The DIY'rs Cookbook
Dec 16, 2019 at 12:11 AM Post #1,486 of 1,974
This is akin the 'Mr Science Show' where he has 'Jimmy' (not his real name) hold 'those' 2 wires, followed closely there after by a staccato vibratory yelp for help.

I just thought I'd throw in a complimentary example of methods to discharge a power supply cap, with similar 'learn by doing' results…
hahahahahahahahahahahaa

JJ
Was Jimmy related to Mr. Bill from SNL?
 
Dec 16, 2019 at 1:26 AM Post #1,487 of 1,974
That's a good question, one that should be asked of either of the unwitting 'test' subjects in order to find out…
hahahahahahahahahahaha

JJ
 
Dec 16, 2019 at 5:09 AM Post #1,488 of 1,974
Tube Amp Designs, 2 Types…

Recently it occurred to me that there are 2 distinctly different approaches to tube gear design.
And while neither is strictly observed, there is a fundamental orientation to the design process itself that is significant.

And really this isn’t limited to just tube based audio gear, but because very few classic SS designs, using all vintage parts, (especially the silicon based parts) are high on the list of ‘projects I want to do’ these days, the tube based gear stands out as a more substantial example.

These 2 approaches begin at the very start of the design process itself.
To wit, does the design intent begin with using parts that are at the ready and already on hand, or are desired to be used, vs. a ‘clean sheet of paper’ approach, where the design is based upon achieving functional and operational goals to solve known circuit ‘deficiencies’.

So in a nut shell, approach…
#1) select the tube or tube(s) you want to use and design a circuit around them.
&
#2) design a circuit and then match tubes to properly run that circuit.
And the differences between these 2 approaches would seem to be relatively minor, except that the resultant SQ, will be different.

And to be sure, neither #1 nor #2 will be strictly followed, for instance some tube combinations just aren’t going to work well together, and in some cases won’t work at all…

And manufacturers have additional needs such as tube availability which in many cases precludes option #2 because they need to have a sufficient supply of tubes for their designs…

So approach #1 would be in the majority of the available tube gear, mostly because manufacturers can produce lots of examples of all kinds of tube circuits.
Also #1 is more ‘marketable’ in that the tubes chosen are already well known and are plentiful, so familiarity and acceptance is less of a stumbling block to purchasing ‘new’ tube gear.

Still the advantages of approach #2 is that the design comes first with the tubes being matched to perform as the design dictates.
Instead of the, dialing in the circuit to match up the tube(s) to work with each other approach, design approach #2 aims at optimizing the choice of tubes to function at (or near) their optimal operating points, so the entire circuit (not just the tubes) get dialed in, in order to function ‘properly’ in the 1st place.

Am I splitting hairs here?
Perhaps in some eyes, yes, but the differing results between these 2 design approaches can be quite significant in terms of SQ.
And for some, the very nature of the SQ is the primary and final determination of the desirability of an amp.
And in some cases so much so that some will go to extremes (read $$$$$$$$) in order to achieve SQ that meets their requirements.

And to be sure approach #1 is what most manufacturers employ, and realistically they are limited to those tubes that are either currently being manufactured, or are plentiful enough for production, with a few exceptions of course.

Which means it’s the DIY’rs and home builds where #2 can be found, although certainly not exclusively.
Many DIY’rs have specific tubes they want to use and so build around it (my BIG 45 amp is an example), and so the number of tube based circuits that fall into the #2 category is a rather small percentage of the overall number of available choices that are and have been made in the 1st place.

And it should be noted that approach #2 does require access to specific and customized designs by a ‘real’ designer and in an ongoing basis as the sought after final end results are pursued and do require.

So why bring all of this up in the first place?
Well, this distinction between these 2 approaches is generally unknown, except to designers and those who know not just how to design a tube circuit, but also those who would want to explore diverse kinds of design choices in the first place.

And to add yet another layer to all of this, most designers efforts are aimed at the production of these designs.
These designs are after all meant to be used and enjoyed and so ‘completing’ the design by having a working and fully functional ‘prototype’, ready to be replicated for sale to others, is often the point of spending the time to ‘perfect’ the transformation of a design idea into a fully functional working audio component.

In addition there are some designs and their implementations that have reached near cult status and lots of them are nearly 100 years old. They can sell for 100’sK$ or more. These designs use an ever diminishing supply of parts and not just tubes.
These designs were the result of research into all phases of sound reproduction and for all manner of commercial, industrial, and institutional uses. From recording studios, to sport stadiums, theaters to school PA systems and much much more.

