Hey Chris ,
Well I had a quick look.It’s nothing too exciting , not much headroom by the looks of things.
I really don’t like the meter being there , it only makes things worse.
A separate buffer is always preferable.

Here are some quick instinctive thoughts….(hopefully we’re talking about the same diagram here….) :

 R3 could be bigger
 Things like C8 C10 C2 C3 don’t really need to be there if it is made well. If the spacing is right , you won’t have oscillation problems.
 R5 should be a bit bigger say , 1.5 k or 1.8 k . R4 should equal R5.
 I wouldn’t load the circuit with the meter.
 P1 should be bigger…..maybe 500k.
 R6 could be a bigger….maybe 10k or even 47k.
;)

Thanks Sebatron,

Interesting thoughts...

Originally posted by Sebatron:

 Things like C8 C10 C2 C3 don’t really need to be there if it is made well. If the spacing is right , you won’t have oscillation problems.

I always wonder about extra decoupling caps, and the "safety in numbers" approach is often taken in analog and digital design. Is it really down to just a matter of layout, or do you have to take into account varying external conditions that the circuit/device may be connected to? And does this take into account the (possible)use of and external PSU?

And does the type/quality of the decoupling cap itself come into play? i.e. can a single cap take the place of a multiple cap array in all situations? I know a lot of designers must just bang in a bunch of caps because it's a "quick fix" to any instability problems, rather than spend more time in layout/design stages.

...and about the buffer for the meter driver, I've read a lot about the effects of loading/distortion that a passive VU meter can cause. Is this just due to the resistive/reactive loading at the output of a circuit, or is it due to the combination of a load consisting of a resistance/inductance/diode junctions? Does this cause a non-linear "ringing"? :confused:

Mark

--------------------------
"Oscillators don't, amplifiers do....."
Anon.

Is it really down to just a matter of layout, or do you have to take into account varying external conditions that the circuit/device may be connected to? And does this take into account the (possible)use of and external PSU?

A few factors Mark.
Layout is very relevant if the circuit is running high gain and high impedance.Where everything is channeled is important.
Point to point or P.C.B.?
The latter is worse for this disease.The tracks must be well thought out with good distance between ‘give’ and ‘take’ type tracks.Thicker for earth return loops and cathode supplies.
Oh….i forgot….avoid regulators for filament supply , beef up caps and RC network for a more reliable and organic rail.
Shielded cable where possible will also help with oscillations.

And does the type/quality of the decoupling cap itself come into play?
i.e. can a single cap take the place of a multiple cap array in all situations?

Not all situations obviously , but if layouts are right it’s ‘normal’ to be that way than to solve it by adding extra excrement …if you know what I mean.

I know a lot of designers must just bang in a bunch of caps because it's a "quick fix" to any instability problems, rather than spend more time in layout/design stages.

It should be a last resort , put it that way.

...and about the buffer for the meter driver, I've read a lot about the effects of loading/distortion that a passive VU meter can cause. Is this just due to the resistive/reactive loading at the output of a circuit, or is it due to the combination of a load consisting of a resistance/inductance/diode junctions? Does this cause a non-linear "ringing"?

All of the above , Yes. It’s a very ugly scene when the meter eats into your audio.
In my vmp-2000e VU , I use a completely separate buffer stage for the meter.I couldn’t settle for anything less .There is virtually no loading.

…oops… gotta get back to work…….
:w:

Hi Sebatron!

Thank you for your thoughts. The idea is not that I want to build this pre, instead it is a proposition to start discussing the pros and cons of various topologies.
You set the rules by saying only tubes, transformers and high voltages alowed, so lets talk about tube mic preamps : this dual SRPP, Jakob's G9, Pultec MB1, Tubetech MP1A and maybe other designs. I did not included the REDD47 and other pentode based pres because they use, er... pentodes [maybe next time]. Also the input transformer is a must because very few transformerless designs are very good [like Fred Forssell's].
So we have input transformers, triodes in various configurations and the option of output with or without transformers.
What you like or dislike, what would you make diferent and why.
What tricks have you used in your designs [if it is not a secret] to obtain a specific sound.

chrissugar

P.S. I know about the fact that the VU meter introduce some distortions but many of the old preamps had a meter like this and I read about the fact that it can produce some interesting [in some situations desirable] effect. So if I would use an unbuffered VU it would be for the sound effect produced and a switch can disable the VU [also the effect]. I think about the VU meter option as an aditional sound design tool like the input and output transformer and the variable feedback.

What you like or dislike, what would you make diferent and why.

I had a quick look over the tubetech and G9.

I think I like the Tubetech the most.Just straight off that is….although ECC83’s can be a bit rough.
Bit too much gain maybe.Quite a ‘valvey’ circuit…May have a tendancy to ‘honk’ or ‘squelch’.

The G9 seemed quite similar really.
It’s a pity about C2 , C3 and C13 in the G9.
Maybe C2 could be larger?
Also , maybe some grid resistors to stablelize things , small ones like 1k in series with grid always helps …
…..Actually apon further inspection I don’t like C4/ C5 and C8/C9 area at all….or even C13/C10.
The Pot could be bigger too.
Try 100k ( or even 500k ) pot , elimante C10 and increase R45 to 22k …maybe up to 47k.
Ummm…not sure.Try it and let me know.

