Addressing John Curl's concerns on low noise designs

Status
Not open for further replies.
Jump to Latest
#362
Well, it disappointed me at the time, since it was destined to be used in a proven circuit, and I had gone out of my way to get the devices through formal introductions in Japan.
Still, it won't work in my patented circuit. Check the circuit out, yourselves, if you are curious. This is only for experts, amateurs need not bother.
 
#373
This is absurd!

The 2SA1084, you can get them cheaply and in quantity?
I am impressed.
Of course, when I sent one to Yishay Netzer, some years ago, it couldn't have been for any good reason, could it? And when Honeywell engineers replied back and wanted more, it must have been a fluke as well.
Now, because it was so hard to get, I gave Dr. R.G, Meyer an alternative part: a Fairchild PE8550, 2A TO-92 type device for his project, since they were available from 50 mi away, with a phone call. The 2SA1084 would have worked in his input stage, but he didn't need the very high beta, high voltage, and low input capacitance, and they were harder to get. Rbb' was about the same in both devices.
 
#375
john curl said:
This is absurd!

The 2SA1084, you can get them cheaply and in quantity?
I am impressed.
Of course, when I sent one to Yishay Netzer, some years ago, it couldn't have been for any good reason, could it? And when Honeywell engineers replied back and wanted more, it must have been a fluke as well.
Now, because it was so hard to get, I gave Dr. R.G, Meyer an alternative part: a Fairchild PE8550, 2A TO-92 type device for his project, since they were available from 50 mi away, with a phone call. The 2SA1084 would have worked in his input stage, but he didn't need the very high beta, high voltage, and low input capacitance, and they were harder to get. Rbb' was about the same in both devices.
Click to expand...

http://bdent.com/search/part.jsp?partnum=2SA1084
http://bdent.com/search/part.jsp?partnum=2SC2547

Don't know what you mean "in quantities" but you can certainly get them in the hundreds. I just ordered 150pcs. each.
 
#378
john curl said:
At this moment I am getting 0.42nV/rt Hz at 3ma on my Quantech with a 2sa1084.
Click to expand...

That's pretty dang good! Are you going to try replacing your matched jfets with them and see how they sound?

My opinion and experience of bp's against jfets says they will sound better than the jfets, with more resolution, more detail and just overall more musical feeling. At least as cascodes under tubes, but also in x1 buffers and other different all SS circuits.

Regards, Allen (Vacuum State)
 
#380
Wavebourn,
transistors have a linear region, appointed by biasing.
The nonlinearity of BJT s chiefly caused by base-emitter irregularities and the reverse transfer capacitance. Keeping Ice and Vce constant stabilizes several parameters. A higher Ic flattens variations in Gm, thereby in the impedance, which relates the input voltage to output current. (It can be mentioned here, that the tight I/Gm ratio of BJTs is not advantageous in transconductor applications, as nonlinearities in the current-voltage characteristic become large for small voltage swings. In addition, FETs have better low-level signal amplifying ability due to lower intrinsic noise, the current flowing through the channel uninterruptedly without crossing a P-N junction, also a much wider operating frequency range).
It is obvious that Vbe must be small. Emitter degeneration reduces it, alongside with Gm and the hurtful nonlinear temperature dependency.

More can be done, without using diodes, they don`t linearize anything, just supply distortion.
 
Status
Not open for further replies.
Top Bottom