Western Electric, to this day stands far above and far ahead of the vast majority of all current designs, and to the extent that there companies trying to replicate the original parts and to fully replicate this highly sought after gear.

I raise this because there are those who still, to this very day, continue to attempt to replicate the SQ that was achieved back ‘in the day’, only using todays available parts and components.
And because WE followed path #2 to such a marked degree, to the extent that they created, from scratch, component parts to match their designs, tubes, resistors, inductors, wire, and more.
And today those very same parts can bring eye-watering prices.

And as some know, pursuing SQ that beckons from just ever so slightly beyond our grasp, can be a life long obsession, hobby and avocation, and obviously one worthy of such life long ongoing pursuits.
Even if it gets excessive…

JJ
 
Dec 25, 2019 at 3:10 AM Post #1,489 of 1,974
So I just got my 2nd batch of 6BQ7A's in and am now running a Cossor variant of a Brimar made tube.
Apparently these Cossors are hand selected/pick of the batch, Brimars.

Thus far the SQ is roughly the same as the Brimars (as you'd expect) but as they accumulate hrs they do seem to be moving in an upwardly mobile direction in terms of SQ.
And the 1st Brimar tube had stabilized it's SQ at ≈ 150hrs +, which had just reset the bar to new heights in overall SQ…

And I have a set of NOS 1960 Mullards made in GB incoming as well, to use as an additional set of tubes for comparison in a 3 way battle for my stairway to heaven winner.
hahahahahahaha

I also have a set of tube dampeners that are being tested as well.
The first set was unimpressive to say the least.
So I have a change of pads inbound to see if they will improve the net results.

I am rather curious to see if in a HP setup, where the acoustic space itself is not being modulated and thus no acoustic energy is impacting the tubes, whether or not they will make any difference.

And now that the SQ of the amp itself has reached 'elevated levels' with increased inner details and nuances adding new levels to the acoustic presentation, it should be all the more telling of any changes that result from using tube dampeners, or not.

JJ
 
Jan 2, 2020 at 12:30 AM Post #1,490 of 1,974
So I fussed with the 71A's bias resistors and changed them to Allen Bradley carbon comp's from the 60's.

Yet another 'mutha may I' step up in SQ.
Smoother and more detailed everywhere along with an increase in REALNESS as the acoustic space and all instruments attained even tighter C3 (Cohesion, Coherence, Coupling).
This somewhat surprised me that a mere resistor change could have such a beneficial change.
And granted these resistors are in a critical portion of the audio path, still it was most insightful.
And I have an even simpler (use just 1 resistor instead of the 3 I'm now using) experiment just waiting for parts to show up.

And my 3 way driver tube shoot out has now expanded to 6 tubes, not counting the development sets of RCA/GE.
Hytron (CBS)
Brimar (NOS BVA)
Cossor (NOS BVA)
Mullard (NOS BVA 1960's)
Phillips (NOS Miniwatt Holland)
Bentley (NOS British?)
(BVA = British Valve Association)

The Mullards, Phillips, and Bentley's have yet to show up, so it will be interesting to see if the current king of the heap (Cossor) is bested…

And this expanded list of tubes is thanks to bcowen for continually feeding my addiction compulsion to experience the ultimate in audio perfection, as in 'damn the torpedoes full speed ahead'. hahahahahaha

JJ
 
Last edited:
Jan 3, 2020 at 11:43 PM Post #1,491 of 1,974
So yesterday we tried an experiment to see if changing the bias circuitry from led's to a battery bias scheme would improve the SQ of the Purp-Amp, in preparation for our upcoming local Head-Fi meet later in January.
And after implementing this change, the B- being fed to the 6BQ7A driver tube dropped far to low for 'optimal' operation, so this morning I reverted back to the LED bias circuitry and then measured all the 'critical' voltages to see that they had returned to their previous amounts.

I relate all of this as a means of letting those who think that we only make forward progress with our mods and tweaks, to know there is always a degree of uncertainty in any design and the 'proof' is in the execution and subsequent evaluation (measurements and listening tests) to know of the value and veracity of these experiments.

Sort of a variation of 'learn by doing' and take nothing at face value, while refining and tweaking these designs to reach their highest SQ, as well as optimize the functional aspects of the basic design itself.

This approach to dialing in the circuit yields 2 types of useful info.
#1 What does work and how and in what ways there is an improvement.
#2 What doesn't work and why, which in most cases is more valuable and significant, because by reducing all of the possible options and methods to just those that yield 'useful' results makes the entire design process more 'streamlined'.

