Cascode Design Using 2 Different Triodes?

[drmachismo]

Hi All - I am designing a mic preamp and looking at doing a cascode first stage. Anybody have any thoughts on the pros and cons beyond the advantages of lower noise and no Miller effect? I have Merlin Bledcowe's book on preamp design and have worked through his design process, however it focuses solely on designs that use the same triode for the upper and lower sections. I've seen posts that talk about higher gain using a high Gm triode (12AU7/6SN7) for the lower section and a higher mu triode (12AX7/6SL7) for the upper section but cannot find any more details on the design process or associated baggage that may come along with this approach. Would love some input from the community. Thanks in advance!

 

[ruffrecords]

Miller effect is usually not a problem in the first stage of a tube mic pre. The reason is that, even with a 1:10 input transformer, the typical source resistance seen by the tube is about 15K. Miller effect rears its ugly head in the second stage which is driven from the much higher output impedance of the the input stage which can easily be 10 times this value.

Cheers

Ian

 

[drmachismo]

Miller effect is usually not a problem in the first stage of a tube mic pre. The reason is that, even with a 1:10 input transformer, the typical source resistance seen by the tube is about 15K. Miller effect rears its ugly head in the second stage which is driven from the much higher output impedance of the the input stage which can easily be 10 times this value.

Cheers

Ian

Thanks Ian - good to know. I'm also going to be driving this from a DI, which could be a magnetic pickup with a much higher output impedance so I suspect the cascode would be a benefit here.

 

[ruffrecords]

Thanks Ian - good to know. I'm also going to be driving this from a DI, which could be a magnetic pickup with a much higher output impedance so I suspect the cascode would be a benefit here.

Yes it would definitely benefit a DI. On the other hand most guitar amps (which are driven by magnetic pickups) include quite a large value resistor (typically 68K) in series with the grid of the firs tube stage to deliberately tailor the frequency response using the Miller effect. And having said that, if you were using a piezo pickup then the cascode would definitely be an improvement.

Cheers

Ian

 

[Wilhelm]

Hi All - I am designing a mic preamp and looking at doing a cascode first stage. Anybody have any thoughts on the pros and cons beyond the advantages of lower noise and no Miller effect? I have Merlin Bledcowe's book on preamp design and have worked through his design process, however it focuses solely on designs that use the same triode for the upper and lower sections. I've seen posts that talk about higher gain using a high Gm triode (12AU7/6SN7) for the lower section and a higher mu triode (12AX7/6SL7) for the upper section but cannot find any more details on the design process or associated baggage that may come along with this approach. Would love some input from the community. Thanks in advance!

Good idea - But the 12 AU 7's you mentioned, can't really be considered low-noise. In hi-fi applications, the E188CC, E288CC and similar "steep" / high mu tubes have long been preferred in low-noise designs. There's also the extreme, special-socket, hard to find ECC2000 ( can't remember the American designation for any of these) . You could probably even stack a compatible ( idle-current-wise) low noise pentode on top of two triode-sections running in parallel ( remember separate cathode resistors for each triode's cathode - though their plates get to be directly connected- ) and so achieve even better gain/noise figures. But as the output Z / impedance would be close to that of the plate-resistor itself ( of your choice), a buffer- ( perhaps just a simple cathode-follower ) might be needed right after the gain-stage (naturally, the buffer stage grid could be DC-coupled to the gain-stage pentode-plate) But all this would depend on the actual needs /length of cables etc.

 

[gyraf]

Imo, there is nothing wrong with the ECC82 when it comes to real-world noise figures - I don't think we should pay too much attention to what the hifi-crowd are lusting after this month - over there it's only good if it's stupidly expensive or made unobtainable by hoarding

/Jakob E.

 

[Wilhelm]

Nah - 12AU7 is OK good for phasesplitters, driving power-tubes and cathode-followers. But you get almost no voltage-gain in exchange for the added noise of yet another active component in the signal-chain. Absolutely not an "input-tube", for a microphone-module ( Lest the "clever" designer relies on high-quality ( = expensive) input transformers to bring the signal-voltage- up to standard line-level, before current-buffering via those 12AU7's

 

[Wilhelm]

For Dynamic microphones with the pre-amp integrated - it's possibly a trail to walk down/test. ( A simple diy ribbon mic, would most likely be capable of manifesting the necessary current-modulation. To avoid a huge electrolytic cap/ or an expensive transformer on the input, the ribbon-mic could simply "sit" in the cathode-to-ground string. Good luck with it !

 

[NovaPA]

How about usinc cascode, combining tube+BJT transistor?

 

[abbey road d enfer]

Ok, time to be a bad influence....

The only time when you don't have the miller effect is when you don't use the grid as the input.
Instead of a typical cascode, how about a grounded grid to a cathode fallower?

So your proposal is to trade Miller-induced HF attenuation for wide-band attenuation due to to heavy loading of the mic?
This would not sit well with dynamic and ribbon mics.
Only reminding you that recommended load for mics is 10x nominal impedance, which, of course you can consider as being "the limitations they made for themselves."

 

[abbey road d enfer]

Hi All - I am designing a mic preamp and looking at doing a cascode first stage. Anybody have any thoughts on the pros and cons beyond the advantages of lower noise and no Miller effect?

Unless I have forgotten a significant part of my past, I don't remember that cascode improves noise.

I've seen posts that talk about higher gain using a high Gm triode (12AU7/6SN7) for the lower section and a higher mu triode (12AX7/6SL7) for the upper section but cannot find any more details on the design process

I'd be curious to read about it. It may be beneficial on the standpoint of gain, which is rarely an issue, but probably would be detrimental to output impedance and noise, compared to a combination of a high-gain lower-section and a high Gm upper section.

