Stepped attenuator

[Bas Horneman]

I'm going to build a Cleo V for a customer.
http://www.triodedick.com/cleo_5_deel_1.htm

I have offered a noble pot (You know the ones by Bill Fitzpatrick) with the detents. But was thinking...if the customer gets a DACT or other brand stepped attenuator...what kind of gains wil he get...in other words .......

is a stepped attenuator really a step up......?!?

Does anyone have any experience in upgrading to a stepped attenuator? And which one is best for the money?

Asking here because the premium is very high..DACT = 180$,
Noble is 25$

Cheers,
Bas

 

[Eli Duttman]

Bas,

The noise factor of the metal chip resistors DACT uses is superior to conductive plastic. However, the Noble detented controls are NOT junk.

FWIW, Gordon Rankin prefers Penny & Giles conductive plastic to stepped attenuators. P&G pots. carry a hefty price tag.

 

[ABO]

Hi Bas,

I just recently changed from ALPS bluepot (I now call it the Blue Basstard) to a stepped attenuator from Schuro (Elma switch with Dale resistances).

The increase in low-level detail was astonishing. It changed my sound from flat to 3D. Basses were somehow deeper, background- and dubbed vocal can now be easily followed, voices sound more natural.

I can only say that is a major step up. Much more than I had expected. Now, when I see a picture of a DIYer using a bluepot, I can't help to feel sorry for him, especially when he's build an otherwise perfect amp (like an expensive Aleph).

Hope this helps.





Beware: This only a comparison between ALPS bluepot against Schuro stepped attenuator kitset.

 

[Bas Horneman]

The noise factor of the metal chip resistors DACT uses is superior to conductive plastic. However, the Noble detented controls are NOT junk.

Thanks Eli

The increase in low-level detail was astonishing.

That confirms what others have told me! Thanks.

 

[Ralph]

Same experiences with a blue Alps to a DACT. Very much worth the extra $$.

 

[Bas Horneman]

Thanks Ralph. I appreciate all the input by people that have actually upgraded their pots. So much more than the marketing.

Cheers,
Bas

 

[fdegrove]

Hi,

Bas,

Since this is for a valve circuit I'd recommend one of the series type attenuators as these provide a constant impedance to the circuit.
This can be critical if the designer had this in mind and chose the value of the potentiometer accordingly (value of g' in formulae).

If you want to avoid putting all resistors in series as you attenuate further down (lumped Johnson noise), add the R values of such series attenuator as you'd move position (R2' = R1 +R2, R3' = R1+R2+R3 and so on) and use a double deck switch.
Most likely you'll need the E96 series range of 1% resistor values for a precision steps of 2-3 dB/step.

By doing so you'll never have more than one resistor in the signal path but instead you'll have a second switching contact.
Given the high contact integraty of the ELMA, Shallco and Seiden switches I have not noticed any degradation of sonic quality because of that.

If money's no object I'd choose the naked Vishay bulkfoils, followed by Dale, Caddock, Beyschlag (Dutch ), Resista, old style Holco...plain good quality carbon films work surprisingly well too.

The shunt part resistors aren't really an issue here, use decent MF and you're done.

Also keep in mind that those switches and resistors should not carry more than say, 20 VDC for highest reliability.

On a final note, in real life and with my system, I hardly ever use more than 3-4 positions. So you know where I spend the hard earned Euros.

Use of Ag doped silver (2%, preferably 4%) is highly recommended especially since the solder lugs of the switches are goldflashed in case you use the hardwire version, not the PCB. Besides the better sonics it's a doddle to use too.

Naturally, the above description applies to a single channel only.

Cheers,

 

[stvnharr]

I have just built a Goldpoint Shunt attenuator, with Vishay S102 as series resistor and Dales as shunt resistors.
No pot can come close to this!!
No need for a Dact, as there are kits from Percy and others that cost way less.

