The DIY'rs Cookbook
Nov 21, 2018 at 7:38 AM Post #1,351 of 1,959

johnjen

Headphoneus Supremus
Joined
Apr 13, 2010
Posts
3,418
Likes
2,928
Location
Well Grounded-here and now
So I fired up the BIG eyeyrrrrnnnn amp and have been listening to it, on and off, during the day and now at night.

I added 2) line level 1:1 input xfrms to the circuit, which in turn means I now can 'properly' use the balanced output from my ROK amp or the JggyB dac direct.
There is less hum, it is now nearly equal in both channels, it has dropped to where I only hear it during the quietest passages or when nothing is playing.

There is still some tweaking to do to the bias circuits, and the filaments circuits, and finish up the AC line voltage step down xfmr setup.

But the amp is 'listenable' now, is engaging, has 'enough' of the Magic in the Mids, without the distractions of annoying hum and noise and a lack of MitM.
Which brings the eyerrrnnnn count up to 17 with future plans now to replace 4 plate chokes with 2 alternatives and also add 2 grid chokes.

And as soon as my 2" wide rubber bands show up I'll start using the AC mains step down xfmr setup for even more ac power stability.

JJ
 
Nov 24, 2018 at 7:23 PM Post #1,352 of 1,959

johnjen

Headphoneus Supremus
Joined
Apr 13, 2010
Posts
3,418
Likes
2,928
Location
Well Grounded-here and now
A BIG eyerrrrnnn amp update.

I nailed the 45 filament circuit voltages to within 0.001Vdc between the left and right channels regardless of what the ac line voltage is.
This has resulted in yet more balance between the 2 channels such that the SQ has risen (although it is subtle) to new 'Better' heights.

Next up is seeing if I can dial in the plate chokes that feed the 2) 6J5 driver tubes by arranging how each pair is wired together.

And we have begun the process of designing the 'final' build's layout which includes the size of the case etc.
And of course no matter how we arrange the layout at this point, we KNOW it will change as we progress AND again as we actually install all the eyerrrrnnnn (and remember to use that texas drawl for the full effect) :ksc75smile:

We are using the z axis (height) to an even greater degree while trying to shrink the footprint.
We aim to optimize the 'flow' of the power supply and perhaps even more significantly the analog circuits, such that we avoid problems and optimize the results (lowered hum and noise, which yields increased SQ).
Which means we should be able to even further match the channels together to a greater degree for even more of that Magic in the Mids.

And right now Bonnie Raitt is sounding really good with an extra dollop of REALNESS as only triodes can provide.
The amp is starting to blossom, just a bit as the MitM peeks from around the corner.
I can hear Bonnies voice with moisture on her breath and it is starting to smell of what she last had a drink of… :) hahahahahahahahahahahahaha

JJ
 
Nov 30, 2018 at 7:03 AM Post #1,353 of 1,959

johnjen

Headphoneus Supremus
Joined
Apr 13, 2010
Posts
3,418
Likes
2,928
Location
Well Grounded-here and now
BIG eyerrrrnnn amp update.


So after several more tweaks (adding .22µfd bypass caps to the output caps, creating a faux center tap for the 6.3vac filament supply, adding the ac power voltage drop xfmr stack, adding more 100µfd caps to the PSU, and removing yet more resistorators from the audio path) we are now figur'n on adding interstage xfmrs instead of using plate and grid chokes for the 6J5 tubes, and changing the solid state rectifiers to tube rectifiers instead.

But the amp's SQ is blossoming with MitM (Magic in the Mids) which is not just peeking out from around the corner anymore, but is staring me directly in the face, AND singing into my ears.
And I'm smiling as I notice as more and more of these SQ changes, that are highly sought after from tubes, start to manifest themselves.
And I'm still using my circuit development tubes, not my tweako set (the 45's from 1938 and the made in 1944 6J5's).

And that range of line voltages from the dedicated branch circuit for my audio system that runs from 120 to 125Vac, has been reduced to 120 to 123Vac, then down to 115 to 117Vac, due to the final step down xfrmr.
This has stabilized all the working voltages in the amp and brought them into their optimal operating ranges.

