Can someome help me to understand the feedback loop on the KT66 (R10) and how it can be calculated
JORGETronic said:
This is voltage feedback over the output tube (parallel feedback). The calculation is difficult, because it depends of the plate resistance of the driver and output tubes and the ampflification factor of the output stage.
This feedback is not so effective as a voltage feedback over the driver and the output tube; more than 6 dB NFB increases the distorsions (more second harmonics, generated by the driver).
Regards Andreas
many THANKS Andreass, any comment about the circuit, I was doing some calculation over the 6AU6 and as the Ig2 current is dificult to calculate I do not know exactely the max input voltAGE, as I want to attact the amp directly from a cd palyer, lets say 2 volts....
JORGETronic said:
That calculation isn't particularly difficult since it's just a voltage divider. The feedback signal appears across the parallel impedance of the r(p) and Rp of the 6AU6, and the grid resistor of the KT66. The actual value isn't particularly critical, and O. Schade recommends a 10 : 1 voltage division.
Given this circuit, I wouldn't change it since the coupling is DC, and changing the 330K resistor is likely to upset the DC operation of the 6AU6 stage.
Thanks for the answer, the book also mention that if you change the KT66 for EL34 yuo must change the resistor to 470K so i was traying to do some paralles with the different rp(KT66) and the rp(EL34) with no succes, someone can tell me if the 6AU6 con handje 2 volts in its imput easely...
I have notice that you daid that the amp is DC coupled, the KT66 is coupled to the 6AU6 via a capaitor of 0.5 uF...no DC coupled right?
JORGETronic said:
The local feedback from the KT66 to the 6AU6 is DC coupled. The stages are AC coupled.
I just modeled the EL34 version, using LTSpice, out of curiosity (I don't have a model for KT66). I used 470k for the NFB resistor, as you said. I also used 3.5k plate load and 190 ohm cathode resistor for the EL34, as per Mullard's recommendations with 300v B+. All other values were as per your schematic.
The amp seems to be designed for 250mv rms input signal for maximum output. 2 volts input signal would be far too high.
The amp seems to be designed for 250mv rms input signal for maximum output. 2 volts input signal would be far too high.
How about this one, instead of some old dusty curcuit - should be possible with KT66
http://www.tubelab.com/SimpleSE.htm
http://www.tubelab.com/SimpleSE.htm
Attachments
I wouldn't discount that 'dusty old circuit' quite yet. I have built a few local feedback amps, and they can sound very, very good. See this link for more info:
http://www.tubeaudio.8m.com/807/807.html
pm
Miles Prower said:
Yes, this is a voltage divider, but there is the input impedance of the output stage in parallel: the driver "sees" R10 "smaller" by the amplification factor of the output tube. The result is a decreasing amplification and increasing distorsions in the driver stage (lower Ra).
Better determinine R10 by measurements.
Regards Andreas
jane presented a method of calculating the local feedback in his thread on "Just another PP...", with an Excel model here: http://www.diyaudio.com/forums/showthread.php?postid=918353#post918353
It is a little complicated for me to understand, so I second the suggestion of making a ROM estimate, and adjust by measurements.
Svein
It is a little complicated for me to understand, so I second the suggestion of making a ROM estimate, and adjust by measurements.
Svein
Hi Svein,
thank you for the link.
My recommedation was to evaluate the best ratio of second and third harmonics for varius degrees of feedback by "try and error".
Regards Andreas
thank you for the link.
My recommedation was to evaluate the best ratio of second and third harmonics for varius degrees of feedback by "try and error".
Regards Andreas
Could u pls provide details on o/p transformer.
If I were to wind it locally in Mumbai, will need sec and pri details.
Thanks in advance
Regards
Rajesh
If I were to wind it locally in Mumbai, will need sec and pri details.
Thanks in advance
Regards
Rajesh
Thanks you all for take the time to answer, but going back to the topic, how can I modify the imput to handle 2volts, I think i have to decrease Rp in the 6AU6 then get more current in the first stage, but all the info I find inj data shets is for RF and very low imput signals....any suggestion for Rp and Rk in order to increase the imput voltage range?
thanks
thanks
Hi Jorge,
pls. look at the data sheet of the EF94, this ist the european equivalent of the 6AU6.
http://frank.pocnet.net/sheets/062/e/EF94.pdf
Best regards Andreas
pls. look at the data sheet of the EF94, this ist the european equivalent of the 6AU6.
http://frank.pocnet.net/sheets/062/e/EF94.pdf
Best regards Andreas
Andreas, many thanks for the info, I didnt know the equivalence with EF94, in fact I have four of them (VALVO) German too I guess I got from a very old GRUNDIG tape reel...thanks again I will have to gat a Geman diccionary too....but the date sheet seems to be easy to understand
AndreasS said:
Yes, yes, yes: there's a whole bunch of theory on I to V conversion that goes along with it. However, in the final analysis, the value of that resistor really isn't all that critical. O. Schade doesn't mention any complex theory, and neither does Crowhurst. A first approximation is good enough in almost every case. That's also mentioned by O. Schade, and verified by moi. Unlike sandy circuits, hollow state is very "forgiving", and extreme accuracy is seldom required. Which is a good thing since during "the day" you didn't have 0.1% ultra-precision, laser trimmed metal film resistors, or metal film polypropylene capacitors. It was all +/- 20% resistors (+/- 5% was considered to be "ultra-precision") and leaky, lossy, wax paper capacitors (and those things were sticky and nasty to handle). There weren't any fancy toroid xfmrs, or exotic core materials for ISTs and OPTs. Yet they were able to come up with some damn fine equipment despite component deficiencies.
In the 807 amp, whatever increased distortion was coming from the driver due to decreased Ra was more than compensated by the reduction in distortion coming from the 807s themselves in pure pentode mode operation, which could go from barely noticeable to truly nasty depending on the program material. Connecting the local feedback made a great improvement in bass performance, and all but eliminated that pentode nastiness. Whatever nastiness remained, from whatever cause, was cleaned up nicely by adding 4.0db(v) of gNFB.
Miles Prower said:
Hi Miles,
...the distorsions of the 807 (third harmonics) are more masked by the distorsion of the driver (second harmonics); the result is a more "euphonic" sound.
A NFB of 4 dB is to small, this effect is by feedback degrees of 6 dB an higher.
Best regards Andreas
- Status
- Not open for further replies.