[Maxx134]

I bet the middle is same as our front end differential,
Then it feed two common cathode. .

 

[coinmaster]

No because you only need 4 triodes to create a balanced common cathode stage and the common cathode stage would not counteract the distortions caused by the differential amplifier.
 
I suspect it is not a common cathode because putting any other tube in the path would only add distortion rendering the common cathode stage pointless.
I'm willing to bet one of those tubes is a phase splitter.

 

[coinmaster]

I think for the time being someone should try the battery or LED bias mod since it is so easy anyone and their mom can do it.
 
The battery mod consists of simply replacing the 1.5k resistor with a AA battery. Positive terminal to cathode negative terminal to ground.
The LED mod consists of replacing the resistor with an LED with a 1.5v forward voltage drop (infared probably) and throwing a resistor between cathode and B+ so the LED gets the current it needs.
Both are claimed to sound better then resistor.
 
Edit: That only counts for the Mk6. I don't know what bias voltage the MK8 needs.

 

[Maxx134]

Bias voltage I measured at 2v for mk8 with a 6sN7 on driver stage.
The resistor values are the same on both amps.

 

[coinmaster]

2v LEDs and batteries can be found as well. LED will probably be red or green, the best ones are the vintage "dim" variety. The new bright ones get mixed reviews.
Overall batteries get better reviews, I've only heard good things. Lead acid batteries come in 2v and so do deep cycle batteries if you want to go balls to the wall.
Both are among the best sounding batteries if I recall.

 

[Maxx134]

This weekend will let you know how much fire

2v LEDs and batteries can be found as well. LED will probably be red or green, the best ones are the vintage "dim" variety. The new bright ones get mixed reviews.
Overall batteries get better reviews, I've only heard good things. Lead acid batteries come in 2v and so do deep cycle batteries if you want to go balls to the wall.
Both are among the best sounding batteries if I recall.

I will see this weekend how much of an explosion I can make..
Ha


I have heard a rumor little dot prototype amp using 300b tubes. .

Also, look at the tube types used by Eddie Current (my only choice if not DIY)

6sL7, 6sN7, 2a3, 300b

Those are the popular tubes of choice.

Something to think about. ..

 

[MrCurwen]

Coin,
Now that you have your CCS I believe you should bring your attention to possible clever ways to achieve the effects a balanced topology does

I don't get this. As far as I know all of the recent examples in this thread have been of balanced topology. No SE designs at all.

The source follower systems driving output tube grids I've discussed would necessarily be SE, but since they have a gain approaching 1, it's inconsequential. One buffer per phase, so... it's a balanced topology as a whole.

Anyway. Please explain.

 

[coinmaster]

He explained earlier that he was mixing up balanced topology and balance by definition. Which lead to the last few pages of discussion.

 

[MrCurwen]

OK I see,
So you will need at least this CCS to stabilize the triodes, in order to implement the cathode follower / grid drive mod...

I'm not entirely sure what you are talking about.

The grid drive system I was talking about is input stage output -> cap of soviet quality -> source follower loaded by a CCS -> direct coupled to output tube grid.

It has nothing to do with the input stage as such, other than relieving it of all responsibilities to provide current to cap, or output tube grid.

Now count the amount of components (maybe custom chip, or mosphets, transistors, diodes,caps , and their separate PSU stage) in order to implement this.

Yep, it's much.

Remember this is all to replace one coupling cap and one cathode resistor in our current amp..

Just to re-emphasise: These are 100% different mods or designs.

1) The tail CSS

2) Source follower driving output tube grid

Number 2 doesn't in fact replace the coupling cap. It needs to be there. It just makes it's quality rather irrelevant (or, supercharges it's quality to glorious heights, however you wish to look at it).

Also the issue of the facts at hand on the grid side of the coupling cap. Currently they are left completely unaddressed. Of course there is NFB...

Now this is where design choices are made.

Wich path will we decide to choose?.

Yep, that's DIY. =)

Many paths.

 

[MrCurwen]

This is exactly the point I am trying to make about the distortion cancelling abilities of the common cathode design.

No common cathode designs being discussed in this thread. Please use universal terminology for ease of understanding of everyone.

LTP is even faster to type out! In addition to being factually correct.

As far as I have checked,
this driver stage implement almost textbook spec.

Yet you and MrCurwen kept pointing out amp by stating the driver stage is poor and non optimal.

Tube textbooks are not HIFI, and by no means high end (whatever that means to anybody).

1% distortion is ok and actually incredibly good, if you live in 1960 or 1950.

To my ears it's bad.

Not as bad as 0.001% achieved via NFB, but still bad.

Open loop 0.001% (which none of you have heard unless you go to DIY meets) on the other hand, is good.

In reality nothing is optimal.
We have to deal with what we have.

True. But, some things are more optimal than others.

So our amp driver stage may not be perfect, but I have not notice anything poor in driver circuit except the tubes themselves.

I see. Well, 'poor' is a subjective term. Hard to argue.

Put a 10% mismatched signal into the input LPT and see what comes out.

Looking over the values it does seem this circuit is more for triode balance than the "common cathode" benifits of noise cancelling.

Please elaborate.

 

[MrCurwen]

As an example about Sweet Spots in tubes,

I was able to increase my tube performance in sound stage with a slight increment of bias to rear power tubes..

I do believe that. I remember using a lot of time optimizing bias points and I did on a number of occasions obtain great sonic benefits. They did of course vanish quite fast if I changed the tube and got unlucky.

The sonic differences in bias points are not because of the curves. Sometimes you get these optimization benefits even if you move within the linear regions of the curves.

They are because of plate resistance changes.

Imagine the tube is a resistor inside the circuit. A variable resistor with a center point it returns to when there is no signal.

