Little Dot MK8SE / MK6 Super Mods (All verified mods are on first page)
Sep 15, 2018 at 4:00 PM Post #3,481 of 4,154
Now I choose tubes based on linearity, gain and looks. I don’t build amps for tube rolling.

Good criteria.

If I may add, maybe also variance. Meaning, trying something different for a new build. Choose a less known tube type and take it as a challenge.


If you have the space, why sell tubes! I have thousands in my shelf. Someday I might need them for a project. Maybe I'll build a hundred guitar amps or something.

As an ironic sidenote, my workhorse uses 6E5P tubes. They've been in the amp about a year now, they see over 10 hours of use every day. I just have the amp on most of the day (turn-on and turn-off cycles are worst on the tubes). Anyway, the op point is pretty harsh for them, especially since the triode mode 'anode' is just a tiny tiny mesh grid.

It's started to get microphonic, sometimes if there is a loud transient in the sound there is an echo after it. It comes from electrostatic movement of the grids and their support structures.

I've got thousands of tubes in my shelf, but no more 6E5P. I have to buy some more soon or find time to solder another configuration. Maybe EL81, I've got dozens of them and they should be more rugged anyway...
 
Sep 16, 2018 at 5:13 AM Post #3,482 of 4,154
So I am going to try the CCS, then swap out the output tube cathode caps with some others (Muse caps), and then WCF caps I have (are to romantic) for more clear ones .

Then give up lol.

I just came across this link, don't know if you've see it questioning the Muse caps soundstage:
http://funwithaudio.blogspot.com/2011/12/capacitors-brief-review.html

While I'm here, these are the other links on 6v6's and 6L6's:
https://www.watfordvalves.com/cgi-bin/documents/testreport_45.pdf
https://www.watfordvalves.com/cgi-bin/documents/testreport_44.pdf
https://www.watfordvalves.com/cgi-bin/documents/testreport_36.pdf

The authors are mainly involved in valves for guitar but also mention audio in their reviews. They put Mullards at the top invariably, but also like the GE's for clean, Fender type sound, and also soundstage, and Mazdas which I love on the other amp.
 
Sep 16, 2018 at 7:31 AM Post #3,483 of 4,154
I may do this too, after the improvement they've made to the other amp I think those Kaiseis must be something special
I have recently been experimenting with various caps on the Eddie Current Aficionado...
IMG_4093(1).jpg
and it was instrumental to me finding out exactly how the popular different electrolytic (cathode) caps affected the sound.

I ended up removing the Kaisie due to characteristics of slight lower end bloom & richness, and slightly lowered but liquid trebles.
Overall a good choice but not for ultimate transparency.

I just came across this link, don't know if you've see it questioning the Muse caps soundstage
That was an error on part of the writer.
His negative observations were on the lower FG line, not the Muse (KZ) line.
The Nichion Muse is not affected in soundstage.

Also those impressions did not state the Elna Silmic2, only the original Silmic.
Posts from around the web referring to the Silmic2 better than original, but having treble edge or elevation, so I stayed away.

I also tried the latest Jensen audio grade caps and found them to be clear and authoritive in lower range, but also very dark and so not suitable for the cathode cap position.

So far the Nichion Muse (KZ) cap gave me most neutral sound and clarity so they are used in place of stock caps, and jupiter cap for coupling .

This elevated the Aficionado above the Cayin HA300, and the solid state Benchmark HPA4 amp in different ways.

In stock form, all of these amps are a "lateral-move" preference choice.
All over $3k and with tubes over $4k.
I do not own any so had unique perspective of testing without owner bias.

I could not compare to my own amp as I am not finished my relocating of residence.
But from memory its not far off lol.
 
Last edited:
Sep 16, 2018 at 8:35 AM Post #3,484 of 4,154
This is a stab at establishing the connections for the filter. It's the best I can do with my present (lack of) understanding, I hope you can put me right.

upload_2018-9-16_13-24-58.png


  • the existing connection to B- is disconnected and the new B- goes from M1 drain to CCS instead
  • the M1 gate stopper is 1K R7
  • the voltage source is connected to M1 source via 10K R8
  • the ground connection is indicated by the down arrow
I'm not sure how to connect to the voltage source.
 
Sep 16, 2018 at 2:17 PM Post #3,485 of 4,154
V4 is the old B-.

Ground is ground, 0V.

From the CCS point if view simply substitute new B- (output of filter) in place of old B-.

Look up FET pinout from datasheet. From memory it's "the same" as n-channel FETs, simply think of the middle pin as 'drain' (rail voltage, raw voltage input) and right pin as 'source' (output of filter). Left pin should be gate.
 
Sep 17, 2018 at 1:49 PM Post #3,486 of 4,154
they have mentioned that the soundstage is largest with GE and Mazda tubes
Yeah I tried the Mazda very nice yes but nothing special.


V4 is the old B-.

Ground is ground, 0V.

From the CCS point if view simply substitute new B- (output of filter) in place of old B-.

Look up FET pinout from datasheet. From memory it's "the same" as n-channel FETs, simply think of the middle pin as 'drain' (rail voltage, raw voltage input) and right pin as 'source' (output of filter). Left pin should be gate.
Where does the R6 go?
So V4 (old B-) goes to R6 then, correct?
That diagram has V4 typed on the bottom of the symbol..
 
