cascodes

[enthalpystudios]

quick question.. I'm trying to slowly learn this stuff, thought perhaps you guys may have some insight.

must a cascode utilize the same two triodes? and further more, must those triodes be a) matched or b) of the same valve?

I was thinking about something I could do with what I have lying around, which happens to be ef86's and 6sn7's, some cinemag input transformers, and some stancor outputs. the stancors primary inductance is perhaps not as high as it should be, so I thought that the amp should have a very low output impedance as is.

so I thought about an ef86 rc gain stage into a 6sn7 cathode follower.

but this cascode design intrigues me, so I thought about using the ef86 as a triode, cascode with one triode of the 6sn7, and the other triode from the 6sn7 as a cathode follower.

are there any blaring holes in my initial plan?

thanks to anyone who may have some input for a tube newb

 

[Guest]

Cascode minimizes Miller capacitances loading the tube on very low dynamic resistance. The upper tube (actually it may be a transistor) works with a "common grid" (or common base in case of transistor). It is good for very high frequencies, however I don't use them on AF because I don't like their distortions.

 

[enthalpystudios]

yeah... i think that it's also just unnecessary for my goals in addition to that, and other things...

i think i'm going to go for ef86 rc triode, gain control, to 6sn7 rc triode, and then cathode follower with the other 6sn7 triode section. i think this could be 'neat'

 

[solder_city]

[quote author="Wavebourn"]Cascode minimizes Miller capacitances loading the tube on very low dynamic resistance. The upper tube (actually it may be a transistor) works with a "common grid" (or common base in case of transistor). It is good for very high frequencies, however I don't use them on AF because I don't like their distortions.[/quote]

ok, well the guy already knows what a cascode is, and he didnt ask for your opinion on their sound or 'distortions'. if you dont know the answers to his questions, youre under no obligation to reply anyway!

for someone who claims to have a masters degree in electronics, you sure dont display much knowledge beyond what a hobbyist or anyone who reads 'audiophile' web sites would know. i wouldnt bring this up except you prance around here acting like you know everything- and some poor souls who dont know better might even believe it.

moderators can go ahead and scold me for the 'attack' but i believe its everyones responsibility to call out ********- since this is after all supposed to be a technical forum.

 

[burdij]

I have to disagree with the statement on distortion. Think of the cascode combination as a current injector and a fairly linear one at that, followed by a current to voltage converter driven by the cathode. Both of these elements take advantage of the fairly linear response of the tube's grid to cathode (or transistor's base to emitter diode).

They are successfully used for RF because of the reduction of the interelectrode capacitance in the second gain stage element. They have also been used successfully in applications where a very high input impedance is needed such as in the input stages of phono preamps and condenser microphones. Often, in the cascode configuration, the circuit is realized with an JFET input transistor driving the cathode of the triode or pentode gain stage.

I think your plan of a cascode input stage followed by a cathode follower should work well enough to merit further development.

 

[enthalpystudios]

well, perhaps i can drop my idea of using the ef86... i could find use for them else where i'm sure.

i am thinking two channel, 300v b+, and have 4 6sn7's available, so perhaps i could do it all with 6sn7's.

i read a bunch of stuff, and i was reading the section on cascodes, cathode followers, white cathode followers, etc, in 'valve amplifiers' last night, the author states that most dual triodes wont work well as a cascode, and that the 6DJ8 was the only tube designed to be a cascode.... is this true?

i do in fact ask questions here because there are good people that refuse to type ********.... other wise i wouldn't ask anyone... i find ******** on my own without much trouble at all.

anyway, the original question was asking if an ef86 as triode could be cascoded with the triode in 1/2 of a 6sn7.... and i guess i'm still wondering, and wish i had the experience and insight to understand what about a cascode might beg of its valves matched specs.

anyhow.... also, based on some of my reading last night, i got the impression that the ef86 isn't the greatest tube of all.... any truth in that one?

and hey, maybe i'll just see if i can make a single rc triode work before i try tackling complex sh*t


thanks for the replies guys

 

[Kit]

since this is after all supposed to be a technical forum.

Ah yes.

Thank you, soldercity for gracing us with replys like the one above.

The amount of technical info in your post is overwhelming.

 

[burdij]

I don't know if an EF86 operating in triode mode would be the best way to kick off the cascode. A "real" triode would be better in that position.

The 6DJ8 was designed as an high frequency amplifier and mixer for the tuner and RF amps of TV sets. This is one of the first "frame grid" tubes which allowed the reduction of the cathode to grid space. Often these circuits where constructed with cascode circuits, too, to take advantage of the low capacitances and (one advantage I forgot to mention) the lower noise offered by the cascode stage over a pentode.