Thus far I have ≈ 39hrs and I'm adding ≈ 5-7hrs/day of run time, and the amp is responding well to our careful tweaking and experimenting with various sections (PSU, bias circuits, etc).

And it appears that I will have to build a pre-amp as well.

My ROK amp has a 'feature' where it won't shut off the speakers AND send the audio signal out the pre-amp outputs, at the same time.
This means I can't listen to just my 800-Jmod HP's from the BIG amp and not have the speakers on at the same time.
Fortunately the speakers run at a MUCH lower volume than the 800's, still, this is cause enough for a 'proper' dedicated pre-amp.
And with all of the 'extra' parts I'm accumulating from this BIG amp build, I should have plenty left over to build a decent, simple, and 'proper' pre-amp.

But first things first, as we continue to dial in this BIG eyerrrrrnnnnn amp.
And I AM enjoying the SQ as it continues to blossom and tantalize me with more and more of that MitM.

JJ
ps Next up will be Part 2 of my power cable series.
 
Last edited:
Dec 2, 2018 at 7:30 AM Post #1,354 of 1,959

johnjen

Headphoneus Supremus
Joined
Apr 13, 2010
Posts
3,418
Likes
2,928
Location
Well Grounded-here and now
Recently things have been changing much more quickly, and for the 'better', as we have hit several 'licks' in a row.
And with a few more in the works, that should help solve a few nagging problems, that still need to be 'put to bed' so to speak.
And one of those tweaks will make this amp 'battery powered', sorta kinda.
I figure this is in keeping with the 'OLD skewl' theme for this amp.

The most niggling of them all, is the all to common for tube circuits, the hum and noise levels are still higher than we want.
Right now those levels are hovering in the -40 to -45dB level and we have improved them such that the hum (the most audible contribution) is 'softer' and less objectionable.
Still when the source of this type of sonic veil is lifted, the SQ should take a considerable step up.
And we have a plan, or well, we're working on a plan now that we have a clear goal and direction to pursue.

But one of the most recent tweaks was adding more isolation for the source of B+ that feeds the tubes, from each other.
IOW I added a 'separate' cap that feeds the 6J5's so that both tubes are getting more 'customized' voltage sources.

This was sort of the icing on the cake so to speak.

Last night and again tonight I simply CAN'T take my 800's off my head.
No matter what music I play.
Nor how 'loud' (how much I crank on the MOAR knob) I play this system at.

A case of SDSG, and with the blossom action happening right into my ears, my ability to want/need to leave my 'listening area' has diminished, until the very last possible second.
Like going to bed.
(last possible second as in fall asleep in my chair) hahahahahahahahahahaha

And this is still on my 'development' set of tubes, the 'expendable ones'.
My pre-WWII 'vintage' tubes are just waiting until the circuit is truly ready for them.

I figure that too will be a big step up, just because these matched sets are really closely matched.
And were made during the height of the tube era.

But for right now typing this and listening to Paul Simon or Joe Satriani or Pink Floyd, or The Pachelbel Cannon, etc, or any of the rest of the 10 hrs of music running in a continuous loop, has got me by my ear hairs and won't let go.

Yeah that MitM is parked in this amp and has started setting up permanent residence.
And I like the results thus far, and it will only improve from here onward.

JJ
 
Last edited:
Dec 10, 2018 at 5:46 AM Post #1,355 of 1,959

johnjen

Headphoneus Supremus
Joined
Apr 13, 2010
Posts
3,418
Likes
2,928
Location
Well Grounded-here and now
So we just installed a different bias system for the 6J5 tubes and it has improved the SQ yet again, albeit in a small way.
I used a 9v battery instead of using UV LED's, so now my BIG amp is battery powered.
And we have plans to do the same for the 45 tubes as well.

Next up is an interstage xfmr to help stabilize the 6J5's and directly feed the 45 tube, without any coupling cap.
This should increase the SQ of the amp with easily noticeable changes.

And in related news…

We are now working on a pre-amp to use to feed this amp instead of using the ROK as a pre-amp, which has an idiosyncrasy of not being able to turn off the speakers AND send the analog audio signal to an external amp (my BIG amp).

And currently the design for this pre-amp will also be able to drive my 800's (or just about any set of HP's for that matter) as a 'portable' HP amp.

