Little Dot MK8SE / MK6 Super Mods (All verified mods are on first page)

[baronbeehive]

Straight line in graph is not advisable as that would not taper off below 20hz so your in now suspectable to circuit oscillations or anything outside the audible range.

OK, this is the best I could come up with, still a bit of a slope at around 100Hz which may or may not affect mid bass:

Same figures but for 6SL7, (5751):

May be possible to refine it but listening tests would be key here.

I have a feeling that the cathode bypass cap is even more important than I originally thought which makes the work of CopperFox that much more useful. I certainly found it made a big difference when I swopped to the Nichicons.
.

 

[baronbeehive]

I've tried to relate what @CopperFox found here: https://www.head-fi.org/threads/lit...e-on-first-page.782183/page-259#post-16430754
The problem is that with the calculator you need to enter quite large differences in capacitance to see differences, either that or the frequency curve is not that fine grained as it appear in the graph. For example below are 3 settings, on the left with a low capacitance of 100uF, in the middle is 440uF, and to the right is 570uF and as you can see the curve is more pronounced in the first graph with little differences between the second 2 graphs.

The overall gain, (vertical axis), is unchanged.

I don't remember what figures @CopperFox is using for his anode/cathode resistors.

 

[Maxx134]

OK, this is the best I could come up with, still a bit of a slope at around 100Hz which may or may not affect mid bass:

From what I recall, usually the average design is to start roll off at 20hz.

Also, another point is that the
formula is showing for frequency responses. It is not showing the main parameter we are also looking for to decide on the cathode cap size. ..
Which is to keep the volt drop stable across the cathode resistor with varying load.

So we know film cap will respond faster than electrolytic (better transients)
So we know effective size would be smaller than electrolytic.

There should be a minimum size.
I don't know what formula is available for that. Would have to calculate the drain over time according to the load resistance.
Then again we don't need to reinvent the wheel and can assume and test real world usage.

 

[CopperFox]

Well that was a lot of work. I now have replaced the signal wires with SPC wires and all of my extension leads with pure silver wires. 14 meters of wire in total, some of which was bound together.

The output signal wires going from the PCB (or actually some of them from the tube sockets?) to the 2*3 pin XLR outputs and from those to the 4 pin XLR headphone out were not connected to anything else so I also routed the output wires straight to the headphone out, skipping the XLR outputs in the rear. Also removed the single ended headphone output as I'm not using that for anything. I will probably later remove the RCA inputs as well but decided to let them stay there for now. I decided to keep the old transistor wires as well.

Edit: Removing the rear XLR connectors and/or RCA inputs would also make some extra room inside the case for switches/capacitors/whatever.

Sound is improved but no more comments on that for now as I've been changing the cathode resistors with different brands/models as well and on Friday got a package with four 0.56uf Duelund JDMs and I'll still be changing those (and likely more different resistors) soon. The amp probably also needs some re-burn-in and such after having been largely dismantled for days.

It is surprising that the 100R resistors in the output leads were not mentioned before in the thread and I also now removed them.

Last year I rembember modding a Zishan DSDs dap that also had 100R resistors in the signal path just before the outputs. It was thought that they were there for some sort of protection but people would just remove them or replace them with far lower value resistors (resistance in the single digits). With that particular dap I put them back in there as I felt sound would be too brigh twithout them. But that's off topic of course.

 

[CopperFox]

The point about whether it would fit in the chassis is irrelevant to me as my chassis is choc full anyway, so I would have to extend it .

How large a chassis extension are you planning on?

Fun?????

Isn't modding fun? If doing hifi stuff is like wine tasting, then modding would be like wine tasting where the wines can be made from lego bricks.

Haha, don't know how you managed to come up with this!

Google image search with "little dot mkvi+ schematic"

 

[CopperFox]

Here is an online tool I've found that will show ideal values for the cathode bypass cap which should take the sweat out of decision making on the size of cap: https://www.ampbooks.com/mobile/amplifier-calculators/cathode-capacitor/calculator/

So whether to use 100uF or 200uF should become obvious and we might be able to dispense with guesswork on ratios. However it doesn't deal with the differences between lytics and film caps but it should show the point at which it is overkill in going any higher in values of microfarads for example. For example it would show up on the graph whether there would be anything to gain having reached an optimum point, on going any higher in value because the graph would stay much the same if that makes sense. It would also show optimum values for bass, mids and treble so depending on your preferences you could see instantly what cap to use.

That calculator seems interesting. Its website states that its function is based on the page 484 of Radiotron Designer's Handbook from 1953, which can be found in its entirety as .pdf here. On the previous page (483), there is also a formula for the voltage drop caused by the cathdode bypass capacitor, which includes the total impedance of the capacitor where its resistance would also be a factor.

As for the results of that calculator, it seems the 200uf would not have a lot of bass roll-off and it would be something in the 0,1 to 0,2 dB range. With the 100uf it would be higher than that... and to lower it perhaps something higher could be also tried but that would probably not fit in my extended chassis.

 

[Maxx134]

so I also routed the output wires straight to the headphone out, skipping the XLR outputs in the rear.

That's what I did as well. A straight run from the tube pin to the headphone XLR socket.

Also removed the single ended headphone output as I'm not using that for anything

That actually good place for putting your switch (impedance or tube change switch)

I decided to keep the old transistor wires as well

Nooooooooowaaaa!..
Actually I don't know if they changed the wire, but in past the transistor (in power supply) wire used become very brittle and breaks off easily on board, causing PSU failure.
I would check/replace those wires.

It is surprising that the 100R resistors in the output leads were not mentioned before in the thread and I also now removed them

Yeah I don't believe they were there before...

