[whubbard]

I'm going to be putting in a TKD stepped attenuator (with the Bob Crump Vishay S102 mods) into my GK-1m. The big problem is that Michael Percy doesn't have any more of the 25k (which hugh recommends). He only has 10k, 100k, and 250k. I went ahead and told him that 100k would be okay, as I figure higher is better than lower. Will there be any loss in sound detail because of the 100k, or will the GK-1 not really be able to tell much of a difference?

Hopefully someone will get back to me soon, as its a bit time sensative.

Thanks,
West

EDIT: I called Michael back up, and he was very helpful. I am going to be using S102s of 100k, instead of 250k, which he explained will bring the impedance down to around 50k, which seems like it will be fine, right? Especially considering the input on the GK is around 45k, and then my input on my ASP is probably going to be 47k (but im modding it so I can make it pretty much anything).

I was wondering however if anyone thinks putting this into a t-pad setup would have better sonics? If so, would someone mind pointing me into the right direction of doing this?...is it worth it?

[AKSA]

West,

Go LOWER, not 100K.  I'd suggest 10K.  The Zout of the SS section of the GK1 is very low, around 30R, so 10K will be fine.  25K is optimum, but 10K will be quieter than 100K.

With a 10Vpp signal, that's 5V peak, the current through a 10K pot is 5/10mA =0.5mA.  The output stage of the SS section of the GK1 runs at 12.6mA, and 0.5mA is just 4% of this, which is fine as the margin reaches up to around 25%.

Sorry for my silence;  I'm advancing as usual on four fronts.

Cheers,

Hugh

[whubbard]

Thanks Hugh.
I will go with the 10k, but still use the 100k vishays as I assume it will be pretty much the same as 250k vishays in this case.
How exactly would the sonics be different with a 20k attenuator? Lower noise floor? Noticable to a human?

-West

p.s. I totally understand you've been busy, so take your time. But when you get a chance, would you mind looking at my email from 4/20? Thanks Again.

[DSK]

West, the resistors need to be 10 to 12 times the attenuator rating, so 100k S102's for the 10k TKD are fine.

[LM]

Would anyone care to elaborate for an electronic illiterate like me on what constitutes a good value for money stepped attenuator.  For example, Theo has said a simple cheap as chips DIY Audio attenuator gave him a big improvement.  Admittedly it uses reasonable/good Vishay Dale resistors and is of ladder type construction. 

However, names like TDK, DACT and Goldpoint etc. seem to offer better potential quality but at a commensurately greater cost.   Is a ‘cheapie’ worth it over the standard pot as it seems to me that Hugh has optimised the circuitry beautifully for a simple pot?  Or should I save up for a far better implemented solution and is ladder really best  V series or shunt in the case of a GK1?
 

[whubbard]

I am certainly no expert, but here was my thinking.

I could have gotten a Shallco attenuator from Michael for $130 with all the PRP/Holco resistors. It would have been an improvement from the supplied pot. However, I would rather spend $250 (+$100 for the Vishays) and do it once, instead of getting the update bug later on. It the same way that I'm going to be putting in Sonicap Platinums. There are caps in the middle that would make a difference, but I would rather do it once, and do it right. I would assume that the difference form the supplied pot to the Shallco would be bigger than the difference from the Shallco to the TKD, but again, I wan't to do it once, and do it right.

The other fact is that the TKD + Vishays was Bob Crumps selection for the BEST attenuator. He claimed nothing was better. Therefore, it really won't be something I will ever need to repurchase.

-West

[jules]

LM,

aside from the choice of brand/type which I'm loath to offer an opinion on, the mechanical design of the switch is important. The pricey ones have high quality contacts, ball bearing shaft supports,  heavy gold plating of critical contact points, wipers with spring mechanisms that don't weaken and are guaranteed for a number of life-cycles. This means they'll be working as well in 15  or more years time as they are now.

jules 

[gerado]

At the time I got my stepped attenuator from a Sydney supplier I did not know what I was getting but for about 60 dollars I thought it was worth the gamble. I was thinking I did not want to spend near half the cost of my GK(I have the very basic unit of the GK) on the TKD.
When I installed it I was surprised I could hear the improvement so I stuck with it. I then installed the sonicaps platinum at the 2 important spots - again the improvement was noticeable, however buck for buck the attenuator was a better buy, close to the same degree of improvement (?Change) for much much less money.

I have not gone back and taken it out and left the caps in to see how they affect the sound but sonic memory is such a deceiving thing I probably should?

Nor have I compared it to the TKD..........but I mustn't


Lyn you sound like you researched the attenuator a little and turns out it  has half decent components in it.

Theo

[andyr]

The other fact is that the TKD + Vishays was Bob Crump's selection for the BEST attenuator. He claimed nothing was better. Therefore, it really won't be something I will ever need to repurchase.

-West

Absoloootely!  aa  What satisfys (or I should say "satisfied"! ) the late, great Bob Crump is certainly good enough for me!  aa

Many thanks to DSK for researching this great tweak! 

