John Curl's Blowtorch preamplifier part II

Status
Not open for further replies.
Jump to Latest
Joshua_G

Joshua_G

Member
Joined 2008
gomeh.com/
ThorstenL said:
I can tell you how I did it, if that helps?

E88CC parallel cascoded, run at loads of current. Probably around 2nV|/Hz A Weighted... The WE 437A also comes clsoe to those noise figures, arguably tubes need to be selected.

If we can use more exotic tubes we can do better. The E810F run at zero bias and around 40V Anode voltage (cascoded) gives similar trasconductance and noise as a 2SK170...
Click to expand...

Hi Thorsten, I'm glad to see you back.

You also did JSK147 cascoded by 5842 tube on your Toccata MKII MC front-head.

How about 2SK170 cascoded by E88CC/6922, or E810F/ 7788 triode connected, or C3g triode connected?
 
V

VladimirK

Member
Joined 2007
Joshua_G said:
Hi Thorsten, I'm glad to see you back.

You also did JSK147 cascoded by 5842 tube on your Toccata MKII MC front-head.

How about 2SK170 cascoded by E88CC/6922, or E810F/ 7788 triode connected, or C3g triode connected?
Click to expand...

Some RF jFETs, like CP650 (US part) or KP903A (similar russian part) are much better than 2SK170
For the KP903A the datasheet says
typical 0,7 nV/sqHz
Idss 300... 500mA
transcondactance 85 mA/V
Ciss = 18pF
dissipated power 6W
I have used these KP903A successfully both as an input transistor (cascoded) in preamp, and as an output transistors (8 matched pcs in parallel) in SE power amp.
 
ThorstenL

ThorstenL

Previously: Kuei Yang Wang
Joined 2002
linkedin.com/in/thorsten-loesch-studio-raumklangg/
Hi,

Joshua_G said:
How about 2SK170 cascoded by E88CC/6922, or E810F/ 7788 triode connected, or C3g triode connected?
Click to expand...

Yes, can all be done. If using the simple direct cascode it is necessary to watch the current of the J-Fet (means selection) so it gets enough voltage to work. It's basically all in the various datasheets.

Joshua_G said:
Do you have the data for E810F/7788 and/or C3g in triode mode?
Click to expand...

It's all in the datasheets.

Duncan's Amp Pages

click on TDSL

Ciao T
 
ThorstenL

ThorstenL

Previously: Kuei Yang Wang
Joined 2002
linkedin.com/in/thorsten-loesch-studio-raumklangg/
Hi John,

john curl said:
Good to see you back, Thorsten. Just in time! Now we have some new material to work with.
Click to expand...

Thank you, the new material is interesting (I had not realised just how rusty my Russian has become). I do think we still have some old material to cover, but I'm very much pushed for time ATM...

Ciao T
 
V

VladimirK

Member
Joined 2007
PMA said:
Would you kindly attach or link the datasheet? The combination of parameters does not seem real.
Click to expand...

Here is the CP650 datasheet

I would also like to mention, that kp903a are stable at SE output stage with 36V power supply, in spite that 20V are in datasheet. Even in industrial equipment, they work at 40V. I do not know, why the datasheet is so conservative. Moreover, I use the military specified version of these transistors, the 2P903A. Important difference between CP650 and kp903a, the last has all pins isolated from the stud mount case , while CP650 is not.
 

Attachments

  • CP650-653 n-chan 20V 0,6A 8W.pdf
    98.5 KB · Views: 161
Last edited:
ThorstenL

ThorstenL

Previously: Kuei Yang Wang
Joined 2002
linkedin.com/in/thorsten-loesch-studio-raumklangg/
Hi,

Joshua_G said:
I have all those datasheets for a while.
There is no info on neither the E810F/7788 datasheet nor on the C3g datasheet about them in triode mode.
Click to expand...

Telefunken C3g has triode mode info.

The E810F Datasheets also have triode curves, there is just one grapch and ot much detail, but it should be enough for the kind of work we do here (DIY).

Ciao T
 
scott wurcer

scott wurcer

R.I.P
Joined 2004
simon7000 said:
There ain't no micro diodes. If you measure any component to an extreme enough level you will always find some sort of distortion.
Click to expand...

I've decided not to bother with this, someone sent me an article by Doug Self from a long time ago wherein he found nothing at all.

Secondly Ed, the plots you posted months ago and your description of "conduction thresholds" or dead zones _ARE_ rectification. Any VI characteristic that shows, as at least one plot you posted, a flat spot at the origin will rectify.

I was asked Tuesday to basicly stop criticising someone who wanted to make a vacuum microphone (so it would be noiseless). Please don't take it that I am comparing this to that, but consider what I said.
 
F

fireworks

Member
Joined 2010
VladimirK said:
What is interesting, attention to this paper has been drawn by one strong supporter of NFB at one russian forum. But he was trying to simulate and to measure with all possible details just a transfer function of his amps. And he has catched it very fast, that if we look at the TF from dynamical point of view (their is nothing infinitely fast), then even GNFB can cause a kind ot TF jitter. I would add, after-action of electrolytic caps can also add. So we may deal with "thickening" of TF line, and it cause some loss of signal microdetais.
What everybody always note with deep GNFB designs.
Chumakov speaks not about pure kinds of noise generated by parts, but about "effective noise" due to all kinds of imperfections of dynamical origine.
In SE tube designs we have no zeners, no bad electrolytics, no transistor distortions due to chip temperature variation under signal, no or short feed back path - vuala, good pure sound, plenty of microdetails.
In a russian forum, they have reported listening impressions from three designs
1) tube SE amp by Makarov (russian tube guru, makes amps of 500kg weight)
2) deep GNFB SS amp by Sergei Ageev (russian SS guru, improving implementation of the same schematics for the last 20 years, not to confuse with Ageev-academician and his theorem)
3) low power but very fast SS amp, with very short and fast current NFB path (something like 500V/uS in the FB loop)
Main claim to the amp 2) - loss of microdetails (of cause, in comparison with the amp 1), amp 3) was closer to amp 1) )

Other system's components were
CDP Mark Levinson, near 30 000 USD price tag
speakers Montana, near 80 000USD price tag
Before claiming about microdetails, they have listen also during a night time, at very silence
Click to expand...

Very interesting.
Can you post some links to the listening/measuring tests themselves ?
Also links to the schematics or details on the three amplifiers would be useful.
 
scott wurcer

scott wurcer

R.I.P
Joined 2004
fireworks said:
Haha, interesting ! How would that work ? I mean how did that person say it would work ? :p
Click to expand...

Well we got as far as evacuating just one side (the side with the noise of course). I pointed out the 194dB SPL offset and he suggested a polarized grid to suck the diaphragm back out. I then gave up, you would need 7700V at 50 microns to get 1 N/m^2, and figure out how to turn on both the HV and vacuum without destruction by either. It would arc over in any case. You would only find that the noise has migrated to the other side in any case.
 
Last edited:
Status
Not open for further replies.
Top Bottom