It was a long time i made this one and i know some that made it also. Works perfectly. Its not a mistake in 68000uF. you can use 6800uF (and no need for 500V 
) Big value was offered by one audiophile..
Mine looks like that:
) Big value was offered by one audiophile.. Mine looks like that:
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
It is very SIMPLIS ...
An externally hosted image should be here but it was not working when we last tested it.
I attach to this post my personnal symbols and models library if it can help ...
Alain.
The mistake is using big capacitors like that for absolutely nothing ...It was a long time i made this one and i know some that made it also. Works perfectly. Its not a mistake in 68000uF. you can use 6800uF (and no need for 500V
It is a total nonsense and like I said in my last yesterday post, this schematic suffer from many problems. It can be improve a lot using a 2500 ohms load, a 195V Vak and -70V bias to get more output power and less distortion ...
You have to take care with the schematics found on the net, anybody can have a web site to show their schematics, sometimes they are good and sometimes they are very bad ... Personally, I never assemble a circuit found on the net, I just look at them to get some new design ideas and calculate mines myself.
When they don't mention exactly what transformers brands and models to use, it is already suspect to me because this change many importants things : Voltages, currents, bias, power, distortion, etc ...
Alain.
P.S. In this schematic, all supply filter capacitors must have a working voltage of 500V or more, the ECC88 bypass capacitor can have a 10 or 16V working voltage but the 6S19P one must have a minimum 250V working voltage ... This is a very serious warning, a big capacitor can explode violently if the DC voltage across is too high ...
Oh, i found mine different. Not the same circuit actually.
Anyway, like it was said, after building a circuit, it has to be measured and calculated. Its not enough to connect and just play without any adjustments. Power supply must be a diode rectifier, but not a tube. Too big cap values.
An externally hosted image should be here but it was not working when we last tested it.
Anyway, like it was said, after building a circuit, it has to be measured and calculated. Its not enough to connect and just play without any adjustments. Power supply must be a diode rectifier, but not a tube. Too big cap values.
Yes, this one is very different, not a direct coupling ... A capacitor coupled amplifier like that don't need a very high supply voltage ... I can see the power tube is not a triode but a triode connected tetrode ...
But, the rectifier don't have to be a silicon rectifier, most audiophiles prefer a good rectifier tube like the 5Z4 or 5U4, they said the sound is better ... Of course, the transformer must have a higher voltage because the voltage drop in a rectifier tube is much higher than in a silicon rectifier (0,6V) ...
Alain.
P.S. I am not really an audiophile myself, I prefer deep silence at night time ... What I like is electronic circuits design and maths, not really building them ... I am not just interrested in audio electronic but all kind of circuits, power control, microcontroler and microprocessor, industrial automatisation, electric motor control, etc ... I work 15 years as professionel PCB designer and schematics drawer but I return to school 3 years ago to get an electromechanic diploma to try getting a good job but unfortunatelly, I get it at the same time this terrible crisis begin ...
You have to play around a lot with it, try everything you see ... I download mine about 6 month ago and it took me many days just to understand how to create my own symbols and models library and I still discover new things every week ...
This software is very powerfull and have almost unlimited possibilitys but it takes months to learn how to use it really efficiently, like for any professional design software.
Alain.
Pentode, actually.
http://www.magictubes.ru/sprav/pdf/6f5p.pdf
Surprise, but I have not seen such audiophiles here. I know some diy-audio folks who say that, also they prefer paper in oil capacitors because electrolytic capacitors are actually oxide-semiconductor capacitors. It is not kosher to use semiconductors for them.
First, I have to said I under evaluate the maximum power of the original russian schematic in my first post ... With a 5K transformer, the power can reach 2,5447W with my 6S19P Spice model and a Edcor CXSE25-8-5K output transformer ...
My simulator calculate a 3,5568% total harmonic distortion with a 0,9 second 1Khz sinewave sample at full power but this can be improved like you will see in my next posts ...
The "normal" maximum power is reach just before the signal between the grid and the cathode cross the 0V bias ... Over that, the grid start to pull current and if the driver cannot supply it, clipping occur and the distortion raise fast.
I use the SIMetrix "differential voltage probe" between the grid and cathode to ajust the input signal to get the maximum output power.
Alain.
An externally hosted image should be here but it was not working when we last tested it.
My simulator calculate a 3,5568% total harmonic distortion with a 0,9 second 1Khz sinewave sample at full power but this can be improved like you will see in my next posts ...
The "normal" maximum power is reach just before the signal between the grid and the cathode cross the 0V bias ... Over that, the grid start to pull current and if the driver cannot supply it, clipping occur and the distortion raise fast.
I use the SIMetrix "differential voltage probe" between the grid and cathode to ajust the input signal to get the maximum output power.
Alain.