Regards,

Andy

[stvnharr]

LM,

aside from the choice of brand/type which I'm loath to offer an opinion on, the mechanical design of the switch is important. The pricey ones have high quality contacts, ball bearing shaft supports,  heavy gold plating of critical contact points, wipers with spring mechanisms that don't weaken and are guaranteed for a number of life-cycles. This means they'll be working as well in 15  or more years time as they are now.

jules 

This is a very good point as most tend to focus on the brand of resistors used and whether the attenuator series, shunt, or ladder attenuator, rather than the switch employed.   
Elma uses 2 different 24 pole switches.  The BV series is made for attenuators and does not exhibit noise bounce like the other Elma switch.  DACT and Goldpoint both use Elma switches.  Goldpoint uses the noisy one, not sure about DACT.  Shallco and Seiden switches are more like Jules has mentioned.  For the price, TDK had better be using very good switches!

Remember, the whole audio signal goes thru both the resistive element and the switch contacts in these attenuators.  Which is the weakest link? 

[PSP]

Steve,
Your post prompted me to check the DACT site, since I have a DACT attenuator in an old Bottlehead Foreplay preamp that I might audition in the GK-1.  DACT has some very specific instructions regarding cleaning the contacts of their switches and attenuators.

See:  http://www.dact.com/html/contact_cleaning.html

Peter

[stvnharr]

Peter,
I just read the DACT cleaning page.  Interesting.
I suppose this could work, but may depend on what is causing the noise.
In my GP attenuator I have always had a bit of noise on one of the channels and not on the other.  On the noisy one it has seemed dependant on contact pressure of the 2 contact points.  At the moment it is just occassional, and not a big deal.
The BV series of Elma's come with an attached pcb.  I think Goldpoint and Dact cannot use these.  I've had a switch apart and looked at the contact points, and they look smallllllllllllllllllllllllllllllllllll lllllllll and thinnnnnnnnnnnnnnn.

But the only ways I know of to get around the switch contact issue is to use something like the Placette Audio attenuator, or GeorgeHiFi's lightspeed attenuator.  I would think either of these might be a little difficult to implement into a GK-1.

[AKSA]

Steve,

Years ago I decided that a major weak point in any preamp was the level control.

That's when I figured that something of moderate cost would be best, but with the electrical environment optimised with low source impedance and high target impedance.

That was the genesis of the GK1 back in 2001;  the other benefit of this approach was lower noise from the gain block since it's always working at full line level.

I discovered that SFernice made ceramic pots which sounded very good because the ceramic substrate was much harder and the multiple contacts thereby eliminated over the conventional carbon type.  But SFernice (a small company established in Nice, the South of France) was absorbed by Vishay, the Philly based US company which also bought Beyschlag, Roederstein, and a host of other passive component manufacturers in the last 15 years, and the ceramic pots are no more.

Dammit.

That's when I went for the Panasonic EUWM, only available in motorised form and used in my Swift.  These are very good.

I have been experimenting with shunt wire pots.  They seem very good too, and this configuration can tolerate carbon pots without reduction in sound quality.

The switched attenuators are, some say, the ultimate, but they are a mechanically complex, miniaturised component, with so many moving parts, wear and friction is an issue and they do wear out.  AND they are VERY expensive.  The cheap Asian versions from Taiwan (MY3C.com) are really only durable for a few short years, as you would expect, but they look very nice.

Transformer attenuation to me is ridiculously expensive, and I cannot see the advantage, and I suspect that RF pickup could be an issue.  Very bulky.

Cheers,

Hugh 

[jules]

I can't resist. I've no idea how much these cost but they look the part ....



the above is a Jos van Eijndhoven attenuator

and here's an Elma in an AKSA



jules

[AKSA]

Jules,

Those implementations are amongst the best I've seen.  LOVE the way the GK1 board is cut in two, then joined at the SS sections;  wonderful - and the tubes are beautifully set up, LOVE it!!

The attenuator uses relays - an option I briefly considered but dismissed for its digital complexity and potential for hum.  It's beautifully designed, I must say.........

Hugh

[jules]

Thanks for the compliment Hugh.

jules

[gerado]

Jules,
that is truly jaw dropping. What a wonderful layout.

Is that some windows at the front of the tubes to see those babies glow?

Theo

[jules]

Thanks Theo,

yes, they're windows using curved pyrex glass ... must take some photos of the outside.

jules

[LM]

Jules,

Sorry I've been away for a few days and couldn't reply till now to your earlier post but quality switching does make sense.  What does this item look like to you then http://www.acoustic-dimension.com/ (click on the 41 step attenuator menu item).  Here is a review http://www.tnt-audio.com/accessories/stepped_attn_e.html  Another item that looked interesting was here http://www.hificollective.co.uk/components/shinkoh_stepped_attenuator.html .  I find this all rather fascinating but would have a very difficult time making any sort of informed decision on what to get if it wasn't for this thread.  Perhaps Theo has the right option with a good value improvement even if it might not last forever.

[jules]

Lyn,

I'm away from home myself at the moment so there's a couple of details I can't put my hands on.

If you choose to build one it does save a little on costs. The one I chose, an Elma, is relatively easy to work on for soldering.
It's a slightly different model to the one from Percy Audio and the one you've found. I bought it direct from Elma USA. There are two layers of resistors in there [a total of 90 odd] and it can get fiddly.

The Acoustic Dimension looks great though both it and the model of Elma you've found look a little hard to work on. With 41 steps you're going to have to buy a stack of resistors and while it will give very small steps, my own experience is that 24 is more than enough.

There's quite a bit in the older Aspen posts on attenuators, including the suggestion that I might have overdone it  . A search could be fruitful though.

Theo is probably right though 

jules