I mean that the electric field E is proportional to applied voltage, then hysteresis loop will be thicker and losses bigger, regardless of capacitor type, but totally agree with you that for ceramic and electrolytic capacitors the effect will be much more noticeable.
Totally agree, curiously, the best PP capacitors I found, was on old Grundig and Telefunken TVs, also could see very often small 63 V, electrolytics quality was also amazing, and you can found more than thirty years old electrolytics like new and still working !
Exactly, and that is the key of my reasoning, as I said before, seems to me that you was the only capable to understand it, even more, you said it better than I did.
Let me relate my experience, once upon a time, existed extremely good Siemens capacitors, you could buy them with closed eyes, later Siemens did a fusion with Matsushita, only for capacitors, later Siemens launched Epcos brand, still very good, but now Epcos was buyed by TDK, and datasheets are a little bit deceitful, them refer tan(δ) up to 100 nF @ 1 KHz and 100 KHz, but beyond 100 nF only at 1 KHz, so let me distrust.
Totally agree, seems that 6CG7 is a 6SN7 in a different bottle in regard for linearity, even my next project includes one, but, instead of with ECC82 you cannot go too crazy with applied +B.
Agree, but I have another explanation, but it is weird and more later we can speak about it.
Well, I'm not a big fan of mu-follower, surely I am very prejudiced, but two reasons
1) AC coupling is prone to blocking.
2) Enormous resistors are noisy.
That would be great, however, I can't buy decent instruments, even if I sell my old car, it cost less than a Fluke multimeter !
have you built the amp you wanted to improve or still busy writing equations?
Yes, I did, and not only one, and I can assure you that my speculations really work in real world.
As you would know, I'm poor and I have no a digital camera, so the only pictures that I have, taken by my cousin, I already sent to my friends of the forum, sometimes I have the impulse to post them on the sticky, however they are commercial designs and I won't to be petty if people ask for schematics, so, because I don't need to show up, simply I don't post them.
By the way, if you want I can send them to you by email.
it this forums there are builders, folks who actually built something and then there are talkers, people telling other people what to do, what to think and yet themselves have nothing to show in terms of something they actually built....
In this forum also there are designers, folks who actually research and develop something, and want to share it with the community.
Every time that I share something, I am fiercely attacked, mostly with false assumptions/misunderstandings.
However I don't think be myself a charlatan, even less I want to tell people what they must think, people can think what they want and I don't care, after all this is a public forum, isn't it?
By my calculations, I think double than that, about 6K, but maybe I'm wrong.
Almost nothing. IMHO
Yes the higher E the ticker hysteresis loop but the nominal values given basically make you stay away from this, at least for quality capacitors.
Forget Epcos and similar brands. Use Solen/SCR capacitors. These are still fairly accessible in terms of cost and up to 1uF losses are the same, for higher capacitance values losses are only slightly worse but nothing comparable to electrolytics and ceramics.
About blocking, I assume you are referring to the upper triode mainly. It can be an issue when you reach the grid current. If you make sure that for the max Vout you want grid current is negligible then it is fine. This might need some real valve selection.
Otherwise you can make an hybrid mu-follower using mosfets although these have their own "issues" as well. This solution is probably even better. Ask Ale (Moglia) on this forum he has done a really a great job with the 46 driver. You can see here: 46 triode-strapped | Bartola Valves
Forget Epcos and similar brands. Use Solen/SCR capacitors. These are still fairly accessible in terms of cost and up to 1uF losses are the same, for higher capacitance values losses are only slightly worse but nothing comparable to electrolytics and ceramics.
About blocking, I assume you are referring to the upper triode mainly. It can be an issue when you reach the grid current. If you make sure that for the max Vout you want grid current is negligible then it is fine. This might need some real valve selection.
Otherwise you can make an hybrid mu-follower using mosfets although these have their own "issues" as well. This solution is probably even better. Ask Ale (Moglia) on this forum he has done a really a great job with the 46 driver. You can see here: 46 triode-strapped | Bartola Valves
Do we use the same formula?
An externally hosted image should be here but it was not working when we last tested it.
Those are unobtainable here, what do you think about WIMA?
Yes, I mean that.
Although I dissimule it, I am quite dumb, so a more forgiving topology like SRPP is ideal to me and my knowledge, not to mention my prejudices.
What about this?
http://www.diyaudio.com/forums/tubes-valves/207544-ecc82-12au7-line-preamp-18.html#post3172658
I myself visit that page quite often, it's great, technically and aesthetically, even cartoons !
Here we call Ale to people named Alejandro, so I conjectured that Ale Moglia could be Argentine.
If so, I have the suspicion that we are like a plague on the forum.
Edit: Here we have this saying to mean that we are a lot of people.
Last edited:
I don't think so, when I am too lazy to struggling with Herr Vogel notation, I use this
http://www.google.com.ar/url?sa=t&r...f8956oWDbut0ebg&bvm=bv.66330100,d.cWc&cad=rja
But keep it quiet...
no one is calling you that, and i was not referring to you in particular.......just that there are posters here who are hard on newbies, argues too much, and when challenged to post their builds come up with lame excuses, really nothing to show for it in comparison with what they posted....
I am sorry about your financial situation, but you're not exactly playing with a cheap hobby either. It is expensive to build your own equipment and expect it to perform as well as commercial products. Sure, you can build junk, but who wants that? I would suggest that you find a vintage low cost preamp of some kind locally and modify it instead of trying to build from scratch.
no one is calling you that, and i was not referring to you in particular.......just that there are posters here who are hard on newbies, argues too much, and when challenged to post their builds come up with lame excuses, really nothing to show for it in comparison with what they posted....
OK, cleared, no problem.
It is not my purpose to victimize myself, but explain why all of you can't see pictures of my amps/preamps.
I built my first amp when I was fifteen, since then I couldn't built my second one for myself, and I am still building it, very very slowly.
I never did an amp/preamp from junk, and I would never do that, so my solution is the OPM method: Other People Money, i.e. make audio equipment for customers under request.
I have little or no money to invest on research and development, so my cheap solution exposed on this kind of threads is simple: paper and pencil...and a lot of equations.
It is very useful to me, and my hope is that would be also useful to others, so I share it.
Without any intention to offend you, I would suggest you that make your amps/preamps as you want and let me do mine as I can.
That's my source too (see page seven) and I tripple checked the math but I'm still at 3K, not 6K.I don't think so, when I am too lazy to struggling with Herr Vogel notation, I use this http://www.google.com.ar/url?sa=t&r...f8956oWDbut0ebg&bvm=bv.66330100,d.cWc&cad=rja
But keep it quiet...
the numerator= 180K
the denominator=60K
result: 3K
for
Ra= 3.000Ω
Rk1= 1.500Ω
Rk2= 1.500Ω
μ= 16
Seems to me that your Ra data is wrong.
http://frank.pocnet.net/sheets/127/1/12AH7GT.pdf
http://frank.pocnet.net/sheets/049/1/12AH7GT.pdf
I did my calculation with Ra≈6K6
Alright.... is the term referring to "anode resistance"? Seems plausable coming to think of it
Reading the text under Fig. 6 (page 7) I understood it was referring to the anode resistor Ra in the picture.
Glad you pointed that out!
So, a plate resistance / anode impedance of 6K6 would result in Zout= 322,6K/67,2K = 4.800Ω
Indeed, that would produce F-3dB 1/(6.28 x 1x10^-6 x 4.8x10^3) = 33Hz.
I'll drop in a 4,7μF film for that position, leaving F-3dB at 7Hz.
Low enough for a natural bass fundament at 32Hz?
Many base drums are tuned to this frequency I once red.
Onother question arises... the load is not in the formula for the output impedance.
Doesn't it play a role in the stage output impedance?
Reading the text under Fig. 6 (page 7) I understood it was referring to the anode resistor Ra in the picture.
Glad you pointed that out!
So, a plate resistance / anode impedance of 6K6 would result in Zout= 322,6K/67,2K = 4.800Ω
Indeed, that would produce F-3dB 1/(6.28 x 1x10^-6 x 4.8x10^3) = 33Hz.
I'll drop in a 4,7μF film for that position, leaving F-3dB at 7Hz.
Low enough for a natural bass fundament at 32Hz?
Many base drums are tuned to this frequency I once red.
Onother question arises... the load is not in the formula for the output impedance.
Doesn't it play a role in the stage output impedance?
Last edited:
Alright.... is the term referring to "anode resistance"? Seems plausable coming to think of it
Reading the text under Fig. 6 (page 7) I understood it was referring to the anode resistor Ra in the picture.
Glad you pointed that out!
So, a plate resistance / anode impedance of 6K6 would result in Zout= 322,6K/67,2K = 4.800Ω
Indeed, that would produce F-3dB 1/(6.28 x 1x10^-6 x 4.8x10^3) = 33Hz.
I'll drop in a 4,7μF film for that position, leaving F-3dB at 7Hz.
Low enough for a natural bass fundament at 32Hz?
Many base drums are tuned to this frequency I once red.
Onother question arises... the load is not in the formula for the output impedance.
Doesn't it play a role in the stage output impedance?
Following your notation, eq. at page 7
Zout(unbypassed) = (Ra + 2Rk) [Ra + Rk(μ+2)] / [2Ra + 2Rk(μ+2)]
This can be simplified to
Zout(unbypassed) = (Ra + 2Rk) / 2 ≈ 4K8
You are right, I was wrong, quite typical...
Zs is the output impedance seen by the load, and ZL is the input impedance seen by the source.
Then output impedance remains the same, but output voltage is reduced due the voltage divider formed by Zs and ZL.
Attachments
graphics
I'm dropping a third of the output voltage here through the devider Zs/Zl.
There's still headroom for some voltage feedback.
Would 5% f.b. be derived from Rfb = 20 Zout ?
Any HF issues to be expected?
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.
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.
I'm dropping a third of the output voltage here through the devider Zs/Zl.
There's still headroom for some voltage feedback.
Would 5% f.b. be derived from Rfb = 20 Zout ?
Any HF issues to be expected?
Please, don't do that, global negative feedback will add higher order harmonics.
If high gain is an issue, e.g. for CDP or DAC, you can put a resistive divider at the input, you could start with two 10K resistors, not an issue for CDP/DAC output impedance and not too much noisy, then you can use direct input for standard line level.
BTW, which software do you use to modify images? Very often I have no time to make a new one and I use the more close to each case.
Last edited:
Pft, another audiophile myth. I have yet to see this effect in an FFT, and even if it does do that, they are so far down that it doesn't matter. All single ended Class A amps, both pentode and triode, solid state or vacuum tube, will normally have the second harmonic dominant, followed by the third. All of the rest is just "grass" farther down.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.