Thank you Charles for your detailed explanation of your results. May I ask you was it mono or stereo setup? What genre was used? Did you try to change direction of resistor? Did you change both resistors in your L-pad or only one? Where can I find any info about these legendary Jocko Homo designs?
Excuse me for being obtrusive.
Thank you
john curl said:
Any room in the JC2 budget for a couple of input trannies? There's certainly plenty of room in the chassis. And if purchased in decent quantities, you could probably get the CineMag 15/15Bs for about $60 per pair.
se
GRollins said:
Mr Hansen,
looks like you're on a roll again, transconductance volume attenuation no less.
Mr Rollins,
pardon the coastal neighbor post, some folks are still passionate about the audio products of the Ernst Roederstein fascility in Bavaria. ERO=E-Ro=Ernst Roederstein.
(overhere, the basic number of foreign language classes in highschool is 3 to 5, optional extras)
This is enough to drive a simple fellow like me to drink.
When I win the lottery, I'm going to start a parts factory.
Grey
When I win the lottery, I'm going to start a parts factory.
Grey
Steve Eddy said:
Probably less. I scored some Jensen JT11P from a quantity purchase at something like $25 each. It's a great way to accommodate balanced- hey, you, John, and I are all agreeing about something, it MUST be right!
SY said:
From Jensen? Or from someone selling surplus? Jensen's listed 48-99 price is $51.23.
Along with other benefits of course.
😀
se
Yes, this was an OEM purchase. I don't know their quantity, and I don't know how firm Jensen's quantity pricing is.
I'm using the Jensens in two preamps now, and at the input of a power amp I built for someone else. I blame you.
I'm using the Jensens in two preamps now, and at the input of a power amp I built for someone else. I blame you.
dimitri said:
Stereo, all genres (folk, rock, classical, bluegrass -- no jazz for those tests), only used one orientation, changed both resistors together.
Jocko Homo had a company called "Analog Research". Just a one man show, but really nice and professional designs. He mostly made DAC's but when the whole hi-rez thing came along the market for those dried up. He was semi-retired for a while, but now has started a new company called Analog Research Technologies (ART).
Search the archives here for information on his designs. He also posts at:
www.diyhifi.org
jacco vermeulen said:
When you think about it, the standard preamplifier configuration of an attenuator followed by a gain stage is sort of at cross-purposes.
PS Audio makes some products with variable gain circuitry. As near as I can tell, he puts VCA chips from That Corp. inside a potted module. (I could be wrong, I often am. But the specs are suspiciously similar.)
But this got me to thinking about things. With a variable gain circuit, then there is no attenuator. This brings about several advantages:
a) The input impedance is not restricted by the value of the volume control. Once the volume control goes much beyond 10 or 20 kohms, the frequency response varies a lot with volume. For a good example of this, take a look at Stereophile's measurements of any Conrad-Johnson preamp. The input impedance is 100 kohms, meaning that the worst-case output impedance is 25 kohms. This interacts with the input capacitance of the active circuitry to make a -3 dB point somewhere around 20 kHz or so. In contrast the KX-R has an input impedance of 1 Mohm per phase and the frequency response goes out past 250 kHz regardless of volume setting.
b) The signal path is simplified and there are fewer switches in the way.
c) The signal-to-noise ratio is constant regardless of the volume setting. With a conventional design, the active circuitry outputs a constant noise voltage. Increasing the volume increases the signal, increasing the S/N ratio. The max S/N occurs at max volume. But with the variable gain transconductance volume control in the KX-R, lowering the gain also lowers the noise voltage. Since a typical listening level is something like 20 dB below full gain, this translates into an effective increase in S/N ratio of 20 dB. This is on top of an already quiet circuit and improves the resolution of low-level detail audibly.
There are some disadvantages:
1) The circuit I developed only works with a zero-feedback design. This is no problem for us as we have 15 years of experience with zero-feedback circuitry. But it might take a while for others to catch up.
