[baronbeehive]

OK. I have sound. It was panic stations when I inadvertently used the volume knob of the other amp and still couldn't hear anything. Then I realized my mistake!

Everything appears to be working OK, both meters are reading normal, about 65mA, all good. So I assume when they were both reading around 80 previously the current was being shared somehow between circuits compensating for the bad circuit.

Now for the rest of the mods, then maybe I can actually listen to it for more than a minute before something else happens lol.

I plan to add resistors in parallel to the driver stage rather than resolder the whole shooting match, so that resistor calculator has been invaluable Sonic, it would have taken me days to work that out by hand!

 

[coinmaster]

but the PSU improvements are the lesser,
compared to the most important noticeable upgrades,
which are:
1- the coupling caps
2-the WCF caps
3- and the bias adjustment to the driver tubes...

I would argue that the power supply cap upgrade is one of the biggest upgrades if you use film caps. The power supply bypass caps are basically giant coupling capacitors for the entire amplifier. The current loops for the whole shebang have to go through those bypass caps in the supply.
Lytic caps have measurable distortion, they are just bad for audio. Especally if you use off the shelf power supply lytics like the stock mk6 has.
Replacing mine with film improved every aspect of the sound and gave my HD800s powerful bass.
That being said, there's no way you're going to fit film caps of similar value inside the chassis.

 

[baronbeehive]

......

MK6 users like you don't have to be left out.
I would also try it (same tube model type, just different name ) ...
putting your brighter/clearer tube on left and your thicker/richer tube on right (of each channel) as done above.
Get the best from both tubes..

No I plan to try that with my RCA 6AS7G's combined with some 6080's, probably Thomson, I like the RCA's but they need livening up a bit. But I still have 3 matching 421A's and looking forward to trying those out when I get another one to complete the set.

 

[coinmaster]

Just make sure to check both positions when you mix.

 

[baronbeehive]

Wonderful to have the amp back!

The baby amp however has held it's own while the LD was out of action, it is shockingly good for practically no money.

 

[coinmaster]

Do you know what the schematic for the "baby amp" is?

 

[Maxx134]

I plan to add resistors in parallel to the driver stage rather than resolder

That's what I did..
Just added parallel resistors to increase bias..
Going to have to add it to first page eventually..

No I plan to try that with my RCA 6AS7G's combined with some 6080's

I dunno if they are equivalent to mix ask Coin.

Do you know what the schematic for the "baby amp" is?

There is some of the schematic in the APPJ thread.
It using a switching supply, triode strapped pentodes(beam-power tetrode) surface mount parts...
There are TO-220 transistors by each of the driver tubes which not in the schematic here:
https://www.head-fi.org/threads/any...une-appj-pa1502a.780178/page-24#post-13480908

 

[baronbeehive]

Getting the LD back again has had one undesirable side effect after being used to listening though speakers. The soundstage of my HE-500's is not so good so everything is now pointing to the HEK V2's at some point in the distant future. This is the one area that is so much worse than listening with the APPJ. The other thing that the APPJ does better IMO is the bass response is slightly sharper, although quantity and impact is similar. However that could be down to the RCA power tubes because they do have that characteristic sound.

It's difficult to compare speaker amps with headphone amps though but they are remarkably similar in performance, and the APPJ is single ended of course.

 

[MrCurwen]

There are TO-220 transistors by each of the driver tubes which not in the schematic here:
https://www.head-fi.org/threads/any...une-appj-pa1502a.780178/page-24#post-13480908

Interesting. The FET in the output tube cathode node is part of a "fixed bias" system. It's there to make sure the output tube conducts the same DC bias current at all times. However it does nothing at AC, and therefore does not in any way improve or decrease sound quality.

Sound quality of that particular module is determined by C10.

The extra transistors you mention are probably doing something similar, since if the amp has anode resistors included, that part is taken care of, and the input tube grid is the least likely place to utilize grid drive. So, PSU related probably.

It's a nice retro amp, I've built several of these. 6V6 however is not that linear in SE mode; I guess that's why there's NFB included (not usually used in SE amps).

