Ah! Okay, then adding the script for watts average also produce a result?
Neat! 😀
I saw a mention of the error log for something else, so put it out of my mind. Now it is clicking.
Yes. I add that script to all my power amp schematics, makes my life easier 🙂
It is not very intuitive that you have to display an error log to see calculated values, but hey, this is free, good enough for me 🙂
You can customise what a window is showing. Let's start with the output voltage. Clicking on OUT give you this:
Right click on Vout:
You will get a prompt where you can modify the source of data for that window. Just enter the formula for power, which is current*voltage (Vout*I(RL))
After that you will get the instantaneous power curves
Neat! 😀
Just finishing my last bite of six grain cereal. Sure can't understand why Roman Meal decided to quit producing rye cereal (sometime in the late 1990s), miss it a lot, don't know of any other source.
And then the program crashes. Thankfully re-starting the program fixed both the bugs. For now... 😆
Last edited:
I am saying this with love, with awareness of compassion and generosity: sometimes it is better to slow down and analyze. Remember how there is/was a reasonable concern about the Elektor Labs amplifier having too much gain because of an extra tube? Well, LTSpice has produced some interesting results!
Here is the L1 clone with ECF80 tubes.
Then ECF82 tubes. Interesting what a difference it made for the better.
The Elektor Labs running ECC83 tubes. Surprise! 😀
Let's get rid of those nasty ECC83 and roll in some ECC81. I think this is a great result despite the distortion going up for some reason, anyone have opinion(s) or thought(s)?
Here is the L1 clone with ECF80 tubes.
Then ECF82 tubes. Interesting what a difference it made for the better.
The Elektor Labs running ECC83 tubes. Surprise! 😀
Let's get rid of those nasty ECC83 and roll in some ECC81. I think this is a great result despite the distortion going up for some reason, anyone have opinion(s) or thought(s)?
After I looked at the Elector simulations I knew I was in trouble after having a go at the ECC83 🙂 . It looks good indeed, I tried high frequencies and it still there, distortion only goes up a little. Credit to the EL84, very easy to drive. And I guess big the open loop gain is helping the negative feedback to keep things clean.
Carefully with tube rolling. The ECC81 and ECC83 are very different tubes, with different optimal working points. You would have to do some modifications to the circuit, mainly resistor values, to optimise it for something else. But I guess there is not much wrong in this particular schematic with that ECC83 after all.
Word of advice, careful with LTSpice, it's extremely addictive, time flies 🤣
Carefully with tube rolling. The ECC81 and ECC83 are very different tubes, with different optimal working points. You would have to do some modifications to the circuit, mainly resistor values, to optimise it for something else. But I guess there is not much wrong in this particular schematic with that ECC83 after all.
Word of advice, careful with LTSpice, it's extremely addictive, time flies 🤣
Ha.
You have a favorable opinion of it?
Interesting. I will look into the ECC81 at some point.
What do you mean by "big and open"?
My guess based on a quick glance is regards to less resistance.
When the concern about the ECC83 was brought up, I easily did a hour or two of research. Folks were doing nothing to the circuits, though I know this is not ideal.
Taking a cursory look, I can see it is resistors, back to @Markw4 second favorite formula (Kirchhoff's formula is his favorite). 😛
Pulled the data sheets and see differences besides gain:
1) Typical operation is about 10 times more current through the ECC81 than the ECC83 (1,2mA versus 10mA on Phillips tubes)
2) Higher transconductance (Gm), 1,6mA/V versus 5,5mA/A
3) And the ECC81 has far less internal resistance, 62,5KΩ versus 11KΩ (about 82 percent less)
My guesses:
1) More current is beneficial, maybe reduces distortion (still reading the books)
2) All I know is for phonograph amplifiers, the higher the better (PC900 for example).
3) The higher the resistor value, the less voltage and more heat.
I would appreciate correction and/or instruction on this. 🙂
By the information I read, which know is dangerous, I have this question of if can be any better.
I am very much finding this out... There will have to be a time where I set down the mouse and pick up the soldering iron.
Though I did try and find the information with so far no success before going to men's get together, got home not to long ago, how to write an input for giving damping factor value.
Tuning the parts values for ideal results is fun. Slowly getting it as close to perfect as possible is addictive at times!
It's like jetting a carburetor. The dual weber 40IDFs on my daily driven 73 Beetle took some dialing in, and it's giving nice power and good MPG now. Simply going "textbook" on the values doesn't always work or account for the big picture and the rest of the setup.
In the past I found optimizing each section of an amplifier for minimum distortion, then adding them together didn't always yield the lowest distortion from input to output once everything was added together. Sometimes you optimize one section at a time while the entire thing is all together and it will yield better results. Especially in push-pull there is some interesting distortion cancelation that happens.
It's like jetting a carburetor. The dual weber 40IDFs on my daily driven 73 Beetle took some dialing in, and it's giving nice power and good MPG now. Simply going "textbook" on the values doesn't always work or account for the big picture and the rest of the setup.
