Switched attenuator

[Wendymott]

Hi Peeps.. Another thread in Test Equipment re how to measure 50uV, prompted me to "blather" about the switched attenuator I made as an accessory for the Spectrum Analyser .
All resistors are SMD 0805 types and the switches are minature DPDT toggle.
I copied the values from data already published..... I just added my slant to the proceedings.
As can be seen from the photo's its pretty flat to 150Mhz, but after 175 Mhz it goes a bit iffy. (Less lossy)
The PCB artwork was generated on a previous package which is not now available, so if anyone is interested I will need to re do it on Diptrace.

There is a discrepancy with the lower values of switch position which need to be looked at.

[kellys_eye]

Just curious - why did you use those specific steps - as opposed to (say) a 1 - 2 - 2 - 5 or even 1 - 2 - 4 - 8 etc?

I started an attenuator project that used a rotary selector to step 0-128dB (1dB steps) with LCD readout based on an Elektor design that I 'must' complete some time....

[Skywave]

The sequence, in ohms, of 1,1,3,5,10,10 requires 6 switches and gives a range of 1 dB to 30 dB in 1 dB steps.

Al.

[kellys_eye]

6 switches in a 1, 2, 4, 8, 16 and 32dB setting gives 1-63dB in 1dB steps (0r to 31dB in FIVE).

Just sayin'

[Bazz4CQJ]

Interesting...but I've made a promise to my one remaining good eye that no SMD components shall cross our threshold .

B

[Wendymott]

Haa Haa Baz.... I use a bench magnifier, and also a closeup lens at times....To the others...I am quite happy with the arrangement, you have your ideas,..Al.. I wanted up to 60dB.. The nice thing is... the published info.. allows any choice.
I re checked the atten at 10 Mhz and the lower ranges were "spot on"... obviously a frequency thing..still I use the 20 + 20 +10 mostly, and they are good up to +175 Mhz

[kellys_eye]

Not poo-pooing your design/choice at all Wendy - sorry for the distraction! However I do understand that losses can be more severe in the actual switches if you don't get the right part? How much of that is true and what version/manufacturer did you choose/use?

The attenuator I have (binary-step sequence) uses 1% thru-hole 1/4 watt in deference to my eyes! and miniature double-pole relays so switch choice is pretty much irrelevant and can even be remote from the board.

The last section (64dB) made from two series-connected 32dB pads given how the higher atten ranges are more difficult to 'resistor-up'.

What I do lack, however, is the test equipment to prove its quality. Maybe on next years Christmas list :lol:

[Wendymott]

Hi Kelly's eye... Its ok.. if we post we / I always expect, or should, additional comments that add to the discussion. I did have an Adret Signal generator 0.1 - 1 Ghz... the Attenuator was a series of switched attenuators, using Dual micro reed relays..... with LOTS of compensation trimmer caps between each section. As the attenuator was way too big physically I repurposed it. I never intended this attenuator to be anything but a "go to" for my Siglent Analyser, to prevent serious overloads that can occur...I am not in the professional league by any stretch of imagination

[rjmusto]

Hi Wendy,

I’d like to make myself one of these, so may i ask if you have any Gerber files available?

Many thanks,
Ralph
EI2KU

[Wendymott]

Hi Ralph. Sorry for the delay. Is it just typical... the pcb you requested I cannot find...? What I can do.. is Re generate it for you.. but that will take a couple of hours.. So hang fire and it will be with you.

[rjmusto]

Brilliant, thanks