Considerations for cascaded channel design and review of 3 well-known circuits

[CascoSieg]

I'm considering building a 5F6A-ish, 2-channel amp with mods, mainly from Rob Robinette's site. I have sooo much to learn, but I'm revved up to do so. Because I'd like the tonal options including getting something close to 5F6-A vintage stock tone, as well as others including higher gain, I'd like to implement a cascaded channel mod. Rather than just copy Rob's or some other kit's design, I'd like to understand some of the fundamental elements at play for any possible design, especially related to gain stages that can work both an an initial input stage as well second gain stage. So, I've been looking into variations that have been done, and there are things I don't quite understand about some of them, and the likley reason for the choices made. The Amps I've looked at that implement a double-duty gain stage are the Park-1210, the JCM-800, and the BE-100. The attached doc is my own narrative of stepping through each pre-amp circuit, which I welcome feedback on if you're interested in looking at it, but I'll cut to my main quesitons for those who are already intimate with these circuits:
- Assuming one has already addressed all the questions about why to use a cascading circuit in the first place, and what else one might want the circuit to do, it seems like perhaps there are just two main questions that need to be answered to inform a cascading channel design strategy: how big a signal do you want to present to the second gain stage, and how do you want to shape that signal tonally before it gets there?  Of course there are lots of ways to address those questions, and knowing what's going to happen next in the circuit also matters, but, is this basically the crux of it, or are there other big considerations?
- With my basic understanding of HP and LP filters, I'm having trouble sorting out the Resistor/Capacitor networks filtering the signal prior to the second (or first, when using the lo gain option) gain stages of the Park and JCM (see circuit snips in attached doc figure ). What's a good resource for advancing my understanding of these particular filter networks?
- Two of the circuits dramatically reduce the output voltage of their initial gain-stages with a voltage divider prior to the second stage (the Park-1210 and BE-100) and one does not (the JCM-800). Does that simply mean that the JCM is always going to overdrive the second stage, except at the lowest of output levels, and the others will have more clean headroom in the second stages?
- I see a 100pF cap coming off the plate of the first JCM gain stage, but without a resistor to ground it doesn't look like a LP filter. Aside from coupling caps, what do series-wired caps do to an AC signal without a paired resistor like you'd see in a LP/HP filter?
- Both stages of the the BE-100 have a 220k plate resister in series with a 500pF bypassed 100k resistor. What value would I use when trying to calculate the bias and other qualities of the working tube? Since caps dont pass DC, I'd assume 330k in this case... but wow, that would mean a very warm bias, and with a 2.7k cathode R, really shifts the load lines left, which is unusual.

I've learned a lot already on this forum - Thanks for sharing knowledge and assistance.
~Sieg

[pdf64]

how big a signal do you want to present to the second gain stage, and how do you want to shape that signal tonally before it gets there?

With RC coupled common cathode (CC) stages, the main issue caused by overdriving them is bias shift, leading to blocking distortion.
Here is a great resource written by a skilled, competent electronic engineer / guitar amp designer / manufacturer http://www.aikenamps.com/index.php/white-papers
specifically http://www.aikenamps.com/index.php/what-is-blocking-distortion and http://www.aikenamps.com/index.php/grid-resistors-why-are-they-used

- With my basic understanding of HP and LP filters, I'm having trouble sorting out the Resistor/Capacitor networks filtering the signal prior to the second (or first, when using the lo gain option) gain stages of the Park and JCM (see circuit snips in attached doc figure ). What's a good resource for advancing my understanding of these particular filter networks?

See https://www.ampbooks.com/mobile/classic-circuits/soldano-slo/

- Two of the circuits dramatically reduce the output voltage of their initial gain-stages with a voltage divider prior to the second stage (the Park-1210 and BE-100) and one does not (the JCM-800). Does that simply mean that the JCM is always going to overdrive the second stage, except at the lowest of output levels, and the others will have more clean headroom in the second stages?

Note that the 2nd stage of the JCM 800 2204 is the cold clipper. The large unbypassed cathode resistor is in series with the g-k diode, which makes it very  resistant to blocking distortion.
I suggest not to use the term 'headroom', as within this field folk use it ambiguously, describing various, often conflicting, characteristics; rather use clearly understood and defined terminology.

- I see a 100pF cap coming off the plate of the first JCM gain stage, but without a resistor to ground it doesn't look like a LP filter. Aside from coupling caps, what do series-wired caps do to an AC signal without a paired resistor like you'd see in a LP/HP filter?

