Cascode advice?

[murrayatuptown]

Hi:

I found some older posts that looked suitable to append questions to but the forum suggested a new post due to thread aging...

15+ years ago I bought a variety of tubes, sockets & a budget temperature-controlled soldering iron. We moved and the stuff ended up in the garage...I found it maybe a year ago (problems starting & finishing projects...)

I was probably leaning toward phono preamps & home 'hifi' things, and some of the schematic websites I was looking at are extinct.

So now I'm thinking small guitar amps, and my interest in cascodes just shifted to this application.

I read Merlin's chapter in cascodes, then bought the preamp (guitar/bass) book...good book, but I think I already had all the cascode content...

Found Ulrich Neumann DRRI Weber kit redesign and then Ken Gilbert preamp.

But I'm thinking 12AU7's and 5687, so that brings me here for opinion and advice (please).

I actually have more 5687's (GE JAN 5697WB) and  they appear capable of SE and PP output work (schematics exist).

I wonder if it's a 'waste' to use such 'robust' tubes in a preamp, but cascode mic pre's and Broskie Aikido variants show plenty of satisfied users.

I'm wondering if anyone has tried 5687 in cascode preamp stages and roughly what kind of bias. I am working through Merlin's load lines for 12AU7 before I try my own for 5687 cascode (haven't found this path blazed yet).

I think I am thinking too hard about both preamp and power amp with the same tube & need to get a handle on what is a suitable bias range for 5687 small signal use.

I don't think I need to load line them for 70-100% plate dissipation like one might for output stage. All the mu, gm, etc. datasheet info is at near max. dissipation (for logic gates?). 10-30 mA seems goofy for preamp stage but I haven't figured out any gm traits at low current over usable plate voltage range. Maybe there is a different operating point that's beneficial for each stage.

Another thought was 5687 cascode front end, 5687 outputs and TBD PI.

I also see merit in Ulrich's 12AX7 interfacing with guitar and cascode on subsequent stage.

Thanks for reading and any thoughts.

One thing at a time, I guess, but I was really intrigued by someone's post about cascode PP output stage too!

[vibrolax]

I don't have any specific advice about how to use cascodes in a guitar amp, but I have some thoughts about how I end up designing/building what I build.

On starting point is to try to make a preamp/amp that will allow us to produce a certain range of tones with our instrument.  An opposite starting point is to choose some particular set of parts, e.g. 5687 triodes, and figure out how to turn them into some kind of useful guitar amp.  I've started from both ends on various projects.

For me, it's not a waste to experiment with non-canonical circuits (for instrument amps) as long as:
1) I get the tone I wanted OR
2) I learn something about circuits, and why I didn't / couldn't get the tone I wanted.

Broskie's a wonder.  I have been reading Tubecad Journal since he started publishing it, but I haven't found anyone using his ideas to make instrument preamps/amps.  I actually did refer to TCJ when I used a DC coupled optimal white cathode follower in some guitar/bass preamps I built.

It's time for a cascode Aikido guitar preamp!  You can be the one who makes it happen.
Jon

[drgonzonm]

A quick peek at valvewizard, tells me you could make an all 5687 pp amp, probably in the 10 to 15 w range.

A couple things that concern me include,
1. is the 0.9 amp heater draw on each tube.  say you use 4 tubes, that 3.6 amps at 6.3 volts
2. The possible need to elevate the heater voltage, not hard, just one more complication.   

[murrayatuptown]

Hey, maybe I need to look harder at Valve Wizard to see what you found...or maybe you read between the lines & can figure out the inferences.

I was intrigued by alot on the TubeCad site but it just encourages bouncing from obe udea to another...some seem like concepts and not finished ideas. I need to either build something from finished design (good idea) or start simply for a scratch concept.

I really hate to admit I am a transformer hoarder, so I'll just say I have, to use that figure of speech, several manure-tons of power transformers, many with robust filament current capability.

I have very few output transformers but they are available...

I was reading Blencowe'schapter on cascode Saturday and he has an example of a cascode graphical analysis of a dissimilar triode construction.

Maybe a hare-brained idea, and I should try something already published, but I would like to digest the statement he made about what traits are important abd what are less so for a dissimilar cascode...grid-leak variant.

In addition to the 5687's, I have a carton of 6AS7's (have to look; might be 6080's). Anither filament current hog, they have a lit of frustrating traits deterring output device use. Nain one people cite is very low mu and needing massive gate drive as a power tube.

I think they have fairly high gm, however, so a weird idea, no, make that two weird ideas, is
to A) make a cascode from 1/2 a 5687 and 1/2 a 6080/6AS7. Probably no point in paralleling them...two stages of cascode follows conceptually, if there is a point. (I'm at the crazybidea stage, drunk on my overabundance of filament transformers.

and B) I barely caught wind of people allegedly making a cascode power stage amplifier. If that makes any sense, 5687-6AS7 power cascode.

If those are not the stupidest ideas possible, I get to use tubes I bought before realizing they were not the most useful choices. Regulated power supply doesn't appeal to me at the moment...

One thing at a time...but I also have 6C33C-B's...no, I am not thinking what you think I am thinking...I don't think so, anyway.

Thank you

[drgonzonm]

Tube cad shows 200v b-rail, this is to minimize that cathode to heater  delta voltage limitation of 100v without resorting to heater elevation circuits. 

[murrayatuptown]

Thanks.

I searched 'dissimilar cascode' today and found an archived Audio Asylum thread that discussed & expanded on Morgan Jones book remarks.

