stepped attenuator

[onno]

new in this forum I have to ask:
is there anybody, who knows

NSF Controls (a brand in UK)

I'm in planning a BZLS and I'm on the search to get a rotary switch
with more then 24 steps.

Take this one or is it too rough ?

regards

Onno

[hifimaker]

Hi Onno,

This guy in Taiwan makes a great 24 step attenuator, with one step as mute. 23 steps is darn good. The preamp gain could be adjusted to have the 23 steps to fall within your volume sweet spot.
http://cgi.ebay.com/ws/eBayISAPI.dll...tem=3079706044

Beyond 24 steps, the only reasonable solution I'm familar with is to use a digital controller managing a set of resisters through relays. Combinations of 8 relays/resisters can provide 256 different combinations.

Just a thought,

-David

[labjr]

The best stepped attenuators I've seen are the dual mono kits from Michael Percy. www.percyaudio.com. He uses a great quality 45 position Shallco rotary switch (not made any more) and Holco resistors. Ladder L type. The signal only passes through one resistor. You specify the impedance.

Not cheap but still a good value. $275 US.

[Fox]

Hi Onno,

I recently bought a 23-step make-before-break switch at steinmusic in Germany. It costed 56 €. Sounds as a true bargain to me .

Steinmusic uses it as a ladder-type attenuator but of course it can also be used as a shunt-attenuator at the output of the BOSOZ. I have made a 2k2 shunt to lower the output impedance of the attenuator.

Hope this helps,

Fox

[Variac]

When using it as a shunt, do I remember Nelson saying that you put it between the plus and minus ? If so, whats a good size pot?

In the plans I think 5k is specified. Does that mean that if it is a shunt as I mentioned that 5k will still be the best choice?

I think some saidd this doesn't sound as good. Anyone have that opinion?

I just bought a 26 position switch. Maybe we need a group resistor buy!


Mark

[labjr]

The Kits from Percy may seem expensive, But the switches are great and include 176 Holco resistors.

[Duck-Twacy]

How about these (41 steps)

http://www.acoustic-dimension.com/at...uatorsmain.htm

http://www.acoustic-dimension.com/

Or maybe this one
http://www3.alps.co.jp/cgi-bin/WebOb...Rdr0va94w/18.2

infinite steps (and expensive, >500 euro)

[Fox]

Hi Mark,

When using it as a shunt, do I remember Nelson saying that you put it between the plus and minus ?

The option you describe is also possible but is not a shunt-type attenuator. According to my simulations in Simetrix, 1k is the maximum pot-value for this kind of attenuation. Beware, however, that at (very) low settings the signal is not very well attenuated.

A shunt-attenuator relates to a device with one fixed resistor in line with the signal and after that a resistor between the signal and ground. With a multi-step switch you can change the resistor between signal and ground.

In the plans I think 5k is specified. Does that mean that if it is a shunt as I mentioned that 5k will still be the best choice?

5k is a good value but the lower the better because the output impedance lowers accordingly. In the BOSOZ-article Nelson states that 5k is the lowest you can go. However this 5k-value relates to a ladder-type attenuator. With lower values than 5k you will loose too much gain.

I just bought a 26 position switch.

This offers more-than-enough steps. The Aleph P attenuator offers some 48dB attenuation (see the instruction manual). You can attenuate the signal in steps of 2dB and have 52dB attenuation.

Hope this helps,

Fox

[Variac]

Well, now I'm getting some good information.
OK,

1. If I put the switch between the plus and minus then at lower settings it is almost dead short, Since Nelson mentioned this I guess it doesn't hurt things? Especially the BOSOZ puts out a lot of juice, so I'm worried about hurting it.

2. 1K is the max value? so I should make my switch to mimic a 1k pot?

3. My Shallco switches have 6 poles or wafers, whatever. so my choice is to use 4 wafers only to create 2 ladder type attenuators
which would need to be (shorting) between the plus and minus outputs of two channels. You said that the max value should be 1k, so you are saying that the max output I'll get is a 1 k shorting across the outputs. The minimm output would be like a 1 ohm across them?

Or are you saying that the lowest shorting resister I can use is 1k across the outputs and everything else has to be grater than 1k, up to 5k?


OR

4. use 4 wafers to make 4 series attenuators, 2 for each channel
but make 'em 5k and each terminal gets it's own attenuator


5. I'm leaning toward option 3. That way I have a ladder switch, only one resistor in the signal path, and I can use less resistors, so better quality ones. The question is wots better : series attenuators but one for each output, or ladder type that shorts the outputs?

6. I think I'm pretty confused still.....
Mark

[Fox]

Hi Mark,

1. If I put the switch between the plus and minus then at lower settings it is almost dead short, Since Nelson mentioned this I guess it doesn't hurt things? Especially the BOSOZ puts out a lot of juice, so I'm worried about hurting it.

The dead short is not going to hurt anything. The + and - signal are 180 degrees shifted in phase and therefore cancel each other out. Beware however that due to the design of the BOSOZ the + and - signal differ in amplitude. As a result, a zero volume setting (max. attenuation) is virtually impossible.

2. 1K is the max value? so I should make my switch to mimic a 1k pot?

Yes.

3. My Shallco switches have 6 poles or wafers, whatever. so my choice is to use 4 wafers only to create 2 ladder type attenuators

When you opt for attenuation by shorting the + and - signals, you do not need a ladder-attenuator. Just put 1 resistor between the two signal-polarities.

Or are you saying that the lowest shorting resister I can use is 1k across the outputs and everything else has to be grater than 1k, up to 5k?

The highest will be an infinite resistor, which of course cannot be bought in stores . Just use air! The switch position before this infinite resistor should contain a 1k-resistor. This implies that, apart from the infinite-resistorless resistor, 1k is the highest resistor value in your attenuator.

For this type of attenuator you only need two wafers: one for each channel.

4. use 4 wafers to make 4 series attenuators, 2 for each channel

I am lost...

5. I'm leaning toward option 3. That way I have a ladder switch, only one resistor in the signal path, and I can use less resistors, so better quality ones. The question is wots better : series attenuators but one for each output, or ladder type that shorts the outputs?

Where are the trees in this forrest?

6. I think I'm pretty confused still.....

So am I This stuff can really drive you nuts. In case you want to dig deeper into this problematic, you should check the website of Goldpoint. Goldpoint offers a good explanation of series, ladder and shunt attenuators. Unfortunately, the "balanced-attenuator", which shorts the + and - signal of one channel, is not elaborated upon.

Hope this helps,

Fox