It's difficult to compare speaker amps with headphone amps though but they are remarkably similar in performance

They are both retro type circuits in essence (elaborate DC bias schemes notwithstanding), and while dissimilar in topology, after NFB is applied they are bound to sound more than a little bit alike.


The most interesting part for me is this:

However that could be down to the RCA power tubes because they do have that characteristic sound.

That amp uses 6V6 type tubes. The LD does not, nowhere near even. By what is this "RCA sound" transferred?

 

[MrCurwen]

Oh how did I miss this, the 6V6 is operated in TETRODE MODE, so NFB is required, not optional. THD will be thru the roof otherwise.

Just for fun, here's what I'd do to improve sound quality of that amp:

1) Remove NFB. This will decrease sound quality by increasing THD and output impedance, and worsening freq response.

Now then to gain back these things and possibly much more:

2) 6V6 goes triode mode.

3) Cathode bias of output tube goes out the window. No reason to play with C10 rolling. Instead use fixed grid bias; cathode is connected to ground, negative bias voltage is delivered to the other end of R15.

4) 6V6 is found too unlinear in SE mode without NFB. So, go for a linear tube. True triode 1626 is perfect replacement (even if more expensive these days) if 12V heaters can be supplied (which are apparently available). Others are available, such as EL36 / 6P31S, which is linear enough in SE and takes 6V heaters.

5) Fixed grid bias for input tube as well. It really is the most transparent bias method. Cathode bias always messes up dynamics to some degree, esp. bass.

6) Change R19 - 22 to one FET, two resistors and one cap. A simple gyrator costing a few pennies.

At this point sound quality is probably significantly better as a whole then at the start. Further improvements would be regulated B+ (very important in SE amps, those two 100µF caps at B+ are bound to be very slow and dusty sounding) and grid driver for output tube.

Parafeed cannot be arranged without completely redesigning the PSU (more voltage would be needed), so that would be the end of that road.


I realise there is no space inside the chassis for any of this, but if that amp is found to sound good, why not go adventuring on a new build, starting from about the same place as that amp.

 

[baronbeehive]

..........

The most interesting part for me is this:



That amp uses 6V6 type tubes. The LD does not, nowhere near even. By what is this "RCA sound" transferred?

Just to be clear in case of misunderstanding, my LD has RCA 6AS7G power tubes whereas my APPJ, the speaker version, has EL84's. I've found that for the circuits I've used them in, the RCA's pretty much in general have good quantity bass but are characteristically quite a soft smooth, balanced sounding tube, hence my comment that the APPJ has a slightly tighter bass sound whereas the LD is slightly softer and smoother sounding. I've found in the 2 or 3 circuits I've used them in that these tubes have a characteristic sound regardless of circuit. For example I prefer Tungsols on both amps for their sound. Also I found that in may cases tube makes have a similar sound, for example I can tell a Tungsol from a Sylvania regardless of whether it is a 12AX7 or a 6SN7, therefore tubes by the same manufacturer but completely different construction. This doesn't always hold true but I find that quite strange, especially in view of your comments that in a well designed circuit tube types should have no effect on sound assuming equal load lines characteristics and such like.

BTW I like your ideas to mod the APPJ.... where does it all end!

 

[baronbeehive]

.......

I realise there is no space inside the chassis for any of this, but if that amp is found to sound good, why not go adventuring on a new build, starting from about the same place as that amp.

Maybe there is room, another extension maybe.....

Chassis already extended:

The insides of the APPJ speaker amp to show circuit layout.

 

[baronbeehive]

I've tried the Mullard ECC35's today, nice tubes!

Next is driver bias mod when I get the resistors, then after that the impedance mod when I get the last 421a.

 

[MrCurwen]

Just to be clear in case of misunderstanding, my LD has RCA 6AS7G power tubes whereas my APPJ, the speaker version, has EL84's

Oh, ok. Why is this? The schematic you linked has 6P6 which is soviet 6V6GT.

I can tell a Tungsol from a Sylvania regardless of whether it is a 12AX7 or a 6SN7, therefore tubes by the same manufacturer but completely different construction.

