Attenuators..Advice?

[Jwillis]

Here's my problem.I haven't used attenuators for some time and I want to calibrate a DolbyDX board.So I ordered a 6Db attenuator but when it arrived I found it to be DC blocking .I don't want to have to wait again for another one. I built a step attenuator but I haven't tested it for accuracy so i figure I'll just make a fixed  on.Question is since I don't have a whole lot of resistors to measure for accurate values I was wondering if I could use trimmers instead or would that just create more problems than its worth.

This is the step attenuator i built . The thing on one end is a 50 ohm terminator.

[IconicPCB]

I guess you are working in audio arena.

Why 50 ohms then?

Shouldnt 600 ohm be the golden rule?

[Jwillis]

Just need to apply a 400hz sine wave through  a 6db attenuator  which should be  0.388V at 400hz .So far the right and left biasing has come up correct.I just need to set minimum distortion  then the encoder and decoder.then I'm finished with this.50 ohms is fine for this procedure according to what I've been told from the original manufacturer . 

[BrianHG]

Take a look at this thread here: http://www.diyaudio.com/forums/pass-labs/21198-constant-impedance-relay-resistor-logarithmic-attenuator.html#post246385

The opening post explains about a relay switched attenuation with a fixed impedance with a set fixed load.
Further down the page, there is a .zip of an excel spread sheet which allows you to enter your impedance, output load, and, db attenuation wanted, and it will give you the 2 resistor dividers needed according to 1 stage of the relay switched attenuation.

Read carefully and you will be able to build exactly what you need using 1 switch, or, even none giving you always 6 db drop with your expected 50 ohm load, while the input will still appear to be a 50 ohm load.

[BrianHG]

I get the following values for R1 and R2 in the attenuation schematic.  This works assuming the output is a 50 ohm load.

R1= 24.95 ohm 0.1%
R2= 50.20 ohm 0.1%
Output Load must be = 50.00 ohm.
Input impedance of this attenuator will be = 50.00 ohm.
Attenuation = -6.00db.

(It's been 15 years, I hope I got it right...)

You may need to construct R1 and R2 from multiple 0.1% resistors.  Search around in this forum, if I remember there was a spreadsheet which calculated a wanted resistor value from parallel/serial wired available precision resistors.

For standard off the shelf values, for 50.2ohm, maybe use 51 ohm with 3.24k in parallel.
For 24.95 ohm, maybe use 24.9 ohm.

24.9 ohm in 1% already exists.
51 ohm in 1% already exists
3.24k in 1% already exists.

[Bassman59]

I guess you are working in audio arena.

Why 50 ohms then?

Shouldnt 600 ohm be the golden rule?

No. Impedances in audio designs are bridging, which means low-impedance outputs (< 100 ohms) and moderate impedance inputs (pro line level is generally a 10k input impedance).

[Jwillis]

OH ! I didn't even think to use a "L" pad .I was making Pi pads.That would make things much easy.But back to the original question .If I'm short on suitable resistors can I use trimmer pots instead.The closest electronics store is 700km away so popping down to the local Source is ....difficult.

[BrianHG]

You can use trimpots, but, they are noisy, expect them to drift, and they are very low power, usually.  I would recomend using some close, but slightly larger resistors and put a trim pot ontop to finely zoom in on the value you want.  IE, use 51 ohms resistors in my example with a 5k or 10k trimpot ontop to adjust it to the 50.2 ohm target.

If you are fortunate enough to find old busted VCRs, DVD players, TVs, you can salvage many of these values, especially the ones I mentioned above, from them.

[IconicPCB]

Some more options:

With the constant impedance switcable section attenuators you have a choice of T or Pi sections.

I would recomend you set up a spread sheet which would evaluate the two forms for the range of attenuation factors You are looking at.

The spreadsheet should also be able to evaluate attenuation and impedance given the closest available resistor to hand.

Try to match the impedance as best you can and settle for what ever attenuation you can get.

If for exam[le you need say 10dB @50 ohm and you come up with a resistance combo of 51 ohm 9.5dB and 52 ohm 10dB choose the 51 ohm option it will provide a better outcome overall when daisy chaing other switched in sections.

Yes it wont be 10dB but at least you know it will be 9.5dB at a lower  mismatch with the follwoing section(s).

Dont worry if you have to use a combination of a couple of resistors to get to where you need to be.