Valve mic preamp design incoherent rambling

[Winston OBoogie]

If you've only got 200V and want to do cascode, then a folded cascode makes sense. Since we have no 'P' Channel tubes, you'd be looking at solid-state.
I've done lots of messing around with all solid-state folded cascodes, but not a tube/SS hybrid so don't have any recommendations on the cascoding device. A P-channel MOSFET or a pnp BJT, but which one? dunno?

 

[abbey road d enfer]

6S17k-v can be grid biased at around 47K, with the cathode grounded. A very quite front end.

What are the voltages?

 

[abbey road d enfer]

If you've only got 200V and want to do cascode,

Seems to me a good case for a parafeed cascode.

 

[Winston OBoogie]

Seems to me a good case for a parafeed cascode.

Actually yes. I hadn't thought of that but, if we were staying all tube and, in the tradition of the great German NWDR designers, a choke on top would work
Cool

 

[solkatten]

What are the voltages?

Actually I could get 230vdc with minor mod, but thats about it...

 

[solkatten]

Actually I could get 230vdc with minor mod, but thats about it...

What about a tube cascade? Should the 6sk17 be in the top or bottom or both?
I could do like Bartolo with j-fet and triode on top and anode choke instead of resistor and get away with lower B+

 

[abbey road d enfer]

Actually I could get 230vdc with minor mod, but thats about it...

My question was more specific. What are the anode, grid and B+ voltages?

 

[Winston OBoogie]

What about a tube cascade? Should the 6sk17 be in the top or bottom or both?
I could do like Bartolo with j-fet and triode on top and anode choke instead of resistor and get away with lower B+

I'd need to check the data sheet for the 6SK17... it's quite common to use the same device top and bottom but not necessary of course. A J-Fet on the bottom obviously works for him so that's one you have a starting point for.
The 2SK170 is slightly noisier than the BF862 he used (he mentions this), but not anything I'd worry about if you're using a step up input transformer.

Couple of advantages as far as I'm concerned: the 2SK170 is thru hole so I can see it and solder it without magnifying and, the Gm is lower which would give you a more manageable gain - unless you really do need 55 - 60B on top of what your IP transformer provides?

I do like that he's isolated the power supply for the cascoding grid voltage with an LND150 CCS... one thing that cascode amps are bad at is PSRR so anything that helps is good. **

If you don't have any K170's you could give me a holler and I'll send you a few 'BL' grade to play with. I just sent my last 'V' grades to another member so... nada V's left. Use 2 BL's in // for higher current and lower noise maybe?

** Edit: I also like that his regulator for that grid V is a low current CCS (the 6SK17 cascoding grid doesn't need/shouldn't have any current) feeding a resistor. I've done that too and then fed that reference voltage at the top of the resistor into a MOSFET cap multiplier for higher currents.
Using a resistor to define the voltage seems about the quietest thing to use, but I got some funny remarks when I mentioned it here once so didn't push it. I guess a lot of tube folks don't like too many small three-legged things in their amps?

 

[Winston OBoogie]

P.S. If you were goin to use a choke to load it, you'd need to make sure it had enough inductance for the higher impedance of the cascode's output anode. And at the current you're using. Might be an expensive Lundahl jobbie?

One thing I tried some years ago, which shouldn't really work, is an active CCS loading a cascode. But just treat it as a transconductance amplifier and put a resistor to ground from the junction of the two, converting the current into voltage. You can define the gain with the resistor.

I guess this thread has morphed into a free-for-all concerning preamp design ramblings.
Hope that's OK with Gus.

 

[solkatten]

My question was more specific. What are the anode, grid and B+ voltages?

B+ 200V, Anode voltage ca. 60 V, grid: ca. 0,5V

 

[solkatten]

P.S. If you were goin to use a choke to load it, you'd need to make sure it had enough inductance for the higher impedance of the cascode's output anode. And at the current you're using. Might be an expensive Lundahl jobbie?

One thing I tried some years ago, which shouldn't really work, is an active CCS loading a cascode. But just treat it as a transconductance amplifier and put a resistor to ground from the junction of the two, converting the current into voltage. You can define the gain with the resistor.

I guess this thread has morphed into a free-for-all concerning preamp design ramblings.
Hope that's OK with Gus.

Yep if my anode chokes arent sufficent at 200H 20ma, it will be a very expensive one... Then its cheaper to modify the PSU for higher B+ and use the resistor. Your right about gain u really dont need 60db + IPT. I already have a U3 chassi with 2 channels, and extra sheilding chamber for 2 anode chokes. I could use both for one channel to do the parafeed thing + LL1930 for the line stage. I have a nice Lundahl 1671 SE I can use for the second channel + grid biased 6s17 at the front end as I originally intended.

Each 6s17k-v needs its one filament supply so it make sense to just use one in the cascade.

 

[solkatten]

P.S. If you were goin to use a choke to load it, you'd need to make sure it had enough inductance for the higher impedance of the cascode's output anode. And at the current you're using. Might be an expensive Lundahl jobbie?

One thing I tried some years ago, which shouldn't really work, is an active CCS loading a cascode. But just treat it as a transconductance amplifier and put a resistor to ground from the junction of the two, converting the current into voltage. You can define the gain with the resistor.

I guess this thread has morphed into a free-for-all concerning preamp design ramblings.
Hope that's OK with Gus.

The tread has been hijacked... but its still incoherent rambling thou

 

[Winston OBoogie]

Incoherent rambling on my part for sure
I spent way too long messing around with ideas like this in the past, instead of actually building the usual stuff that people actually wanted to buy.


Sounds like you have enough stuff around to experiment with and have some fun anyway. Yeah, both chokes for one channel if you need to

 

[abbey road d enfer]

B+ 200V, Anode voltage ca. 60 V, grid: ca. 0,5V

OK. It runs at about 4mA.
You could have 60V Vka for both the bottom and top tube, and 80V across a ca. 40k anode resistor (or 10k+27k). That would give plenty of headroom. Gain would just be a matter of choosing the ratio between the coupled and un-decoupled part of the anode load.

 

[solkatten]

OK. It runs at about 4mA.
You could have 60V Vka for both the bottom and top tube, and 80V across a ca. 40k anode resistor (or 10k+27k). That would give plenty of headroom. Gain would just be a matter of choosing the ratio between the coupled and un-decoupled part of the anode load.

Like this?

 

[abbey road d enfer]

Like this?

Why not? Should work. I don't like too much the idea of running teh palte current through a potentiometer. What about this?

 

[abbey road d enfer]

I think you'll exceed the limits for the FET. Max 40V/400mW.
You should decrease the grid voltage.
For better noise performance, the capacitor in seies with the gain pot should be referenced to ground, rather than B+.

 

[solkatten]

I think you'll exceed the limits for the FET. Max 40V/400mW.
You should decrease the grid voltage.
For better noise performance, the capacitor in seies with the gain pot should be referenced to ground, rather than B+.

Jepp I removed that schematics I to realized the bias range was off...

 

[solkatten]

Here we go again pretty low dcr on the anode choke 2k47 a B+ of 100v would do?

 

[Winston OBoogie]

OK got it. I completely misread your earlier post, thought your chokes were 20H.

Voltage across the J-Fet is fine. On this, you could try degenerating the J-Fet a little if gain's too much.

Your next stage, is cathode follower right?

A little feedback from there returned under the input transformer to trim gain?

Edit: Merlin Blencowe states on another site that PSRR will be better without the cap across the lower resistor of the voltage divider