In the past I found optimizing each section of an amplifier for minimum distortion, then adding them together didn't always yield the lowest distortion from input to output once everything was added together. Sometimes you optimize one section at a time while the entire thing is all together and it will yield better results. Especially in push-pull there is some interesting distortion cancelation that happens.
Last edited:
I had to force myself to stop, ideas kept flowing out of me, however, the last was a major failure for distortion, though not learning.
Maybe it be easier on folks if the lesson(s) and/or take away(s) are at the beginning?
They are:
Regarding the EF86 tubes, the most anode/plate voltage possible is 150V and operates at an incredibly low current. My guess be it lives a very long life being under worked.
The reason this is not possible is because: the 100KΩ required, because it follows after the phase splitter, and no way to bypass the phase splitter and feed voltage directly (and because it is DC can't be stepped up).
Furthermore, if use resistor(s) to decrease the voltage to the EL84 tubes, interestingly, increases distortion.
Increasing voltage decreases distortion, of course within reason, to diminishing returns. The revised ECC81 to come close to the revised ECC83 distortion levels prefers a B+ of 350VDC, the ECC83 can get slightly less distortion at this level, however, the higher the voltage to a tube, the shorter its life is; most folks can't hear this distortion at the 0,01 stage so no point.
A couple hours of finagling had to be done to the ECC81 version to achieve close results to the ECC83, resistor values were not making positive changes, rather, resistors had to be deleted and B+ voltage increased. The only reason folks don't like the ECC83 is microphonics, bet a flexible tube base and/or shield address this (disappointingly, this can't be modeled).
Maybe even put the question(s) up front.
My observation on both versions is removing R9 and R10 sufficiently decreases distortion. Is there purpose something else besides voltage reduction?
It is way late, so instead of putting up all the data, including tube currents and voltages, put up the highlights, probably to you all's relief. 😛 Though if anyone was it, hopefully be home in the afternoon.
Here is the ECC81 starting out. Bet most folks would not be unhappy with the sound due to increased distortion.
Then after.
Onto the ECC83.
The starting distortion.
A weird result with the deletion of R9 IIRC.
Fixed with a resistor to the grid, a 22Ω worked, however, chose 47Ω since no effect on distortion and already a number of them, buying more than 10 gets a discount. I think did well figuring this out, though have to be careful not to have pride.
And after.
Maybe it be easier on folks if the lesson(s) and/or take away(s) are at the beginning?
They are:
Regarding the EF86 tubes, the most anode/plate voltage possible is 150V and operates at an incredibly low current. My guess be it lives a very long life being under worked.
The reason this is not possible is because: the 100KΩ required, because it follows after the phase splitter, and no way to bypass the phase splitter and feed voltage directly (and because it is DC can't be stepped up).
Furthermore, if use resistor(s) to decrease the voltage to the EL84 tubes, interestingly, increases distortion.
Increasing voltage decreases distortion, of course within reason, to diminishing returns. The revised ECC81 to come close to the revised ECC83 distortion levels prefers a B+ of 350VDC, the ECC83 can get slightly less distortion at this level, however, the higher the voltage to a tube, the shorter its life is; most folks can't hear this distortion at the 0,01 stage so no point.
A couple hours of finagling had to be done to the ECC81 version to achieve close results to the ECC83, resistor values were not making positive changes, rather, resistors had to be deleted and B+ voltage increased. The only reason folks don't like the ECC83 is microphonics, bet a flexible tube base and/or shield address this (disappointingly, this can't be modeled).
Maybe even put the question(s) up front.
My observation on both versions is removing R9 and R10 sufficiently decreases distortion. Is there purpose something else besides voltage reduction?
It is way late, so instead of putting up all the data, including tube currents and voltages, put up the highlights, probably to you all's relief. 😛 Though if anyone was it, hopefully be home in the afternoon.
Here is the ECC81 starting out. Bet most folks would not be unhappy with the sound due to increased distortion.
Then after.
Onto the ECC83.
The starting distortion.
A weird result with the deletion of R9 IIRC.
Fixed with a resistor to the grid, a 22Ω worked, however, chose 47Ω since no effect on distortion and already a number of them, buying more than 10 gets a discount. I think did well figuring this out, though have to be careful not to have pride.
And after.
Thought finally after 30 45 minutes, can call it a night. Ha. 🙂
And yes, I was getting some really weird results and weird waves, even got a square wave! 😳 😆
All the time! 😆
Interesting another member with an air cooled Volkswagen. Only once did I tune a carbie, 34s getting them synced (they were rubbish, especially the linkage), vast majority has been D-Jetronic and it too needs a little tuning (especially when you are young and dumb and keep the duration while putting in a high lift camshaft). When finally done with building the engine, might have to adjust the 28-1PICT. Bet feels great getting duals all dialed in, little more fiddly than a single.