It will use the same tubes as my BIG amp (2 x 6J5's, and 2 x 45's) and in most ways it will be a Mini-Me of the BIG amp, only scaled way down with a simpler PSU and using variable impedance output xfmr's to drive either the BIG amp or HP's.

This means I can take my pre-amp to meets as a 'portable' HP amp and also use it as a volume control for the BIG amp, at home.

And my BIG amp is settling in nicely as it accumulates hrs (101) and continues to surprise me with subtle and not so subtle acoustic and harmonic nuances.
At times the bass seems to have additional texture and harmonic over and under tones that are most welcome, which adds 'richness' to the extreme bottom end, where you normally don't expect such details because hearing (feeling) that portion of the frequency spectrum is uncommon.
IOW this amp even at this stage of it's development is the equal of the ROK amp in this regard.

And elsewhere, (those magic mids) this BIG amp is already capable of greater resolution and inner details and an added degree of the 3C's (Cohesion, Coherence, Coupling) and we really haven't 'uncorked' it yet.

There are yet more tweaks waiting beyond the interstage xfmrs and the battery bias for the 45 tubes, namely changing the SS diode bridge for a rectifier tube (GZ34 is top of the list) and dialing in the output coupling caps, along with full electrical isolation of all the xfmrs, and the possibility of majorly changing the output xfmr configuration.

But this HP/pre-amp combo deal has us both scratching our chins going, "ya know if this works it'll be pretty frack'n kewl".
To be able to do double duty, and do so with very few parts, (simplicity in the design, which keeps the costs down) and use the same set of tubes, and be an ideal pre-amp to drive the BIG amp, well, it will be an interesting build and basis to compare these 2 designs (Big amp & HP/pre-amp) to each other, since they both can use the same tubes.

JJ
 
Last edited:
Dec 10, 2018 at 2:40 PM Post #1,356 of 1,959

gefski

1000+ Head-Fier
Joined
Apr 29, 2003
Posts
1,263
Likes
976
Location
Pacific Northwest
Portable preamp as "fits in a shoebox" or "fits in the trunk"??????
 
Dec 11, 2018 at 7:44 AM Post #1,357 of 1,959

johnjen

Headphoneus Supremus
Joined
Apr 13, 2010
Posts
3,418
Likes
2,928
Location
Well Grounded-here and now
hahahahahahahahahaahahahahahaaha
Am I that easy to read? hahahahaahahahahaahahahahahahahaha

Yes it’s the "fit in the trunk" option.
So far it looks like 17"(L) x 10"(w) x 5"(h),
for a pre-amp.
Albeit a pre-amp with dual mode uses (HP-amp and pre-amp), but wait there's more…

The feature set is still expanding and morphing as new possibilities and uses for this HP/pre-amp are incorporated into the design.
Like being able to use a bevy of DHT tubes by adjusting the voltages (B+, filament) to suit the needs of each of at least 5 DHT triodes.

And it will handle HP's from 32Ω to 600Ω AND provide a 600Ω balanced line output to match the 600Ω input xfmrs the BIG amp is using.

And a really kewl feature will be the switch that electronically adds the 6j5 (or other) driver tubes, to add more gain to the 45 (or other) output tubes, in order to achieve the necessary output power in order to drive HP's.

It will also serve as a test bed to listen to a variety of tubes to then determine what tubes I want to settle on, to use when we build the 'final' dedicated tweako preamp only, that will drive the BIG amp.
And this HP/pre-amp will remain as a meet rig with whatever set of tubes that sound best in it.

At least that is the plan thus far.
Stay tuned as this design gets further refined etc.

JJ
 
Dec 14, 2018 at 6:49 AM Post #1,358 of 1,959

johnjen

Headphoneus Supremus
Joined
Apr 13, 2010
Posts
3,418
Likes
2,928
Location
Well Grounded-here and now
Plans for the HP/pre-amp have taken a 'just do it', step up in terms of where this build can go.
And when these changes settle down, the impact these design elements can make should be remarkable.
At least if what has just popped up on our radar, turns out to be what I think.
This may be frick'n cool to the max.

