[baronbeehive]

I do not believe cathode cap decrease sound quality at all. The interaction between the cathode cap and tube creates more gain, and more tube harmonics. This is added onto the sound, so doesn't take anything away. But it has to be done lightly, as if you sprinkled salt on food, not overdone.

No and the gain without the cathode cap is so low as to make the sound lifeless unless other things are done to compensate

I see that we haven't even touched the driver stage cathode area, but since you replaced the cathode resistor with
a CCS, you can't place a cathode cap there.

I'm thinking of taking out the CCS as the cathode research looks to be promising. I only kept it in because of the fact that the current was always constant with varying voltage so not having to worry about worn tubes. And the B- filter didn't have any effect on sound anyway.

And I was happy with the anode/cathode values that I found on sound quality before I put in the CCS.

 

[baronbeehive]

One last go with the cathode calculator this time to see the effect of overkill. As you can see the stock amp curve is exactly the same with 10000uF as with lower capacitance, (500uF, right hand graph). Of course there might be oscillations with such a high capacitance.

 

[Maxx134]

It seems that ALL the schematics are from @coinmaster :

The MK6 uses the OPA445AP as well.
Don't worry about what the opamp is doing. I verified it long ago and I ordered my own custom versions for my experiments




This is a straight up copy of the configuration on the MK6.
My initial trials were P2P construction but I did not feel comfortable having the opamps flimsily dangling around.

The jist of it is that the opamps positive terminal is connected to ground.
The negative terminal is connected to the output.
The opamp will try to do what it can to make both inputs the same.
This means that the output of the opamp will bias the top triode of the WCF so there is ground level dc (0v) on the output of the amplifier (also negative opamp pin).
The 100k resistor/1uf cap is an RC filter so the opamp ignores frequencies above 1.6HZ.
The 470k resistors/ .22uf cap is another RC filter at 1.6HZ presumably to get rid of any noise the opamp creates and it is also a high impedance buffer from the signal.
The 1uf caps are just bypass caps to stabilize the supply voltage.
The 20k trimmer adjusts the offset error between the inputs of the opamp so when the opamp tries to keep both inputs the same they actually both measure the same. Basically you use this to tweak the DC at the output of the amp.

I've used what I've learned from the MK6 integrator to come up with much more complex and useful designs.

This is the upcoming PCB for the biasing circuits that will allow me to fully adjust the load line vertically and horizontally while in operation. It's not finished cuz these software designs are a P.I.T.A to make within the "design rules" of the manufacturer but I should have it done soon.


The pinout of the OPA445AP datasheet is correct.
Refer to my schematic for the MK6 design which now that I look at it again is kind of crappy but it works




R75, R5, and C4 in the relay schematic form a 2 second time constant to the "VCC ON mute" pin which from my short reading on "mute" seems to simply mean off/disconnected.
So effectively it seems to just delay the power to the IC for 2 seconds on startup to give the servos enough time to stabilize so you don't blow your headphones. It doesn't seem to be a timer for the protection relay in active duty and I don't see why there would be a timer.

That's actually a good point though, I didn't code my arduino to protect the headphones on startup. Glad I caught that before I used it on the real thing.

I've verified that my arduino protection circuit reacts instantly to offset above the set threshhold so If the analog circuit is really as slow as you think then my arduino solution would work, but I think such a thing is beyond the scope of this modding thread and probably a horrible idea to have people playing around with.

Keep in mind that the IC in the amp a 1990 model. Whether that's relevant or not I don't know.

 

[baronbeehive]

It seems that ALL the schematics are from @coinmaster :

Then I must have had a "moment" with my failing memory, I thought Sonic did a schematic later .

 

[coinmaster]

Anyways, I have stumbled onto some better schematic diagrams from an asian website.

An asian website?
Interesting, because those are my original schematics that I made forever ago before I started using LTspice and before I knew what I was doing really.
I guess the asians are reading head-fi.

 

[coinmaster]

The question that needs to be asked is, what are you guys trying to achieve? Because it doesn't take much deviation before it's more efficient to just to a complete redesign.
The design is so deeply flawed that it's too easy to end up on something completely different when trying to solve the issues.
I can make something dirt cheap that has unmeasurable distortion and sounds as good as the best equipment.
Just ask Maxx, I used a switch to instantly swap between super expensive tube amps and my tiny, cheap but very sophisticated SS one at his place a while back and it sounded identical.
It's all about how it's designed.

I can also make pure solid state designs with tube curves

You name it, I can do it, no hype or gimmicks involved.
The question always starts with, what do you want?

But as far as this amp goes, it's actually more difficult to keep things within the scope of the original design then it is to change it completely.
Unfortunately the amp doesn't leave much physical room for modification.
So it's more efficient to ask what your goal is.

I do not believe cathode cap decrease sound quality at all. The interaction between the cathode cap and tube creates more gain, and more tube harmonics. This is added onto the sound, so doesn't take anything away. But it has to be done lightly, as if you sprinkled salt on food, not overdone.

Well it's very simple. If the cathode caps could be removed without reducing the gain, then it would sound even better. The reason you think they add better sound is because without them, in the current design, the sound will indeed suck.
But the reality is, using tubes for the output stage is just terrible in general, especially with this design.
A simple source follower will sound better. In fact, that is an easy mod to do and ironically one of the best ones that can realistically be done on this design.
But then it reverts to a typical hybrid with a primitive input stage and a crappy power supply.
So I can see the appeal in not doing that.

As for the existing design. You could try using a ccs on the cathodes in place of the resistor to isolate it from the supply ripple and and tie the cathode caps of each of the opposite phases together to maintain the cmmr. I don't know the full implications of this as it's just an idea that popped into my head and I am too lazy to simulate it, but it should work. Although it will mess with the WCF current balance to some degree so the anode would need to be adjusted as well ideally.
But my gut tells me that will most likely improve the sound to some degree.
It would be a simple mod to try.