I can see that, have experienced it with the D-Jetronic (increasing the fuel pressure gains better fuel economy and performance, for example). I am wondering if on some designs if even considered a change, for example, the EF86 plate has a resistor paralleled with a capacitor, however, the phase splitter does not (and does improve though held off posting, figure best in its own separate post, also search the forum).
I don't know how to separate, so didn't, though did make changes to a section (ex. anode of a tube) and work my way out and back to B+. Then I worked my way back. Figured better to have a system so things aren't missed.
And yes, I was getting some really weird results and weird waves, even got a square wave! 😳 😆
Yet another question...
Just realized the Elektor Labs EL84 has no input coupling capacitor, is there a reason why?
If I am reading this correctly, the EF86 likes a grid stopper of 47KΩ. This also then serves as an input high pass filter, incorporating the input coupling capacitor. As can see, did no harm to the distortion.
By the way, for those who notice detail, the EL84s only had 8W of dissipation at 325Vs, so allowed to raise the voltage and decrease distortion; now at 10W.
Just realized the Elektor Labs EL84 has no input coupling capacitor, is there a reason why?
If I am reading this correctly, the EF86 likes a grid stopper of 47KΩ. This also then serves as an input high pass filter, incorporating the input coupling capacitor. As can see, did no harm to the distortion.
By the way, for those who notice detail, the EL84s only had 8W of dissipation at 325Vs, so allowed to raise the voltage and decrease distortion; now at 10W.
My only addition to this thread: If the OP actually wants to build his first tube amp and listen to some music, start with something simple like the Blueglow KT88 SE or the 6SQ7/EL34 SE amp I show how to build. No GNFB or PP to deal with, they put out decent power, just about any brand output transformer works and either will drive something like a pair of Klipsch RP600m easily (which you can get version 1 on sale now pretty cheap). Plus they will sound great.
https://www.youtube.com/playlist?list=PLtEhh3UpOsiCMpG7Haj-EHHN4-rSyCBR1
Given we are at 435 posts, I'm not sure the goal here is to listen to music 😛
https://www.youtube.com/playlist?list=PLtEhh3UpOsiCMpG7Haj-EHHN4-rSyCBR1
Given we are at 435 posts, I'm not sure the goal here is to listen to music 😛
Thank you for the offer. 😀 I am not understanding, which are you offering to give me?
As for Klipsch, don't like the sound and not big enough in my understanding. Besides, be less money to replace a driver then replace the whole speaker.
Yes, already watched the series plus more, ditto for your channel, which is also been a help.
Ha. I am listening to music, just through one speaker. 😛
On a serious note, if had built something, I would have not ended up where I am with the understanding I now have because of LTSpice and the generosity of the forum. Nor would I have received such a generous offer, looking forward as would not have built a SE until now.
"Thank you for the offer. 😀 I am not understanding, which are you offering to give me?"
Simply giving some friendly advice, nothing more. And yes if you already have decided "don't like the sound" of efficient speakers, and further feel 9.8W of power from a proven design isn't enough (Blueglow KT88) but 10W of power you created as a novice on a virtual simulation is what you want to pursue, my advice appears to be misplaced, as I clearly don't understand your goal. And honestly I don't need to!
Just under 10W of power can be made with a simple, low parts count SE circuit, which makes sense for a beginner to learn on. But it appears you have seen and were aware this option exists, yet instead are choosing a very complex design, very few people are familiar with, as your first attempt instead. Carry on.
Simply giving some friendly advice, nothing more. And yes if you already have decided "don't like the sound" of efficient speakers, and further feel 9.8W of power from a proven design isn't enough (Blueglow KT88) but 10W of power you created as a novice on a virtual simulation is what you want to pursue, my advice appears to be misplaced, as I clearly don't understand your goal. And honestly I don't need to!
Just under 10W of power can be made with a simple, low parts count SE circuit, which makes sense for a beginner to learn on. But it appears you have seen and were aware this option exists, yet instead are choosing a very complex design, very few people are familiar with, as your first attempt instead. Carry on.
If the EF86 likes a 47K grid stopper why did you use 470k? And why the shunt capacitor? The purpose of an input coupling capacitor is to block DC. Parallelling it with a resistor makes no sense. It should be in series, or omitted. Your way it's not a 'high pass' filter, it is a high shelf fillter with turnover at 15Hz: far too high, and wrong configuration. And it's not a grid stopper above 15Hz.
I do not deserve to be treated in such manner just because of my condition. I did not choose to be born this way.
Additionally, where were you when I was asking for this advice?
If had looked at the list, you would have seen the PP requirement with preferably EL84s (bought them based on advice and happened on what seemed a good deal). Tubelab's PP was suggested, however, could not find any figures on the performance and honestly got hung up on monoblocks (you said you provide the schematic yet months later still not shown up). Sometimes we get hooked on something and have trouble being reasonable.
Since you want to help, what is the dampening factor of the TubeLab PP? I ask as don't know how to calculate it in Spice, not sure know it well enough to calculate by hand. The rest of the data can be generated by Spice, thanks to José's generosity I have the schematic.
Last edited: