People who can design, study what they made, measure it, listen to it, modify the circuit and retest,trying to improve it, and often they do. 
People who can not, keep old design (they have no other option) but roll parts (including resistors) and self convince they "improved something".
A.k.a. "Placebo effect".
People who can not, keep old design (they have no other option) but roll parts (including resistors) and self convince they "improved something".
A.k.a. "Placebo effect".
Correct. Many seem to feel they’re superior, I’m not sure where the misinformation came from. From Vishays documentation: (Edit: copied and pasted wrong info)
http://www.vishaypg.com/docs/63520/FACTS_122.pdf
When Small Isn’t Better in Precision Resistor Applications
See part 1 excess heat.
Last edited:
Rational arguments like this will forever hamstring your system's development
My personal justification for not having a z-foil attenuator is the presence of too many boxes, all with attenuators. As a cheap-skate solution z-foils are just the series resistors, the shunts are usually nos Holcos. Perhaps it's best to just forgo the 40+ steps and go for sensible 12 before the full scale z-foiling
Lot's of opinions, advice and experience...all good. Whilst you are going to this kind of trouble and expense, my advice is keep lead right out of the equation, ie be certain to use lead free solder. There are many lead free alloys to choose from, some with extraordinary claims and prices. I use 99.3% tin, 0.7% copper lead-free because it is locally available, economical and solders/tins well with right temp. I also have a small amount of another alloy 96% tin, 4% silver lead-free - SOLDER SILVER KIT STAY BRITE | Jaycar Electronics These alloys sound different with the silver alloy sounding brighter, the copper version sounds plain, maybe ever so slightly diminished in the top end, but pleasantly clear. Silver and lead do not play nice IMHO.....avoid
. Dan.
Last edited:
There is nothing in the vishay pdf to hang your hat on from a design perspective for this application, as no resistor would be operated anywhere other than at ambient plus about 1degC, and drift data is all about operating at the max power rating of the part.
Imho, it all falls back to personal confidence in the part not drifting, and whether the OP has a resistance measuring technique that can achieve the accuracy required for matching (and can set their calander for +10yrs to do it all again then) or just wants to live with the initial part accuracy spec. Even then, the attenuator is only one part in a channel circuit that has an influence on channel gain, and one would hope that the rest of the audio chain out to a pair of ears has been similarly calibrated (including getting an auditory response check done).
Imho, it all falls back to personal confidence in the part not drifting, and whether the OP has a resistance measuring technique that can achieve the accuracy required for matching (and can set their calander for +10yrs to do it all again then) or just wants to live with the initial part accuracy spec. Even then, the attenuator is only one part in a channel circuit that has an influence on channel gain, and one would hope that the rest of the audio chain out to a pair of ears has been similarly calibrated (including getting an auditory response check done).
Hi trobbins,
The Dale parts do not need rematching after 10 years or more. They stay put as they do not see any appreciable power. I use an HP 3457A to match resistors with a long average and Kelvin connection. The 34401A is good, but not that good compared to the 3457A.
Others may be as good, but I have very long term experience with Dale parts and also test & measurement equipment.
The other way to make accurate resistance measurements is to set the current level by measuring across a precision resistor, then also across Rx. The 34401A is good enough on DCV scales. Again, long averaging is your friend. Slow work but your results will last well beyond your lifetime. The basic DC accuracy for the 34401A freshly calibrated is about 0.05%. The new little bench tops are better. I would love to have the new little 7 1/2 digit meter. Yes, I would use the extra accuracy.
-Chris
The Dale parts do not need rematching after 10 years or more. They stay put as they do not see any appreciable power. I use an HP 3457A to match resistors with a long average and Kelvin connection. The 34401A is good, but not that good compared to the 3457A.
Others may be as good, but I have very long term experience with Dale parts and also test & measurement equipment.
The other way to make accurate resistance measurements is to set the current level by measuring across a precision resistor, then also across Rx. The 34401A is good enough on DCV scales. Again, long averaging is your friend. Slow work but your results will last well beyond your lifetime. The basic DC accuracy for the 34401A freshly calibrated is about 0.05%. The new little bench tops are better. I would love to have the new little 7 1/2 digit meter. Yes, I would use the extra accuracy.
-Chris
At the rate this is going you may never wake up ;-)
At Ligo they seem fairly content to use 1-0.1% >100ppm resistors from mouser.
Advice is to read the Ligo paper and put it all in perspective. When you compare cost it’s silly to use a Zfoil when you can use an adequately derated metal film of your choice from the list of tested resistors. The cost would be 10 cents - $2 at the higher limit per resistor, versus $25+ per resistor.
Also the bulk foil only goes up to a certain resistance, at which you have to double the cost to get an appropriate value of two resistors sandwiches. Now keep in mind you are buying 40 of them or so. So you’re talking a thousand dollar volume control. What if you also want balance control?! How that isn’t universally viewed as insane thinking is beyond me. Unless of course you are going for pride of ownership in which case I cannot help you. https://dcc.ligo.org/public/0002/T0900200/001/current_noise.pdf
“For frequencies above 1 Hz, the very expensive ultra high precision Bulk Metal[emoji768] Z- Foil and classical wirewound resistors appear to have very little or no excess noise. Wirewound resistors however, do have inductance that may affect the circuit operation. Metal film or thin film resistors have little excess noise in these tests and thick film resistors show large excess noise [26, 27]. It could be observed that the noise decreases with increasing mechanical size or volume of the same type of resistor, which is in agreement with experiments reported in the literature [4, 12, 13, 17]. For the same resistance value and current, a higher power resistor of the same style will generate less noise. Furthermore, the excess noise has a linear voltage dependence as reported in the literature [7]. The temperature dependence of the low-frequency noise has not been measured, but it can be found in the literature that the 1/f-noise in resistors can be temperature dependent [2, 6, 14]. These papers suggest that a change in temperature changes the distribution of relaxation times of the noise process due to changes of the microstructure. This is only important for large changes in temperature, not for operating conditions around room temperature.”
