Eddie, it is not about a sine tone on full power. It is about slew rate limitation by an output stage that causes clipping by a driver stage and other nasty processes like bias shift due to grid currents of output tubes pumping coupling capacitors. That's why a Zobel network right on anode of the 1'St tube is what a doctor prescribed. It slows down both signal and a feedback. Just play with it to get the amp unconditionally stable, that's it.
famousmockingbird - sorry I don't know your name;
I am attaching one of versions of my Pyramid amp driver. As you may see, anode load of the 1'st pentode (6F12P, 19 mA/V) is 18K, Zobel network is 360 pF, 1K. From anodes of output Gu-50 tubes add 240k resistors to anodes of driver tubes.
Transformers were 100W Edcor, 10K P-P, and 20 Hz on full power. I don't think they were better than yours. The amp was stable and slightly ringing on around 35 KHz because of transformers, but frequency response from 20 Hz to 20 KHz was flat.
famousmockingbird - sorry I don't know your name;
I am attaching one of versions of my Pyramid amp driver. As you may see, anode load of the 1'st pentode (6F12P, 19 mA/V) is 18K, Zobel network is 360 pF, 1K. From anodes of output Gu-50 tubes add 240k resistors to anodes of driver tubes.
Transformers were 100W Edcor, 10K P-P, and 20 Hz on full power. I don't think they were better than yours. The amp was stable and slightly ringing on around 35 KHz because of transformers, but frequency response from 20 Hz to 20 KHz was flat.
Attachments
This data is false. I have to redo the charts.......I found my signal generator's amplitude decreases as frequency increases
I did put some square waves in and they look a little better, less ringing at 1khz at 20w.Sorry about that
Have you tried different output transformers for your amp? If so were the results better?
Last edited:
What do you mean by "better"?
The result was already great. The amp drives hard to drive speakers easy. Fools imagination, as if sounds are real. However, speakers have to sound clean for that, but with 80 W per channel it is easier to get nice full range speakers, even if their sensitivity is low, than to construct something extremely expensive for 8 watt per channel.
Ringing is an indication of a marginally stable amplifier. It may or may not be due to the iron.
Cheers
Ian
The requirement for conditional stability is not 'lacks a dominant pole' but 'loop gain locus wiggles around the (-1,0) point in such a way that a reduction in gain will enclose the point'. Most amplifiers, including unstable ones, are sufficiently simple in their gain/phase behaviour that they do not exhibit conditional stability. In any case, dropping forward gain and open loop gain by the same factor will always preserve loop gain unchanged so the stability will be unchanged.Alan0354 said:
Sometimes, conditional stability is accidentally achieved and this shows itself by bursts of oscillation either on signal peaks or when powering up or down.
The 5-20 design does suffer from a minor problem, in that the LTP phase splitter biases the ECC83/12AX7 right on the edge of grid current - especially if the anode voltage gets too low (caused by the grid voltage being too high). A bit of grid current through the 1M resistor then unbalances the LTP. This means that the anode voltage of the first stage must sit in the right region, despite valve tolerances and ageing.famousmockingbord said:
I don't use root locus plot, only Bode plot. I am not talking about conditional stable. As long as you have single pole 0db crossover, it's stable. If you don't change the loop gain ( open loop gain minus closed loop gain), it will not affect stability. But if you change open loop gain without changing close loop gain or vise verse, then there is no guaranty.
But anyway, OP is not interested in this judge from the later post, so it's no point of talking about this
I suspect that conditional stability is harder to spot in a Bode plot, but it would be quite obvious in a root locus plot.Alan0354 said:
You are, whether you realise it or not. An amp which goes unstable when forward gain is reduced is conditionally stable.Alan0354 said:
There is a guarantee if the amp was unconditionally stable; you can be sure that reducing loop gain will not create instability.
Perhaps so, but it was you who introduced this topic by saying something which needed to be corrected (so newbies are not misled).
I know all this too well. The limited power bandwidth is a consequence of slew limiting.
There are many workarounds, including Zobel and Miller compensations. It makes solid state amplifier design seem like a no brainer by comparison. Even a modest chip amp blows away a well sorted tube amplifier, on paper.
