[baronbeehive]

Then you have the 0.0001% newer design approach, getting away from those designs, like the SonicTrance design approach (based on Mr.Curwen).

I am grateful to have an opportunity to hear such a design.

Wow, so good luck with your modern amp build!

.

Edit:
I myself would be interested to try a modern tube amp such as that. With the way we've got the LD now it's difficult to conceive of anything better, if pushed I suppose I could imagine something super realistic especially in the soundstage. I've heard a friends top speaker system and the only way it is superior is possibly in neutrality and resolution of very fine detail, but in may respects I prefer my system for SQ.

 

[baronbeehive]

As I've lived with the recent changes for a while now I just thought I would summarize where I'm at with the last mods which have been tested against the other amp as a control, because I think it would be interesting to discuss. We've all pushed this baby quite a way from stock... travelled to distant galaxies to seek out new life and new civilisations... and all that.

I highly recommend the last 3 mods: CCS with improved operating point over stock, OCC silver output wires, and the Nichicon Muse KZ cathode caps. This has brought the sound up to a very high level. The CCS has made the sound a touch cleaner, and more lively with the revised operating point, the silver wires made the sound brighter and more transparent, and the Nichicon caps made it sharper and more extended. The combination of the silver wires and the Nichicon caps on the high frequencies has brought the high hats and cymbals into the foreground now whereas before they were very much in the background which is a very positive improvement I think. To compare the 2 amps, my APPJ, and the LD, they now sound very similar, due to the mods, in a good way.

For anyone thinking of doing this the APPJ cost virtually nothing, as amps go, but the cost was doubled with the mods. Both amps now are very highly resolved, the APPJ with the 14AF7 tubes, and upgraded caps has been pushed to a high level. The 14AF7 cost a few dollars, the EL84 Amperex tubes are not expensive. For the LD the Nichicon caps are cheap, the CCS was pretty cheap, the OCC silver wire was a bit more expensive but not much was needed. The LD uses RCA 6AS7G tubes which are cheap, and 12SL7 GE's which cost a few dollars. I see now that the 6SL7 GE's have the same silver round plate construction and so would be preferred because they don't need adapters obviously, but I already had the 12SL7's from the APPJ.

To sum up the sound signature, both amps are very resolved, have good soundstage, good rich open sounding mids, tight bass and extended treble with good textural details. The sound is neutral to warm sounding, I would suggest if you wanted a warmer sound to use OCC copper wires instead of OCC silver, this will make it slightly smoother and less incisive sounding. With the OCC silver the sound is more transparent and sharper.

Needless to say both amps sound fantastic! The amps do have slight differences but nothing major. I wanted to increase the resolution and soundstage of the APPJ and for the LD I wanted to get it to sound more dynamic and lively with a heightened treble and tighter bass and I think I've finally managed to lick the problem! It's a tad brigher and sharper, but not harsh, especially in the high frequencies which is the main reason for the improvement, and transients are very lively. Having the 2 amps was good for comparison especially as the APPJ is a speaker amp.

 

[MrCurwen]

It's been a while since I bummed everybody out so here goes:

If you can hear differences between output wires (or any wires) you have some serious issues in your amp design. Not good.

That basically means you have an unreasonably high output impedance in the 10kHz and over range. NFB should help with that but if you hear a difference, it doesn't. Very troubling.

I've done a tiny bit of RF work, and that's where wire inductance can make or break your circuit. Different wire types have different inductances; look up the specs, calculate impedance from inductance, do your RC low pass calculation and see where you're at with regards to output impedance at 15kHz.

 

[baronbeehive]

It's been a while since I bummed everybody out so here goes:

You're welcome!

I can hear high hats and cymbals now so I'm very happy. I don't think it's affected by the NFB in the design because it's something I've been talking...… boring people about for a while, and I've tried high and low gain, if anything high gain sounds better.

Edit: This is recording dependant, some recordings sound muffled and dull, but the best recordings now exhibit all the frequencies very well, so you're right about the wire inductances affecting this design.

 

[Maxx134]

I believe that's why the wires from input socket in back, to the front of amp (input stage) was of a shielded design.

