Originally Posted by Simon - PG Music
Directional cables exist and can be beneficial. Consider ground loops - lifting ground at one end of the cable can remove the ground loop, but in some cases there can be a measurable reduction in noise having the lifted end at one end vs the other end.
The observation is sound, but "directional cables" are not implicated. You could reverse a "directional cable" end-for-end, repeat that test, and the outcome would be the same, all other things equal.

Originally Posted by Simon - PG Music
48v isn't a big issue here, since typically the 48v is run through a pair of 6k8 resistors which drop a chunk of that voltage and also limit to a maximum of ~14ma. If the circuit is drawing more than a few milliamps, that can easily drop the voltage to ~30v which I'd say most decent op-amps can handle - and if not, add some resistance to pull the voltage down a bit more (or a zener, but that can add noise).
6k81 +/- 0.1% is the standard. The value isn't really critical and 6k8 is a very standard value. The +/- 0.1% is important for line balance and hence noise pick-up.

I'm a mostly-retired designer. I'll either use suitably rated components (and margins!) and I'll control the conditions if necessary. For me "most decent op-amps can handle" is not an option. If one makes a mistake and there are 10,000 instances of the product in the field, one has questions to answer! I have made the occasional mistake. They can be very expensive for the manufacturer.

Originally Posted by Simon - PG Music
I'd imagine that having signal on the power rail could be more of an issue for op-amps than for discrete JFETs or BJTs - perhaps this is why they're not widely used in 48v inline preamps.
The 48V is, of course, actually on the signal lines, so unavoidable. It must be taken off in the mic or pre-amp for use and however it's taken off must also not risk feeding noise back into the cable. I suspect the reasons Op-Amps are not more widely used in in-line pre-amps is that they tend to cost more, particularly once one a dealt with the supply issues, and with discrete parts, whilst there may be a little more distortion, it's low enough and typically of the "warm" 2nd harmonic form, that's of very little importance.


BTW, I made a SPICE model of what I believe the FETHead comprises, though using other JFET models with similar characteristics and had the right gain and a nice clean signal as far as I could tell by eye. I didn't try to prove noise levels or distortion. They could have done it with only two FETs, but using the four gives an additional small noise improvement. They do only cost a few cents each, though, so the profit margin is pretty near unaffected. As so often, the mechanicals and assembly are where the costs lie.
Edit: Incidentally, my model suggested that the circuit heads into clipping territory at an acoustic level of around 110 dB SPL, depending on the mic(*), but at that sort of level, noise shouldn't really be a concern. (*) Or about 116 with that rather insensitive SM7B

Hmm ... costs etc., That SSM2019 is listed on DigiKey at £2.50 at 2500 off ($3?), whilst 2SK880-GR is listed a 15p at 3000 (20 cents?). And no need for power supply to guarantee that the supply to the chip doesn't go above it's 36V rating.

Last edited by Gordon Scott; 12/30/23 05:01 AM.
Jazz relative beginner, starting at a much older age than was helpful.
AVL:MXE Linux; Windows 11
BIAB2025 Audiophile, a bunch of other software.
Kawai MP6, Ui24R, Focusrite Saffire Pro40 and Scarletts
.