Goal of upper triode having high gm and low rp if different from lower triode suggested 1) go ahead & try to estimate outcome of the potentially absurd 6AS7/6080 + 5687, and 2) consider less absurd 5687 + 12AU7, vs single envelope pair each type, etc.

I found an online cascode calculator that allows, uh, dissimilarology, and I went nuts entering combinations. 6AS7 didn't produce interesting predicted result. I may have to re-read the fine print & see if that calculator is limited to something specific like 6DJ8...the 9-pin ones aren't that big a project to evaluate (famous last words)...

Maybe make modular cascode blocks with octal plugs so they can be swapped easily! I've got little 2-socket u-chasses that came from something ancient that got scrapped...mayve an HP universal counter with Nixie tubes (earlier stages-Nixies aren't normally 9-pin, if ever).

Maybe I should try that before calling them 'cascodules'. That sounds a little more elegant than dissimilarology...

[murrayatuptown]

OK...found my cannibalized HP AC-4K 120 kcps counter sub-chasses.

I got the scraps. Someone else had already gotten the dual triodes & NE-2 'digits backlights'. 1962 date stamps. 4 x 9-pin sockets in-line, fairly close spacing, riveted in and common chassis grounded. Mine had 12-pin Amphenol plug bases which someone told me were used inside Leslie cabinets. I either tried or failed getting rid of them on eBay or pitched them or they'll turn up in a future basement archaeological expedition. Only found one     
octal plug, so swappable module idea is fading for the moment.

Some potential ideas there...all 9-pin mini amp chassis with remotely located transformers...2-channel mic pre...but I never get too worked up about that...cheapest mics on sale are probably a lousy signal source and no desire to be in front of a mic (perhaps someone else?) reduce the enthusiasm for this...

Revisited Keith-snook.info site's cascode calculator. Uses ra, mu, and peripheral infrastructure resistances (plate loads, grid resistors, I guess... didn't try to figure out yet.

I made a table of ra & mu for each dissimilar upper/lower triode pair candidate, and left the other pre-loaded resistors alone. My logic was that they were 'largish', in the ballpark for other 6DJ8 or 12AU7 cascodes (conventional/similar configuration), and that I don't know yet what alternative values might be of interest.

As a first approximation (or 3/4 of the way there), I just entered the ra and mu for various upper/lower triode and operating point values to see 'what if?' trends...sugnificance and practicality to be sorted out later.

If I rationally concluded the intent is to observe the V1 and V2 gains it calculated for the purpose of tweaking the upper and lower triode AND take the combined gain of the two as the parameter of interest (hey, the calculator was fired up & I took it for a joy ride...I don't know what direction I was flying..).

If I look at this combined gain alone, ignoring the output impedance, etc., there are some interesting results. 12AU7 similar and 6DJ8 similar cascode are within a couple dB of each other, low 70 dB range. I KNOW there are some circuit differences that will necessitate re-evaluation.

The 5687/12AU7 dissimilar combinations produced comparable or barely higher 'calculator gain', and the 6DJ8/5687 was the highest gain.

I found I had written some of my results in reverse order (transposed Av1 and Av2), but again, that can be fixed and clarified.

Some of these initial observations may debunk themselves for secondary impracticalities that may be revealed (one of you may say 'duh' (quoting me) to my pseudo-science.

6AS7 + 9-pin dissimilar Franken-cascode simulations ranged from (hmm, a little better than a voltage divider if I ignore the heater power required) to 'less but still substantial gain'.

Not sure if the 'but what hell for?' answer is at hand yet, but if a P-P or SE cascode power
output stage offers any benefits, there appeared to be two combinations with enough gain to take a second look.

A second look at the TubeCad site again (is rabbit hole an appropriate term? Or time-space continuum anomaly? I was lucky to escape and get back to what else I had to do that day).

I don't know how to stop once I start exploring there...or if I'd ever finish a build without changing the goal based on another blog page there!)

Anyway, there is something interesting here.

(To me)

[murrayatuptown]

NO, NO, NO! No power cascode at that power level...pentode-like Zout is not transformer friendly...actually, the converse...

Maybe I learned something.

[Tony Bones]

I can't help a whole lot with cascodes as I haven't designed around them myself. But, the way I understand them is essentially as a composite pentode. The only advantages that I know of are (potentially) high gain and wide bandwidth. Don't really care about bandwidth in guitar amps. In fact, it just increases the odds of oscillation, sometimes at frequencies so high you won't even see them on your scope if you have it set to look at audio frequencies.

But, you should experiment. That means just build stuff and try it. I mean, you can narrow down the number of things to try by thinking about it, but eventually you need to just build something.

As for loadlines and x% of rate plate dissipation, just banish that rule of thumb from your thinking once and for all. The sooner you do the better off you'll be. The only time it's at all useful is if you're trying to get as many watts as possible out of a power amp. But if you think about it, that doesn't need to be a goal, even with power amps; good sound might be a better goal.

Obviously, you don't want to exceed the rated plate dissipation with power amps or preamps (at least not by much) but there's absolutely no reason, when drawing loadlines for preamps, to aim for x% of plate dissipation. It makes no sense whatsoever when designing guitar amps or hifi amps. Just look at the plate curves and consider your B+. Then find a line and operating point that looks good.

I've used a lot of 5687 in hifi amps. I found they sound best when they're passing a lot of current. Don't baby them. That's hifi, but I believe hifi and guitar amps have a lot more in common than most people think. By that I mean what sounds good in hifi often sounds good in a guitar amp too. We just might turn the guitar amp up a whole lot more sometimes.