[Elecraft] new mini-modules: applications

[Geoffrey Mackenzie-Kennedy]

Evening Kevin,

A step attenuator capable of adding up to140db or so of loss is certainly 
very useful when working with receivers, but would be quite expensive I 
believe. If two step attenuators were used in cascade they would have to be 
built and connected in such a way that ensures that the attenuation in any 
sneak path around the attenuators at all working frequencies is >> 140db, 
and that the switched pads that make up the attenuators do not talk to one 
another via spurious paths.

But even with a perfect step attenuator, the crosstalk between the coax 
cables connected to the attenuator must be considered. The leakage from coax 
such as RG-174 or RG-58 would be enough I think to provide a significant 
sneak path around a 140db attenuator, adding measurement error particularly 
so if the cables are physically close to one another. Double screened coax 
helps to reduce this problem. Leakage from the signal generator must also be 
considered.

To get around these problems but still have a useful step attenuator for 
single frequency tests, I use a compromise arrangement consisting of a 90db 
step attenuator and two HP 11708A 30db fixed attenuators or pads with N 
connectors, in a way slightly similar to what you are suggesting. Having a 
fixed 30db pad right at the circuit / receiver under test provides a 
reasonable 50 ohm source, likewise a load for the step attenuator. The 
second 30db pad when used is placed at the generator's output. The step 
attenuator includes five 1db steps which I find to be very useful at times. 
Double screened coax all the way.

73,
Geoff
GM4ESD


----- Original Message ----- 
From: "kbgluxford" <kbgluxford at fastmail.fm>
To: "Elecraft Reflector" <elecraft at mailman.qth.net>
Sent: Thursday, September 28, 2006 8:12 AM
Subject: [Elecraft] new mini-modules: applications


> Wayne and list members,
> The AT1 step attenuator looks very interesting.  However, would we not 
> need an attenuator capable of 140 db or so for receiver testing?  Yes we 
> could cascade four of them, but this would scarcely be very cost 
> effective.  How about supplying the same circuit board and parts but with 
> a different set of resistors to enable increases in attenuation in say, 
> steps of 40 db, which would mean that we would only have to buy two kits.
>
> Best regards from Downunder
> Kevin
> VK3DAP / ZL2DAP