Both in a circuit with a resistive anode load of less than 10x rp --> very small rp differences will be altering the function of the circuit.


Theory time:

Everybody understands how a voltage divider works, right? You have two resistors between + and ground, and the junction node voltage is determined by the resistors.

Now imagine you make the bottom resistor a potentiometer. You wire it as an adjustable resistor. You take your hand and you turn the wiper; the junction node (which is above the pot of course) voltage then varies, because the top resistor stays the same.

Are you following so far? Simplest ohm's law stuff, nothing to worry about.

Now if you make the + voltage really big, let's say 300 volts, and the top R 100k, and the pot max value is 100k. You now have a certain adjustment area.

Let's say your hand moves in tandem with a sine wave you see somewhere. Now as you move the pot wiper, you are in fact controlling the R / pot junction node voltage.

Let's say you move the pot all the way from open to max value during the sine cycle. You get junction voltages between 0V and 150V.

Now if you keep the R same value, and change the pot to a 90k max value pot, you get junction voltages from 0V to 142.105 V. Same input, your hand moves the same amount, but different output.


A tube is actually just an adjustable impedance. The above example can be expanded and clarified to apply more to an actual tube, but it would become much more complicated when you take into account resistance vs. impedance etc. It's a good enough analogy as it is.


So the principle is that if you have anode R that is not very big in comparison to rp, then if rp varies (it varies during operation, I mean statically between different tubes) it will cause the circuit to operate differently.

If anode R (or rather, Z) is large enough to drown out the influence of rp (remember, they are always in series between + and ground, one of them is by necessity dominant in determining the current thru that node and the junction node voltage), then variations in static rp (between different tubes) is meaningless.


What you are hearing is that some manufacturers are "over spec" in some specs, and some are "under spec". Whether or not these are consistent across different tube types within a manufacturer seems a bit odd, but not completely out of possibility.


Anode resistance is key here. A reason why some tubes sound "dark" and some "bright" is because a small variation in rp changes the R in an RC lowpass or highpass filter somewhere in the circuit. Why somebody would want to make a circuit that is so vulnerable to small variations is beyond me. I try to eliminate them.

The tube itself isn't "dark", you just dropped in a component that had "the wrong value". Let's say you'd have an RC highpass filter. You have a C of 100nF. Now you solder in as the R a random resistor. You cannot see it's value. What frequency pole ("sound") will you get? This is tube rolling in an essence.

 

[baronbeehive]

Oh, ok. Why is this? The schematic you linked has 6P6 which is soviet 6V6GT.

.....

The confusion arises from the fact that there are 2 APPJ's, the headphone version which Maxx linked to the schematic, and the speaker version which I have, both with different power tubes, in fact I believe there is a third version now.

Oh, ok. Why is this? The schematic you linked has 6P6 which is soviet 6V6GT.

.........

What you are hearing is that some manufacturers are "over spec" in some specs, and some are "under spec". Whether or not these are consistent across different tube types within a manufacturer seems a bit odd, but not completely out of possibility.


Anode resistance is key here. A reason why some tubes sound "dark" and some "bright" is because a small variation in rp changes the R in an RC lowpass or highpass filter somewhere in the circuit. Why somebody would want to make a circuit that is so vulnerable to small variations is beyond me. I try to eliminate them.

The tube itself isn't "dark", you just dropped in a component that had "the wrong value". Let's say you'd have an RC highpass filter. You have a C of 100nF. Now you solder in as the R a random resistor. You cannot see it's value. What frequency pole ("sound") will you get? This is tube rolling in an essence.

Yes I get that, and it's interesting how that translates into sound, ie, every GE I've heard sounds thin and clean, nearly every Tungsol sounds silky smooth, every RCA I've heard sounds very smooth and laid back, etc. What I wondered about the similarities between same manufacture tubes was something in the manufacturing process that is only known to the manufacturers themselves but it seems a bit esoteric.

Edit:I see there are now 2 new APPJ's, one with 2x 6AD10 tubes, the other has 2x6 JI+ 2x6P14's.