|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
I'm looking for a series step attenuator that will give the best audio results and have shortlisted the following:1) Parts Connection (Elma switch with Holco resistors)
2) Goldpoint (Elma switch with Roderstein resistors)
3) DACT - Danish Audio Connect (Elma switch with surface mount resistors of unknown brand but they said low noise metal film)They are all in the same price range. Which will give me the best sound (most neutral and detailed and lowest noise)
I know a stepped ladder design with vishays will probably be best but it's too expensive.
DACT claims surface mount resistors give better results than discrete resistors because of shorter signal path. But don't surface
mount components usually have lower quality, more variable and noisier?Please let me know your experiences.
and where can I get a program or chart to select the correct resistor values?
Shallcos too expensive. Besides a ladder attenuator is not what i'm looking for since I'm contemplating a shunt design and the attenuator I buy will only be used as the shunt resistor. The series resistance will be fixed by a single resistor.I have put the selection of resistors into a spreadsheet but I need to clean it up before posting since it's got bugs that I still can't figure out. The straight forward spreadsheet is OK but the one that needs to have iterations seems to not work perfectly for some values.
How do I post the spreadsheet here? I have no website.
Hi there,All things being the same (which they never are) I'd not recommend any Series Attenuator. The difference between Series and Ladder is not quite as large as that of Pot vs. Attenuator, but close.
I personally recommend Military/Telecom Surplus Switches, preferably using teflon pieces for making the switches but allways with sild silver "coin" Contacts. The Switch should be a 4-Deck unit.
Then simply wire your favourite Resistors in, directly. I personally use a specific type of industrial metalfilm which sound exceptionally good, better than Holco and Co, though not up with Vishay.
The result is an attenuator better than any commercial part known to me.
Later Thorsten
Thorsten,The Elma switches come in gold plated and silver contacts. The silver ones are cheaper and may be better initially but after time will they degenerate faster and therefore sound worse than the gold ones?
Which military switches and resistors do you use? Where can one get them?
I'm currently using Holcos for experimentation as they are cheap and plentiful at the neighbourhood store. Once I get the design right I will get the best possible resistors.
Have you tried a one step attenuator? ie. just one resistor for series and one for shunt calculated for your desired listening volume - no switches no wires as the resistor is soldered directly from input to output. It is amazing!!!! but of course you have no control over the volume but it's fun to try at least once.
I find the shunt design gets you to or exceeds a ladder attenuator quality without the cost. Are you familiar with this design. It's a fixed series resistor (use vishay since you only need 2) and variable shunt which uses a step attenuator to control the shunt resistance. So you always only have one resistor in the audio signal path. However I find that even though the attenuator/pot is used only as shunt resistance and is not in the signal path, it still makes a difference to the quality of the sound
Hi there,> > > The Elma switches come in gold plated and silver contacts. The silver ones are cheaper and may be better initially but after time will they degenerate faster and therefore sound worse than the gold ones? < < <
My problem with the Elma Switsches is that use some plating over Junk metal. The stuff I use has the entire contact and the solderpin as one solid piece of Silver.
> > > Which military switches and resistors do you use? Where can one get them? < < <
The Switches I use where made in the 1960's by GEC/Pleesy in the Northampton Plant in the UK. The Resistors are UK made precisition Metal Films which I do not know the Manufacturer of.
Neither Items are directly available to the public, but as far as the switches go, similar units can be found as Surplus in the US. It will take a little looking around but is well worth it. Cost for a 4-deck 30-Position for me was about $ 40....
> > > Have you tried a one step attenuator? ie. just one resistor for series and one for shunt calculated for your desired listening volume - no switches no wires as the resistor is soldered directly from input to output. < < <
As I use many sources and many different listening Levels, this is not really an option. But yes, it clears up things a little more over the Attnuator I use. Not very much though, I'd say this is a complement to the exceptional quality of the Switch I'm using.