It also seems to be a sensitive circuit , but let down by the ECC82….a low gain valve.
You see I use ECC81 ( 12AT7) which is between the two …Less gain than 12AX7( ECC83) and but more than 12AU7 (ECC82).It suits me the best….horses for courses.

They’re both quite nice designs…….variable feedback is always handy.

What tricks have you used in your designs [if it is not a secret] to obtain a specific sound.

Well ,, as I have stated , minimalism I feel is quite important.Avoiding too many solid state components in the power supply is a challenge (..regulators , zener diodes ).A lot can be achieved with careful RC networks , but you have to start with quite a high rail.Avoid too many audio caps in series with the path….and always avoid handling your food when handling solder.
:s:

Hey Seb,

Originally posted by Sebatron:
avoid regulators for filament supply

Originally posted by Sebatron:
Avoiding too many solid state components in the power supply is a challenge (..regulators , zener diodes )

Why don't you like regulators?

Peace,
Al.

Greetings!

I designed that tube preamp a few years ago for some nice folks at another forum.

The design is minimal, no fuss and easy to build. It's got _plenty_ of headroom. Dunno what your calculator is saing, but in real life it doesn't clip. Don't like meters myself (actually hate 'em), but someone requested that. Yes, a buffer would be better, but not more simple (a key point for the folks who wanted the design). Besides there's a ton of gear out there with unbuffered meters, like Neve. Don't hear anyone complain about that..

I don't know why you'd wanna change those component values, since they pre is just peachy as it is. Especially V2! Listen, don't look.

So how does it sound? Well, that's always personal. One guy thought it beat his Manley/Avalon stuff.. Funny, you should mention the TubeTech. That does not sound good to me. Brittle and no dynamics. Something is _wrong_ with that preamp, me thinks.

Schematics don't say much, do they?

I too avoid Vregs like the plague. RC/LC is the way to go, imo.

Any Sebatrons in Denmark/Scandinavia yet?

I’m sorry Shalimar.I totally respect your art.
The changes I stated were merely gut feelings , the circuit still rocks!! And always will.
Good to see you down here. I connect with the attitude.Yes, there are a few Sebs up North I believe.
Yeah , someone should shove an ecc81 in the Tube Tech to calm it down. :)

Why don't you like regulators?

Don’t get me wrong ….Regulators obviously have a place in audio electronics but can and should be avoided when designing.
When a circuit is ‘stressed’ the most likely components to fail initially are the Solid State devices.
Amongst these there are of course Regulators and Zener Diodes.While they perform their function beautifully , they can easily be avoided with careful resistive-capacitive or resistive-inductive design with comparable results.

This has a multitude of advantages.
Amongst these:
 Durability to mains sagging…
 Durability to mains spiking…
 Efiiciency in power consumtion…
 Possible tendancy to dynamically soak…organically to high current demand…
 Variability depending on component tolerance…like valve current demand for e.g

Now to get the same regulation figures through these basic components (caps,inductors etc. ) may require a long trial and error period…also an extensively thought out EARTHING NETWORK.
BTW , there are better ways than ‘star earthing’ tracks.

Let me also remind you that the majority ( if not all ) audio that was built pre 1965 ( I think ) totally lacked Solid State based regulation.
It is very well thought out gear that sounds as it looks….earthly and organic. :d:

....blah...blah....blah

Hey, don't be sorry! If I had a Seb, I'd probably mod it too... haha.

Vregs blow. I'm sure all gear before 1965 does not use 'em. I've seen bits'n'peices with transistor/zener regulation from that period.

Sebatron shorted and said;

"....blah...blah....blah"

I agree!

[quote=Sebatron]

The G9 seemed quite similar really.
It’s a pity about C2 , C3 and C13 in the G9.
Maybe C2 could be larger?
Also , maybe some grid resistors to stablelize things , small ones like 1k in series with grid always helps …
…..Actually apon further inspection I don’t like C4/ C5 and C8/C9 area at all….or even C13/C10.
The Pot could be bigger too.
Try 100k ( or even 500k ) pot , elimante C10 and increase R45 to 22k …maybe up to 47k.
Ummm…not sure.Try it and let me know.

It also seems to be a sensitive circuit , but let down by the ECC82….a low gain valve.
You see I use ECC81 ( 12AT7) which is between the two …Less gain than 12AX7( ECC83) and but more than 12AU7 (ECC82).It suits me the best….horses for courses.

They’re both quite nice designs…….variable feedback is always handy.

Very interesting Seb.

I have one of your VMP400E's and, I have also built the G9. Could you please explain briefly WHY you recommend these changes. What effect would they have, OR, why are they not necessarily good ideas these things you mention you dont like. Im keen to get experimenting. Here is the link to the G9 info again for you :

http://www.gyraf.dk/gy_pd/g9/g9pd.htm

Thanks! :D

Too many Caps in series is something to avoid if possible. Having C2 too small could roll off too much bottom end. The highpass filter could somehow be earlier on to increase the headroom of the preceeding stage when engaged.C10 could be eliminated because there is no D.C on the pot already ,, but you’d have to increase R45 to de-couple the loading of the pot onto the valve.
Some of these changes would probably yield differences that are quite minimal really , probably wouldn’t be detected at all.

Cool, thanx seb. Im looking to learn more about how and why than actually trying to mod it. Sounds fine as is. thanks again. 8-)