Dact Attenuator

[Tubetec]

I just spotted these on ebay , a regular 22mm pot shaft and housing , on the inside a series of discrete resistors instead of the usual track .
Dual wafer , which I presume means stereo , might be interesting to create a 'special' dual pot with different values , theres certain circuits that use that to good advantage by having one pot control the feedback to the lower end of the grid resistor while the other forms a build out resistor in series with the input to prevent the previous stage from seeing too low a load when feedback is high .

No data found about db per step etc .

 

[mjrippe]

As far as I can tell these are meant to replace pots, not balanced attenuators, so there is no dB per step. I have seen them in linear and log at many common values. They seem to be modular, so I would think you could order two of each value/taper and combine the parts to make two with the values/tapers you need.

 

[TwentyTrees]

Yeah, I've used a couple of these now, including as main attenuator on my passive monitor controller. The 10k log ones I've used have been tightly matched on both channels (<0.1dB), but I don't know how robust they are - so I bought several replacements! Steps are a little random, but usable. Overall, on the face of it a good cheap way to get stepped log ladders in a compact package.

I've not seen linear ones when I've been looking - do you have a link, mjrippe?

Step values below measured between 100R DAC output and 10k ADC input.

 

[Michael Tibes]

Does anyone know a shop / manufacturer of these that could deliver 'unusual' types - reverse log for example? Or even individually specified ones?

Michael

 

[JohnRoberts]

I vaguely recall Alps making a premium attenuator using SMD resistors years before SMD became mainstream. I never saw any in the wild.

JR

 

[rock soderstrom]

Thanks for posting. This is exactly what I need right now. I will order a few.

 

[mjrippe]

I've not seen linear ones when I've been looking - do you have a link, mjrippe?

Hmmm, perhaps I am mis-remembering. It was a while ago that I was looking at these but I thought that was an option. Can't seem to find any now It is sometimes possible to contact sellers in China via AliExpress or elsewhere to get quotes on custom parts.

 

[Tubetec]

Thanks for posting the extra details Twenty Trees ,
0.1 db tollerance is pretty good really , despite the slightly random nature of the steps .
The INF position should also have much better attenuation than a pot at minimum .

Just wondering , if we took one of these log pots and flipped the wafers front to back would we end up with rev log ?

 

[scott2000]

These monitor controllers use those dact type pots. Had one in the past. Very accurate in that application... Just remember the sound being too edgy for me for some reason.... Was many years ago though...probably could use more edgy now that my ears have softened things up a bit...lol

https://reverb.com/item/20662503-ro...ontroller-for-active-studio-monitors-speakers

 

[Tubetec]

Hi Scott,
Thats a decent mark up for putting a volume control and some xlr's in a hammond enclosure , I cant see any reason these matched dual track attenuators cant be used in balanced stereo config , even though the attenuation range might end up shifted slightly depending on the series resistors used .

 

[scott2000]

Yes the cost of them has risen a bit over the years it appears.

 

[Tubetec]

Back in the old days you might have a big huge Daven attenuator , imagine all the stray capacitance , especially if its high impedence , these little 24mm attenuators could be great for tube gear .

What would also be quite interesting would be figuring out what resistors you could use to slug the attenuator for different situations , different attenuation or over a different range of values .

 

[Monte McGuire]

Does anyone know a shop / manufacturer of these that could deliver 'unusual' types - reverse log for example? Or even individually specified ones?

Just make a spreadsheet and populate it with values that you can purchase and see where you end up. The biggest hassle with these ladder attenuators is finding all of the strange resistor values needed to do anything useful. A true R-2R ladder could use just one resistor value, but then you get a linear ladder, and attempts to bend it into anything else will require every value known to man, and then a few more. (N.B. a linear ladder is a good answer, but you're gonna need 12-16 stages to do anything useful and then you need calibration and logic control to make sense of it.)

Another trick might be to provide two sets of resistor pads in parallel so each "resistor" could be made up of two in parallel. At least then, you're not chasing quite that many unobtainable precision resistors. 2012 is a good size these days. Nobody really uses 3216 for precision work anymore, and 1608 is more popular, but it's getting onto the "creepy small" package size, just asking for thermal effects.

 

[wave]

Blore-Ed makes these pcbs that fit on their PZ swithces. 1206 footprints. You can use as many wafers as you want and populate them as you want.

https://www.blore-ed.com/product-page/pz-pcb
-Dave

 

[ruffrecords]

Just make a spreadsheet and populate it with values that you can purchase and see where you end up. The biggest hassle with these ladder attenuators is finding all of the strange resistor values needed to do anything useful.

There is absolutely no need to use weird value resistors. Sometimes I think this is just a marketing ploy to get people to buy them. The thing is, each step does not have to be exactly 1.0dB or 2.0dB or whatever. The important thing is that the left and right steps are identical. Here is a link to a 12 way 3dB step attenuator I designed using only E24 values. There is a table showing the actual attenuation achieved at each step. It is a simple matter to extend this technique to 24 steps.

stepped attenuator

Cheers

Ian