 

[baronbeehive]

An asian website?
Interesting, because those are my original schematics that I made forever ago before I started using LTspice and before I knew what I was doing really.
I guess the asians are reading head-fi.

Correct.. I just found them amongst some papers which I've had for ages and they are the same!

@Whitigir sorry... I just found those schematics but didn't know what they were until I saw Maxx's post .

 

[baronbeehive]

Ugh ugh ugh..
With help of @Whitigir , I am seeing multiple weaknesses that need to be addressed.

Really... . I can't see anything untoward, but then it's all Chinese to me .

 

[Whitigir]

oh well, like coin master said above. He may as well design another amp lol

 

[Maxx134]

What is that OPAMP chip ? Would swapping it For another kind (discrete) be an upgrade ? Cathode caps are in series ? Interesting

After speaking with coinmaster, he made me realize the opamp chip having the resistors and .22uf cap at output safeguards any sonic influence from the op-amp is neutralized.
Plus this is the only op-amp listed in Mouser with a 90v limit rating.

And I was happy with the anode/cathode values that I found on sound quality before I put in the CCS.

Yes the bias point can be made to "sound" as good, but the CCS guarantee it can be in an optimal point.
So I would probably stayed with it if I did it. Currently I have manually found my sweet spot for a specific tube, and am done tube rolling that stage (after thousands dollars and almost all 12pin & 6pin tubes rolled in, and anything with adapters) so do not want to mess with it.

But the reality is, using tubes for the output stage is just terrible in general, especially with this design

Yes it has a narrow range and limitation for tube out only, but it is great that it works to eliminate expensive transformer output, or expensive large cap output...
Hybrid would have been most optimal.

A simple source follower will sound better. In fact, that is an easy mod to do and ironically one of the best ones that can realistically be done on this design

I rather have the more optimal hybrid of @SonicTrance if I was going hybrid, or one of your new designs that no one did before.

But then it reverts to a typical hybrid with a primitive input stage and a crappy power supply

But in this thread we juiced up the PSU greatly, so not so crappy, just old school & not modern.

You could try using a ccs on the cathodes in place of the resistor to isolate it from the supply ripple and and tie the cathode caps of each of the opposite phases together to maintain the cmmr

No need no supply ripple but that idea of joining the WCF cathodes sounds interesting.

Although it will mess with the WCF current balance to some degree so the anode would need to be adjusted as well ideally

I'm wondering how this would affect the circuit...


In the meantime, optimizing the caps in this output stage is all that's left before a design change...

 

[coinmaster]

I rather have the more optimal hybrid of @SonicTrance if I was going hybrid, or one of your new designs that no one did before.

Sonic's design would be more expensive and it has many redundancies and some components with flaws that cannot be overcome.
This does not mean it will necessarily have any detrimental sonic effects, but I personally I only design in a way where I have complete control over all of the performance variables and distortions. Better to not have to wonder.
Using the loads we are using, I would be surprised if a simple class A source follower sounded any better or worse than any sufficiently designed output stage. Despite the fact that I would never use something so crude myself.

But in this thread we juiced up the PSU greatly, so not so crappy, just old school & not modern.

Even though you beefed it up, there is not nearly enough capacitance in the supply to eliminate ripple.
The design fundamentally relies upon the usage of balanced inputs to cancel out the audio currents in the supply.
This is why balanced sounds so much better than SE with this design.
But only with perfect balance will all of the supply currents cancel.
Is the beefed up capacitance enough to deal with the remaining ripple? I honestly don't remember, probably not in the lower frequencies.
But if you use SE then you are pretty much screwed. No real regulation in the supply.

 

[baronbeehive]

But in this thread we juiced up the PSU greatly, so not so crappy, just old school & not modern.

Yes, and what surprised me in a good way was how sonic made his "old school" Odyssey sound so good according to one member, who apparently likes it at least as much or possibly even better than the "new school" Oblivion for example. So old school is not dead by any means apparently .

 

[baronbeehive]

Yes the bias point can be made to "sound" as good, but the CCS guarantee it can be in an optimal point.
So I would probably stayed with it if I did it. Currently I have manually found my sweet spot for a specific tube, and am done tube rolling that stage (after thousands dollars and almost all 12pin & 6pin tubes rolled in, and anything with adapters) so do not want to mess with it.

Yes, I just feel that if the cathode film cap upgrade can be that much better than what is there atm it might be worth jettisoning the CCS .

But in this thread we juiced up the PSU greatly, so not so crappy, just old school & not modern.

Yes, and what surprised me in a good way was how sonic made his "old school" Odyssey sound so good according to one member, who apparently likes it at least as much or possibly even better than the "new school" Oblivion for example. And the Odyssey relies on quality parts like we have done in the LD as opposed to new school design of the Oblivion. AFAIK it doesn't have regulated PSU, parafeed OT, source followers, for example, but it is an optimised amp which is what we are trying to do here. So old school is not dead by any means apparently .

 

[coinmaster]

"Old school" doesn't mean anything in engineering terms. It's a matter of whether X is sufficient to accomplish Y.
"Old school" power supplies can do their job just fine. Depending on the inherent regulation (or not) of the power supply, the PSRR of the amp, what type of ripple appears in the supply lines, and how the sound quality of the amp is affected by the ripple, you may have a good or bad power supply.
As for the MK6, it does not respond well to ripple and the supply is under designed considering this fact. They did this to save costs most likely.

 

[baronbeehive]

I guess the MK8 is better in that respect in that it only has to deal with much more of a fixed range of output impedance for the Sennheiser type headphone unlike the MKVI which has to deal with the full spectrum of impedances?