So, the 6Fxx series russian tubes mean "triodes/pentodes" ...
I am very interrested using some russian tubes, I should begin studying how to read russian texts because "google translate" give very funny results sometimes ...
So, the 6Fxx series russian tubes mean "triodes/pentodes" ...
I am very interrested using some russian tubes, I should begin studying how to read russian texts because "google translate" give very funny results sometimes ...
Do you mean Google Translator is needed to translate tetrode picture to pentode?
You should count grids. Tertode has 2 grids, pentode has 3 grids.
The curves graphics with my load lines on the preceding page suggest the power will be higher with a 2500 ohms load and that is true. But it also suggest the distortion is lower and that is not true ... In my last circuits simulations, the distorsion is higher at full power but there is some easy ways to reduce it ...
My first simulation with a 2500 ohms output transformer :
The 3,4W maximum power is much higher than with the original schematic using a 5000 ohms transformer but the distortion reach 7,36% and is higher at any power.
Since the 6DJ8 bias was near 4V, I try a LED bias instead of a bypassed cathode resistor and the distortion is about the same. But this is simple and much cheapper than a good bypass capacitor ... Also, I prefer not using any capacitors in the sound path when it is possible.
The LED is actually a "super bright bluish white" you can found in the "made in China" flashlights sold in the "1$ stores" ...
I experiment many negative feedback to reduce the distortion and just adding a simple Schade's local feedback resistor (R4) work very well ... I arbitrary use a 150K resistor and the distortion drop under 6% for about the same output power with a slightly higher input signal.
My last version use a much heavyer Shade's feedback, a 68K resistor, notice the 6DJ8 DC load resistor have to be adjust to get about the same voltage at the plate, then the same bias for the 6S19P ... The distortion drop near 5% at the full 3,428W output power with just 2,519V RMS input signal ...
The distortion drop very close to the original schematic at 2W but the great advantage is this version is more powerfull ... Edcor also make a good quality 3K transformer in the CXSE series, it is in between 2,5K and 5K ... I will also try this last circuit version with this transformer to see what are the differences.
They also made a 1250 ohms transformer that can handle a 200ma DC current, I will try it with two parallel 6S19P to have double power ... Around 7W ...
Alain.
My first simulation with a 2500 ohms output transformer :
An externally hosted image should be here but it was not working when we last tested it.
The 3,4W maximum power is much higher than with the original schematic using a 5000 ohms transformer but the distortion reach 7,36% and is higher at any power.
An externally hosted image should be here but it was not working when we last tested it.
Since the 6DJ8 bias was near 4V, I try a LED bias instead of a bypassed cathode resistor and the distortion is about the same. But this is simple and much cheapper than a good bypass capacitor ... Also, I prefer not using any capacitors in the sound path when it is possible.
The LED is actually a "super bright bluish white" you can found in the "made in China" flashlights sold in the "1$ stores" ...
An externally hosted image should be here but it was not working when we last tested it.
I experiment many negative feedback to reduce the distortion and just adding a simple Schade's local feedback resistor (R4) work very well ... I arbitrary use a 150K resistor and the distortion drop under 6% for about the same output power with a slightly higher input signal.
An externally hosted image should be here but it was not working when we last tested it.
My last version use a much heavyer Shade's feedback, a 68K resistor, notice the 6DJ8 DC load resistor have to be adjust to get about the same voltage at the plate, then the same bias for the 6S19P ... The distortion drop near 5% at the full 3,428W output power with just 2,519V RMS input signal ...
The distortion drop very close to the original schematic at 2W but the great advantage is this version is more powerfull ... Edcor also make a good quality 3K transformer in the CXSE series, it is in between 2,5K and 5K ... I will also try this last circuit version with this transformer to see what are the differences.
They also made a 1250 ohms transformer that can handle a 200ma DC current, I will try it with two parallel 6S19P to have double power ... Around 7W ...
Alain.
Come on, I know what a pentode and a tetrode are, I work with tubes since 40 years, I design and build my first tubes amplifiers when I was 16 years old and sold most of them to some friends ...
The 6F5P power part is not really a pentode (like a 6CA7) but a "beam power" (like a 6V6 or a 6L6) , it is slightly different ... But most american databooks (like my olds RCA ones) show them like a pentode ...
The one on your schematic show a tetrode ... There is not too many real power tetrode (like the 12K5) on the today market, tetrode are a very old tube type and not very used anymore mainly because the "kink" in their curves and very often bad linearity ...
Of course, when I talk about learning russian reading, it is to be able to read their datasheets specifications, recommendations and warnings ... This is the most important when working with tubes ...
I also have a lot of only chinese and japanese tubes and transistors datasheets on my hard disc drive but I will start with russian, it is much easyer to learn ...