Sep 17, 2018 at 2:33 PM Post #3,487 of 4,154
Where does the R6 go?
So V4 (old B-) goes to R6 then, correct?

Yes.

R6 is between unfiltered (old) B- and R8, C3 and M1.

That diagram has V4 typed on the bottom of the symbol..

Yes? V4 is the whole symbol, not for example for some pin. Ground is ground, so it cannot be input for B- filter.

I should've just left V4 out and put a text there. It's not that complicated.
 
Sep 17, 2018 at 4:46 PM Post #3,488 of 4,154
The yellow and green symbols in the diagram, for some reason I thought the direction of travel was reversed in the P-channel FET, that's why I crossed out drain and source and reversed them. I think that I was right in the first place, no doubt you will say if I'm wrong.

So the bullet points should be changed as well:
  • the existing connection between B- and CCS is disconnected and the new B- goes from M1 source to CCS instead. The old B- is connected to the filter via R6.
  • the M1 gate stopper is 1K R7
  • the voltage source is connected to M1 drain via 10K R8
  • the ground connection is indicated by the down arrow
I will try to knock up a dummy filter out of odd resistors just to get the connections.
 
Last edited:
Sep 18, 2018 at 9:26 AM Post #3,489 of 4,154
These are the connections for the filter I think....

It's a dummy run so I've labelled the parts:

upload_2018-9-18_14-26-7.png


…. you can tell me where I've gone wrong.
 
Last edited:
Sep 18, 2018 at 9:52 AM Post #3,490 of 4,154
Looks about right. Remember to orient the electrolytics so that + terminal is connected to ground. Capacitance values are not important, you can double them or halve them, makes no difference. If you have the space (which you don't, but in theory) C4 could be for example 4µF / 160V MBGO capacitor.

Increasing C3 doesn't really help, it simply increases turn-on time. It's there more as insulation (between raw supply and the active filter FET) and oscillation control than as a filter. C4 shouldn't be too big also since that can increase turn-on time too much.
 
Sep 18, 2018 at 5:03 PM Post #3,491 of 4,154
Looks about right. Remember to orient the electrolytics so that + terminal is connected to ground.

Shouldn't that be negative to ground like in the pic below, or is it different for a filter?

upload_2018-9-18_21-53-26.png


How do I connect the whole filter circuit itself to ground?
Also do I need to connect the 2 points marked with a question mark to complete the circuit or just have the B- coming out from old B- to R6?
 

Attachments

  • upload_2018-9-18_21-52-21.png
    upload_2018-9-18_21-52-21.png
    67.6 KB · Views: 0
Sep 18, 2018 at 5:23 PM Post #3,492 of 4,154
Shouldn't that be negative to ground like in the pic below, or is it different for a filter?

Polarized capacitors should always connect the - pin to the more negative side. Which side is more negative here, ground (0V) or the negative rail?

How do I connect the whole filter circuit itself to ground?

You take a wire, connect one end to the filter module ground, and the other end to the circuit ground. Doesn't really matter where, just as long as it's after the B+ decoupling cap. Any ground within the audio circuit itself is fine.

Also do I need to connect the 2 points marked with a question mark to complete the circuit or just have the B- coming out from old B- to R6?

No, disregard that part completely. Once again, V4 is just for the Spice simulation. Forget it once and for all. V4 symbolizes the whole of the original B- supply, with a big ripple voltage included. That's all.
 
Sep 19, 2018 at 4:11 AM Post #3,493 of 4,154
Looks about right. Remember to orient the electrolytics so that + terminal is connected to ground. Capacitance values are not important, you can double them or halve them, makes no difference. If you have the space (which you don't, but in theory) C4 could be for example 4µF / 160V MBGO capacitor.

Increasing C3 doesn't really help, it simply increases turn-on time. It's there more as insulation (between raw supply and the active filter FET) and oscillation control than as a filter. C4 shouldn't be too big also since that can increase turn-on time too much.
Hey, now I recognize it. It's basically a filament reg with a different FET and fixed resistors instead of trimmer. Also much lower value for C4.
 
Sep 19, 2018 at 4:24 AM Post #3,494 of 4,154
Hey, now I recognize it. It's basically a filament reg with a different FET and fixed resistors instead of trimmer. Also much lower value for C4.

Yep. Fixed resistors because output voltage exact value is not important, like a filament voltage is. FET difference reasons are obvious; p-channel for negative side operations and low Rds for high current operations.

Actually it's the same as the plate loads, just a high value for C4.
 
Sep 19, 2018 at 4:27 AM Post #3,495 of 4,154
Polarized capacitors should always connect the - pin to the more negative side. Which side is more negative here, ground (0V) or the negative rail?

Ah, so I've got it the wrong way round in the pic. Of course the negative rail is more negative than 0V.


No, disregard that part completely. Once again, V4 is just for the Spice simulation. Forget it once and for all. V4 symbolizes the whole of the original B- supply, with a big ripple voltage included. That's all.

OK, just wanted to check that.
 

Users who are viewing this thread

Back
Top