 

[abbey road d enfer]

To avoid a huge electrolytic cap/ or an expensive transformer on the input, the ribbon-mic could simply "sit" in the cathode-to-ground string. Good luck with it !

I can see two reasons not to do it.
First, the ribbon would be off-centered because of current going through it, probably resulting in significant sensitivity loss, distortion and possible damage.
Second, the tube's noise voltage would dominate the intrincsic ribbon's noise by at least a factor 10 (actually more like 100).

 

[Wilhelm]

Hi !
As of now i'm not actually in the proces of building.... and so i can only thank you for bringing some more good thoughts to the table.
A) A typical PA/recording ribbon-microphone has a diaphragm that's essentially the size of treble units in speakers. ( So i would not be worried about currents in the order of about 2-3 miliamperes .
B) Regarding displacement, : as i' ( and many others ) contemplate a design that has the active electronics built-in, it would be quite easy to accomodate for some small DC-current-set-off via adjustment to the suspension - and actually, such current would provide a ( small) positive constant electric damping of unwanted residual movements, when in use. C) Regarding noise : if that's the main concern - we need to go digital - . There's no analogue buffering without some noise introduced.
But to prove all this - I need to build the damn thing - the proof of the pudding is in the eating " ! Cheers

 

[Wilhelm]

It gets interesting when something like this begins to form too:
View attachment 119374

Nice - but i think you need a balanced step-down (voltage to current) transformer at your input if you want it to work with standard-mic's

 

[abbey road d enfer]

Obviously you don't know that a grounded grid is used in 20-50 ohm input circuits.

Actually, whatever you think, I know that common-grid or common-base circuits are used in many receivers.
I've never seen them used in mic preamps, with reason.

Rarely I've seen them applied in the audio sector other than a few phonograph preamp circuits.

Moving-coil preamps are just about the only case where the use of common-base ot common-grid circuits can be justified.

Also, you don't understand where that 10 times rule applies and where they got that from.

Do I not?

But maybe you were only exposed to the corporate junk builds of the $5 preamp (the INA217 wonder) they put in a lot of things.

You make a lot of ASSumptions about what I know or don't, don't you?

I suggest you look that up and review what capacitance reactance does and how ten times rule doesn't apply in a dc coupled preamp circuits.

Please tell me, I'm eager.

Btw @gyaf knows I don't play well with moderators that arbitrarily chime in and try to counter with things that they don't know truly about. This is your warning.

Wow! What are you gonna do? Report me? gyraf and JR know I don't play well with members who call themselves doctors and show their basic misunderstandings in just a few sentences.
If I was like you, evaluating your knowledge or lack of it by my own deficiencies, I would say you have no notion of OSI, but I'm not like you.
I'm ready to read your proof that I'm wrong.

 

[JohnRoberts]

The 10x source impedance rule of thumb is just that, a rule of thumb.

warning, another Peavey anecdote... Back when I was working at Peavey I tried to take advantage of expertise I was clearly not exposed to. One day I wandered over to the transducer group and tried to pick the brains of some old senior gray haired microphone design engineers about optimal terminations. They did not have a concise answer for me.

If we look back far enough microphones and preamps were designed together as systems. There was no interchangeability of mics between standard preamps. A benefit we now take for granted. The 10x rule of thumb works with modern gear.

JR

PS: Abbey is the kind one.....

 

[abbey road d enfer]

Hi !
As of now i'm not actually in the proces of building.... and so i can only thank you for bringing some more good thoughts to the table.
A) A typical PA/recording ribbon-microphone has a diaphragm that's essentially the size of treble units in speakers. ( So i would not be worried about currents in the order of about 2-3 miliamperes .

I'm not worried either by Joule effect.

B) Regarding displacement, : as i' ( and many others ) contemplate a design that has the active electronics built-in, it would be quite easy to accomodate for some small DC-current-set-off via adjustment to the suspension

It would need to be designed in. It's feasible but it's an element of complexity.

- and actually, such current would provide a ( small) positive constant electric damping of unwanted residual movements, when in use.

Tell me why a constant force would provide damping. In a resonant system, damping is provided by a speed-dependant term.

C) Regarding noise : if that's the main concern - we need to go digital - . There's no analogue buffering without some noise introduced.

Whatever the rest of the chain, the input stage is crucial in the global noise performance. Low-noise design is governed by adequating the input stage's OSI to the source.

But to prove all this - I need to build the damn thing - the proof of the pudding is in the eating " ! Cheers

Good luck with it!

 

[Wilhelm]

Thanks - interesting and good to know . But please allow me to note that If some software concludes that "the two 50 ohm resistors is the 100 ohm voltage source.", it can't be correct ( the two cathodes facing also ! ) . But don't take my word for it. The setup to measure the z-input is simple : a potentiometer in series with a near 0-ohm signal-generator source, connected to the circuit input, will do it. Adjust the potentiometer until a scope reading indicates the signal-amplitude is exactly the same across the circuit input terminals and the potentiometer - at fx 1khz. The z-input will then be the same as what the potentiometer has been adjusted to - within most of the circuits useable audio-frequence-range.

 

[abbey road d enfer]

Thanks - interesting and good to know . But please allow me to note that If some software concludes that "the two 50 ohm resistors is the 100 ohm voltage source.", it can't be correct

I do agree. The common-grid impedance comes in parallels with Rk. depending on the tube, it may be negligible, but it's anyway real.

 

[Ivan K.]

I'm no tube/valve expert, but you might be interested in www.valvewizard.co.uk/mufollower.html and John Broskie's Guide to tube circuit analysis and design at www.tubecad.com/2009/10blog0173.htm covering SRPP circuits.
Regards,
Ivan