Last edited:
Quite a few of the newer Russian data sheets are bi-lingual with English as the second language. I'd try to find those first and then nicely ask any native tube savvy Russian speaker to translate for you if it proves necessary.
My other comment is not to put too much stock in distortion tests using spice models for pentodes and beam power tetrodes, there are known issues with the models such as G2 behavior at low voltages, note also that many of these models are made by curve fitting, and the model is only as good as the fit and that is assuming that the curves themselves are really good.
My other comment is not to put too much stock in distortion tests using spice models for pentodes and beam power tetrodes, there are known issues with the models such as G2 behavior at low voltages, note also that many of these models are made by curve fitting, and the model is only as good as the fit and that is assuming that the curves themselves are really good.
I use the distortion simulator evaluation just to compare different circuits, of course, it is not exactly the real distortion ... It also depend on how long the waves sample is, I alway do the test on 1000 cycles. SIMetrix calculate the distortion using the Fourier series calculation. Looking at the wave shape also give a good idea on what's going on in the circuits. If there is a bug in the simulation, there will be probably one in the real life ...Quite a few of the newer Russian data sheets are bi-lingual with English as the second language. I'd try to find those first and then nicely ask any native tube savvy Russian speaker to translate for you if it proves necessary.
My other comment is not to put too much stock in distortion tests using spice models for pentodes and beam power tetrodes, there are known issues with the models such as G2 behavior at low voltages, note also that many of these models are made by curve fitting, and the model is only as good as the fit and that is assuming that the curves themselves are really good.
With my SIMetrix simulator, I always generate the curves of the tubes and transistors to compare them to their datasheet curves and triodes usually fit very well. But almost all the pentode models I found was really off the datasheet curves at some place. I think they make them precise just around the usuals operation points area but this often give a poor idea of the real signal swing and output power.
I really like to learn how to make pentodes and transistors spice models, do you know any software to do that ?
I often use the Andrei Frolov "Curvecaptor" program but it work only with triodes ...
Thanks,
Alain.
Alain,
Look here
6F5P SET Amplifier
I built this from scrap parts and it sounds pretty good. There are several pages of results.
He also has several java based tools to find tube parameters - look here:
Model Paint Tools: Trace Tube Parameters over Plate Curves, Interactively
Cheers
John
Look here
6F5P SET Amplifier
I built this from scrap parts and it sounds pretty good. There are several pages of results.
He also has several java based tools to find tube parameters - look here:
Model Paint Tools: Trace Tube Parameters over Plate Curves, Interactively
Cheers
John
It is very strange because I never saw 6F5P datasheet that describes it as a ray tetrode.
European analog ECL85 is triode-pentode as well: ECL 85, Tube ECL85; Röhre ECL 85 ID18599, Triode-Pentode
Well, this is a very basic amplifier, there are hundreds like that on the net ... This can be done with almost any triodes and pentodes (triode connected or not). It is very easy to design one like that ...
The circuit with the 6S19P I am studying is very different because it is a direct coupling one and I am allergic to capacitors, they are a real pain for the sound ... Low working voltage and high bias power triodes and high output voltage swing triode drivers are perfect for this type of circuit because the supply voltage can be low ... If you try to make the same kind of circuit with a ordinary power pentode, the supply voltage usually have to higher ...
I think the Curvecaptor for Windows XP program from Andrei Frolov is much better and userfriendly than this one but it can be use only for triodes ...
But those Java scripts will be usefull for me because they contain all the mathematical codes to create the tubes models, triodes but there is also a link for the pentodes and that's what I am looking for ... I don't like having to install codes interpreters to run those programs, I prefer real ".exe" Windows program.
I got all the C and C++ compilers for DOS, Windows 3.1 and 95 home and I use to work with them many years as analyst programmer for electronic companys, so, I can incorporate those mathematical Java codes to a "real" Windows program ... All old Windows versions program can run very well on the latest Windows operating system, there is no problems with that.
I have a quick look the Model Paint Tools Web site and they have a different approach than the Andrei Frolov program. The user have to move the parameters adjustment cursors until the generated curves on top of the datasheet curves image fit and the job is done. I will look at those codes closely when I have more spare times, maybe next year.
Thank you very much for the link.
Alain.
There is nothing strange about that ...
It is true that many peoples call the "beam power" tubes "beam power tetrode" but they are not "tetrode" or "pentode" like most people think when they see the "beam power" symbol in some american tubes datasheets.
Personnally, I just call them "beam power tubes" like in the tubes databooks, it is less confusing that way ...
An externally hosted image should be here but it was not working when we last tested it.
"Beam power tetrode" is a very old appellation, I have a 1956 RCA tubes databook home and they already call them "Beam tubes" or just "Beam power" ... This image come from the '30 tubes electronic books ...
Alain.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.