> > > I find the shunt design gets you to or exceeds a ladder attenuator quality without the cost. Are you familiar with this design. < < <
We have tried that. Surprisngly (or actually not surprisingly) it does not give the same performance. My personal Take is that we hear the effects of the quite severe loading (low input Impedance) of the Shunt-Attenuator scheme on the preceeding stages.
Anyway, in one of the Euridice Preamp's (we build two of them recently) my Friend Jon used a Shunt Attenuator using a single 20k Vishay Series Resistors and suitable (mostly Holco) shunt Resistors. The other used a classic 100k Ladder Attenuator with Holco Resistors. All other Details of the Preamp's (Parts, Valves, output transformers and Powersupplies) where at this point identical between the two Preamp's.
The Sources where both an Audio-Synthesis DAX-2 DAC (intended to drive Audio-Synthesis own Shunt Attenuator (aka "Passion") and a Accoustic Precisition EIKOS DC-Player. Eithe runit is in the absolute Top-Class for Digital Equipment and has VERY beefy Output Stages.
Still, the Preamp with the Ladder Attenuator consitently sounded SUBSTANTIALLY better. Since then both Preamp's have been converted to Ladder Attenuation using Holco Resistors for most positiona and with Vishay's in the most often used positions.
> > > It's a fixed series resistor (use vishay since you only need 2) and variable shunt which uses a step attenuator to control the shunt resistance. So you always only have one resistor in the audio signal path. < < <
This is a very common misconception (or indeed from some manufacturers misrepresentation). The Switch and the shuntresistor allways remain in the Signal-Path (so are BTW Powersupplies and many other items usually claimed not be part of the Signal-path)....
Later Thorsten
> > > I find the shunt design gets you to or exceeds a ladder attenuator quality without the cost. Are you
familiar with this design. < < <We have tried that. Surprisngly (or actually not surprisingly) it does not give the same performance.
My personal Take is that we hear the effects of the quite severe loading (low input Impedance) of the
Shunt-Attenuator scheme on the preceeding stages.Anyway, in one of the Euridice Preamp's (we build two of them recently) my Friend Jon used a Shunt
Attenuator using a single 20k Vishay Series Resistors and suitable (mostly Holco) shunt Resistors. The
other used a classic 100k Ladder Attenuator with Holco Resistors. All other Details of the Preamp's
(Parts, Valves, output transformers and Powersupplies) where at this point identical between the two
Preamp's.Still, the Preamp with the Ladder Attenuator consitently sounded SUBSTANTIALLY better. Since
then both Preamp's have been converted to Ladder Attenuation using Holco Resistors for most
positiona and with Vishay's in the most often used positions.------------------------------------------------------------------------
A better comparison would have been to have both ladder and shunt attenuators having the same input impedances. Otherwise you wouldn't know if you are listening to impedance related distortion or other distortions of the signal path like solder junctions.
The true test is when a certain attenuation level is selected, both the series and shunt resistance of the ladder and shunt designs have the same values, then any difference in sound quality would be due to everything else except impedance matching.
That's exactly what I did in a few scenarios using exactly the same series and shunt resistor values. The results in order of quality:
1) just 2 resistors with no switches - best result
2) Shunt design with one shunt resistor only for each selection and ladder design - the same sound - actually tantamounts to the same design if viewed statically in this one controlled case
3) Shunt design with shunt using series step attenuator - in the test case the shunt comprised a series of ten resistors adding up to the same resistance value as the above cases)
4) Series Step attenuator in normal cofiguration (like regular pots)
5) Shunt Design with Conductive plastic pot as shunt
6) Conductive plastic pot in regular configuration.
The point I was trying to make was referring to (2) where you can get the same (sometimes better) sound from a shunt design compared to a ladder attenuator but at a lower cost since you use half the amount of resistors and usually have less solder junctions that the signal has to pass through
I have two Elma HT-series mono attenuators. Can you explain how I can use these in a shunt design with a Vishay resistor? What values do I need? I am a bit lost. Greg
Sorry Greg read your post wrongly the first time. To start you can use Vishays (series resistor) with the same value as your attenuator.However if you are planning to use it as a passive, then your Elma HT attenuator should not be higher than 10k ohm to start otherwise your output impedance might be too high. What's the output impedance of your source and input impedance of your amp?
