So the remaining issue is the servo. We have to look into this ones more.
analog-sa, when you short the input of the Paradise : how is the noise then ?
As i reported many times : the Paradise has approximately the noise of a 10 Ohm resistor.
Connecting the DL103 that has a DC impedance of 40 Ohm worsens the noise 4 fold.
That is 12dB. That has nothing to do with the Paradise and is simply the Johnson ( thermal ) noise of the cartridge coil windings.
analog-sa, when you short the input of the Paradise : how is the noise then ?
As i reported many times : the Paradise has approximately the noise of a 10 Ohm resistor.
Connecting the DL103 that has a DC impedance of 40 Ohm worsens the noise 4 fold.
That is 12dB. That has nothing to do with the Paradise and is simply the Johnson ( thermal ) noise of the cartridge coil windings.
By the sound of it, the decoupling analog_sa implemented is in parallel to the big elcaps (C6 and C7). Thats a good one, thanks for the tip!
As Joachim points out, Paradise is operating at very low noise levels so you may end up seeing some low-frequency output signals that are noise-related. Bipolar transistors don't have a lot of 1/f noise (unlike JFETs), so they have the edge here. Some of that low-frequency output signal can also be caused by thermal variations (try blowing on the input stage transistors.....)
From a "sound" point of view, this decoupling might better be done with foilcaps rather than ceramic caps, but that is pure guessing on my behalf.....
Others may not have a problem. Just stripped a bit of the screen close to the 33ohm junction and soldered a 10n smd between the junction and ground. On both rails of course.
Btw, RC, how stable is your offset?
Btw, RC, how stable is your offset?
It is not fully encased. The dc drift has a frequency in the region of 1-2Hz; isn't that too fast for thermal wandering? Approximately 120mV of wandering with servo on and 2v without the servo.
Apparently my beta matching is not so great as there is a constant disbalance around +2v at output. At max it goes up to maybe +4v, so not using the servo effectively limits the max output swing on one polarity - not so great.
Haven't made a final judgement about the servo yet. Maybe speeding it up will cure the drift. Will do some more listening with and without the servo but if output offset cannot get below a few mV there will have to be an output cap anyway.
How was the dc stability in the beta run?
Good point. Will try a Kontrapunkt B as well. I think coil resistance is around 7ohm.
You could also try to degenerate the positive input transistors some more so raise the emitter resistors somewhat in value until the +2V is cancelled.
You could also try this on the mirror emitter resistors but this part of the topology is extremely sensitive to even very small fluctuations of the resistor values. I think it is better to tune in the input stage.
You could also try this on the mirror emitter resistors but this part of the topology is extremely sensitive to even very small fluctuations of the resistor values. I think it is better to tune in the input stage.
Instead of raising the positive side it could be easier to lower the emitter resistors on the negative side because then you do not have have to solder the resistors out but just parallel from the top.
I am tempted to glue a piece of thin copper foil on the input transistors and see if a more even temperature spread will have an effect.
Sorry for misleading everyone this but the oscillation issue remains. It behaved fine with an open input but as soon as the input is short circuited it goes bonkers.
DC condition at output is probably related. Will proceed with Hesener's suggestion to remove the helper transistors from the current mirrors
DC condition at output is probably related. Will proceed with Hesener's suggestion to remove the helper transistors from the current mirrors
Put a cap of ca. 1nF parallel to the input. That will of cause not help when the input is short circuited. I have more ideas.
1-2Hz is OK for small thermal masses (like in this case); for power transistors it would be a very different story.....
It may not be HFE mismatch but base current mismatch. Changing offset voltage with different input impedance should reveal this... maybe its just one "bad apple"....
Rock solid (in both my R1 and R2 builds). With your oscilloscope, can you see oscillation in the 40-50MHz range? (just checking....) Please also check the supply rails (AC coupling, of course)
Not need for apolologies this is Diy so time and waste of it are part of funn.
As on PM my batch of transistors from Fairchild mesure prety Hi (I have not oppened goodies bag so far)
I be happy to help as I have no space for 1800 transistor to sit on bench gatering dust.
But IMO start posting what you need on the tread and start swaps
There may be somebody near you with oposite problem as you have
As on PM my batch of transistors from Fairchild mesure prety Hi (I have not oppened goodies bag so far)
I be happy to help as I have no space for 1800 transistor to sit on bench gatering dust.
But IMO start posting what you need on the tread and start swaps
There may be somebody near you with oposite problem as you have
Using an unexpected day off I started stuffing my boards ...
(GB parts missing come from my parts Box and Reichelt (C's mainly))
However:
When I came to the the Heatsink I noticed that I have no more Isolation Pads...
...and the Manual only says "remember to put isolation if needed" <grin>
Question: Do I need to isolate the Power Transistors on the Heatsink?
(sry for beeing lazy 😉 but I am in for a late lunch break and hope the answer is ready when I am finished ... while the answer may be useful for all of us still soldering.)
(GB parts missing come from my parts Box and Reichelt (C's mainly))
However:
When I came to the the Heatsink I noticed that I have no more Isolation Pads...
...and the Manual only says "remember to put isolation if needed" <grin>
Question: Do I need to isolate the Power Transistors on the Heatsink?
(sry for beeing lazy 😉 but I am in for a late lunch break and hope the answer is ready when I am finished ... while the answer may be useful for all of us still soldering.)
