis, the decade dividers and other logic is on two stripboards, and there is a switch band of five Yaxley wafers, each with 11 positions to be wired.
I housed it in an aluminium case I made. It still works fine, 38 years after I built it, so my initial outlay and the effort involved has paid dividends over the years!
Time has moved on – I later built the 1 GHz ‘PW Robin’ frequency counter by the late Mike Rowe, noted for the ‘Sussex’ valve tester, but nowadays it makes little economic sense to build this sort of stuff as they’re so cheap to buy. EG. 3 GHz 10 digit LCD readout for £90.00, EG:
http://www.wsplc.com/acatalog/FC-130_Wat...unter.html
Shop bought stuff has rather take the fun out of the hobby of amateur radio, and turned it into little more than a niche of the consumer electronics market - ‘toys for boys’. Apart from which - unless you are a home-brewer - why would you want such test gear anyway, if all your stuff is ‘out of the box’ 'plug ‘n play’? It’s a mystery!
Hope this trip 'down memory lane' is of interest.
This was the first fairly complex project that I built – a counter/timer designed by two radio amateurs, G Firth, G3MFJ, and David Pratt, G3KEP who later became Chief Examiner for the C&G Radio Amateurs Exam for many years until the RAE was made more elementary by Ofcom and dropped by C&G. The design featured in the RSGB’s Radcom magazine in March 1976. Back then, such a commercially made instrument was far outside the pocket of the average amateur, and as it was, it cost me £35.00 in components, which equates to £240 today. It had quite a high component count – 35 ICs, 33 caps, 36 resistors. It has a seven-way push button switch bank, and 5-pole 11 way rotary switch. The mains transformer had three windings of 9V, 6.3V and 250V to supply 250V HT for the nixies, +5V and -6V stabilsed for the ICs via the PSU.
I built it to satisfy the terms of the Amater Radio Licence at that time (then administered by the GPO). Your station was periodically inspected back then, and you had to demonstarate to an inspector that you had the means and the abilty to measure the frequency of any signals you transmitted. This counter satisfied that requirment. (There has been no such requirement or routine inspections by Ofcom or its predecessor, the Radiocommunications Agency for many years now).
Although the counter will measure up to 200 MHz, unusually, it was built on strip-board – not usually associated with high frequency applications. The secret was the input circuitry. The HF input circuitry was build ‘dead bug style’ with the components wired directly to each other, including the two input ICs - long since obsolete - (one - an NE529K - has a circular 10-pin base like a large transistor) with wiring as short as possible. One IC divided by ten, the other by four, so whatever the RF input, the output to the rest of the counter was divided by forty. Thus, if 200 MHz was input, by the time it was passed to the rest of the counter, it would be down to 5 MHz. A further precaution was the removal of all the unused copper tracks of the strip-board – a laborious delicate task, but it had to be done.
The counter used five small nixie tubes, but would display more than five digits of a particular frequency. For example, on one of the attached pics the counter can be seen displaying 100MHz from my Heathkit RF1-U signal generator, showing two decimal places, giving a readout of 100.00. However, if the display is switched to KHz, the counter will then display the KHz beyond the decimal point to five digits, so will read the frequency to more than just two decimal places, and would read to within 10Hz (Albeit within the limitations of accuracy of the 1 MHz crystal in the timebase).
The nixie tubes had short leads intended to be plugged into sockets rather like valve pins. Longer leads had to be soldered to 12 of the 14 pins on each tube, taking care not to fracture the glass seal by excessive heat. So, 12 soldered joints to each tube, and 12 soldered joints at the other end to each of the five boards which drive the display. That’s 120 soldered joints for the five tubes alone! Each of the five boards had 22 wire links, so that’s another 44 soldered joints per display board - 220 joints, and 46 IC pins for the three ICs on each display board - another 230 joints. Hence, just the five display boards had almost 600 soldered connections.
The ‘clock board’ has seven ICs which divide the 1MHz signal from the crystal oscillator to produce 100kHz, 10kHz, 1kHz, 100Hz, 10Hz and 1 Hz. There are another six ICs on the main control board. All in all, I estimate that there are well over 1,000 soldered connections, but as each board was built separately, it could be tested and any faults highlighted.
The underside of the chassis is quite spacious and uses several construction techniques – a tagboard and tagstrip for the power supplies as well as the ‘dead bug’ input sections in two little screened boxes made from tinplate. Above the chassis, the decade dividers and other logic is on two stripboards, and there is a switch band of five Yaxley wafers, each with 11 positions to be wired.
I housed it in an aluminium case I made. It still works fine, 38 years after I built it, so my initial outlay and the effort involved has paid dividends over the years!
Time has moved on – I later built the 1 GHz ‘PW Robin’ frequency counter by the late Mike Rowe, noted for the ‘Sussex’ valve tester, but nowadays it makes little economic sense to build this sort of stuff as they’re so cheap to buy. EG. 3 GHz 10 digit LCD readout for £90.00, EG:
http://www.wsplc.com/acatalog/FC-130_Wat...unter.html
Shop bought stuff has rather take the fun out of the hobby of amateur radio, and turned it into little more than a niche of the consumer electronics market - ‘toys for boys’. Apart from which - unless you are a home-brewer - why would you want such test gear anyway, if all your stuff is ‘out of the box’ 'plug ‘n play’? It’s a mystery!
Hope this trip 'down memory lane' is of interest.
{Quote: "£240 in today's money": ouch!}.