The shunt design connection is as follows:
Input Hot----VISHAY Resistor----Output Hot
|
|
Attenuator Wiper (usually middle)Attenuator Ground
|
|
GroundThe inpot and output RCA ground sleeve is of course connected to ground.
Your attenuation range will be slightly different from using it in the traditional method.Let me know what value your attenuator is and I should be able to work out the attenuation range for you.
If you tell me the value of your Elma attenuators I can work out the values for you unless you want the formula and work the values yourself.What I need would be the Input imput impedance of the attenuator and the attenuation steps in dB and I can work from there.
I presume the Elma HT is a series attenuator and not a ladder.
I would like the formula for a ladder pot if possible.
I would like the formula. However, I have all but given up on the shunt design. I was planning on using a series Vishay or Caddock with a 25k 24 position series attenuator. I am sure this would be a good upgrade for anyone who was using just the series attenuator. But now after reviewing Thorston's comments, I have put my order on hold for the Gold Point HT-series attenuator mono kits from Arn Roatcap. I am confused by all the different comments. The Welborne Labs ladder attenuators seem to have a lot to offer. Thier contact points are gold plated silver, and seems to be well made. If I put Caddock Or Vishay's in the listening positions, I don't see how this could be beat for the price. Any comments? Thanks for your help, Greg
If you used the formula that I posted, you will note that while what Thorsten said is technically true, it is not the whole truth (don't read this as an attack on him, I've read many of his posts and have the utmost respect for him).While what Thorsten said is true about the output impedance of a shunt design at little to no attenuation, it is unlikely that you will ever use your attenuator near max setting.
What he said about input impedance can be overcome by selecting a higher series resistor of say 5k to 10k ohms depending on your equipment. You will note that input impedance of a shunt design rises with reducing attenuation (ie. louder settings) which makes it easier for the source to drive.
It is true that a shunt design has input and output impedances changing with settings, but with proper selection of series resistor and shunt resistor values you wouldn't have any problems. A ladder attenuator is usually designed so that input impedance is constant but its output impedance changes with setting as well (in the same way as a regular pot). Whether fixed input/variable output impedance or variable input/output impedance is better depends very much on how you have chosen your resistor values in either design to match your components.
A shunt design can be made using 3 methods of varying the shunt resistance.
1) A volume pot (worst)
2) A series step attenuator
3) A selector which selects only one shunt resistor each time (best)You will note that in case 3, at each vol setting there is only 2 resistors involved and one contact point. In a ladder attenuator, you also only have 2 resistors but you now have 2 contacts points. In the ladder design you also end up having to buy twice as many resistors as in Case 3.
So for a given price (or at any price if properly designed) the case 3 is a superior design. You could spend more money and improve on the design by using switching relays instead of a mechanical switch but the ultimate (in sound quality - you should try at least once) is to calculate you favourite volume setting and just hard solder one series resistor and one shunt resistor in place. This is my control case to which I compare all other designs to see how they degrade the sound.
Now I see your points more clearly! Yes, it is unlikely that I will ever use my attenuator near max setting, since my speakers, Aria5/Raven1's are very sensitive, 91.5db. A selector which selects only one shunt resistor each time so that, at each vol setting there is only 2 resistors involved and one contact point, seems to be the best sounding and least expensive control system, that I have ran across. I would need dual selectors for a balance control set up. I am trying to think of some great selector switches to use, can you reccomend some? Also, the formula you provided gives me a headache,> grin <, could you provide the values needed for the shunt resistors? Right now, I am using a Linn Linto phono preamp which has an output impedance of 100ohms. The solid state poweramp has an imput impedance of aprox. 40 or 50k, I would guess. When I switch to a tube poweramp, I guess I could replace the resistors. Are you planning on using a selector switch too? Thanks for your help, Greg
One thing to note on the shunt design that selects only one shunt resistor each time. You MUST get a selector switch which makes the next contact before it breaks the previous (can't remember the technical term for it). If the selector switch breaks the shunt resistor contact before it makes the next contact then for that split second or more you will have NO attenuation ie. FULL volume....which might blow your speakers....so becareful!