When you move the bias point, or op point, you change the center point of the virtual resistor. This might make it fit the circuit more, or make it fit the circuit less.

The trick to eliminating this effect is to design the circuit so, that the tube's rp has no significant importance to the function of the circuit.

As it is, if you can hear benefits from op point optimization (inside the linear regions of the curves) it means your circuit is designed so that it is just barely able to fullfill it's current demands, and any change to the rp is thus critical.

I no longer like to do that stuff, so I make the rp irrelevant. Best just have the tube doing voltage amplification and that's it. That's what it was designed to do, and it does it superbly. Have it provide current to load, and it starts to fail at it's main job, voltage amplification.

 

[MrCurwen]

Meanwhile,
I will listen to my amp,
which "shouldn't work" as good as it does.

No, it works just as well as it's supposed to. A collection of good solutions and bad solutions, design compromises which leads to the level of fidelity that it leads to.

Pride of the customer and engineer mentality are not, as I said earlier, very compatible.

If you do DIY long enough, you will build several THE BEST amps you've ever heard, and spend many nights marveling how you couldn't even imagine how any sound reproduction system could sound any more realistic and musical. Then you design a couple of failed things, and then you design the thing that is again THE BEST, even better than the last one. Then you spend many nights.... etc. Continue until you run out of cognitive ability to understand more electronics. That will be your best then. For many people it will be the classic schem with one or two mods they copied off of internet.

Good and best are two different things.

Your amp is no doubt good. Is it special in the grand scheme of things? Not at all in my opinion. I haven't heard it, but simply looking at the output impedance, I can tell I wouldn't be super impressed with it, even though I'm sure I'd enjoy listening to it.

Any previous systems with that high (or even much lower say 10 ohms) output impedance have sounded unimpressive in the long run.

These are my opinions and yours may vary. I haven't stated you shouldn't be able to enjoy your amp, you totally should be able to enjoy your amp. If you find criticism of it unbearable, I don't know what to tell you. Why mod if improvement is not interesting?

 

[MrCurwen]

Also the decision of the designer to a place the bias point lower. .
why?
I think because of heat and longevity and no fans on bottom MK8 unit!

No, he didn't want to pay for parts to a higher voltage B+ for the input stage.

Calculate the differences in heat. Completely insignificant, not an issue in the slightest.

Also longevity not an issue at all in the case of 6SL7. Again, calculate the watts and compare to datasheet.

 

[MrCurwen]

A common cathode tube stage is one where the cathode is at AC ground.

In a LTP the cathode is not at AC ground.

The Broskie circuit should be called a cathode coupled stage. Common cathode was in common use before WWII and still is.


A LTP is not a LTP unless the cathode node voltage (top of the tail) can float. It cannot be biased with a fixed or semi-fixed voltage system such as LED or battery.

If you prevent the cathode node from floating, you have something else, not a LTP anymore.

The balance of LTP is dependent on the impedance ratio of the tail versus the brances that leave from the cathode node upwards (tube, plate load, on both sides). If you put a LED as the 'tail', it has an impedance of around 10 to 100 ohms, depending on make of LED. What's your ratio then?

The tail then ceases to be the tail, since it cannot do the functions of the tail. The functions of the tail (forced balance or attempted balance, power supply noise reduction) are dependent on the impedance ratio. The tail impedance must be the biggest impedance in the whole system.

The CCS in my schems, with the 2N2222 darlington pair and the FET, achieves a 2 to 6 megaohm impedance, depending on FET choise. The gyrators are about 500k, tubes' plate resistances are so small you can leave them out of the equation (bye bye tube rolling, not missing you personally).

500k + 500k = 1M total both of the brances combined. Tail let's say 4M. Perfect balance.


I've done stuff with battery bias. Batteries do have a sound. Some sound better than others. They always colour the sound, so I quit using them at the cathode. Grid bias can be provided with batteries, but it's too much of a hassle for no added benefit.

An ideal battery cannot be used as the 'tail' in a LTP. It has constant voltage, so the cathode node cannot float. Impedance 0 ohms.

In reality it is at least hundreds of ohms, and very highly

1) current

2) frequency dependant

This is how it colours the sound. Good for effects seekers, bad for transparency and realism.

Also it changes with age, a lot. Batteries are only good (in HIFI) for fixed grid bias, nothing else.

 

[coinmaster]

The trick to eliminating this effect is to design the circuit so, that the tube's rp has no significant importance to the function of the circuit.

Unfortunately this can't be done in the MK6, or with any tube output stage where the tube feeds the load, at least not that I can think of.

 

If you do DIY long enough, you will build several THE BEST amps you've ever heard, and spend many nights marveling how you couldn't even imagine how any sound reproduction system could sound any more realistic and musical. Then you design a couple of failed things, and then you design the thing that is again THE BEST, even better than the last one. Then you spend many nights.... etc. Continue until you run out of cognitive ability to understand more electronics. That will be your best then.
 

I hope I don't end up like this. I want the best now 

.
I certainly hope my upcoming designs function like they do in spice. Way to expensive for failures and the common cathode inspired design I'm working on has 20 feedback loops per channel for biasing, I don't even want to think about troubleshooting it.
 
 

Open loop 0.001% (which none of you have heard unless you go to DIY meets) on the other hand, is good.

I would love to know how you do that with open loop.
 

The balance of LTP is dependent on the impedance ratio of the tail versus the brances that leave from the cathode node upwards (tube, plate load, on both sides). If you put a LED as the 'tail', it has an impedance of around 10 to 100 ohms, depending on make of LED. What's your ratio then?

Yeah I was thinking about that, but since it's an easy experiment to try regardless I thought it might be something to experiment with. A learning experiment if nothing else.