I like split bobbin too, I would choose 22+22VAC 100VA. So to be sure I would end up with no less than 25VDC Reg in for plus/minus 15VRegDC out after the drop on the chokes and some possible drop on the secondaries due to 1.2A total constant current draw. How are the mains drops during the day in your non industrial remote country?
at the cost of having no lifestyle to talk about we have reliable rules and technology so do not be yealows
no commend but the shootout was very enjoyable. i hold the Paris in high regard. It is very affordable and sounds much better then most offerings costing 5 times as much. the enginering is creative and is a good example of elegant simplicity. Hugh is a great guy too and very pleasant to work with so who am i to critizise his work ?
I made two small modifications to it. I bypassed the Auricap with a 0.1uF Vishey/Roederstein MKP and put it in a shielded box. The shielding made tho tone more robust and the MKP made the treble more to my taste. It sounded more open and natural to me. Hard to explain but i found the treble had to much "audiophile" character to me before the modification, an experience i make with many "boutique" capacitors and the Auricaps are really one of the best and not crazy expensive. I am awaiting some of their teflon caps. they make 200V versions that are not to big and are 1% tolerance. i whould like to try them out in the RIAA of the MPP.
I go for the low ESR version and will tell you if it is any good.
Another thing. I found a very stage part in the Interfet catalog : the IF3601.
It has a staggering forward transconductance of 750mS so we can afford to put two in series without big concern about the noise. Long Tail Pairs are possible that way but i find them a bit boring. Enough has been written about them and you can easily come up with your own circuit idea. an interesting design to study is the AD797. just supstitude the input bipolar pair with a pair of those fets. This part has quite a high input impedance of 300pF so cascoding helps. I do not think that is a big problem with low impedance MC´s but still i think it´s good to reduce it for more general use. just out of my head i present here a circuit idea that uses the first Fet as a source follower coupled to a grounded base stage followed by a folded cascode so input capacitance should be well controlled. it has feedback to the input so resistors in the feedback loop should be very low value for best noise performance. the diamond buffer is shown in it´s simplest form and you can find other designs with less and more complication in this thread that should work too. i forgot the plus-minus supply but i think you get the idea.
Another thing. I found a very stage part in the Interfet catalog : the IF3601.
It has a staggering forward transconductance of 750mS so we can afford to put two in series without big concern about the noise. Long Tail Pairs are possible that way but i find them a bit boring. Enough has been written about them and you can easily come up with your own circuit idea. an interesting design to study is the AD797. just supstitude the input bipolar pair with a pair of those fets. This part has quite a high input impedance of 300pF so cascoding helps. I do not think that is a big problem with low impedance MC´s but still i think it´s good to reduce it for more general use. just out of my head i present here a circuit idea that uses the first Fet as a source follower coupled to a grounded base stage followed by a folded cascode so input capacitance should be well controlled. it has feedback to the input so resistors in the feedback loop should be very low value for best noise performance. the diamond buffer is shown in it´s simplest form and you can find other designs with less and more complication in this thread that should work too. i forgot the plus-minus supply but i think you get the idea.
as far as i understand it this capacitance appears in the case of a grounded source circuit where the voltage amplification is taken from the drain with high impedance loading. the circuit shown does not do that. the first stage is a source follower, the second a common base stage and that is loaded by the folded bipolar cascode so everything is low impedance. anyway, i will not explore that root in the forseeable future. i was more interested if anybody has tried that part. i did not even know that 750mS is posible. for what do they use that part ? to measure something with extremely low noise i suppose.
Not in your case, I understand you use another topology. I just wondered how it would do if common source and if it would be controlled easily by cascode. With such Ciss they quote, the noise must be as they say and the gain too. With such gain the internal resistance must be something like 10 Ohm or less. And you can get them in Europe? You know that brand being distributed?
You can get them from Rhopoint. They have other very interesting parts. For example i always wanted to try their precission wirewound resistors. At least in speakers i always come back to wirewounds. They have a very low temperature coefficient and subjetivey very good resolution of microdetail. they make low watage types that could be used in ampifiers too. Concerning this special fet i think with that design noise could be even lower then 0.3nVqHz because they specify that value at 5mA idle. so raising the standing current could reduce the noise even more.
Sorry, but 41550 and 41750 are basically the same cap and documented in the same datasheet (at least in the versions that I have), except that the ..570 has threaded stud mounting, the 550 does not. ESR values "should" be exactly identical. This looks more like a typo in the datasheet.
FWIW, Sikorel have the lowest ESL of all can type electrolytics, especially compared to snap-in electrolytics, at least from all the datasheets that I have seen, and hence have the highest resonant frequency and bandwidth. Add their exceptional long life specs and they are certainly value for money despite the price.