But we still love our tube amps. Subjectively, they can sound cleaner and more powerful than a solid state amplifier. I think there's several reasons for this; the distortion profile (with and without global feedback) being subjectively cleaner is the big reason I think. And some of us grew up when the typical household had some kind of tube audio system; typically one of those big consoles. Many of these consoles were marketed as "hi-fi" (owner's manuals were peppered with the phrase "STEREOPHONIC HIGH FIDELITY") when in reality all they provided was room filling sound with "good tone." Modern "hi-fi" did begin in the 50s but it was embraced only by quirky tinkerers (like us) and was utterly impractical by any reasonable standard. It wasn't just the big, watt sucking space heaters we called amplifiers that made them unappealing to typical consumers; but speaker technology was still in the Fred Flintstone era until the early 70s - big bass meant huge speakers; no way around it.
Yet the state of the art still advances. McIntosh never stopped pushing the envelope, and their products still fill a niche market.
Bode plot can spot what you called conditionally stable. I don't know your term of conditional or unconditional stable. If you refer to an open loop plot that you can increase or decrease the open loop gain while keeping the closed loop gain constant or keeping open loop gain constant and change the closed loop gain. The ONLY way to have it unconditionally stable is IF you have a single pole role off in the whole open loop gain. That is a dominant pole compensation.
Bode plot served me very well for all the closed loop feedback systems I designed. Very few system can have a simple dominant pole. It is very obvious that OP original circuit has more than one pole already, or else it will not oscillate on the first place. So your talking about unconditional stable is not practical.
From my experience, if I design with single pole roll off ( dominant pole to guaranty what you called unconditional stable), performance is not optimize. Slew rate( which govern rise time) and settling time suffer.
Last edited:
Oh well, I figured it would be a fun project to build a classic tube amp, boy I was wrong. I apparently lack the intelligence and proper test equipment to figure this out even with all your help. I think if I could afford a proper signal generator and scope it would be a little easier on me but my computer based gear isn't cutting it. Sorry for wasting everyone's time.......I will just put them in the closet until I am less of a "newbie".
Go to surplus store in your area and see whether you can find a scope really cheap. Invest a few hundred dollars and it's very rewarding. You plot out the open loop gain graph, people here can help you. It's fun. I've been an EE for almost 30 years, still my passion after retirement.
Sorry, there's nothing simple about this. I am still learning. Stability is just something I have a lot of working experience. But the rest of the audiophile, I am just as green.
Ha ha, a year ago when I decided to get into guitar amps, I thought how hard can this be, it's only audio frequency. Was I wrong!!! I spent so much time on it that my wife actually said if I am so interested, go get a job again, at least I get paid!!!!
Last edited:
I mainly dabble in guitar amps, I like that there is more of an art to it because it is suppose to add something to the sound where as Hi Fi isn't. Do you play guitar or study music?
Playing music is the greatest hobby...........maybe I should just stick with that and be grateful
In the meantime I will keep my fingers crossed I can find somebody willing to let me use their test equipment. It sounds stupid but I don't have a steady income so it was probably pretty dumb of me to spend a grand on parts to build a stereo, my priorities are pretty messed up I guess lol.
I mainly dabble in guitar amps, I like that there is more of an art to it because it is suppose to add something to the sound where as Hi Fi isn't. Do you play guitar or study music?
Playing music is the greatest hobby...........maybe I should just stick with that and be grateful
In the meantime I will keep my fingers crossed I can find somebody willing to let me use their test equipment. It sounds stupid but I don't have a steady income so it was probably pretty dumb of me to spend a grand on parts to build a stereo, my priorities are pretty messed up I guess lol.
I played seriously for 10 year back in the 70s, but I quit since. I am designing guitar amp as a hobby.
The most important thing is whether you are interested in electronics. I started out not satisfied with my Twin Reverb in 76, so I started dabble into modding. I did not know electronics at the time. I even designed a power scaling in 78 using variac to control just the power stage and the PI stage. It worked beautifully. But at the process, I found my true calling.....electronics. I was so into electronics that I quit music and study, become an EE in 1980 and never looked back. After almost 30 years, electronics is still my first passion. That's the reason I go full circle and started working on guitar amp last year.
