Hi John,
I think the problem is that you quite sensibly decided that the lowest load this Phono would ever drive would be 10K//1nF and that only when JA or other 'scope jockeys test it.
So someone tested it at 600 Ohm, for which it was not designed... I deal with such rank tomfoolery by writing 10K minimum load impedance into the Spec's and doing a Steve Jobs on anyone who decides to ignore that spec.
JA also tested the AMR PH-77 with 600 Ohm and got 3% 2nd order THD, I'm not worried (the output is directly from a Tube and not from a Follower), as my Spec says "Minimum Load Impedance: 10KOhm".
Ciao T
PS, something someone send me recently, I would like to share with you John:
Don't try to
win over
the hater's
You are not
the JERK
whisperer
I think the problem is that you quite sensibly decided that the lowest load this Phono would ever drive would be 10K//1nF and that only when JA or other 'scope jockeys test it.
So someone tested it at 600 Ohm, for which it was not designed... I deal with such rank tomfoolery by writing 10K minimum load impedance into the Spec's and doing a Steve Jobs on anyone who decides to ignore that spec.
JA also tested the AMR PH-77 with 600 Ohm and got 3% 2nd order THD, I'm not worried (the output is directly from a Tube and not from a Follower), as my Spec says "Minimum Load Impedance: 10KOhm".
Ciao T
PS, something someone send me recently, I would like to share with you John:
Don't try to
win over
the hater's
You are not
the JERK
whisperer
This is funny reasoning. I just shared the experience that I have had for 10 years with both solutions of output stages, i.e. plain opamp and buffered output. Walt Jung also described buffered link stage outputs in his articles - will you also call it differentiation and plain marketing?? I wonder why John is open to parts selection, capacitor types, input stage devices, power supply solutions, transformer types - would you call it differentiation as well? But he does not admit the possible effect of output buffering vs. simple opamp, and uses very simple reasoning - this is surprising! In case that everything except for NE5532 + uA7815/7915 is a differentiation, then I understand. You SY use tubes. Differentiation?
Thorsten, the 600 ohm plot of JC 3 in the Stereophile has shown CROSS-OVER DISTORTION (not 2nd and 3rd order THD). Maybe not everyone seen it immediately in the plot. Just for this reason I asked what is the design of the output stage, and several posts later I read it is the simple opamp. That's why I pointed to a possible problem. I would not expect a simple opamp output in any of John's design, for the stand-alone unit (not integrated inside another product).
Well, I thought that todays IC's were free of crossover distortion. That is what everybody tells me. However, this design was NEVER made for 600 ohm operation, so testing it in that way will show distortion. Everyone using the JC-3, please refrain from using less the 5K ohms as a load, 10K or more preferred.
PMA on my test bench I have an augmented HEADPHONE AMP design for Parasound consisting of an op amp and a discrete mosfet output stage. It HAS to drive 600 ohms or less, (30 ohms actually), so I am not relying on the IC itself, for optimum drive. Satisfied? (I've been playing with it for the last 6 mo or more)
Last edited:
You're getting old enough to appriciate the advances in low power circuits, that will go into portable patient monitoring, etc. You are the one who said "Universal". Don't use them for audio.
There are plenty of op-amps that will take a 16mA current source at the output, and you have your no crossover at 600 Ohms.
BTW Walt was always convinced something like the AD815 was needed for the best line stage.
For the record, looking at a number of examples of op amp loading into 600 ohms, I can't find a suitable 'improvement' without trading off another design advantage.
It is a MYTH that op amps, in the vast majority, are designed for effortless performance into a 600 ohm load. The idea of adding a discrete output stage or a high current IC buffer would complicate my JC-3 design, and take space that I didn't have. However, my choice of the OPA2134 as the output op amp still looks like the best compromise in order to have a dual jfet input general purpose op amp.
It is a MYTH that op amps, in the vast majority, are designed for effortless performance into a 600 ohm load. The idea of adding a discrete output stage or a high current IC buffer would complicate my JC-3 design, and take space that I didn't have. However, my choice of the OPA2134 as the output op amp still looks like the best compromise in order to have a dual jfet input general purpose op amp.
Last edited:
This is sometimes called a 'strawman ploy'. That is to make a false obstacle, knock it down, praise yourself, and then accuse others of not doing so. The 600 ohm load is a FALSE requirement. Why 'Stereophile' tested the JC-3 this way was consistent with their more expensive phono designs, but not really applicable for the JC-3.
The idea that an external output stage might be more tolerant of RFI is an interesting one, perhaps PMA you should tell Scott what he needs to do to improve his products.
The idea that an external output stage might be more tolerant of RFI is an interesting one, perhaps PMA you should tell Scott what he needs to do to improve his products.
It's only a myth if you make it up. Where on earth did you get that idea? Most amplifiers now are designed for rail to rail micro-power applications or 1V into 50 or 75 Ohm RF/video applications. The 24dBm into 600 Ohms is a shrinking niche market. The AD815 was designed for ADSL and 400+mA output current, this is a very specific application that coincidentally has spawned a bunch of nice line stages and headphone amps (some commercial products). Stay tuned some new stuff is coming.
Last edited:
What about driving its own low value feedback resistors, Scott? How can we get rated noise, unless we use lower value feedback resistors? It has always been a problem for me, including designing with the AD797, as I did originally. LT1028 is even worse.
For the record, I am now using 50-1600 ohms for my R1/R2 ratio on my input stage, AND it is a compromise. In fact, I lost about 3dB in noise with this combination, yet I don't dare load it with anything lower.
For the record, I am now using 50-1600 ohms for my R1/R2 ratio on my input stage, AND it is a compromise. In fact, I lost about 3dB in noise with this combination, yet I don't dare load it with anything lower.
I have finally fixed the problem. The case was UHF oscillations in the circuit application (ADA4898 + high power buffer) that occurred after clipping. It (oscillations) was not seen at the output. After change of frequency compensations this effect completely disappeared and I can confirm that to me the ADA4898 is one of the best sounding audio opamps I have ever heard.
- Status
- Not open for further replies.