Valve mic preamp design incoherent rambling

Help Support GroupDIY:

Winston OBoogie

Well-known member
Joined
Jun 3, 2004
Messages
2,466
Location
UK.
Yep, I think that's not a bad idea given you're using a 4:1 OPT

If your anode voltage on the 6SK17 is 100V and your supply is 200V, you might think about just directly coupling the 6n6 cathode follower to the 6SK17. It's a couple of resistors, a cap, and a low freq pole cheaper.

Feedback as you have it around the cascode and follower is what I meant above 👍
It's possible you may still want a small (10 - 22R) degeneration resistor under the K170 source rather than running at IDSS but that's easy to just try or play with as you build it.
 

solkatten

Well-known member
Joined
Apr 3, 2021
Messages
121
Location
sweden
evb
The second stage's response is very much load-dependant. I would insert another CF there. Or make it a WCF.
does it help to raise the impedance of FB network like in the v72. With another IPT you could match thouse values better. I rather not add more stages
 

Winston OBoogie

Well-known member
Joined
Jun 3, 2004
Messages
2,466
Location
UK.
Miscommunication or misunderstanding I think :).
Sol, your feedback values are fine. Abbey meant the output stage into the 4:1 output transformer. The secondary load on that transformer will have an impact on response.
If you'd rather not add another stage, then you can just try it as is.

Since this is for you (I assume), then at least you'll always know or be in control of the load impedance it sees so you could just see what the lowest load impedance it's comfortable with is and go from there.

However, frequency response of your OPT will be better if driven from a cathode or White follower
 
Last edited:

solkatten

Well-known member
Joined
Apr 3, 2021
Messages
121
Location
sweden
Miscommunication I think :).
Sol, your feedback values are fine. Abbey meant the output stage into the 4:1 output transformer. The secondary load on that transformer will have an impact on response.
If you'd rather not add another stage, then you can just try it as is.

Since this is for you (I assume), then at least you'll always know or be in control of the load impedance it sees so you could just see what the lowest load impedance it's comfortable with is and go from there.

However, frequency response of your OPT will be better if driven from a cathode or White follower
I don't get it... There is no NFB loop in the output stage.
The input stage is good (I think).
I don't get it... There is no NFB loop in the output stage.
The input stage is good (I think).
ok missunderstanding👍
 

solkatten

Well-known member
Joined
Apr 3, 2021
Messages
121
Location
sweden
Miscommunication I think :).
Sol, your feedback values are fine. Abbey meant the output stage into the 4:1 output transformer. The secondary load on that transformer will have an impact on response.
If you'd rather not add another stage, then you can just try it as is.

Since this is for you (I assume), then at least you'll always know or be in control of the load impedance it sees so you could just see what the lowest load impedance it's comfortable with is and go from there.

However, frequency response of your OPT will be better if driven from a cathode or White follower
I also have a cathode feedback winding on the OTP to play with if I want a cleaner output👍
I don't get it... There is no NFB loop in the output stage.
The input stage is good (I think).
Or what about a v72 inspired cascade😀👍
 

Attachments

  • 560D3EBF-4EDA-4B44-B924-F0DF7878761E.jpeg
    560D3EBF-4EDA-4B44-B924-F0DF7878761E.jpeg
    305.7 KB · Views: 54

Winston OBoogie

Well-known member
Joined
Jun 3, 2004
Messages
2,466
Location
UK.
There's probably enough gain in that input stage with transformer that you'd be fine without your additional 6n6p output stage so, why not.

I've never used the E81L but, assuming it's OK at 15mA which I'm sure it is, then biasing the cathode off the choke would put it at 37V for the cathode for 15mA. I'll have to look at the data sheet but, If it'll do 20mA then I'd do that with 50V on the cathode.

Give it a shot.

Edit: try it with scaled down feedback elements too, there's no need be that high I don't think... although I don't doubt the V72A designers had good reason to do it that way so, not sure I guess, hmm?


I think it has potential to be a damn fine mic amp :)
 
Last edited:

solkatten

Well-known member
Joined
Apr 3, 2021
Messages
121
Location
sweden
There's probably enough gain in that input stage with transformer that you'd be fine without your additional 6n6p output stage so, why not.

I've never used the E81L but, assuming it's OK at 15mA which I'm sure it is, then biasing the cathode off the choke would put it at 37V for the cathode for 15mA. I'll have to look at the data sheet but, If it'll do 20mA then I'd do that with 50V on the cathode.

Give it a shot.

Edit: try it with scaled down feedback elements too, there's no need be that high I don't think... although I don't doubt the V72A designers had good reason to do it that way so, not sure I guess, hmm?


I think it has potential to be a damn fine mic amp :)
Yes it can do 20ma Jonte Knif uses E81L in his therapre design
 

solkatten

Well-known member
Joined
Apr 3, 2021
Messages
121
Location
sweden
There's probably enough gain in that input stage with transformer that you'd be fine without your additional 6n6p output stage so, why not.

I've never used the E81L but, assuming it's OK at 15mA which I'm sure it is, then biasing the cathode off the choke would put it at 37V for the cathode for 15mA. I'll have to look at the data sheet but, If it'll do 20mA then I'd do that with 50V on the cathode.

Give it a shot.

Edit: try it with scaled down feedback elements too, there's no need be that high I don't think... although I don't doubt the V72A designers had good reason to do it that way so, not sure I guess, hmm?