And to be sure, the refinement and dialing in of the circuit is a process, not simply a 'build it and your done' affair, by any means.
That is if the goal is to tweak and optimize the SQ as much as possible.
Of course "as much as possible" has certain inherent limitations, such as $$$$+ and even certain practical considerations such as size and lethality, which can all to easily become limiting factors.

I mean when 2 or 3 THOUSAND volts are needed, certain additional safety precautions are necessary (no pets allowed) etc. hahahahahahaha

But yeah I'm having some fun now, again, still…

JJ
 
Jan 3, 2020 at 11:59 PM Post #1,492 of 1,974
And as a compliment to 'the experiment gone bad'… hahahahahaha

I swapped out the '3 resistors makes the value I wanted', to a single Allen Bradley carbon composition from the 1980's.

This experiment is still settling in so the full impact has yet to be determined, but thus far this simplification has further refined the previous change to using carbon comps, instead of a wirewound resistor in the bias circuit for the output tubes (71A).

The entire soundstage/acoustic space is yet more refined, smoother, more precise and articulate in its 4D presentation.
And the bass as in ALL of the bass has taken a 'mutha may I step up' yet again.

Now these changes aren't as dramatic nor as slap you upside yer head obvious, as the previous resistor change was, but are as welcomed and reveal yet further musical relationships and information, and in delightful ways,
already.

A decidedly good sign.

JJ
 
Jan 5, 2020 at 6:19 PM Post #1,493 of 1,974
And the Hits just keep coming…

So tonight we went over several 'layers' of mods in prep for the upcoming head-fi meet later this month, and beyond.
These mods are a series of refinements and alternative circuits which will provide a 'better' electrical environment for the tubes to operate in, in many ways.
And they make the audio circuit itself simpler, but with a bit of added complexity in the PSU.
And tonight we took a step towards the next stage (of 3) mods and in the process we fixed/eliminated the last come (and mostly stay) and go, niggling (as it turns out) power supply issue.

We are back to -120dB s/n (0.000mvac) noise (or lack there of) level, on BOTH channels.
For the longest time I had -120dB on the left channel but only ≈-66dB (0.4xxmvac) on the right, even though at times it would measure as 0.000mvac, it didn't stick around for long.

And in the process we considerably simplified the PSU, and took 2) 100µfd caps out of the circuit, which will be relocated to the analog chassis when stage 2 is implemented.
My initial impression was, we hit yet another home run, and that was after only a few minutes of being on.
I need to give it more time to settle in (as per usual) and this is but the beginning of the break in for this mod.
And there are those new bias resistors that are still settling in as well.

And this whole series of mods, which should improve all aspects of performance, (power output, distortion, dynamics, FR bandwidth, circuit stability, etc.), will get implemented in layers as each mod is then used in the next mod, or is removed entirely.
And wind up with a direct coupled amp with only 2 caps in the audio path, with one in the bias circuit for the output tubes, and the other is the final cap in the power supply feeding the output tubes.
And wind up with an audio circuit with fewer parts (down from 6 to 5 components) in the direct audio path itself.

And tonight is just the 1st step of 3…
Yep, even more fun I tells ya…

JJ
 
Jan 7, 2020 at 6:40 AM Post #1,494 of 1,974
Why yes the fun does indeed continue, even further, still, yet again,

We figured we should 'clean up' the ground circuits and isolate the PSU ground from the analog ground, in terms of any type of 'cross linking'.
The 1st attempt was one of those one step forward and 20,000 steps backwards.

Then we down shifted and punched it…
So To Speak.

What happened was those 2 caps I removed from the PSU (mentioned above) were installed in the analog chassis and wired in.
The immediate result was a complete loss of all the s/n gains we had made, to excess.
I didn't even bother to try and listen to the amp, it was that bad.

Then I substituted 2 other caps (temco's, our favorite), and the s/n and voltage stabilized a whole bunch, not quite as good as before this 'clean up' began, but it still needs to settle in before a final decision is made on what the next steps should be.

And as I listen to the amp now, it is settling in nicely and there are many aspects that are surprising me, yet another good sign…

JJ
 
Jan 7, 2020 at 7:06 AM Post #1,495 of 1,974
Yes Indeed we are having yet more fun and excitement, while at the same time further refining the circuit and dialing in the specific wiring configuration that takes full advantage of the simplicity of this circuit.

This time we wanted to dial in the ground circuits and eliminate any 'cross linking' of the psu vs analog stage grounds.

