It is there to provide feedback from the output tube's plate to the cathode's driver. Another example here https://www.diyaudio.com/community/threads/output-tube-plate-to-driver-cathode-feedback.356952/
Hi Todd,
the EF184/EL36 schematic from the ETF 06 lecture you are referring to is courtesy of Gregg van der Sluys - I didn´t design, build or measure through it. Gregg used to read & post here, too, but I am not sure he still does. If so, he certainly can elaborate. I showed his schematic because it used "cheap TV tubes" like EF184 and EL36, strapped as triodes. Okay, EF184 / 6EJ7 is not cheap anymore. So, some years later, I used EF95 / 6AK5 as driver instead, and, in general, a different approach, except also using triode stapped EL36. The point of my amp shown above is to show how cheap one can go with a two stage SET using cheap tubes and a rather cheap OPT, by using an uncritical feedback scheme and "hammering flat" its frequency response deficiencies by the very low trioded EL36 plate impedance.
For those who are interested in or don´t know the feedback scheme shown above: Taking the feedback from the power tube anode to a pentode driver cathode in a two staged design was quite common in fine German broadcasting and studio level amps. So, no news here. For example, have a look at the Telefunken V69b schematic. Although push pull, one immediately gets the picture. I presume nobody would accuse Tfk to have utilized that FB scheme because their OPTs were inferior, though ;-)
Regards, Tom
the EF184/EL36 schematic from the ETF 06 lecture you are referring to is courtesy of Gregg van der Sluys - I didn´t design, build or measure through it. Gregg used to read & post here, too, but I am not sure he still does. If so, he certainly can elaborate. I showed his schematic because it used "cheap TV tubes" like EF184 and EL36, strapped as triodes. Okay, EF184 / 6EJ7 is not cheap anymore. So, some years later, I used EF95 / 6AK5 as driver instead, and, in general, a different approach, except also using triode stapped EL36. The point of my amp shown above is to show how cheap one can go with a two stage SET using cheap tubes and a rather cheap OPT, by using an uncritical feedback scheme and "hammering flat" its frequency response deficiencies by the very low trioded EL36 plate impedance.
For those who are interested in or don´t know the feedback scheme shown above: Taking the feedback from the power tube anode to a pentode driver cathode in a two staged design was quite common in fine German broadcasting and studio level amps. So, no news here. For example, have a look at the Telefunken V69b schematic. Although push pull, one immediately gets the picture. I presume nobody would accuse Tfk to have utilized that FB scheme because their OPTs were inferior, though ;-)
Regards, Tom
If you do the 150k to the B+ there is an extra resistor+capacitor needed for feedback.
And with an S=5mA/V of the EF95 the feedback to the cathode is rather small.
There is no problem with gravity here. The Rp of the EL36 triode is less then 1kΩ and I suspect a OPT of 4kΩ. So, accoding to the formula you showed 5W out is possible.
Mona
Attachments
Hi Ketje,
measured rp is somewhat less than 900 Ohms at the operation point shown, reflected impedance is about 4500 Ohms. Even assuming an unusually bad efficiency factor of the cheap OPT it is possible to get there, indeed.
Kind regards, Tom
measured rp is somewhat less than 900 Ohms at the operation point shown, reflected impedance is about 4500 Ohms. Even assuming an unusually bad efficiency factor of the cheap OPT it is possible to get there, indeed.
Kind regards, Tom
Before somebody objects, this is not exactly the operation point from the schematic shown above, but it is close enough:
That is not true , as in your "example" G2 is at +115V fixed Voltage and the feedback goes just to the cathode .
EF95/6AK5 pentode driver stage operation point of the schematic shown above. Don´t underestimate these sh*tty & cheap RF and IF strip tubes, although you most likely will have to select them for microphony. Another interesting candidate is EF80 / 6BX6, used in gazillions of European TV and radio IF strips.
I linked that example to show the plate to cathode feedback principle. At DC both examples are "different", at AC they operate the same. I liked seeing Tom's solution very much, as it cleverly provides the G2 voltage and the feedback with one resistor. (One could also use a resistor from the El36 straight to the driver's cathode and a separate resistor from B+ to G2).
Above was the question? It is supplied by the anode of the output tube so that the swing at the output tube's plate is sent to the driver's cathode, for feedback.
OK, I see what that 150K resister is doing now. The FB factor looks like 0.001 if the cathode impedance of the EL95 is ignored.
Plugging that into the formula for NFB I got a gain reduction from 225 to 184, 1.76 db.
That reduces the internal impedance at the plate connexion to the OPT primary, an advantage.
But it also makes the output stage more susceptible to power supply hum & noise.
The same degree of NFB from the secondary would improve both those problems to a degree.
I've used up to 20 db NFB on FB pairs such as this with no stability problems at all.