 

[Peter Daniel]

The shunt part resistors aren't really an issue here, use decent MF and you're done.

Don't underestimate those. In my tests they were as important as series resistors (if not more).

 

[DrDeville]

stvnharr's flicking phalanges formulated:

No pot can come close to this!!

No comment...

Seriously--thanks for the tips. The percyaudio.com
and goldpt.com sites are groovy.

Best,

George "Cheech" Ferguson

 

[fdegrove]

Hi,

Peter,

In my tests they were as important as series resistors (if not more).

In the context of a tube preamp (high impedance) and as series type attenuator, you'll find it far less critical.
If at all audible.

For what you've tried, low impedance, shunt, I'd agree.

Cheers,

 

[Peter Daniel]

In that case his 50k Nobel might be no good at all for that preamp.

What shunt imp range are talking about here?

 

[fdegrove]

Hi,

In that case his 50k Nobel might be no good at all for that preamp.

If it's a log type pot it'll work fine but I'd still prefer a higher impedance type of say at least 100K to avoid HF roll-off.

What shunt imp range are talking about here?

The original diagram shows a 100K log pot.

Cheers,

 

[Bas Horneman]

If it's a log type pot it'll work fine but I'd still prefer a higher impedance type of say at least 100K to avoid HF roll-off.

Why would you get HF roll-off?

Cheers,
Bas

 

[DrDeville]

Why would you get HF roll-off?

I too am interested in this topic.

I have my own ideas from my studies of
inter-stage bandpass effects, but would
rather hear from someone more expert.

Best,

George "Grasshopper" Ferguson

 

[Bas Horneman]

Got a mail from Triode Dick..

It is actually the opposite...If you have a 500k and a tube with high gain...like a 12ax7 then that would cause hf rollof.

Cheers,
Bas

 

[analog_sa]

It is actually the opposite

Of course it's the opposite. The higher the resistance, the lower becomes the corner frequency of the low pass filter formed out of the pot resistance and input capacitance of next stage + parasitics. In practice it's not much of an issue unless the pot connecting wires are too long.

Choosing a low value for the pot has no hi-freq penalties but may limit both gain and linearity. Ideally the preceding stage will be very load tolerant which will allow a lowish pot value. Some find perceived dynamics to be better if the pot value is low.

 

[dhaen]

The lower the value, the better - as long as the preceeding stage can successfully drive it.

 

[fdegrove]

Hi,

It is actually the opposite...If you have a 500k and a tube with high gain...like a 12ax7 then that would cause hf rollof.

True in the case of high gain triodes and high Rg. (Miller effect)

In practice the value of the attenuator isn't too critical unless you're dealing with a serious mismatch.

The main reason I prefer to use 100K or even a little higher is that I'd rather use smaller value caps than bigger ones, plus the reasons mentioned by Analog_sa.
Slight deviations of calculated ideal resistive values also have less influence on the gain of the attenuator.

To summarize, you'd sooner have a LF roll off than a HF roll off most of the time.

Some find perceived dynamics to be better if the pot value is low.

I noticed the same but I'd rather use lowish (50K iso 100K for instance) bleeder resistors at the output to obtain the same effect as quite often the volpot/attenuator doubles as rg'.

For those of you wanting to toy with a 12AX7A and see what the Zout of the previous stage and some stray capacitance can do there are some nice examples to be found at this site:

AIKEN AMPS.

Digging a little deeper and playing with Zout, Rl and C values it's not hard to see where things can get really awkward.

Cheers,

 

[DrDeville]

To Quote Bill Shatner

"But why?

I hope you folks will wrestle with this until
you identify the important factors.

Here's a webpage from Goldpoint:

http://www.goldpt.com/vcb_sa4.html

At the bottom it states that higher impedence
potentiometers are better before tube amps,
because of their higher input impedence.
But it doesn't explain... why?

Best,

George "Darn It Spock, I Want answers!" Ferguson