It is a LPF; the R term is formed by the output impedance of the CC stage.
See http://www.aikenamps.com/index.php/designing-common-cathode-triode-amplifiers
I suggest that you research Thevenin and Norton equivilent circuits, eg http://hyperphysics.phy-astr.gsu.edu/hbase/electric/thevenin.html
If you really want to understand the nuts and bolts of it all, there's no substitute for a good background education in physics, leading to a more specialised course in electrical and electronic theory and practice.

- Both stages of the the BE-100 have a 220k plate resister in series with a 500pF bypassed 100k resistor. What value would I use when trying to calculate the bias and other qualities of the working tube?

The dc and ac conditions of a stage are typically different. The dc operating point is set by the 320k total plate resistance. The ac conditions are A LOT more complex 

[shooter]

Thevenin and Norton equivilent circuits

It's as fun as reading about Alice when she went down the rabbit hole 

[CascoSieg]

Pdf64 - Thanks ever so much for your concise reply and direction to some useful and definitive resources.

I suggest not to use the term 'headroom', as within this field folk use it ambiguously, describing various, often conflicting, characteristics; rather use clearly understood and defined terminology.

Thanks - I admit to seeing it used to mean a variety of things, so appreciate your suggestion.

It is a LPF; the R term is formed by the output impedance of the CC stage.

Ahhh, yes, I have read about that, but it didn't sink in apparently.
Lots to read and think about! 
~Sieg

[CascoSieg]

OK - I've read, reviewed and worked through examples (curriculum, really) in the materials referenced: the Aiken resources are good and I've looked at many including the specific ones mentioned. Thevenin/Norton equivalent circuits is certainly necessary for unraveling component sequences (tho I'm still tripped up in some situations, mainly to do with managing low vs high frequency impacts on voltage, and which one is most relevant in a specific case) and the Richard Kuehnel talk-through of the SLO circuit shows how one can think through EXACTLY the stuff I'm wrestling with - really great.

As I look through his 5F6-A analysis (calculating the range of voltage present on the second stage grid) I'm confused by why the two mixing resistors do not seem to act as voltage divider? Taking a voltage at the wiper of either channel volume control, we see:

V1 (stage 1 output)  --->  R1 (270k mixer) --->  V2 (voltage at stage 2 grid) ---> R2 (other 270k mixer + some R component of the volume pot) ---> Ground

Doesn't that divide V1 by as much as 2 when measured at v2 grid?

[sluckey]

Doesn't that divide V1 by as much as 2 when measured at v2 grid?

That's correct.

[CascoSieg]

Thanks - a relief. :-)

[pdf64]

Also bear in mind that if the channel inputs are linked, and their volume controls suitably adjusted, the same signal voltage appears at the outputs of both volume controls. In that scenario, there's no signal loss across the mixing resistors.

[CascoSieg]

Ok, at the risk of making a total fool of myself and producing what is likely a square-ish re-invention of so many perfectly good wheels already out there, here's my first pass at solving a number of requirements/desires for my amp design - addressing just the input stage.
My requirements/desires most pertinent to the input stage are:
- A 5F6-A -like channel
- A JTM-like second channel
- Switchable cascading additional gain stage
- Switchable Jumpering/combining of channels

The attached doc shows my input circuit and a tracing of the signal path for each possible configuration of channel and mode (borrowing liberally from Rob Robinette, Doug Hoffman, Weber Kits, and many original schematics).
Replicating Richard Kuehnel's process of characterizing signal and gain changes referenced earlier in this thread, and after digesting all that I could about blocking distortion (which is not enough), I believe I have configured my cascading mode in a way that provides reasonable attenuation options and grid stoppering at both V1B (as a second stage) and V2, while preserving the basics of standard 5F6/JTM designs as well.
I don't like that I need a lot of switches to pull it off, but that seems part and parcel with versatility, and I don't like leaving the unused grid un-grounded in one scenario even if that scenario is unlikely to ever occur, but I just don't have enough switch "bits" to cover it.
Questions:
- Are there examples of well-implemented pre-amp cathode resistor adjustment switches out there, that are much different than mine?
- Is there some straightforward solution to the un-grounded Ch 1 grid in the "Cascade Mode, Ch 2 in use" scenario?
- Any other pitfalls I'm not seeing?

With much gratitude!
~S

[shooter]

Any other pitfalls I'm not seeing

 

I breadboarded a  3 channel ONCE, never made it to production 

[SILVERGUN]

Holy superfluous switching matrix PDF. I tried Bro.
 