 

[baronbeehive]

Nooooooooowaaaa!..

Haha .

Nooooooooowaaaa!..
Actually I don't know if they changed the wire, but in past the transistor (in power supply) wire used become very brittle and breaks off easily on board, causing PSU failure.
I would check/replace those wires.


Yeah I don't believe they were there before...

No I didn't come across these and I replace those wires too.

 

[baronbeehive]

How large a chassis extension are you planning on?

It depends on the size of the cathode film caps. I was thinking of keeping the large capacity lytic PSU caps and the 47uF film decoupling caps but these may go too so when I've decided on that I will get dimensions and look for the appropriate size heatsink plates and drill through for the standoffs and screw straight into the bottom edges of the chassis.

Isn't modding fun? If doing hifi stuff is like wine tasting, then modding would be like wine tasting where the wines can be made from lego bricks.

Yes, but my head hurt when it all went wrong, I was without the amp on two occasions initially when I couldn't find the fault and then again when if developed another fault it took me some time myself to put it right.

Google image search with "little dot mkvi+ schematic"

Good old google . I did remember after you posted it that I saw it on that thread after all.

 

[baronbeehive]

Also, another point is that the
formula is showing for frequency responses. It is not showing the main parameter we are also looking for to decide on the cathode cap size. ..
Which is to keep the volt drop stable across the cathode resistor with varying load.

Exactly! We need to know how the cap handles the transients too and as I said above the listening tests would be crucial to hear what differences in the curve have on sound, and of course to see what the transient handling is like. These might appear inconsequential on the graph but might be important.

The gain shown on the vertical axis I'm sure influences how lively and dynamic the sound is. I looked at what sonic, you and me did and you can see differences in this shown on the graph.

As I thought when I did my tests a while back the cathode resistor value doesn't affect the sound as much as the anode and so it would be more for fine tuning.

So, the cathode bypass values for the overall curve, then the anode values to modify the curve and finally the cathode resistor value for fine tuning.

Where the graph was useful was in finding the point at which no further increases in capacitance are necessary or are completely overkill.

 

[baronbeehive]

Well that was a lot of work. I now have replaced the signal wires with SPC wires and all of my extension leads with pure silver wires. 14 meters of wire in total, some of which was bound together.

The output signal wires going from the PCB (or actually some of them from the tube sockets?) to the 2*3 pin XLR outputs and from those to the 4 pin XLR headphone out were not connected to anything else so I also routed the output wires straight to the headphone out, skipping the XLR outputs in the rear. Also removed the single ended headphone output as I'm not using that for anything. I will probably later remove the RCA inputs as well but decided to let them stay there for now. I decided to keep the old transistor wires as well.

Your wiring diagram was obviously flawless . I quite enjoyed rewiring, it wasn't so fiddly as the other work, but I still managed to make a mistake in one of the connections, easily put right fortunately. That's where the DIMM came in handy.
.

 

[baronbeehive]

There should be a minimum size.
I don't know what formula is available for that. Would have to calculate the drain over time according to the load resistance.
Then again we don't need to reinvent the wheel and can assume and test real world usage.

The formulae for the calculator are in the book by ampbooks.com, and I think that all the formulae are there for that.

 

[baronbeehive]

That calculator seems interesting. Its website states that its function is based on the page 484 of Radiotron Designer's Handbook from 1953, which can be found in its entirety as .pdf here. On the previous page (483), there is also a formula for the voltage drop caused by the cathdode bypass capacitor, which includes the total impedance of the capacitor where its resistance would also be a factor.

As for the results of that calculator, it seems the 200uf would not have too bad bass roll-off and it would be something in the 0,1 to 0,2 dB range. With the 100uf it would be higher than that... and to lower it perhaps something higher could be also tried but that would probably not fit in my extended chassis.

Yeah, just read your post and yes all the formulae are there, that pdf looks useful.

I think with the bass roll off it would need to be listened to to see the effects. You can see that even with the stock amp there is very little roll off and the rest of the curve is flat and linear so LD obviously thought they had got it right for the combination of tubes that would be used. A steeper bass roll off and a shallower one could be tried to see the effects on sound. As 20Hz is at the limits of hearing a steeper roll off might not be too bad and might result in a flatter curve on the rest of the frequency band.

 

[baronbeehive]

It is surprising that the 100R resistors in the output leads were not mentioned before in the thread and I also now removed them.

Last year I rembember modding a Zishan DSDs dap that also had 100R resistors in the signal path just before the outputs. It was thought that they were there for some sort of protection but people would just remove them or replace them with far lower value resistors (resistance in the single digits). With that particular dap I put them back in there as I felt sound would be too brigh twithout them. But that's off topic of course.

If you knew the current and voltage going through those wires you could calculate the voltage drop across the 100R resistor using Ohm's law.

 

[Maxx134]

From my experience, most amps have .22uf as the standard value for the coupling cap for many designs, and that must mean that it is so gradual a change, that many design make use of that value.

Yet the Cathode cap is another story.
I always see wide range of values.
Just the other day I worked on a Cayin HA-300, and it has only 100uf(!):

So there is definite intentions in the design choice of the Cathode cap.
I believe it is where most all of the tube interaction begins.

Also, if you take note there are specific type of resistors for the tube stages. They are not films, as in the PSU.
This is yet another example of design choice for specific sound, like the ingredients in a soup.

I regard Cayin the absolute best, at making euphoric sounding amps, without the harshness of unwanted harmonics.
They are genius, when you observe their parts choices...

While I regard Little Dot path, is to use more of "design choices", rather than parts choices, for their sound. Just looking at the bias points of the driver stage, is a prime example.