But these will give me 20pF without having to parallel 10 devices (ultimately 20 devices seeing as this is a push-pull topology).
As for ultra low noise, I'm not going to be using them as common source amplifiers, but rather as source followers, fed by ultra low noise 1:10 microphone transformers.
Hasn't anything to do with mysterious Russian military technology. Far as I can determine, these are basically Crystalonics CP650's in stud mount cases. And it's the latter that makes them more attractive to me than the CP650's (well, that and the fact that the CP650's go for $17.50 a pop).
That remains to be seen. That's why I want to get some to try for myself.
Yeah, they do have the fact that they're JFETs going for them. 😀
se
Patrick,
the cascode does not reduce capacitances. A rather high value of Rdson is another condition of linear behavior, 5 Ohms is fine for devices with this power handling capacity. And yes, a fairly low but flat transconductance.
the cascode does not reduce capacitances. A rather high value of Rdson is another condition of linear behavior, 5 Ohms is fine for devices with this power handling capacity. And yes, a fairly low but flat transconductance.
About input capacitances, as I see it, in common source stage, it is roughthly a sum of Ciss + Crss x voltage gain.
Since the cascoded device does not experience voltage swing, it will have input capacitance without multiplication of Crss, i.e. simply Ciss+Crss.
My subjective view, the less input capacitance - the more tiniest sound detailes are resolved, the more close we to a tube stage.
Vladimir,
Crss is incomparably the most detrimental capacitance, always look at that value first, it must be small for all (also bipolar) devices.
Crss is incomparably the most detrimental capacitance, always look at that value first, it must be small for all (also bipolar) devices.
Sometimes it can play a role. Erno Borbeley has measured nearly 600pF input capacitance for K170 common-source stage, near 20dB voltage amplification.
This was because of Crss multiplication.
> a fairly low but flat transconductance
Do we have a language problem ?
The KP903 has a transconductance at Idss (400mA) of 0.1A/V, or 0.1S.
The LU1014D has a transconductance at 1A bias of 4S, or 4A/V.
I assume 0.1S is lower than 4S and not Vice Versa ?
Kindly show me a JFET, even MOSFETs that has a Yfs of 4S at 1A.
And please also show me another Id vs Vgs curve that is anywhere near as flat at 1A.
Patrick
Do we have a language problem ?
The KP903 has a transconductance at Idss (400mA) of 0.1A/V, or 0.1S.
The LU1014D has a transconductance at 1A bias of 4S, or 4A/V.
I assume 0.1S is lower than 4S and not Vice Versa ?
Kindly show me a JFET, even MOSFETs that has a Yfs of 4S at 1A.
And please also show me another Id vs Vgs curve that is anywhere near as flat at 1A.
Patrick
I have some personal building and listening experience, of several Zen9-like amps. I played with 40pcs of LD1014DA and 10pcs of LD1010DA (the last has 10S transconductance at 2A and Ciss close to 1600pF).
I reported stereo power amps, also two monoblock versions, the last was with near 5A idle current, with many high-quality parts, with various degrees of modulation of Aleph CS.
Now I am neutral to LU1014 parts, since in spite of several years efforts I have not managed to get equal with SE 300B amp, in vocal reproduction.
Now I look only at HF parts, MOSFETs and jFETs, with minimum possible Ciss.
And now I am proud about three last designs, all my friends say that they are superior now to both SE 300B and another push-pull tube amps.
I reported stereo power amps, also two monoblock versions, the last was with near 5A idle current, with many high-quality parts, with various degrees of modulation of Aleph CS.
Now I am neutral to LU1014 parts, since in spite of several years efforts I have not managed to get equal with SE 300B amp, in vocal reproduction.
Now I look only at HF parts, MOSFETs and jFETs, with minimum possible Ciss.
And now I am proud about three last designs, all my friends say that they are superior now to both SE 300B and another push-pull tube amps.
Member
Joined 2009
Paid Member
with some devices that maybe true, but I feel it is a compromise. The zen-9 article states that the cascode is more or less a slave device that adds nothing of it'self to the sound. I have no experience to refute that, but I can't help feeling that it's another device in the signal path, just happens to be in common base amplifier configuration. I'd rather do without a cascode if I can.
Yes, as per datasheet (in russian) for KP903A
Noise EMF at 100kHz, Uds=10V, Id=10mA typically 0,7nV/sqrt(Hz) (possible span is from 0,5 to 1,0)
Right, more correctly, Ciss = Crss + Cgs. Actually, the gate capacitances are composed of two parts: linear fixed overlap capacitances that are independent of applied voltage and nonlinear charge storage capacitances that are associated with power handling so they cannot be made too small.
The large and useless Idmax of LU1014 (50A) brings about a lot of capacitance and a thousand times lower Rdson than that of KP903A.
Last edited:
Yes, therefore I do not think that the LU suites preamp or headphone amp schematics, the Zen9 is the most perfect place of its implementation, and Nelsonn Pass has indicated this ideal place.
Unfortunately, even in the best application, the resulting sound is very good, but far away from champions.
I dougbt that it was issued by russian manufacturer (parts discontinued in 1992) but US manufacturer of substitutes (CP650, CP651) - CRYSTALONICS | semiconductor manufacturer of Small Signal Transistors, JFETS, DUAL TRANSISTORS, QUAD TRANSISTORS, POWER TRANSISTORS, CURRENT REGULATOR DIODES, VARACTOR DIODES and custom HYBRID MICROCIRCUITS - still offer these parts, and SPICE model could exsist.
Without further explanation, it has been claimed that the statement that class A amplifiers biased with a constant current source draw constant current from the power supply is wrong, even more mysteriously, with the exception of the design of this thread.
The description of the "concept" with its "special features" is to be found below - sad to say, it`s all mumbo-jumbo to me.
The description of the "concept" with its "special features" is to be found below - sad to say, it`s all mumbo-jumbo to me.
Attachments
Last edited:
- Status
- Not open for further replies.