I think it has potential to be a damn fine mic amp :)
Here we go a russian/german planar 72 mic pre👌 one practical thing to add input pad 0, -20, -30 and -40
As far as transformers goes I will use a LL1538 IPT and LL1930 5,8:1 OPT

It Will be really fun to build that thing....
 

Attachments

  • 6DBA58DB-511E-486C-96E0-10B4C8ED8685.jpeg
    6DBA58DB-511E-486C-96E0-10B4C8ED8685.jpeg
    274.2 KB · Views: 47
Last edited:

Winston OBoogie

Well-known member
Joined
Jun 3, 2004
Messages
2,466
Location
UK.
I don't think I've ever seen that scheme with your output valve before, where it's self biased but riding on the choke. Cool idea. Have you used it before?
 

solkatten

Well-known member
Joined
Apr 3, 2021
Messages
121
Location
sweden
Actually my chokes are rated At 35ma! The cascade doesnt need that huge choke
 
Last edited:

solkatten

Well-known member
Joined
Apr 3, 2021
Messages
121
Location
sweden
Here we go a russian/german planar 72 mic pre👌 one practical thing to add input pad 0, -20, -30 and -40
As far as transformers goes I will use a LL1538 IPT and LL1930 5,8:1 OPT

It Will be really fun to build that thing....
Haha! You'll find out. I'm putting the idea in my brain's library for future playing around with too.
Sometimes you are just lucky...
I think I resistor load the cascade them I got chokes for 2 channels👍
43923DBF-CE17-404F-A513-7A7B0DA590A9.jpeg
 

Winston OBoogie

Well-known member
Joined
Jun 3, 2004
Messages
2,466
Location
UK.
Actually my chokes are rated At 35ma! The cascade doesnt med that huge choke

It's more the inductance rather than current rating (size of air gap) I was concerned about earlier.

What's the output impedance of your cascode?

With that you can see what you need inductance wise for your low end:

L = Xl/(2 * pi * f)

I've seen Hi-Fi parafeeding folks tending to go for 20Hz as f, but I think that's too high. I go an octave lower if not more.
 
Last edited:

Winston OBoogie

Well-known member
Joined
Jun 3, 2004
Messages
2,466
Location
UK.
Missed your last post. Cool. In that case, no worries.

To finish off:

Output impedance of a cascode is (mu^2)/gm in // with RLoad
Just like a pentode, R Load ends up being your output impedance.
With a choke, you don't have an R Load and you end up needing a fair chunk of inductance for your low end.

That's why a V72 needs a choke in the 500H range to load a pentode.

In your scenario, I don't know the details of the 6SK17 ... Gm of a 2sK170BL is (I think) in the 30mS area without degeneration but you have degeneration so... working out your output impedance would require a sit down and looking at the data sheets etc. 🤪


Anyway, that's moot now since you'e going with a resistor and 2 channels 🙂

Edit: I couldn't remember the formula for Gm of a J-Fet so found it online:
 

Attachments

  • Gm of J-Fet.png
    Gm of J-Fet.png
    9.5 KB · Views: 25
Last edited:

solkatten

Well-known member
Joined
Apr 3, 2021
Messages
121
Location
sweden
Missed your last post. Cool. In that case, no worries.

To finish off:

Output impedance of a cascode is (mu^2)/gm in // with RLoad
Just like a pentode, R Load ends up being your output impedance.
With a choke, you don't have an R Load and you end up needing a fair chunk of inductance for your low end.

That's why a V72 needs a choke in the 500H range to load a pentode.

In your scenario, I don't know the details of the 6SK17 ... Gm of a 2sK170BL is (I think) in the 30mS area without degeneration but you have degeneration so... working out your output impedance would require a sit down and looking at the data sheets etc. 🤪


Anyway, that's moot now since you'e going with a resistor and 2 channels 🙂

Edit: I couldn't remember the formula for Gm of a J-Fet so found it online:
And Thank you very much Mr Winston and Mr Abbey for good edvice and feedback 👍👍😀
 

solkatten

Well-known member
Joined
Apr 3, 2021
Messages
121
Location
sweden
Missed your last post. Cool. In that case, no worries.

To finish off:

Output impedance of a cascode is (mu^2)/gm in // with RLoad
Just like a pentode, R Load ends up being your output impedance.
With a choke, you don't have an R Load and you end up needing a fair chunk of inductance for your low end.

That's why a V72 needs a choke in the 500H range to load a pentode.

In your scenario, I don't know the details of the 6SK17 ... Gm of a 2sK170BL is (I think) in the 30mS area without degeneration but you have degeneration so... working out your output impedance would require a sit down and looking at the data sheets etc. 🤪


Anyway, that's moot now since you'e going with a resistor and 2 channels 🙂

Edit: I couldn't remember the formula for Gm of a J-Fet so found it onlin
Carnhill has a 400H 5ma at a good price
 

solkatten

Well-known member
Joined
Apr 3, 2021
Messages
121
Location
sweden
Maybe a more optimized version. The IPT is 1:10. The cascade is choke loaded for high headroom. The line stage has a SE trafo. I got a LL1671se 35ma. Best choice would be LL1660 18ma. A lundahl trafo is cheaper than a big ass choke and cap + parafeed trafo.Plus this version got Hifi credentials because of no electrolyts😜
 

Attachments

  • 5E3727D0-863A-4D4C-9DB7-5C9DF079DAAC.jpeg
    5E3727D0-863A-4D4C-9DB7-5C9DF079DAAC.jpeg
    291.2 KB · Views: 22
Top