In any case, I would not start any kind of low level, and hopefully linear, preamp stage with a pentode, due to its distortion issues (yes, there, I said it) particularly the predominance of odd harmonic distortion. The EF86 is a fine tube but I don't think its dynamic range is very good for audio applications where the input level is going to vary widely (such as in a preamp that will accept a range of signals from microvolts to volts).

The Radio Designer's Handbook (F. Langford-Smith) offers a number of preamp designs using the tube configurations you are proposing. You have to keep in mind that it does reflect the European/Australian bias toward using pentodes vs. the American bias toward triodes in these types of circuits. It does summarize very well the issues of interstage coupling and designing cathode follower drivers for transformer outputs, though.

Here is an interesting article regarding Miller capacitance, cascodes and designing guitar amps:

http://www.aikenamps.com/MillerCapacitance.html

 

[enthalpystudios]

aaahh.. cool link.

and i thought i was going to get some work done.

ok... so the whole story comes out, not only was the cascode designed for hf and tuners, but so was the 6dj8. how would a 6sn7 fare?

i'm going to get into the rdh more now that i have my feet on the ground a little bit. last time i got into i was way over my head.

i'll have to print off some more chapters.... but i have for output a stancor wf30, and that is just like a utc a20, designed for 600max to 600max, each winding going down to like 50 minimum, and all kinds of stop s in between.

because it probably has kind of low primary inductance, thats why i'm thinking of using cathode follower,

i was interested in the cascode because it seemed to alleviate a lot of bandwith by reducing the effect of miller capacitance

thanks for the reply ;]

 

[bcarso]

One detail to consider: maxima of heater-cathode voltages in a dual triode. The upper tube's cathode is operating at the lower's plate voltage, the lower tube's cathode typically close to ground, but the heaters are usually common internally. To complicate matters further it's often desirable to have the heater biased somewhat positive with respect to the cathode for lowest noise. Overall the lower tube's heater-cathode voltage difference can be large.

With two separate tubes you can use two different heater supplies.

You also want to select tubes where the peak plate current of the lower doesn't exceed that of the upper by much.

As far as distortion, on the one hand it's true that triodes have lowest low-frequency distortion operating into high-Z plate loads in common-cathode mode. However the voltage gain and hence the Miller effect under these conditions is maximized and thus bandwidth is compromised, unless the source impedance is low. In a preamp your input signal is apt to be small, so the distortion of the first tube will tend to be fairly small as well. So the circuit can work pretty well, and is definitely wideband.

 

[pstamler]

First of all, yes, you can use different tubes for a cascode.

Now about bandwidth. Cascode operation will minimize the Miller capacitance at the input, so it will have minimal effect on bandwidth when driven by a higher-impedance source. However, a cascode also has a high output impedance, so whatever it drives needs to have low capacitance as well, or bandwidth will get narrowed there.

What Brad says about filament-to-cathode voltages is spot on; usually, though, it's possible to find some level for the filament that will work if you do a single tube if you want. For example, if the bottom tube's plate and the top tube's cathode sit at 150V, then floating the filament to 75V will maintain safe margins to both cathodes; assuming the bottom tube's cathode is around 3V, you see 72V heater-to-cathode, while you see 75V to the top cathode. All happy.

Finally, the EF86 makes an excellent triode if you wire it for triode operation; plenty of dynamic range, reasonably low distortion, and quite low microphonics.

Peace,
Paul

 

[Guest]

[quote author="burdij"]I have to disagree with the statement on distortion. Think of the cascode combination as a current injector and a fairly linear one at that,
[/quote]

Sorry, I can't think of a fair current injector and fair I/U convertor, I got used to think of them as of real beasts with own characters. If it is Ok for transistor to be nearly fair current source being loaded by low dynamic resistance, for tube it is bad idea because it's Ri depends on anode current. I prefer to load vacuum triodes by constant current sources instead of constant voltage sources because distortions are less obviously for everyone who knows what is the vacuum triode.

So, answering the question: "May I use pentodes, triodes, BJTs, FETs. etc... in cascodes" one need to think of what is cascode, what advantages/disadvantages it gives, and what kind of character each particular beast has if it being used such a way.

The answer is obvious: "YES YOU CAN!" but it it calls for another engineering question: "WHY DO YOU NEED IT FOR?"

Engineering is about achievements of goals using combinations of resources. If the goal is to use some lovely combination of resources it is not an engineering.

PS: solder_city, what was your goal when you started flame and off-topic?

 

[Guest]

[quote author="Kit"]

since this is after all supposed to be a technical forum.

Ah yes.

Thank you, soldercity for gracing us with replys like the one above.

The amount of technical info in your post is overwhelming.[/quote]

Kit, I assume it is a provocation, I now supposed to piss off and to prove my diploma reveal some my secrets about how do I use cascoded CCS that has also a servo in it that stabilizes anode voltage in my tube preamp. :green:

I can show it without any provocations...

 

[enthalpystudios]

good post paul, this stuff in the last week or two has really started to setup in my mind.

i was working on a p2p champ today with my dad in his breakroom at work (he wasn't working;] and i started to really grasp some more things, while explaining them to him, a beginner in electronics, as simply as I could...

anyway, i think such a complex design is a) too far advanced for me while just starting out with a scratch built amp, and b) i'm really not sure why i wanted to do this, other than it seemed 'neat' and the names for the compound devices and what not are 'catchy' and i felt like i had stuff in drawers that would work

;]

so yeah, i believe i'll try something along the lines of ef86 as triode into gain control into 6sn7 as simple rc triode and then the other 6sn7 triode as a cathode follower. i think that this seems more along my lines for the time being...

thanks for the replies guys

billy

 

[Guest]

Billy;

you may try everything for yourself and figure out what do you like better.
Don't be afraid of stealing interesting schematical solutions from gear that was in manufacturing to try and gain experience. If you want to try cascode, try it. With pentodes on bottom, triodes on top, pentodes in triode mode, may combine tubes and transistors, without touching everything it is hard to learn, even with modern computer programs that simulate reality.

One experiment worth many discussions. :wink:

 

[enthalpystudios]

Yeah, I probably won't be messing with any simulators anytime soon. I'll most likely try what I see around and look first into the RDH for some ideas. The cascode ccs hybrid stuff is all a little much for me to take in however... still don't have much of an idea how transistors do what they do at this point.

 

[Guest]

[quote author="enthalpystudios"]Yeah, I probably won't be messing with any simulators anytime soon. I'll most likely try what I see around and look first into the RDH for some ideas. The cascode ccs hybrid stuff is all a little much for me to take in however... still don't have much of an idea how transistors do what they do at this point.[/quote]

Transistors are roughly speaking devices that generate current through collector-emitter that depends a little on collector-emitter voltage, but depends much more on base-emitter current.

So, varying the current between base and emitter we may vary current that flows from collector.

If you vary voltage between collector and emitter, measure current variation, and apply the Ohm Law, you will get very high dynamic resistance, while if to divide voltage between collector and emitter by current you will get lower resistance!

So, you can provide for yout vacuum triode needed current, but load it by high dynamic resistance if to use a transistor!

Now, let's go back to the Miller capacitance that is capacitance between base and collecter multiplied by voltage amplification between base and collector.

If we stabilize voltage on baze using a Zener we don't have a voltage amplification between base and collector, so we don't have a Miller capacitance (multiplication of capacitance between collector and base). But current in collector reflects current in emitter, so if we provide this current from one more transistor we will have a current source with big dynamic resistance, but with low Miller capacitance.

Now, in my case a voltage feedback from bottom transistor collector (top transistor in cascode, actually) stabilizes voltage on it. Without a capacitor such feedback would make dynamic resistance much lower, but with the capacitor dynamic resistance of such load depends on frequency, so on DC where the feedback works dynamic resistance is very low, but on audio frequencies it is very high!

This load looks like a very high quality choke load that has low DC resistance, but high AC impedance. Triode is happy with such a load, much-much happier than loaded by a cathode of another tube in cascode arrangement.

Without such tricky feedback (called sometimes servo) dynamic resistance will be high both on AC and DC, it means that working point (voltage on anode) will be unstable and depend much more on heater temperature and tube emission.

 

[solder_city]

kinda defeats the main purpose of the cascode doesnt it? input capacitance (at tube grid) will be high, actually higher than with a simple plate resistor. i knew that right away and i dont even have a masters degree!

i was wondering why he decided he needed a cascode, too, but someone (tablebeast?) busted our balls recently for always asking 'why do you want to do that' instead of answering the question, so this time i didnt ask. :green:

 

[Guest]

The amp with 47K input and RIAA correction, is something a kinda little bit different from UHF antenna preamp-mixer, right, soldercity?

 

[solder_city]

yes, but he was asking about cascode amplifiers. not phono preamps with common cathode triode loaded by cascoded current source. the current source may be cascoded but the amplifier is most definitely not.