My point is you never know what you might end up. I thank my guitar playing that started my career of electronics. So, if I were you, ask whether you have the interest in electronics. If so, it's not wrong to invest a little in equipment. I did, I was so poor those days, I went to the flea market and bought my first scope!!!! Let's face it, you really cannot make much money playing music, don't look at the one that have big names and earn mega bucks, look at to every one successful player, tens of thousands did not make it and have to have a day time job to support their music. I made my transition from music and I'll do it over again.
I mainly dabble in guitar amps, I like that there is more of an art to it because it is suppose to add something to the sound where as Hi Fi isn't. Do you play guitar or study music?
I response to this in separate post because I don't want to get too long on the other one. I know I'll be rebutted on this big time. But I question about people just want minimum distortion in hifi.
You play music, you should know, the sound in CD or any recording of the highest quality sounds NOTHING like the live music playing without going through mic and PA. The biggest distortion is the mic, the recording and the speaker that play the music. If anyone can actually reproduce 100% fidelity, I am not sure people will like it. Listen to the Sax, violin, human voice and drums without going through the mic, they are harsh, cracking and very raw. I remember I went to listen to Paul Mauret? in the late 60s( the one play Love Is Blue) because I like the music so much at the time. The violin sounded silky smooth in the record. In the live performance, the violin sounded harsh, not that nice sounding. Fidelity to me is an illusion. People need to listen to live music without mic and see. Maybe that's the reason people mic all the instruments since the 70s and let the PA reproduce the sound. then it sounds like when you listen to the recording.
To me, I still believe coloring in the right way enhances the music since you don't have true fidelity. So back to your point, I feel there are still some arts in designing audiophile electronics. Like speakers, in audiophile, that is the biggest factor and it's the biggest source of distortion. Or else, why every brand sounds totally different. If there is no distortion, all speakers sounds the same!!! People can say fidelity as much as they want, ultimately, it's how it sounds to them and it's all subjective......just like guitar amp.
Last edited:
Yes this topic always comes up and people will argue about it till the end of time. I have no problems with coloration or "effect boxes" in the home stereo. If it sounds good to you then who cares about numbers. Engineers like numbers though.
I haven't played guitar through a solid state amp that felt or sounded as good as a tube amp.
I haven't played guitar through a solid state amp that felt or sounded as good as a tube amp.
For a guitar amp you should think more about how it sounds when overdriven than about stability and fidelity.
Speaking of home stereo, for me coloration is the one obstacle that I am getting rid of. Learning to design guitar effects and analog synthesizers I better realized how to make distortions disappear, i.e. what to fight with, and what to leave alone because some errors are audible, other errors are not.
Speaking of home stereo, for me coloration is the one obstacle that I am getting rid of. Learning to design guitar effects and analog synthesizers I better realized how to make distortions disappear, i.e. what to fight with, and what to leave alone because some errors are audible, other errors are not.
All I am saying is "fidelity" is only what people define in their mind. The real raw music sounds very different from any of the recording. If people really want fidelity, then none of the recorded music even come close. So, it's better to talk about what pleases you rather than talk about no distortion because what you hear in any system are way distorted from the real music.
Don't believe me, go listen to live music particular performance that doesn't even go through the PA. Even simpler, if you have someone that sing good enough. Have the person sing straight to you without mic, then sing through a karaoke machine, you'll hear the voice sound so different. To me, fidelity is if you can reproduce the raw voice of the person or the instrument.
As for me, I just want a system that pleases me, distortion or not.
Don't believe me, go listen to live music particular performance that doesn't even go through the PA. Even simpler, if you have someone that sing good enough. Have the person sing straight to you without mic, then sing through a karaoke machine, you'll hear the voice sound so different. To me, fidelity is if you can reproduce the raw voice of the person or the instrument.
As for me, I just want a system that pleases me, distortion or not.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.