Welcome to
Azur Electronics

0dB to -130dB RF Attenuator
August 2007

This is as easy as it gets! Making use of the attenuators and BNC connectors out of a scrap HP 8443A Tracking Generator Counter and fitting them into an instrument case. The hardest part was finding a SMB male to male adaptor (thanks Francois) to join 2 coax cables together. One attenuator is marked +10dB to -110db in 10db steps and the other attenuator is marked 0dB to -12dB in 1dB steps. The actual attenuation is 0dB to -130dB of course!

RF Attenuator
The 2 attenuators and 2 BNC connectors just fit into an instrument case which is of similar dimensions to the HP 8447D RF Amplifier and HP 435B Power Meter.

The frequency bandwidth of the 8443A is 100kHz to 110MHz, although these attenuators have a far wider bandwidth and have been used up to 1.3GHz.

I still need to measure the bandwidth / attenuation characteristics accurately when the necessary test equipment is repaired and calibrated. Also one of the BNC connectors includes a dc blocking capacitor, I will remove this when I find a spare BNC to SMB cable.

June 2020

Only 13 years later!  Started modifying the wiring to remove the dc blocking capacitor which is listed as J2 Capacitor Assembly HP 08553-6063 and used in both the HP 8443A and HP 8553B.

Inside the Attenuator
The two attenuators are very similar to the HP 355C 0 to 12dB in 1dB steps, labelled as 355W 08443-60002 and HP 355D 0 to 120dB in 10dB steps labelled as 355X 08443-60001.

Unfortunately the coax connection from the Input to the 355X is intermittant. This cable was replaced with a SMC - SMB Adaptor carefully soldered in and a SMB to BNC cable. This removes the SMB male to male adaptor and SMB to BNC cable.

Replacement Connector
New Wiring

The dc blocking capacitor was replaced with a BNC to BNC panel mounting adaptor and a BNC to SMC connector.  This dc blocking capacitor will be evaluated for other use.

SMC connectors are threaded and a smaller version of SMA. SMB connectors are push-on.

Performance testing shows that the frequency response is fairly flat up to 1GHz. Although considered to be tested ok, there are losses associated with the adaptors, cabling and connectors.  It is still a useful attenuator up to 1GHz, but the actual attenuation needs to be checked at the desired frequency before definitive measurements are made.