I always had a soft spot for step attenuators. They are, however, notoriously expensive and my pockets are notoriously shallow!
Imagine my surprise when I was fortunate to pick up a nice unit made by Advance on eBay last year for a song…
OK – so the operating impedance is 600 Ohms – but I work in acoustics and audio and a 600 Ohm unit is still useful for AF work. What's more, it is a precision instrument and it is built like a battleship - in fact it is painted like a battleship!
Modern equipment tends to operate at higher input impedances than the “old school” 600 Ohms, so I set to work making a 10k unit, which became known affectionately as “dial-a-dB”…
Those awfully nice people at Maplin only stocked 2 pole 6-way rotary switches, so “dial-a-dB” was built in “Base 6” with six unit dB steps 0 – 5 and six dB steps 0 – 30., but it has served well in the lab. This attenuator also had a switchable internal 10k terminating resistor (just in case it was operated into a load much higher than 10k, in which case the dBs wouldn’t be as dialed).
My new-found interest in RF was crying out for a 50 Ohm version – but I was reluctant to make a copy of the “dial-a-dB” concept as the rotary switches weren’t going to work (interested readers can learn about the layout implications of step attenuators at RF by reading the ARRL Handbook (2009 edition, page 19.44) etc).
I was stuck – I am just too tight-fisted to shell out a pound a throw for some 2-pole change over switches, as I would need at least 6. Imagine my delight when I stumbled across a bag of 18 suitable push-button switches at a recent rally at the entirely reasonable price of a pound!
A moment’s work on Eagle produced a repeating pattern of the fundamental “pi” attenuator network...
You can either copy a set of resistor values from the schematic in the ARRL Handbook (op.cit.) or you can brew up your own values for custom attenuations (and operating impedances) using an online calculator.
Alan Yates has an ideal calculator on his excellent website here (look for “Resistive pads”) . I chose to build a six-section attenuator with 1, 1, 3, 5, 10 & 20 dB pads, giving every integer dB attenuation between 0 and 40. Of course the attenuation is only as good as the resistor accuracy (I chose nearest E12 value – which Alan’s calculate automatically generates) but – hey – this is only supposed to be a hobby!
I’m now the proud father of a pair of switched attenuators - all that remains is to encapsulate them inside a box with some appropriate connectors at either end.
The "dial-a-dB" name was both descriptive and alliterative - I'm looking for a similarly potent tag for the new model. Best I've come up with to date is "punch-a-pad". If you can do better, please let me know!
...-.- de m0xpd