I don't know.I use 103db/W/m sensitive Horns and an SE Amplifier. As I listen on many different Times (between quiet late night and afternoons blasting it out) and with many different sources (which have different sensitivity) I use most of the range of my attenuator.... (it has > 60db).
My friend in who'se system we tested different attenuators has similar habits.
So yes. if you ever need two or three Volume Settings, never canhge the source and so on, it won't matter if you use a shunt or ladder design.
But then, you could do without a Volume control anyway....
If you don't need much range, go for Mercury wetted Reed-Relais with Vishay Resistors in Shunt Configuration....
If you need much range, go for the Welborn Kit, start with cheaper (Holco) resistors, not the positions you use most often and then replace the resistors there with Vishay....
Me? I need my Ladder, and so does (IMHO) anyone else who needs a Volume Control with more than 6 Steps....
Later Thorsten
One! I'm getting by fine with resistors on my poweramp inputs, but this is starting to get a bit old. Where could I find mercury wetted Reed-Relays the USA? I have never heard of these. I think what I want are two 24 position selector switches. Greg
Hi there,I can't help you with US Sources. I live in Europe. Go looking at the usual Places (Mouser, Digi-Key, Allied)....
Alternatively, have a look at:
www.steinmusic.de
They are still working on the English pages, but just e-mail Holger Stein (Address somewhere on above pages)....
If you want 24-position, go for it. As said, for me and my own needs I need a REAL Volume Control. If you can get by without one, go for it....
BTW, the TOTAL BUDGET TRICK is to use a chap silver contact 12-Position switch (make befoire break) and to fit one position with the chap Alp's Pot for continous Volume Control for "uncritical Listening" and the other positions with Resistors for your favourite listening Levels....
Yup, many ways to get there. For me the Issue was to make something once and forget about it afterwards.... One thing I know, my preamp has the best Volume-Control possible, my Girlfriend can use it to watch TV and that's that.....
To me all that matters....
Later Thorsten
Thanks Thorsten, Your ideas make a lot of sense. Reading your post's kind of brings me back down to earth. Sometimes one can get so caught up in Tweeking, that they can't see the forest, for all the trees. I think the most sensible thing to do, is to buy two mono kits from Welborne, put Vishays in the listening positions, and just forget about all of this and listen to the music. Since I started this quest, I don't have much time for enjoying music. Thanks for you advice, Greg
Relevant Formulas for an attenuator:ATTENUATION:
dB=20*LOG10(shunt resistance /(shunt resistance + series resistance))INPUT IMPEDANCE
= Series resistance + Shunt resistanceOUTPUT IMPEDANCE (simplified estimate)
= Series Resistance / (attenuation in dB / 3)NOTE
----
Series Resistance is resistance between hot input and hot output RCAs
Shunt Resistance is resistance between hot output RCA and ground
Put this into a spreadsheet and you will be able to look at the relationship between the input and output impedances vs attenuation for various designs. Don't worry about the theoretical max output impedance because you never turn your component that loud. Pay attention to the region where you listen to most and design a best match of your connected components for that region
DG, Please E-Mail me. I have posted 4 times to this Formula post of yours and none of them showed up! Greg xover7@aol.com
This makes the third time today I have posted this reply. They are getting lost or misplaced... I am trying to think of some great selector switches to use, can you reccomend some? I don't know exactly what type to buy. I am looking at the Welbornes and think I need a 24 position switch. Also, the formula you provided gives me a headache,> grin <, could you provide the values needed for the shunt resistors? Right now, I am using a Linn Linto phono preamp which has an output impedance of 100ohms. The solid state poweramp has an imput impedance of aprox. 40 or 50k, I would guess. When I switch to a tube poweramp, I guess I could replace the resistors. Are you planning on using a selector switch too? E-Mail me if you can. Thanks for your help, Greg xover7@aol.com
