Thanks,
Bit of a task to set up all of those cartridges accurately, listen, keep notes and publish. The Ortofon and your new cartridge alone should suffice. NOT looking for a DB test or anything of that sort...just your natural impressions. Being a technical guy you will obviously attempt to avoid the over subjective pitfalls.
Not sure what that is? Anyway, I don't necessary want to push for this power solution, when I have the known as working Jung + cap multiplier thing.
Speaking of heat, I have on the back burner a Peltier cooled BF862 stage
Great. I am planning to become annoying with that one
I already have a working prototype; the problem I still have to find a final solution for, is sealing the thing. Got some suggestions, working on it.
As in putting a wall between the head amp and the heavy dissipation parts, dividing the board in two sections and ample ventilate.
I don't quite follow that you knew you were setting up a heat cluster bomb, and then abandon it so swiftly. (sp?)
(not that i'm in the mood to duplicate the 4.0, assembling the HPS3.1 is more up my alley to compare to the phono stages that i have, but it is interesting reading)
There were tons of things to investigate, outside the thermal stuff, certainly worth an iteration. Noise, stability, general AC prformance, new layout, shunt regulator AC performance, etc...
That pair is certainly not really "low noise", they seem to be the standard double diffused 40ohm Rbb'.
Compare the datasheets.
HPS 4.1 is ready
See the pictures.
This is a full bipolar version that uses Jung regulators and local bipolar buffers. It is also high impedance (default 47kohm), while the input current can be trimmed to under 50nA. No caps in the signal path.
Full measurements to follow, but here's the first noise measurement. RIAA correction is not installed, gain is a little over 80dB (will be 60dB @ 1KHz after RIAA correction). Bottom line - 0.28nV/rtHz.
Distortions are much better than in the HPS 3.1 JFET, I suppose because the large (8 x JFET) nonlinear Ciss is now missing. Overall, about 70ppm at all frequencies and output levels, up to +/-20V. HPS 4.1 has the same headroom of 32dB.
.
Full revised schematic and all construction data (including the PCB Gerbers) to follow on my web site.
See the pictures.
This is a full bipolar version that uses Jung regulators and local bipolar buffers. It is also high impedance (default 47kohm), while the input current can be trimmed to under 50nA. No caps in the signal path.
Full measurements to follow, but here's the first noise measurement. RIAA correction is not installed, gain is a little over 80dB (will be 60dB @ 1KHz after RIAA correction). Bottom line - 0.28nV/rtHz.
Distortions are much better than in the HPS 3.1 JFET, I suppose because the large (8 x JFET) nonlinear Ciss is now missing. Overall, about 70ppm at all frequencies and output levels, up to +/-20V. HPS 4.1 has the same headroom of 32dB.
Full revised schematic and all construction data (including the PCB Gerbers) to follow on my web site.
Last edited:
Any idea on the current noise? I did some plots a long time ago using a cart with stylus removed as a source impedance. I suppose one could do THD that way too. Interesting experiment.
I'll plot that ASAP. Anyway, the low noise trannies are biased at the optimum Ic, for a 10ohm source impedance Ic=Vt*SQRT(Beta)/Rs
RIAA errors
The measured RIAA errors. The ideal RIAA response was generated by synthesis (including the 50KHz pole) then the measured frequency characteristic is normalized to the ideal RIAA response. It's the same process as using a reverse RIAA, only that the reverse RIAA implementation errors are avoided.
The Y scale is 50mdB/Div. The measured response is within +0.1 -0.12dB in the 20Hz - 45KHz range. Target is reached.
The measured RIAA errors. The ideal RIAA response was generated by synthesis (including the 50KHz pole) then the measured frequency characteristic is normalized to the ideal RIAA response. It's the same process as using a reverse RIAA, only that the reverse RIAA implementation errors are avoided.
The Y scale is 50mdB/Div. The measured response is within +0.1 -0.12dB in the 20Hz - 45KHz range. Target is reached.
HPS4.1 current noise
Here's the HPS4.1 current noise plot. Unfortunately the RIAA correction is already installed, so the increase of the current noise with frequency is masked by the RIAA gain drop at high frequencies.
Anyway, the equivalent input current noise @1KHz is about 5pA/rtHz. This maps to a negligible current noise contribution to the total noise, for a 10-20ohm MC.
However, for a MM cartridge having 500ohm, the current noise contribution is 2.5nV/rtHz, which adds to the 0.3nV/rtHz input voltage noise and the 2.8nV/rtHz cartridge noise. This clearly shows that HPS4.1 is not recommended for MM cartridges. OTOH, using something of the HPS4.1 complexity for a MM cartridge doesn't make much sense anyway. A simple, low noise JFET construction with a gain of 40dB (like HPS2.0) will do just fine.
Here's the HPS4.1 current noise plot. Unfortunately the RIAA correction is already installed, so the increase of the current noise with frequency is masked by the RIAA gain drop at high frequencies.
Anyway, the equivalent input current noise @1KHz is about 5pA/rtHz. This maps to a negligible current noise contribution to the total noise, for a 10-20ohm MC.
However, for a MM cartridge having 500ohm, the current noise contribution is 2.5nV/rtHz, which adds to the 0.3nV/rtHz input voltage noise and the 2.8nV/rtHz cartridge noise. This clearly shows that HPS4.1 is not recommended for MM cartridges. OTOH, using something of the HPS4.1 complexity for a MM cartridge doesn't make much sense anyway. A simple, low noise JFET construction with a gain of 40dB (like HPS2.0) will do just fine.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.