But actually this post is to mark the 'discovery' or realization that this BIG amp REALLY likes to get cranked on, as in rock and roll levels, leading to getting swept up in the music.
Yeah SDSG is still at peak levels and where tLFF has dropped to new levels of low.
Translated that means the sound is so insistent and compelling that I simply can't take my 800's off my head, and after cranking the MOAR knob for MOAR volume,
it just doesn't get loud,
but it does get MOAR.

The bass continues to surprise me with the depth, texture, coupling and extension that this amp is tossing at my ears.
The mids have a degree of solidity and depth that is startlingly captivating, not to mention intonation and the harmonic structure of each 'voice' is, if not more 'real', is far more captivating and engaging and compelling, as in, the thought of stopping the music simply does not occur, at all, no matter what time it happens to be.
And this is an interesting tid-bit, it doesn't seem to matter what music I play, it all has 'stick', it all keeps my 800's stuck on my ears.

And tonight I'm experimenting with the 3 different sets of 6J5's I have on hand.
All 3 sets were NOS (New Old Stock) which means they have never been used, except on a tester after being manufactured back in the the 40's and 50's.
That's quite a remarkable statement it seems to me.

Anyway I like the GT versions (Glass Tube) more than my development set of metal cans that I've been running thus far.
Which means I already have the driver tubes for the above mentioned HP/pre-amp.

MOAR now and More later.

JJ
 
Last edited:
Dec 19, 2018 at 7:58 AM Post #1,359 of 1,959

johnjen

Headphoneus Supremus
Joined
Apr 13, 2010
Posts
3,418
Likes
2,928
Location
Well Grounded-here and now
Lately I've been fussing with the amount of capacitance that is used to take the music signal from the 45 output tube and then pass it along to the output xfmr.
It has to block ≈ 295Vdc yet still let the ac audio signal thru to the parafeed output xfmr.
This is a 'tuned' circuit and dialing in the 'proper' amount of capacitance here is part of the process of tweaking the circuit for maximum bandwidth and coupling, especially coupling from well below 20Hz all the way up thru the audio band.

Right now it seems that ≈ 5.35µf is close to optimal, but it might go a bit higher, say 5.6µf or so.
This is the fun stuff as new refinements are dialed in and new acoustic and sonic nuances are heard.

JJ
 
Dec 22, 2018 at 7:11 AM Post #1,360 of 1,959

johnjen

Headphoneus Supremus
Joined
Apr 13, 2010
Posts
3,418
Likes
2,928
Location
Well Grounded-here and now
So right now I'm preparing to post Part 2 of my power cable article.

And tomorrow I will probably start the tear down and rebuild of the analog section to install the interstage xfmrs and isolate all of the xfmrs from the top plate and then tie them all together into a 'star' ground.
I have all the parts on hand and have figured a means and method to yank out the stuff no longer needed and hopefully be able to install the new xfmrs in the same or perhaps better locations.

Fortunately these new interstage xfmrs are smaller than the rest of the Lundahl xfmrs I have used thus far.
Which should help in shoehorning them into place.

JJ

And next up is…
Why DO power cables and such make ANY difference to our headphone gear?
Part 2 Measured Results and Observations.
 
Dec 22, 2018 at 7:33 AM Post #1,361 of 1,959

johnjen

Headphoneus Supremus
Joined
Apr 13, 2010
Posts
3,418
Likes
2,928
Location
Well Grounded-here and now
Why DO power cables and such make ANY difference to our headphone gear?

Yet another experiment focusing on power delivery and some not generally understood factors to consider.

Part A Why Would I Want To Do This In The First Place?

https://www.head-fi.org/threads/the-diyrs-cookbook.781268/page-85#post-14294952

Part 1 Theory and Expectations Based Upon My Initial Research
https://www.head-fi.org/threads/the-diyrs-cookbook.781268/page-89#post-14593223

Part 2 Measured Results and Observations.
Part 3 Analysis and Conclusions.


Part 2 Measured Results and Observations

But first a Re-Cap.
Bits are bits, wire is wire and as long as the music plays there should be enough voltage and current, so what’s the problem, right?
And especially with our HP gear that uses so little power to begin with, as such, ac power cables should make no difference right?
I mean a 9w/ch amp driving high Ω HP’s there should be no effect when trying different power cables.
Except that that isn’t what I heard…

This was a challenge to try and figure out.