Greg, below is a suggested configuration based on a 8k ohm series resistor and a 10k ohm DACT step attenuator as the variable shunt. If you use a different attenuator as shunt, the values below will change. Changing the series resistor within a 30% range of 10k ohm (ie. 7k to 13k ohm) is OK. But you should try to match the value of the attenuator that you use for shunt to the series resistor so that you get nice smooth attenuation steps.Series Shunt ATT Input Output
Res Res (db) Imp Imp
8000 0 -inf 8,000 24
8000 10 -58.1 8,010 413
8000 32 -48.1 8,032 499
8000 50 -44.1 8,050 544
8000 79 -40.1 8,079 598
8000 126 -36.2 8,126 663
8000 200 -32.3 8,200 744
8000 251 -30.3 8,251 791
8000 316 -28.4 8,316 845
8000 398 -26.5 8,398 906
8000 501 -24.6 8,501 976
8000 631 -22.7 8,631 1,056
8000 794 -20.9 8,794 1,149
8000 1,000 -19.1 9,000 1,258
8000 1,259 -17.3 9,259 1,385
8000 1,585 -15.6 9,585 1,535
8000 1,995 -14.0 9,995 1,715
8000 2,512 -12.4 10,512 1,930
8000 3,162 -11.0 11,162 2,191
8000 3,981 -9.6 11,981 2,508
8000 5,000 -8.3 13,000 2,892
8000 6,309 -7.1 14,309 3,374
8000 7,941 -6.1 15,941 3,965
8000 10,000 -5.1 18,000 4,701I don't have the output voltage of your source or the input sensitivity of your amp (and wattage) and speakers so I don't know which range you will most likely use. The general rule is that input impedance should be 20 times that of output inpedance.
The above is what I worked out for my setup. Let meknow if you need more exact calculations
DG, Thanks for your help! This is a real brain teaser, and your infomation has helped me and others to understand some of the complexities involved with the shunt design. Greg
DG, I posted this earlier today, but I think it was getting lost in this thread, so hear it is again. The formula you provided gives me a headache,> grin <, could you provide the values needed for the shunt resistors? Right now, I am using a Linn Linto phono preamp which has an output impedance of 100ohms. The solid state poweramp has an imput impedance of aprox. 40 or 50k, I would guess. When I switch to a tube poweramp, I guess I could replace the resistors. What type of switch would you reccomend? Welborne has some good ones. I think I may need a 24 posistion switch, but I am not sure which one or exactly what type. Feel free to E-Mail me. Thanks for your help, Greg
Hi again,> > > A better comparison would have been to have both ladder and shunt attenuators having the same input impedances. < < <
That unfortunatly is (as you should know) impossible.
A Ladder Attenuator has a Variable Output Impedance which is maximally 1/4 of the nominal Impedance (So a 100k Attenuator will have a worst-case Z-Out of 25kOhm).
A Shunt attenuator on the other hand has a variable Input Impedance and a Variable Output Impedance which can be at the maximum equal to that of the Series Resistor.
> > > Otherwise you wouldn't know if you are listening to impedance related distortion or other distortions of the signal path like solder junctions. < < <
Well, the Impedance Levels where fairly similar. Due to less solder joints and so on, the Shunt Attenuator should have sounded better, but alas did not. To change the Shunt Attenuator to have 100k Series Resistance would have induced a massive and undue HF Rolloff.
> > > The true test is when a certain attenuation level is selected, both the series and shunt resistance of the ladder and shunt designs have the same values, then any difference in sound quality would be due to everything else except impedance matching. < < <
You don't get my point, do you. My point is that in practice, in the real world, in that place where use Equipment, that there the Ladder-Attenuator sounds superior.
