John Curl's Blowtorch preamplifier part II

Status
Not open for further replies.
But whether even it is significantly better overall, given the other noise sources like Brownian motion of air and thermomechanical noise of the structure of the capsule is questionable.
Well I do have a small interest in capsule design. :D

With evil 21st century digits, it's possible to re-visit some of the compromises we had to make in da previous millenium. Excuse me for swearing :eek:
 
The 862 is fine except for the rather high current, given one of Scott's constraints being battery operation. But whether even it is significantly better overall, given the other noise sources like Brownian motion of air and thermomechanical noise of the structure of the capsule is questionable.

Which reminds me: I wonder if Wave ever tested those photonic mic capsules?

2SK222 of Sanyo is the best because the Idss is only around 1mA. There is no need to run a FET in a mic at Idss so this does not matter anyway.

In any case I stated my case in my articles and have tired of correcting the misconceptions. The JFET in a large capsule mic has little or nothing to do with the noise performance. The obsessive Neumann rebuilders search for authentic 1960's 2N3819's for goodness sake.
 
Disabled Account
Joined 2012
Mr Marsh, Error correction works in comparing two signals. I am afraid there is noway to use that kind of thing on an oscillator, where there is nothing to compare.
I was hunting for something along the lines as conceptual... --- Demian has suggested the method Amber uses which phase shifts the harmonics and feeds it back somewhere/how to cancel them. I'll look into that. Thx-RNMarsh
 
Last edited:
I was hunting for something along the lines as conceptual... --- Demian has suggested the method Amber uses which phase shifts the harmonics and feeds it back somewhere to cancel them. I'll look into that. Thx-RNMarsh
The two problems i had encountered with this kind of cancellation is both phases and levels. Hard to be better than 60db in the best situations (low frequencies <1000Hz).
I used this kind of canceling circuit to evaluate dynamically (With musical signals) the distortion of amps (and listen to-it). It is the base of my protection circuit.
Levels are hard to set too.
Did-you want to design an analog distortiometer ?
 
Disabled Account
Joined 2012
The two problems i had encountered with this kind of cancellation is both phases and levels. Hard to be better than 60db in the best situations (low frequencies <1000Hz).
I used this kind of canceling circuit to evaluate dynamically (With musical signals) the distortion of amps (and listen to-it). It is the base of my protection circuit.
Levels are hard to set too.
Did-you want to design an analog distortiometer ?
I'm trying to get the oscillator in an HP339A as low in H2 and others as is humanly possible. It allows for multiple freq and the built-in analyzer is a bonus (I have picked up an ShibaSoku AD725D anaylzer). They use a jFEt for agc and it is the limitng factor -- espec H2. Thx-RNMarsh
 
Disabled Account
Joined 2012
FYI -- anyone else buy one of those nano scopes? i just got one (4 channel DSO: DS203) cause its so cute. And, it takes up zero space on the crowded work bench and good for following a signal around to see where it gets lost or goes wrong. Size of deck of cards and 1/2 the thickness. Nice color display rez, too. USB and signal gen port output. www.minidso.com -RNM
 
Last edited:
OT

How do you show that someone can distinguish DUTs when he believes that he cannot?

As mentioned before, a good start could be to set up a listening test without revealing the effect under test.

A good scientist would not tell lies in such a test, but how would the tester know?

Normally it is not a matter of lies; basically it (at least as a underlying hypothesis in these discussions) is a matter of "uncontrollable" bias.

If the contention is true, then even a good scientist will not work as a good detector if he is biased against the audibility of a certain effect.

Much better to demonstrate that people who think they can distinguish really can do it, then the scientist will be sufficiently intrigued to ask 'How?' and 'Why can't I do it?'.

While that seems to be plausible, it seems that a strongly biased scientist will never accept any unwanted listening test result.
 
Jakob2 said:
While that seems to be plausible, it seems that a strongly biased scientist will never accept any unwanted listening test result.
Perhaps that explains why some people seem so unwilling to risk offering an unwanted listening test result; they don't want to embarrass the scientists, as they know that scientists are delicate individuals with large egos who don't like to be proved wrong by experimentally established facts. To protect the scientists, they selflessly forego all opportunities to confim that they really can hear what they claim to hear.
 
Perhaps that explains why some people seem so unwilling to risk offering an unwanted listening test result; they don't want to embarrass the scientists, as they know that scientists are delicate individuals with large egos who don't like to be proved wrong by experimentally established facts. To protect the scientists, they selflessly forego all opportunities to confim that they really can hear what they claim to hear.

Such a good demonstration about the effect of a strong bias. :) (SCR)

If you remember the short discussion on the "Fremer listening test results on amplifiers" then you will understand what i really meant.
 
The 2SK152 device may still be around... just have to go to Texas to get it at InterFET Corp. They say the 2SK152 is same process as their IF1320 and IFN152. Also, same process as 2N6451,2,3,4 (process NJ132L). Thx-RNmarsh


The BF862 has a Yfs of 38mS @ 10mA and 10pf Ciss
The 2SK152 has a Yfs of 25mS @ 10mA and Ciss of 7.5pF

So the BF862 has a slight shade better gm to C ratio.

EDIT -The specs on IF site are 2X different in some cases, Ciss 15pF and Idss range of 5 to 20mA vs 9.5 to 45mA. Same is loosely interpreted.
 
Last edited:
I might talk about the origin of the 2SK152, what it was made for, and why it has been considered 'special' over several decades.
In analog video recording there is a special area where the S/N limit is significantly based on the 'quality factor' of the input device.
When I was at Ampex up to 1970, the chosen device was a specially selected 2N4416, chosen for low 1/f noise.
Later, Sony, in competition with Ampex, made their own ideal part, the 2Sk152, for this application.
I saw the part highlighted in a tech mag as the highest 'Q' part in the industry, so I thought to use them for hi Z applications. Independent testing with a QuanTech noise analyzer showed slightly disappointing 1/f measurement, but I have recommended them over the decades for condenser microphone use. At least you can get the parts in a T0-92 leaded package, and that makes handling easy.
Now the BF862 has a slight advantage in 'Q' and maybe even 1/f noise.
 
I might talk about the origin of the 2SK152, what it was made for, and why it has been considered 'special' over several decades.
In analog video recording there is a special area where the S/N limit is significantly based on the 'quality factor' of the input device.

Yes they are both RF FET's that can be re-purposed to audio. The fo of the BF862 is 750MHz which makes it want to oscillate unless you take extra care. I'll bet the 2SK152 has similar problems.

I alreay posted on the Pass forum plots of the BF862 showing its almost total lack of 1/f.
 
Status
Not open for further replies.