2) Since the signal is not attenuated before entering the active circuitry it becomes more difficult to handle source components with extremely high outputs. We had to do a fair amount of work to ameliorate this potential problem. With a typical source of 2 to 4 volts balanced output, the distortion is around 0.001 to 0.002%. Increasing the input to 8 volts RMS (like some Theta or Wadia digital products) brings the distortion up to around 0.008 to 0.01%. This assumes that you have the gain low enough to prevent clipping. The circuit will put out around 16 or 18 Vrms balanced before the distortion starts to climb. This is plenty for any known power amp.
As you can tell, I don't like doing things the "normal" way...
Has anyone tried these? I'd love to hear impressions (as well as for the through hole versions--and how the two compare).Steve Eddy said:
I'm wondering if any of you mix different types of resistors, which could allow for a given price limit to use more expensive resistors where they'll make more impact. For example, if a resistor doesn't see significant signal-correlated current (vs the DC current through it), how could it impact sound much?
abzug said:
A reasonable assumption.
But in the crazy world of high-performance audio, there are a lot of real-life results that really don't seem reasonable. Nonetheless, the results are the results, whether or not we think they are reasonable.
For example, if you find that a certain model of resistor has certain sonic characteristics, you will find that when they are used in *digital* circuitry those same sonic characteristics are there. Makes no sense whatsoever, but there you have it.
At the other extreme are those who make claims that are obviously wrong on their face. For example, some companies make a shunt volume control, whereby there is a single series resistor and then a method to select different value shunt resistors to ground to vary the attenuation. Then they claim that since only the series resistor is "in the signal path" that they will use an expensive resistor there, but cheaper parts for the shunt resistors since they "aren't in the signal path". I really hate to disappoint them, but the shunt resistors really are in the signal path...
Yes, but I'm wondering to what extent their effect would be the same in these positions.Charles Hansen said:
For a capacitor, for example, I'd expect a shunting one in the lowpass filter after a DAC to have less effect than a DC blocking series one.
In my own DAC's attenuator, I shunted between inverted and non-inverted with Roederstein MK3, but spent more on the fewer in number series resistors, which are Caddock MK-132... whether any possible improvement by the Caddocks would be masked by the Roederstein shunts I don't know... could be worth replacing the few most frequently used positions on the shunts.
By the way, not to go to far off topic, but how exactly did you motorize your attenuators to get sufficient torque? I used a motor with a gear box and a shaft flexible just enough to allow the attenuator to snap into position as long as the motor stopped close to it (slotted wheel and simple IR pair to guarantee that). I guess one could use a stepper motor, but it looks like you have some sort of belt from the photos above.
Any reason you chose rotary switches instead of relays?
20Kohm attenuator
Charles Hansen,
"Then the hots were connected with a 10 kohm series resistor, with a 1 kohm shunt resistor to ground at the output"
At the risk of showing my complete ignorance; is there a direct connection between the ground of the input & the ground of the output?
dobias
Charles Hansen,
"Then the hots were connected with a 10 kohm series resistor, with a 1 kohm shunt resistor to ground at the output"
At the risk of showing my complete ignorance; is there a direct connection between the ground of the input & the ground of the output?
dobias
abzug said:
Both the original K-1 (released in 1996) and the new KX-R use stepper motors to move the rotary switches in fixed increments. They are connected via toothed timing belts to ensure proper synchronization. The stepper motor we use has quite a lot of torque when driven with the proper motor controller. The torque is multiplied by a factor of two or three (I don't remember exactly off the top of my head) by the pulleys.
In the older K-1, the stepper motor had to overcome the force of the detent mechanism (required for manual operation). In the new KX-R there is no manual operation possible, and instead the switches are controlled in a "fly-by-wire" fashion.
We used the Shallco rotary switches because they sound far better than any relay I have ever evaluated. I wish I could find a good sounding relay, as that would simplify (and sharply reduce the cost!) of our designs. But so far I haven't found any relays that even come close to the Shallco switches. I'm open to suggestions, although I have to warn you that one recommendation from another well-known designer wasn't even close to making the grade. He insisted that he could hear absolutely no change with the relay in the signal path. But in our *very* carefully controlled tests the degradation was quite obvious....