If the Shunt Attenuator sounds better in your test at one specific Level Setting but the Ladder Type sounds bettear at all others, it is in practice superior. And do please allow me to note that the tests where level-matched and blind, as well as having included experienced listeners with long experience with both the System used (Eletrostatic Speakers, with Dynamic woofers below 100Hz and floor mounted 15" Subwoofers, 4-Way active using Croft modified Quad 2, Croft OTL and Rotel Monoblocks for the Subbass).
> > > That's exactly what I did in a few scenarios using exactly the same series and shunt resistor values. The results in order of quality:
The point I was trying to make was referring to (2) where you can get the same (sometimes better) sound from a shunt design compared to a ladder attenuator but at a lower cost since you use half the amount of resistors and usually have less solder junctions that the signal has to pass through < < <
That alas is unfortunatly not true. As noted. Due the fact that any Shunt Design MUST present a much lower (much harder to drive) Load to the preceeding Device, it will usually cause a much higher level distortion form the preceeding devices. So while the Shuntattenuator ALONE, might be identical or even better than the Ladder Attenuator.
But used in a conventional HiFi-System, a Ladder-Attenuator having the same worst case output Impedance will present a much higher Input Impedance and as a result the OVERALL performance of the system will be better.
Remember, Audio Design is Holistic, not Atomistic.
Later Thorsten
Hi DGThe DACT attenuator do use a good quality metal film SMD resistor. These are non magnetic and are very good. I would recommend the DACT attenuator. But, this is my opinion.
Regards,
Mogens
Hi Mogens,Do you know what make and part # the SMDs are?
Did you base your recommendation on actual comparisons of the 3 attenuators or feedback from others that have done the comparisons or just based on your good experience with the DACT compared to a normal potentiometer like say ALPS or Noble?
Any idea which resistor (other than vishay) sounds the best - holco, roderstein, caddock - or at least whattheir relative audio characters are?
Thanks
The Borbely site, (http://home.earthlink.net/~borbelyaudio/) has info on the sound on different types of resistors.
Ahhh exactly what I was looking for, but wish they had compared more resistors though. THanks
I am thinking of ordering a few Caddocks from Welborne Labs. They claim they are as good or better than Vishay! Of course, they do not sell Vishay, so I do not know what to believe. Comments anyone? I have been reading about Aunt Cory's Buffered Preamp,and "Modkateer Fever" at, Stereophile.com. See archive. It has a lot of good tips for construction, like star grounding and floating grounds wraped around the hot signal wires. I will have my preamp finished in about 1 week. Thanks to everyone for thier help! Greg
Hi DGI do not know the make and part#, but perhaps I can find out. At the moment I only know that they are manufactored in Germany.
I have not heard the two other attenuators and I have never used an ALPS for that matter. But, compaired to Noble and Penny & Giles the DACT attenuator is better. I say this based on my experience that Noble pots tend to wear out faster than I like and starts making noise when you turn the pot, at least my 3 did, and also some of my friends reported this. Penny & Giles do not seem to have this problem, but the two channels are not tracking that good and this can be some what frustrating. Both potentiometers do not seem to be as transparent as the DACT attenuator.
In scandinavia the DACT attenuator is known under the name NLE, which is actually the initials for the name of the guy who designed it.
I like them because the signal path is very short (17.5 cm) and the two channels is very similar which is more than you can say about most potentiometers. They have performed very consistent over the years. I got my first NLE (DACT) attenuator in may 1994.
On the matter of resistors I can only say that I use either Phillips 1% metal film or Roederstein. I have never used Vishay, Holco or Caddock. As to the sonic characteristics, I would not bet that I can tell the difference between any of these resistors. I think you should choose what you believe is the right resistor for you and just use them. Valves and semiconducters are way more unliniar than resistors.
I will bet that others will not agree with me, but this is my opinion. Use it if you can.
Regards,
Mogens
This post is made possible by the generous support of people like you and our sponsors: