Steel preamp

Lynxtrap

Tele-Afflicted
Joined
Feb 29, 2016
Posts
1,550
Location
EU
A while ago I started a thread on making an amp for pedal steel for a friend. The "order" got downsized to a preamp for plugging into a PA or a separate power amp.

He had become interested in 6SL7 tubes and I thought why not, I've never worked with those. He also likes an Ampeg B-15 type bass preamp that I've built previously. So I drew something loosely based on an Ampeg, added a Framus style mid control to make it easier to get some Fender "scoop" from the James, and a cathode follower for low impedance output.

In other preamps I've built I have put a center biased gain stage prior to the CF to kind of emulate a PI or an output stage, so I did that here too since I had an unused 12AU7 half. I guess it's not technically necessary, but it seems to add something to the sound when running straight into a PA or recording interface.

This thing is meant to stay clean at most of the volume pots sweep, and if it turns out it has too much gain I'll have to make adjustments.


This is the first draft, all comments are welcome!




Steel Pre schematic_rev1.png


Steel Pre layout_rev1.png
 

Lynxtrap

Tele-Afflicted
Joined
Feb 29, 2016
Posts
1,550
Location
EU
I've kind of forgotten to update this thread, but I got life in this amp yesterday. I have done some modifications to the circuit, updated schematic below.

I'm still trying to sort out some noise issues, mostly a static kind of buzz that seems to stem from somewhere around V1. It is affected by the tone controls and the volume.

But the most puzzling thing right now is a very strange voltage reading at the plates of V2A, the first half of the 12AU7.

Supply voltage is around 310V, but I get a constant reading of about 84V on the plate of V2A! The plate resistor measures 100K, and the cathode resistor measures 1.5K. But somehow a whopping 220V seems to be dropped across the plate resistor!

I got the same reading with an old 12AU7 yesterday and thought there was something funny with the tube. I put a 12AT7 in V2 and got a reading of about 135V, still low but at least higher.

Today I found a brand new 12AU7 in my stash and put it in, and I get that same 84V reading at V2A, while V2B is around 130V (V2B cathode is at 4.5V).

Any clues to what might be going on here?

Steel Pre schematic_rev3.png
 

chas.wahl

Tele-Holic
Joined
Dec 8, 2012
Posts
596
Location
NYC
Since you started with a 6SL7, why not use a 6SN7 rather than 12AU7 in the second position? Seems to me that's what Ampeg would have done.
 

andrewRneumann

Friend of Leo's
Joined
Mar 22, 2020
Posts
2,153
Location
Cincinnati, OH, USA
That’s an interesting output stage. Did you find inspiration from an existing design or come up with that on your own? I would have expected the standard cathode follower there, but I’m willing to entertain your idea. Have you run any headroom or impedance calculations on that?
 

Lynxtrap

Tele-Afflicted
Joined
Feb 29, 2016
Posts
1,550
Location
EU
Looks like an AU7 being an AU7. I don’t think there’s anything wrong with that stage. You could change the plate and/or cathode resistors to get a more reasonable load line and bias point.

You mean that 82V on the plates from a 310V supply is normal for a 12AU7? How can that be?

That’s an interesting output stage. Did you find inspiration from an existing design or come up with that on your own? I would have expected the standard cathode follower there, but I’m willing to entertain your idea. Have you run any headroom or impedance calculations on that?

I found it in the effects loop of an Engl amp and used it for a couple of preamps. When I recently bought Blencowes preamp book I found that he also recommends it.

I used to have a 1M resistor to ground from the output jack, but Merlin had replaced it with a 10k volume pot so I put a 15k resistor there instead.

It has advantages. The signal is about all that is needed for a line out, there is no DC to speak of on the cathode, and you could even use it as a plate coupled gainstage while tapping a line signal from the cathode if used as an effect loop. The .1uF cap is basically AC ground on the plate end.
 

2L man

Tele-Afflicted
Joined
Nov 23, 2020
Posts
1,789
Age
62
Location
Finland
You mean that 82V on the plates from a 310V supply is normal for a 12AU7? How can that be?
Thats what comes using that bias, Operating Point and tube. To "lift" OP you have to change anode resistor value smaller and often tune Bias changing cathode resistor value.

Open Universal Loadline Calculator and bookmark it because it is best what internet offer for tube amp designers :) Select tube and for pre amp stages tick Load-box to resistive. Input anode resistanc, operative voltage and OP and control grid voltage appear to box. Now you can calculate cathode resistor value and often it is off of standard values. You can see the effect on loadline when you change parameters and you can test both standard cathode resistance values above and below tuning anode current slightly.

Input headroom and and distortions appear. You can calculate the stage gain dividing headroom using grid line sweep. Use cursor to point grid lines. Often they don't match headroom so you must estimate one decimal which is enough.
 

Lynxtrap

Tele-Afflicted
Joined
Feb 29, 2016
Posts
1,550
Location
EU
Thanks! I should get around to try and master the art of loadlines. I understand them when I see them but still have some trouble calculating them.

In the meantime, I decided to cheat 😉

I found this (below) in a 12AU7 datasheet, and it looks like 3.3-4.4k on the cathode might work with 300V and a 100k plate resistor.

Still amazed by that enormous voltage drop though, an eye-opener for sure. The tube must draw a lot of current for that to happen?

I always assumed that a 12AU7 would be happy with about the same conditions as a 12AX7. I've never actually worked with them before.

12AU7 amplifier.jpg
 

andrewRneumann

Friend of Leo's
Joined
Mar 22, 2020
Posts
2,153
Location
Cincinnati, OH, USA
Thanks! I should get around to try and master the art of loadlines. I understand them when I see them but still have some trouble calculating them.

In the meantime, I decided to cheat 😉

I found this (below) in a 12AU7 datasheet, and it looks like 3.3-4.4k on the cathode might work with 300V and a 100k plate resistor.

Still amazed by that enormous voltage drop though, an eye-opener for sure. The tube must draw a lot of current for that to happen?

I always assumed that a 12AU7 would be happy with about the same conditions as a 12AX7. I've never actually worked with them before.

View attachment 1038958

Yeah, they are different tubes with different internal designs. One of the big differences is the plate resistance. If you compare them, you’ll see that the U has about 1/8th the plate resistance of the X. This doesn’t mean you’ll have 8 times the bias current, but it does give you a feel for whether a tube will draw more or less current in general. The fact that we use cathode bias on all these preamp stages makes the differences between all the 12A*7 tubes less obvious and is a big reason you can usually get away with swapping without changing resistors.

I don’t remember that output stage from Blencowe, but thanks for the reminder. I’ll go back and look for it.
 

Lynxtrap

Tele-Afflicted
Joined
Feb 29, 2016
Posts
1,550
Location
EU
Yeah, they are different tubes with different internal designs. One of the big differences is the plate resistance. If you compare them, you’ll see that the U has about 1/8th the plate resistance of the X. This doesn’t mean you’ll have 8 times the bias current, but it does give you a feel for whether a tube will draw more or less current in general. The fact that we use cathode bias on all these preamp stages makes the differences between all the 12A*7 tubes less obvious and is a big reason you can usually get away with swapping without changing resistors.

I don’t remember that output stage from Blencowe, but thanks for the reminder. I’ll go back and look for it.

Thanks, learning something every day, one of the great things with a forum like this!

You will find that Blencowe does not use the cap across the plate resistor, it would of course not work if used simultaneously as a gainstage. Shorting AC on the plate seems to give a little more signal from the cathode follower. Perhaps it works like an "inverted" cathode bypass cap?

I have previously used a 12AX7 for that duty, I'm starting to wonder if I should adjust the resistors there as well considering how the AU7 turned out to work. I've got around 4.5DC on the cathode here compared to the usual 1-1.5 with an AX7 under the same conditions.
 

andrewRneumann

Friend of Leo's
Joined
Mar 22, 2020
Posts
2,153
Location
Cincinnati, OH, USA
Thanks, learning something every day, one of the great things with a forum like this!

You will find that Blencowe does not use the cap across the plate resistor, it would of course not work if used simultaneously as a gainstage. Shorting AC on the plate seems to give a little more signal from the cathode follower. Perhaps it works like an "inverted" cathode bypass cap?

I have previously used a 12AX7 for that duty, I'm starting to wonder if I should adjust the resistors there as well considering how the AU7 turned out to work. I've got around 4.5DC on the cathode here compared to the usual 1-1.5 with an AX7 under the same conditions.

What’s the maximum signal you are targeting at the line output? The cap across the plate resistor makes it interesting. I think you might be losing a lot of input headroom by doing so. I have to scratch my head a few more times on that one.
 

Lowerleftcoast

Poster Extraordinaire
Joined
Dec 30, 2019
Posts
6,024
Location
california
The cap across the plate resistor makes it interesting. I think you might be losing a lot of input headroom by doing so.
I am trying to wrap my head around this as well. Most CF examples do not have an anode resistor. When a plate resistor is included, why would there be a difference when adding a cap in parallel?

Different questions.
Considering the B+ in this circuit, why would this CF want or need a plate resistor?

Why not use two series resistors, Rb and Rk, on the cathode? Signal out from the junction of Rb and Rk would make it less likely to oscillate than taking the signal directly from the cathode, according to Blencowe.
 

andrewRneumann

Friend of Leo's
Joined
Mar 22, 2020
Posts
2,153
Location
Cincinnati, OH, USA
I am trying to wrap my head around this as well. Most CF examples do not have an anode resistor. When a plate resistor is included, why would there be a difference when adding a cap in parallel?

Different questions.
Considering the B+ in this circuit, why would this CF want or need a plate resistor?

Why not use two series resistors, Rb and Rk, on the cathode? Signal out from the junction of Rb and Rk would make it less likely to oscillate than taking the signal directly from the cathode, according to Blencowe.

@Lowerleftcoast did you see the stage in the effects loop chapter of Blencowe? That’s what the OP is referencing, not the CF chapter. Just wanting to make sure we’re all looking at the same thing.
 

andrewRneumann

Friend of Leo's
Joined
Mar 22, 2020
Posts
2,153
Location
Cincinnati, OH, USA
When a plate resistor is included, why would there be a difference when adding a cap in parallel?

It doesn’t affect the bias point which is found in the usual way. What it does is hold the plate voltage (reference ground) at a constant 130V (based on OP previous measurements). This is similar to a traditional CF where the plate voltage is constant at B+. The issue I see with this is that instead of the huge amount of input headroom seen in a traditional cathode follower, this baby is going to clip easily, especially on the down going swings of signals. This will necessitate large signal attenuation prior to the grid. It will work, but I’m not convinced it’s the best choice. I’ve done some back of napkin numbers on this but I’m not ready to post them. More to come.
 

Lowerleftcoast

Poster Extraordinaire
Joined
Dec 30, 2019
Posts
6,024
Location
california
did you see the stage in the effects loop chapter of Blencowe? That’s what the OP is referencing, not the CF chapter. Just wanting to make sure we’re all looking at the same thing.
Sorry, no I didn't.
The issue I see with this
That is the point of my *other* questions. It seems this circuit will not optimize the CF. (EDIT: Or the signal from the CF.)
 

Lynxtrap

Tele-Afflicted
Joined
Feb 29, 2016
Posts
1,550
Location
EU
Looking forward to the napkin numbers! I'm not able to give any in-depht technical or mathematical details on it, I've just found that it works fine in previous applications.

As to output signal, 1V is considered line voltage and as far as I can see that is about all that is needed for any reasonable use for a preamp.

As per the latest schematic above, there is no audible overdrive except for at the very highest volume settings (tested with a Tele).

I might try to lift the bypass cap to see what happens without it.
 
Last edited:

andrewRneumann

Friend of Leo's
Joined
Mar 22, 2020
Posts
2,153
Location
Cincinnati, OH, USA
@Lynxtrap could you provide me with accurate voltages (for this output stage only) for supply, plate, and cathode at idle please? Just measure voltage to chassis. I will generate some load lines and post them.

Wikipedia says professional audio line level is 1.735 Vpeak, so let's say that's our target. You want this stage to output at least that much clean(-ish). That's not very much, and Blencowe makes a lot of sense when he suggests the common CF, capable of outputting say 100V Vpeak, is overkill.

You have a lot of interstage attenuation throughout the amp, so it doesn't surprise me that it remains clean. But in my mind, you have your foot on the gas and the brake at the same time, which generates excess noise. I feel it is better to reduce the amplification of each stage appropriately before adding voltage divider attenuators, which add noise. Would you be happy if you could remove all those 220K grid stoppers and still have a clean amp? If the noise level doesn't bother you, than don't worry about it. It's your design and if you are happy with it, great!

You can increase the value of the 15K resistor. The reason Blencowe uses such a low value pot is that he doesn't want to increase the output impedance of the line out when the pot level is reduced. Since you don't have a pot there, all you need to do is keep the output circuit at 0VDC, so a 1M resistor is fine there. The coupling cap (a huge 10uF) can be reduced according to the standard high-pass filter equation.
 

andrewRneumann

Friend of Leo's
Joined
Mar 22, 2020
Posts
2,153
Location
Cincinnati, OH, USA
Ok, did some more analysis of the bypassed plate resistor. Here are the effects as far as I can tell.

Clean Gain: doubled from approx. 0.3 to approx. 0.6 at the cathode.
Max. Output swing at cathode (not clean): quadrupled from 8.4Vp-p to about 34.5Vp-p
Input clean headroom: Increased from 16.8Vp-p to 20Vp-p (unexpected result, but I think it’s correct)
Distortion: more… but I didn’t calculate numbers

Load lines below reflect a stage similar to yours, but not exactly the same.

1665619438887.jpeg
 

Lynxtrap

Tele-Afflicted
Joined
Feb 29, 2016
Posts
1,550
Location
EU
Thanks a lot @andrewRneumann !
I will take a closer look and provide you with the numbers when I have the time, hopefully sooner than later. Very interesting with your calculations on the output stage, it kind of confirms what I have experienced.

The goal is clean with as much sweep as possible on the volume pot. I might have used a 100k pot if I had one at home.

Minimum noise, absolutely!

The circuit would probably have been fine with a 12AU7 in the preamp as well, the 6SL7 was a wish of my friend I'm building it for. But if it turns out it doesn't work we will throw it out.

I'm open to any ideas!

As a sidenote, I happened to look at a Vox AC100 yesterday. V1A is a 12AU7 with 100k Rp, 1.5k Rk, 270V supply and 80V on the plate. Seems familiar...
 




Top