Also the bulk foil only goes up to a certain resistance, at which you have to double the cost to get an appropriate value of two resistors sandwiches. Now keep in mind you are buying 40 of them or so. So you’re talking a thousand dollar volume control. What if you also want balance control?! How that isn’t universally viewed as insane thinking is beyond me. Unless of course you are going for pride of ownership in which case I cannot help you. https://dcc.ligo.org/public/0002/T0900200/001/current_noise.pdf
“For frequencies above 1 Hz, the very expensive ultra high precision Bulk Metal[emoji768] Z- Foil and classical wirewound resistors appear to have very little or no excess noise. Wirewound resistors however, do have inductance that may affect the circuit operation. Metal film or thin film resistors have little excess noise in these tests and thick film resistors show large excess noise [26, 27]. It could be observed that the noise decreases with increasing mechanical size or volume of the same type of resistor, which is in agreement with experiments reported in the literature [4, 12, 13, 17]. For the same resistance value and current, a higher power resistor of the same style will generate less noise. Furthermore, the excess noise has a linear voltage dependence as reported in the literature [7]. The temperature dependence of the low-frequency noise has not been measured, but it can be found in the literature that the 1/f-noise in resistors can be temperature dependent [2, 6, 14]. These papers suggest that a change in temperature changes the distribution of relaxation times of the noise process due to changes of the microstructure. This is only important for large changes in temperature, not for operating conditions around room temperature.”
Last edited:
Hi Chris,
Absolutely not!
I was wondering what was wrong with that picture, then I realised that the reference they are using for the metal film parts are really cheap noisy ones. To make it really one-sided, the resistances in use were very low which put the cheap metal film parts at even more of a disadvantage.
You try it using the Dale RN55 series for example, then post the results.
To prove my point, look up the specs on real metal film resistors (Dale RN55 as the standard part - 'cause it is for most knowledgeable people) and whatever flavour of bulk metal foil parts you wish.
I really hate "real world tests" when the deck is stacked. All that did was waste my time once it became apparent exactly what he was comparing to. You were well aware that the metal film resistors in discussion were the Dale RN series, or equivalents. You should have picked up on that before posting the link.
-Chris
Absolutely not!
I was wondering what was wrong with that picture, then I realised that the reference they are using for the metal film parts are really cheap noisy ones. To make it really one-sided, the resistances in use were very low which put the cheap metal film parts at even more of a disadvantage.
You try it using the Dale RN55 series for example, then post the results.
To prove my point, look up the specs on real metal film resistors (Dale RN55 as the standard part - 'cause it is for most knowledgeable people) and whatever flavour of bulk metal foil parts you wish.
I really hate "real world tests" when the deck is stacked. All that did was waste my time once it became apparent exactly what he was comparing to. You were well aware that the metal film resistors in discussion were the Dale RN series, or equivalents. You should have picked up on that before posting the link.
-Chris
Somewhere here, someone did a pretty good test evaluating resistors for distortion. He posted his results here (DIYAudio) somewhere and IDed those with the lowest THD. If I find it again I'll post a link to it. It might provide you with some answers. Oh, that is it, the ligo link: Seifert, F. (2009). Resistor Current Noise Measurements, ibid.
Cheers,
Cheers,
Last edited:
The second mix I believe is noticeably louder, from a quick phone evaluation. Anyone else hear what I’m hearing? If someone wants to take a stab at analyzing the YouTube audio in software and make a real comparison that will give us actually useful information. Anyone want to take a poll?
The change in resistor likely changed the attenuation level. And louder is perceived as better.
Also, if I can hear this on my phone with its crummy speakers and resistors, through compressed YouTube audio, what does that tell you? Hint: not that foil resistors are that audibly obvious.
They also don’t tell you the measured value or wattage of the original resistor nor the value or wattage of the resistor replacing it. I don’t think it’s even remotely a resistor quality issue to be pronounced enough to hear on a dime on YouTube no matter how bad the original was.
YouTube reviews also by and large now have corporate sponsors, and it’s rarely disclosed. Want to guess how he got that big spread of foil resistors? If you look at his other reviews, he also extols the virtues of a $300 plus cable. Tell me the holes you see in this one: YouTube
ABX advocates will have an aneurysm watching that one.
Last edited:
Ladder or series (or shunt) ? It makes a difference costwise (big).
It didn't use to be so expensive to try different resistors and listen to see if you can actually hear a difference. (Surplus was everywhere) Nowadays, it's not as easy, the price spread is significant and soldering up a whole unit can cost more than you'll want to trash and try another..
If you have an interest (and if you don't why are we talking) you could set your current pot at a level you can listen to in a comfortably sustained way, Measure the resistances above and below the wiper, Then get some resistors of those values in various types (metal film, carbon film, metal foil [$] carbon comp [while you're at it]) and try them out.
Ignore the endless debates. Solder in a few and listen. If you can't hear a difference take the cheapest ones, if you can hear a difference , then take your pick!
For a time you could get a switch that would take surface mount resistors but a search just now didn't turn any up. Maybe somebody else knows. SMD R's can be a bit cheaper than through hole axials if your soldering patience is up to it.
Great advice. Sadly, most people are more inclined to go for a "proven recipe" than do any amount of experimentation.