And now,
Part 2 Measured Results and Observations
I’ll include a few oscilloscope pictures of the voltage and current wave forms super imposed upon each other to show when the current starts and stops flowing in relation to the voltage swing.
What I’ve measured and now present is an approximation of the time and amounts of voltage and current that is shown by the scope.

In addition I’ll provide some insights into the dynamics of this relatively static picture.
I call it a static picture because these measurements were taken while the amp is idling and no signal (music nor fixed test tones) are being amplified, and these wave forms don’t change.

Now these pics of the measurements were taken using my Tek 468 O’scope and an Aim iprober 520 contact current probe by my ipad.
I volunteered my Schiit Mojo amp (1st gen) as test bed for these experiments.
It is rated at a maximum of 45watts of power consumption and delivers 8watts into 32Ω, (or 32watts into 8Ω which it isn’t rated to do, but is a relative gauge).
And operates with 2 watts in class A operation as it’s normal bias for each channel, at idle.

These measurements were taken with no input (& the volume control is all the way down) nor is there any load on the output.
The location where the measurements were made was at the ‘output’ side of the fuse holder, after the switch and before any of the rest of the power supply.

The Aim iprober is a direct contact current probe, meaning it reads the magnetic field at the circuit board trace, and then uses the O’scope to display the wave form shape.
As such it does not add any additional load nor does it limit or affect the current pulses.

pic of setup with arm and probe
AIM iprober in place.jpg

Notice the voltage probe ‘seeing’ the voltage that the current probe is measuring at the same location and at the same time.

And since it is a contact probe and isn’t a direct measurement of the current flow, placement becomes a variable in taking measurements. This is why I am using approximate numbers and amounts.
Plus exact numbers aren’t really necessary at this point in these investigations.
I’m looking for ‘Big Picture’ views and not focusing in, nor drilling down onto, the specific details themselves.

It turns out that while operating at idle, in those ≈ 8+ms windows of time, current passes in ≈ 3+ms slices of time, followed by ≈ 5+ms of being off, followed again by a ≈ 3+ms ‘ON’ window of time.
This means that the current actually flows for only ≈ 37% of the available time, into the downstream power supply.

pic power pulse V&I
Power Pulse V & I w:time.jpg

And in that ≈ 3+ms of time, the amount of current that I measured is ≈ 0.4amps(PtP) in a peak time window of ≈ 1+ms.
This means that in one full cycle of 60Hz the peak current delivered is ≈ 0.8 amps(PtP) during two ≈ 1+ms peak windows, and all current flows in 2x ≈ 3+ms windows for each full cycle.

pic power pulses I PtP
Power Pulse I PtP.jpg


Now, a peak 45watt load needs ≈ 0.375amps on a 120v supply voltage, in a 100% duty cycle environment with a symmetric periodic wave form.
Huh? (I can hear you respond…)

100% duty cycle means the current flows continually (like a battery) with no stops nor starts. So a 50% duty cycle means the current flows for 50% of the time and doesn’t for the other 50%. And in this case has a ‘rep rate’ of 60/sec or 60Hz.
And the symmetric periodic wave form (sine wave) means there is symmetry to the flow of voltage and current.

But the Mojo amp, at idle, is operating in a ≈ 37% duty cycle which means that if our peak 0.8amps(PtP) were scaled up to 100% duty cycle. the current would be equivalent to ≈ 2.1 amps or a 267% difference, to the added time that current flows.

As was mentioned this is an expected result for the need for current (power) but with a limited amount of time to deliver that power.
Which in turn means the power comes into the downstream storage / filtering / regulation portion of the power supply in a series of ≈ 3+ms pulses followed by ≈ 5+ms of ‘off’ time.

So in effect in this light duty situation while driving a ≈18watt amplifier, the amp is ‘asking for’ peak current flow of 0.8amps(PtP) yet a 18watt load should only need 0.17 amps if it were a 100% duty cycle.
This represents a 470% increase in the amount of current from a ≈37% duty cycle vs a 100% duty cycle, and assuming the current flows in a symmetric waveform (which it doesn’t because it’s asymmetric, see below).