Even where the OP stage is a pentode & the OPT a replacement type.
Seely's power output formula assumes better than 5% distortion for the triode SE power amp.
Looking at the very long loadline in the 'poor man's 300B' we can see a large difference in the grid spacing
at the extremes, a sure sign of distortion.
Plugging that into the formula for NFB I got a gain reduction from 225 to 184, 1.76 db.
That reduces the internal impedance at the plate connexion to the OPT primary, an advantage.
But it also makes the output stage more susceptible to power supply hum & noise.
The same degree of NFB from the secondary would improve both those problems to a degree.
I've used up to 20 db NFB on FB pairs such as this with no stability problems at all.
Even where the OP stage is a pentode & the OPT a replacement type.
Seely's power output formula assumes better than 5% distortion for the triode SE power amp.
Looking at the very long loadline in the 'poor man's 300B' we can see a large difference in the grid spacing
at the extremes, a sure sign of distortion.
I start from the premise that the television tubes allow an excess of dissipated power without affecting them too much so i set a working point over the dissipation hyperbola with 8w situation in which i managed to get 8w with1% THD and only at 9,4w the distortions suddenly increase. It probably be useful to simulate this scheme for which i think the data is sufficient. My experience with these tubes makes me belive that they last a few years in operation and are very cheap anyway
Attachments
I am planning to build an el 36 se amp. Your amp looks great. Found a couple of schematics on the internett. Is it this one you have built your amp after? https://www.diyaudio.com/community/threads/pl36-triode-se.276028/page-2
popa marius:
Your findings are in line with my experience, EL/PL 36 can take some abuse. What is the primary impedance of the output transformer you are using?
I will try the feedback scheme (Anode of PL36 to G2 of EL183) in my amplifier.
Happy New Year!
Peter
Your findings are in line with my experience, EL/PL 36 can take some abuse. What is the primary impedance of the output transformer you are using?
I will try the feedback scheme (Anode of PL36 to G2 of EL183) in my amplifier.
Happy New Year!
Peter
Hi Peter,because I have an OT with multiple sockets in the secondary and primary I can accurately determine the impendance of the optimal OT depending on the static point of operation of the tube so for the variant tested by me it resulted 4 Kohm . The Schade type reaction is very useful but I recommend using ARTA software for optimization
HAPPY NEW YEAR !!
Dear Tom,
I am going to build up a PL36 based SE amplifier. I did like very much your one, which diagram is shown in Post #23, especially for the anode feedback. For reason of availability, I would like to use another driver tube, the E80F (very similar to EF86) instead of the EF95.
My problem is that I am not so good in designing. Could you please suggest which changes should I do for adopting that tube?
And could yo please provide the type of the output transformer, or at least its primary winding impedance?
And, did you make some modification in the meantime?
Thank you very much in advance for your help.
Kind regards
Giovanni
I am going to build up a PL36 based SE amplifier. I did like very much your one, which diagram is shown in Post #23, especially for the anode feedback. For reason of availability, I would like to use another driver tube, the E80F (very similar to EF86) instead of the EF95.
My problem is that I am not so good in designing. Could you please suggest which changes should I do for adopting that tube?
And could yo please provide the type of the output transformer, or at least its primary winding impedance?
And, did you make some modification in the meantime?
Thank you very much in advance for your help.
Kind regards
Giovanni
Sample E80F VAS stage.
If you want to implement it to #23 post schematic, must to fit R6 (now 1Meg) value as R-GK feedback resistor to reach suggested G2 voltage on EF80 tube.
If you want to implement it to #23 post schematic, must to fit R6 (now 1Meg) value as R-GK feedback resistor to reach suggested G2 voltage on EF80 tube.
Dear Giovanni,
unfortunately I am away from home for some days and only have limited access to my stuff. So I can only give some thoughts to consider at the moment.
Kind regards, Tom
unfortunately I am away from home for some days and only have limited access to my stuff. So I can only give some thoughts to consider at the moment.
- EF86 and similar tubes usually are operated at much lower currents (and much higher anode resistor values) than f.e EF95. One should at least check if the output impedance of an EF86 driver stage is low enough to drive the Miller capacitance of the power stage to the rather wide swing required without slewing by triode strapped EL36. EF95 at the given operation point certainly does.
- To start with, EF95/ 6AK5 is extremely linear at the given operation point and required wide swing even before application of any FB (see post #67). Also, designed and used in millions as an RF / IF strip tube it still is cheap and plenty
- The OPT was a rather cheap and simple one; actually this amp was a testbed for it to show that the very low plate impedance of triode starpped EL36 helps in "linearizing" it. IIRC reflected impedance is about 5k.
- I didn´t do any mods to the schematic as shown above. I still use this amp about 100 almost daily since years.
Kind regards, Tom