[CascoSieg]

... never made it to production.

Was it because it's just not possible to make it sound good as both an input and second stage? Or... ?

[pdf64]

Well done for working all that out, it looks like it might work ok.
The issues being whether switching between those options will useful, and that every option is a somewhat compromised version of what it’s intended to be.
The way these things usually go is that after the initial trial period, one option is preferred, perhaps it gets fine tuned a little, then it stays on that setting forever
But that may not stop it being a useful exercise for you.

FYI grid stoppers don’t reduce the audio signal level, so no noticeable difference per se between 68k and 33k.

And cascaded stages with bright caps on their volume controls may get far too bright if both at lowish settings, as the treble boost gets ‘squared’ / compounded.

[CascoSieg]

Ah yes, that makes a lot of sense. 
It's true I don't yet have the experience to know what's likely to be useful, and what are the tonal/dynamics costs of compromise, but that will come with time. And I think you're right-on, precisely because I AM learning, that  it will be useful to pursue a variety of options. If I get stumped or too frustrated I can always simplify things.
And the info on input grid stoppers as well as bright caps w/cascaded stages is very helpful.
Thanks very much!

[shooter]

I find tinkering a great way to learn.... remembering, documenting, understanding,
those take some discipline 

[ThePanda]

Necro-posting!
@CascoSieg How did this pan out?
Looking to implement this on my 5F6a build to have the option to Jump and Cascade but go back to normal.

Your DRAFT pdf is very helpful.

[CascoSieg]

As far as that selector switch design goes and the overall concept, it all works as intended. I really like the flexibility of the whole input section on this amp. I've gotten better overdrive sounds on amps I've built since, so I don't use the cascade mode much anymore.
Good luck!

[ThePanda]

That is good to know.

I have started adding it to my schematic. Are there any other special considerations that you learned by adding the Cascading preamps that I should look out for?

What was it about the overdrive that wasn't as good?

[CascoSieg]

As other wiser folks here mentioned to me, configuring a triode to perform two distinctly different roles well, can be tricky. I'd recommend leaving yourself room in the chassis to do a lot of component swapping, and optimize V1a as the driver of the cascaded mode, rather than as a second pure Marshally clean input... meaning, tweak Rl, Rk, and Ck for greatest gain in the frequency range you want, to drive V1b. For me it still does a really good job boosting mids in parallel mode, but I don't tend to use that channel by itself.
The overdriven sound in cascaded mode is a bit harsh or ragged to my ear I guess, compared to the Dumble-ish OD (derived mainly from Tubenit's COS and related designs) I've put into other builds. Just my preference. I think if you like a dirty tweed sound, you can get something you like out of this design.
I'm still pretty green with this stuff, so others may have more helpful insights.
~S

[ThePanda]

I looked into those others and there is a lot to take in.

My idea is to keep the 5F6a channels as is for the most part and offer parallel/normal operation/cascading as options. The bright and normal channels are toggled by a switch, and eventually there is going to be a bright switch on the bright channel to exclude the bright cap.

The goal is to keep it clean while keeping the original voice as much as possible. I have been taking
ideas from the Dumble builds to keep it clean.

Here is what I came up with based on my reading and using your switching diagram.

[johnnyreece]

It's not my circuit, but I believe you're going to want to split the cathodes and give them each their own resistor/bypass cap.

[ThePanda]

It's not my circuit, but I believe you're going to want to split the cathodes and give them each their own resistor/bypass cap.

This is so I can voice the channels independently?

[CascoSieg]

Yes, to separating the V1 cathodes, for the reason you specified.
I had basically the same motivation for my 5F6A/JTM build: keep something like stock voicing, and add some additional tones.
Is this a post-build mod, or are you in the planning phase, or somewhere in between... ? The reason I ask is that if you had room in your chassis, you could add a switchable OD circuit/tube.
Using a switchable single input jack complicates things... I think I started down that route and changed to a 2 input plan which seemed more doable for me. I don't have the bandwidth at the moment to sort through the logic of that switch plus the other switching, but I found it useful to write out what I thought was happening logically in each switch condition, as it helped me to catch my own errors... and then also helped others to more easily see what I was thinking.

 

[ThePanda]

This is a ground up build using a 5F6a kit.
I am going back and forth with one or two inputs. I don't see myself using more than one.

[tdvt]

I am going back and forth with one or two inputs. I don't see myself using more than one.

Those kinds of reality-checks are a good thing.

Easy to get lost in would you could do, but most of the time it's best to edit & simplify.