IOW these pulses are best thought of as short, sharp ‘demands’ for instantaneous current flow. Followed by longer duration ‘off’ times.
And not, steady, even amounts, of uniform, symmetric periodic current flow, as many might think.

end Part 2

Next up Part 3 Analysis and Conclusions.


JJ
 
Dec 24, 2018 at 1:42 AM Post #1,362 of 1,959

johnjen

Headphoneus Supremus
Joined
Apr 13, 2010
Posts
3,418
Likes
2,928
Location
Well Grounded-here and now
Ok so I lied.
I still haven't started the surgery/transplanting of new xfmrs and bias circuits yet.
For a couple of reasons, one being another headphone guy might get a chance to hear this amp tomorrow and if it's all torn apart, that won't happen.
And the 2nd reason is this pipe organ at Methuen has convinced me to continue to listen and not shut the amp down.
It is a compelling reason, not to mention that any other music I play seems to have the same effect.

So surgery was called on account of 'any' excuse to just listen and not start throwing hot solder around.

It's torment I tell's ya. :ksc75smile:

JJ
 
Dec 30, 2018 at 12:26 AM Post #1,363 of 1,959

johnjen

Headphoneus Supremus
Joined
Apr 13, 2010
Posts
3,418
Likes
2,928
Location
Well Grounded-here and now
The surgery went well and no smoke has escaped, nor were there any sparks, loud pops, snaps, or fires. :ksc75smile:

I'm still fussing with the additional power supply caps and isolating the xfmrs from the top plate to see how that affects the hum and noise issues.
And the new interstage xfmrs need more playing time as the SQ is morphing as they 'settle in'.

Next up is to modify the bias circuit for the 45 tubes to use batteries instead of a resistor and bypass cap.
Somewhat like the bias circuit for the 6J5's where we now use a 9volt battery instead of either resistors/caps or LED's.
This means the only caps remaining in the audio path will be the output caps between the 45 tubes and the output xfmrs.
And that all of the resistors in the audio path have been replaced with xfmrs, with the exception of the trim resistors in the filament power supply(s).

And once all of these changes settle down, I'll get back to determining the value of the output caps, among other experiments.

The SQ has risen yet again and it is still rather early, as this new circuit configuration gets dialed in, so there is a degree of anticipation what these new changes will lead to.

Yep I still having some fun now, I tells ya!


JJ
 
Jan 2, 2019 at 5:25 AM Post #1,364 of 1,959

johnjen

Headphoneus Supremus
Joined
Apr 13, 2010
Posts
3,418
Likes
2,928
Location
Well Grounded-here and now
So I was going to try and draw up the full schemo for this amp, but I realized it would take a WHOLE bunch of time and wind being kind of funky and cobby.
BUT…
I did have 2 hand drawn portions of the amp, the power supply and the latest analog version.

So I cut and pasted them into one document.
45 amp schemo.jpg
Talk about a REALLY simple analog circuit.
And yes there are a few portions of the circuit that are 'missing', like the filament circuits, the ac power inputs, and the details dealing with the output xfmrs.

The battery bias for the 45 tube is the latest addition to the amp, and was facilitated by the interstage xfmr that went in recently as well.
And the battery bias for the 6J5 tube was facilitated by the input xfmr.
And after a few screw ups in installing these new modifications and learning what NOT to do, it is working well enough to give it some playing time to monitor the voltages and listen to hear the SQ changes.

Some may wonder where are all the resistors and caps that are normally found in analog circuits…
Well, it's like the designer has oft repeated, "there are no problems that more xfmrs can't solve", like eliminating the resistors and caps with only the output cap remaining.

And the SQ continues to improve, with startling results upon occasion.

JJ
 
Jan 3, 2019 at 6:12 AM Post #1,365 of 1,959

johnjen

Headphoneus Supremus
Joined
Apr 13, 2010
Posts
3,418
Likes
2,928
Location
Well Grounded-here and now
And just like that, the power supply has been modified even further…

45 amp PS mods.jpg

And so it goes.

JJ
 

Users who are viewing this thread

Top