I've tried several different op amps and have not found enough variation in current across R27 to motivate me to change them. I'm running at just a smidgeon under 5mA. It did help to match the J112s.
Is there any sonic benefit in trying to get as close as possible to 5mA instead of being "safe" inside the given interval?
I selected a R27 that gave me just above 3mA. I have a relative large R27 (1k) which also should result in a very stable CCS (a lot of degeneration should give a good CCS according to ZM. I had this advice in my head when I "played" with the J112 CCS at a breadboard). It will be interesting to check CCS "in-circuit") when time comes!
I selected a R27 that gave me just above 3mA. I have a relative large R27 (1k) which also should result in a very stable CCS (a lot of degeneration should give a good CCS according to ZM. I had this advice in my head when I "played" with the J112 CCS at a breadboard). It will be interesting to check CCS "in-circuit") when time comes!
The way to assess "sonic benefit" is to listen. Play music through the Pearl 3 at lots of different bias current values, and write down the sonic benefit of each one. Who knows, maybe your preference will turn out to be 6mA ?! Maybe you'll like 6mA more than you like 6.5mA or 5.5mA or 5mA or 4.5mA or 4mA or 3mA. The only way to find out is to try the experiment.
In reference to some of these discrete op-amps, if they are designed with class A outputs and have their own CCS you might consider the P3's J112 CCS amount. I have some discrete op-amps based off of Nelsons XVR op-amps which the output is already class A biased at 7ma.
The discrete opamps Nelson built for the Threshold NS-10 preamp, ran their output stages at ~ 14 mA of bias current: (Vdiode / 47R).
Hello,
To my $200 ears listening is not sensitive enough to reliably identify a difference.
About the Bias Current set by Q9 and R27; I see more than a couple of possible things going on.
1)
The first and most important thing about the design is to force U1:B into Class A operation. Keep in mind that the Q9 jfet is an active device operating as a Constant Current Source. For this purpose there is no need or desire to source more current, just enough to bias U1:B into Class A operation.
2)
The 2nd thing that comes to mind is that Q9 is an active device that can and will produce distortion of its own? Looking at it, it looks much like a common Gate jfet amplifier. When Q9 draws more than the design 2-4 mA the voltage across R27 increases.
Measurements across R27 will show current / Voltage distortion if any as the bias current is adjusted.
What is the phase of the distortion? With feedback does the distortion add to or cancel distortion from other components of the circuit? Is there a sweet spot?
Thanks DT
Ok. I have not enough experience to know how much current is good to pull out from the output circuit of the opamp I am using. Then for me I felt best to follow the spec (build guide). With 220R I would get 10-12 mA with kit J112. If I had some prof. distorsion measuring equipment then interesting to find the bias where distorsion was minimal. For the riaa I just want lowest noise and lowest distorsion and then let the Iron pre / Lazy Singing Bush doing the "sound".
If my ear should decide 3 mA vs. 4 mA vs. 6 mA bias it would be a very long process. For that I would prefer some prof. measuring equipment 🙂
Would be interesting to know how much of a measurable difference there can be. Also what I should look after in the op amp data sheet to try to find a "hot spot" for the opamp I am using. I guess the 2-5 mA ensures that most opamps is ok with that and output stage is not "overloaded".
If my ear should decide 3 mA vs. 4 mA vs. 6 mA bias it would be a very long process. For that I would prefer some prof. measuring equipment 🙂
Would be interesting to know how much of a measurable difference there can be. Also what I should look after in the op amp data sheet to try to find a "hot spot" for the opamp I am using. I guess the 2-5 mA ensures that most opamps is ok with that and output stage is not "overloaded".
If the guy who designed the circuit says adjust to somewhere in the 2-5mA range, why are we all second guessing it?
Build the Pearl 3. Listen to it, see how wonderful it actually is. It’s going to work great if you never change this resistor. But tweaking it fun, so measure the current over R27. If it’s more than 5mA, make the resistor bigger. If it’s less than 2mA, make the resistor smaller. Once in range, sleep soundly at night knowing that you’ve adjusted it to the operating point that it was designed for.
😀 😀 😀
Build the Pearl 3. Listen to it, see how wonderful it actually is. It’s going to work great if you never change this resistor. But tweaking it fun, so measure the current over R27. If it’s more than 5mA, make the resistor bigger. If it’s less than 2mA, make the resistor smaller. Once in range, sleep soundly at night knowing that you’ve adjusted it to the operating point that it was designed for.
😀 😀 😀
Yes yes......it was just because if seems that people are very keen of trying to reach 5mA instead of just being secure inside the 2-5mA interval.
Mine both came in at 5.2 mA…I do have resistors that would drop them both to 3.5 mA, but is it worth it?
This may be a case of overthinking the whole thing…
This may be a case of overthinking the whole thing…
My experiments on breadboard with J112 CCS was that with 1k vs. 220R resistor the voltage drop over the resistors looked much more stable using 1k resistor vs. 220R. That could also be a design goal to have the CCS as stable as possible. Then J112 with high Idss would be best to use as those would require higher resistor values. My J112's are at 12mA Idss. That is my observations from the breadboard.
I also observed that circuit was very sensitive if my hand was close to circuit. I guess I modulated the current with 50 Hz noise as input impedance is a bit high.
I also observed that circuit was very sensitive if my hand was close to circuit. I guess I modulated the current with 50 Hz noise as input impedance is a bit high.
It’s probably less than optimal to listen to records with your hand near the circuit board…
So I’d advise against that.
So I’d advise against that.
Biggus Dickus and Incontinentia Buttocks...
Yeah, I know, nothing to do with the P3.
BUT, Monty Python released the ONLY three sided record I know about....
https://www.discogs.com/release/1042493-Monty-Python-The-Monty-Python-Matching-Tie-And-Handkerchief
It was funny....depending where you dropped the needle you were hearing a different "side"..
..OK, back to discuss bias voltages.
Yeah, I know, nothing to do with the P3.
BUT, Monty Python released the ONLY three sided record I know about....
https://www.discogs.com/release/1042493-Monty-Python-The-Monty-Python-Matching-Tie-And-Handkerchief
It was funny....depending where you dropped the needle you were hearing a different "side"..
..OK, back to discuss bias voltages.
Also probably less than optimal to listen to records with your hand near the circuit…
I like the circuit board pressed in turrets and grounding buss that is tucked in next to the corner of the chassis next to the twisted heater conductors.
Thanks DT