I replaced this also with a better shielded cable.

Also I see the many red caps to be placed along the output stage (power to ground)for this purpose.

So I do think this amp would have an issue because I did notice a slight difference when I used old wire-wound inductive power resistors at the anode of the output stage.

I since replaced them with non-inductive wire-wound type.

So yeah this I beleive is a common issue that plagues most designers in many circuits and why circuit board trace design is sometims critical.

 

[Maxx134]

I wondered why you stopped there at the XLR lineouts and I am therefore considering completing the wire run from the XLR's on to the headphone socket with 6 more pure silver wires, I think it is perhaps possible to achieve yet more transparency

Agree, from the output jacks to the front headphone socket is a good idea which I actually did replaced as well..

All my wiring was replaced.
The only issue here was space concerns.
No stock internal cabling was left inside. All was removed because they are all of a very low grade.
On unit I had, none looked to be copper(!).

Especially the wiring from the PSU power transistors to the board.
They were brittle and can break off solder joint very easily.

After inspection, the wire has what looks like a high tin content. No copper at all (visually, maybe it was the dark stuff).

Although the input wiring was nicely shieled, it too was of a poor tin composition and strands looked dirty inside(copper?).

I replaced with a "shielded" silver plated cable, as that was the best shielded that I saw the purpose called for.

I suppose you can use other cable, but as MrCurwen states there can be interference.

 

[baronbeehive]

Although the input wiring was nicely shieled, it too was of a poor tin composition and strands looked dirty inside(copper?).

I replaced with a "shielded" silver plated cable, as that was the best shielded that I saw the purpose called for.

I suppose you can use other cable, but as MrCurwen states there can be interference.

No, I haven't been able to find any copper wires either.

I might try the shielded wire, or maybe experiment with this or the silver. I only intend to go from the XLR outputs to the headphone socket, I'm not replacing the wires from the inputs, because I imagine the benefit will be greatest to the signal travelling from the tube sockets direct to the output sockets, ie after amplification. I would rather stick with what I have already tried, ie. silver wires than use shielded copper as you did, at least in this area, because of what I have already found though.

But the improvement to be had just from changing wires and cathode caps to Nichicons is such that it has now solved the remaining things that needed improving with the amp IMO all in one go, which is surprising. The Nichicons were a really great find of yours!

Edit: The wire inductance problem on the LD must have been known by Sword-Yang which is probably why he suggested the wire upgrade as part of the factory mods originally.

 

[baronbeehive]

Agree, from the output jacks to the front headphone socket is a good idea which I actually did replaced as well..

All my wiring was replaced.
The only issue here was space concerns.
No stock internal cabling was left inside. All was removed because they are all of a very low grade.
On unit I had, none looked to be copper(!).

Especially the wiring from the PSU power transistors to the board.
They were brittle and can break off solder joint very easily.

This looks like another area of potential upgrade that I've not bothered about, maybe I should. I've still got existing, resoldered, wire to the transistors.

Did you notice any sonic improvements with your wire changes? I imagine the input wires would benefit from upgrading to make sure the signal from the DAC is clean.

After the improvement I've had with the silver output wire, this looks like it might be worth doing for all stock wire as you said.

 

[klnglim]

Agree, from the output jacks to the front headphone socket is a good idea which I actually did replaced as well..

All my wiring was replaced.
The only issue here was space concerns.
No stock internal cabling was left inside. All was removed because they are all of a very low grade.
On unit I had, none looked to be copper(!).

Especially the wiring from the PSU power transistors to the board.
They were brittle and can break off solder joint very easily.

After inspection, the wire has what looks like a high tin content. No copper at all (visually, maybe it was the dark stuff).

Although the input wiring was nicely shieled, it too was of a poor tin composition and strands looked dirty inside(copper?).

I replaced with a "shielded" silver plated cable, as that was the best shielded that I saw the purpose called for.

I suppose you can use other cable, but as MrCurwen states there can be interference.