The 1st attempt failed, as in failed so bad I didn't even put on the HP's to even try and listen.
The s/n went sideways AND backwards all at the same time, as did the B+ stability and it's noise level.
Were figur'n it's a bad cap or two since alternative plan #2 (using our old standby temco caps) works rather well.

It still needs more settling in time to be sure, but this is showing all the signs of being on the right path.

More later as this tweak gets fully dialed in.

JJ
 
Jan 9, 2020 at 11:33 PM Post #1,496 of 1,974
So I reassembled the PSU using the temco caps and wired it so I could use 2 different grounds for the last 2 caps in the voltage feed to the output tubes.
One was the PSU ground and the 2nd was the analog ground.
I measured s/n and voltage stability and the amount of mvac on the dc feeds, using both types of ground.
There wasn't much difference when I first turned the amp on and gave it a chance to run for 10-15 minutes and then switched the grounding scheme.

So I left it in the 'proper' configuration (to the analog ground) to further settle in.
And after another day or 2 I'll measure the same variables again just to see how much things settle down.

JJ
 
Jan 14, 2020 at 6:03 AM Post #1,497 of 1,974
So a few experiments later…
We're back to the -120/-68dB on the s/n numbers.
This here, and now not here, lack of noise has been vexing us for a while.
And it's back to the original grounding scheme and the SQ is back where I remember it.

But wait there's more…
And when will I learn to leave well enough alone?
Uh, probably never… hahahahahahahaha

We're gunna go all in (well at least for mod level #2) and install the 2nd pair of CCS's (Constant Current Sources) to run the 71A output tubes, and in the process pull out chokes, and caps, and resistors, and then let'r rip.
The plan is to find out if this puts all those low frequency gremlins to sleep, once and for all, AND see what happens to the s/n numbers.

Stay tuned for the next exciting episode of Toobz-'R'-Us.

And the for the upcoming seattle head-fi meet I have a special surprise.
A set of HP's that few know about and even fewer (including me) have ever heard.

They should be arriving later this week and they should match up just about ideally to the Purp-Amp (assuming it has enough poop to sufficiently drive them) and those at the seattle meet should get a rare opportunity to hear a nearly finished but still in development set of SotA HP's.
And they use a driver unlike any other, so this will be Very Interesting.

And comparing them to my Gen-6 800J-mods should prove to be a rather interesting task in and of itself.

JJ
 
Last edited:
Jan 15, 2020 at 8:07 PM Post #1,499 of 1,974
The 'package' has arrived, and after the formal 'unboxing' ceremony had been completed replete with offerings to the HP gods and a careful examination of the tid bits of note, (cable, orientation on the head, adjustability, etc.) and using a 4-pin xlr to TRS adapter, they are on my noggin and playing music.

Initial impressions are that they are quite nice and do reach down deep with extension and finesse up top, and the mids are particularly nice.
The factory mentions that no 'burn-in' is required and thus far (≈2hrs of playing time) I haven't heard their SQ signature morph, although my brain may be 'getting used to' this new signature.

Physically they are big and a bit 'clunky', in comparison to the 800-Jmods weight and form factor, but they are not to heavy and don't want to 'fall off my head' if I make sudden moves etc.

I am making a list of 'changes/improvements' I'd like to see and will pass them along to the factory as a 'report from the field'…

More later as I continue to listen.

JJ
 
Jan 15, 2020 at 9:29 PM Post #1,500 of 1,974
The 'package' has arrived, and after the formal 'unboxing' ceremony had been completed replete with offerings to the HP gods and a careful examination of the tid bits of note, (cable, orientation on the head, adjustability, etc.) and using a 4-pin xlr to TRS adapter, they are on my noggin and playing music.

Initial impressions are that they are quite nice and do reach down deep with extension and finesse up top, and the mids are particularly nice.
The factory mentions that no 'burn-in' is required and thus far (≈2hrs of playing time) I haven't heard their SQ signature morph, although my brain may be 'getting used to' this new signature.

Physically they are big and a bit 'clunky', in comparison to the 800-Jmods weight and form factor, but they are not to heavy and don't want to 'fall off my head' if I make sudden moves etc.

I am making a list of 'changes/improvements' I'd like to see and will pass them along to the factory as a 'report from the field'…

More later as I continue to listen.

JJ

So is there a prize for the correct guess? Eh, no matter...I'm goin' with these. Just hoping you got them in red instead of yellow. :relaxed:


GuyGiantHeadphones.jpg
 

Users who are viewing this thread

Back
Top