Re: 20Kohm attenuator
Yes. In the case of the single-ended (RCA) connectors, the rhodium-plated bodies of the Cardas male and female connectors could be soldered directly to each other.
dobias said:
Yes. In the case of the single-ended (RCA) connectors, the rhodium-plated bodies of the Cardas male and female connectors could be soldered directly to each other.
I wonder what part of the relay can cause distortion.
Silver contacts in a hermetically sealed relay shouldn't cause a problem until worn down, which is unlikely within the typical number of switch cycles that a volume control will see in its lifetime.
The frame of power relays are made of metal that can magnetize, but this is not the case in small signal relays.
The current running through the coil can induce noise, but this is not an issue with latching relays.
So what potential mechanism of distortion may be left?
Silver contacts in a hermetically sealed relay shouldn't cause a problem until worn down, which is unlikely within the typical number of switch cycles that a volume control will see in its lifetime.
The frame of power relays are made of metal that can magnetize, but this is not the case in small signal relays.
The current running through the coil can induce noise, but this is not an issue with latching relays.
So what potential mechanism of distortion may be left?
I hope nobody is reading this thread and thinking that someone has taken control of their TV set or something of that nature.
Speaking from experience, I used to think that stories of flying saucers were a bunch of BS. However, this all changed after I was abducted by aliens. Then my whole life changed, and I became much more open-minded. Just remember, in the crazy world of alien abductions, there are a lot of real-life results that really don't seem reasonable. Nonetheless, the results are the results, whether or not we think they are reasonable. I can and will attest to that, in perpetuity if need be.
By the way, I'd like to use this opportunity to plug my new book on alien abductions. It costs only $15,000 and the pages come only from actual ancient Egyptian papyrus scrolls. You might think it's silly that the book is made this way, but in the crazy world of high-end publishing, many things are still not understood - even by me.
For more information on this subject, be sure to check out my upcoming article in Weekly World DIY Audio News, in which space aliens whose foreheads look like big butts report their finding on resistors sonics.
Speaking from experience, I used to think that stories of flying saucers were a bunch of BS. However, this all changed after I was abducted by aliens. Then my whole life changed, and I became much more open-minded. Just remember, in the crazy world of alien abductions, there are a lot of real-life results that really don't seem reasonable. Nonetheless, the results are the results, whether or not we think they are reasonable. I can and will attest to that, in perpetuity if need be.
By the way, I'd like to use this opportunity to plug my new book on alien abductions. It costs only $15,000 and the pages come only from actual ancient Egyptian papyrus scrolls. You might think it's silly that the book is made this way, but in the crazy world of high-end publishing, many things are still not understood - even by me.
For more information on this subject, be sure to check out my upcoming article in Weekly World DIY Audio News, in which space aliens whose foreheads look like big butts report their finding on resistors sonics.
Charles Hansen said:
Had you tried the Seidens? Don't know about sound as I haven't compared them to the Shallcos, but they sure do look better. 😀
An externally hosted image should be here but it was not working when we last tested it.
se
Charles Hansen said:
I wonder what type of resistors were used during the creation of these recordings? How were they oriented? I wonder if any NFB was embedded? Hmmm...
cheers,
AJ
abzug said:
There are a lot of possibilities:
a) The contacts are not self-cleaning as on a rotary switch.
b) The geometry of the conductors is not optimal.
c) There are several junctions in the signal path. Some are spot welded, some are platings, some are just physical pressure. Bust open a relay and see how the signal gets from one pin to another and tell us what you see.
d) There are magnetic materials near the signal path.
In the end, it doesn't really matter "why" though. Unless you are in a position to manufacture your own relays that solve all of the problems, the only thing to do is decide if you can accept the degradation they cause.
- Status
- Not open for further replies.