Hi Maxx! It's been a long time, I m still troubleshoot and hard work on my MK8SE. I had completed several times unsoldered, resoldered and cleaned the board, unfortunately till now still fail to turn it on.
Left meter light on with "0" reading, Right meter no response at all, of course no sound at all, furthermore unstable voltage reading everywhere and sometimes one 120R/5W resistor near power supply smoke.
Now I try to do PCB on board prototype connection, have already done half way, the whole board seen wire spaghetti all around (By the way, never give up! Just keep on search for the issues) LOL

May I ask can Tube NOS Voskhod 6N1P-EV replace those 6H30P tube?
Here's the website https://www.partsconnexion.com/nos-6n1p-6h30-vacuum-tubes.html

 

[MrCurwen]

No, not at all. 6N1P has anode resistance of 11k, and 6N30P has anode resistance of under 1k. They are meant for completely different things.

 

[MrCurwen]

Why do you need to replace the 6N30P?

 

[klnglim]

Why do you need to replace the 6N30P?

I didn't want to change 6H30p, just want to get some knowledges differential among those tubes

 

[MrCurwen]

When looking for an equivivalent tube, first look at anode resistance (plate resistance, rp). It's expressed in ohms. For the tube to work properly in the circuit as intended, the rp should be not too far off. 1k and 11k are worlds apart.

Second look at µ or mu. It should be in the same area. 20 and 30 can be close enough, but 11 and 30 are not. 8 and 20 probably are not. This is not as critical though as the rp.

Of course the physical constraints of the tube; which base, pinout, heater specs etc need to be considered.

If you are concerned with fidelity (low THD), you should of course make sure the new tube has as good curves (or better) as the one being replaced. The goodness of the curves is determined by how evenly the grid lines are speced. Pick a horizontal line at a reasonable current level and go from left to right on that line. See how evenly the grid lines are spaced.

Of course often the curves are drawn poorly, and the real tube performs much better. This is quite common, especially with non-audio tubes like TV sweep tubes. The curves were not considered important so they were made just "close enough, good enough".


6N1P is a small signal tube, use for input stage. 6N30P is low rp low µ output tube. Use anti-oscillation precautions, both were meant for radio frequency duty also.

 

[baronbeehive]

Hi Maxx! It's been a long time, I m still troubleshoot and hard work on my MK8SE. I had completed several times unsoldered, resoldered and cleaned the board, unfortunately till now still fail to turn it on.
Left meter light on with "0" reading, Right meter no response at all, of course no sound at all, furthermore unstable voltage reading everywhere and sometimes one 120R/5W resistor near power supply smoke.
Now I try to do PCB on board prototype connection, have already done half way, the whole board seen wire spaghetti all around (By the way, never give up! Just keep on search for the issues) LOL

Hi klnglim!

I see you are still trying to sort out your amp. I can't be much help as I don't know exactly what you've done but I would go back to basics and double check a few things:

your traces new and old - your repairs to the board look excellent
your parts - are the specs the same. It is easy to misread some parts names, for example I mistook a .68R for a .68K resistor, also I used the wrong type of diode, both of which could have stopped the amp working. Are your parts all up to spec, the right way round etc.
wires all connected correctly.

I double and triple checked mine recently and still missed a few things!

All pretty basic but a mistake at this stage will be critical. It is difficult to know what's wrong when you have done so much.

When you've got the amp up and running you must complete all the upgrades to date, that way your amp will be pretty spectacular lol!
.

 

[baronbeehive]

If there were space, and if the loop didn't become unstable, you could add proper plate loads. That would increase linearity very drastically. It's not a subtle change.

If I understand you at all I think your other amp has an OT. That OT is the primary determinant in output impedance, the gNFB simply helps it a bit. In this LD amp the gNFB is the primary motor in providing a decent output impedance.

If you drive the PRIMARY of an OT with a decently lowish output impedance (some hundreds or tens of ohms is overkill level sufficient), then the OT secondary will have an output impedance lower than 1 ohm. Talk about tight bass then, let me tell you.

Yes, that basically means that the coil inside the element has a tricky impedance curve. Flood it with current, and it's as good as any. Of course there are other overall limiting factors in the quality of the headphone, but you would be surprised at what can be dug out of those cans.

I am interested in what you have said here with these points, could you go into a bit more detail?