Critique my preamp design

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Attached is my first design for a bass guitar tube preamp. The goals of this design are:

-Boost a passive bass signal (~0.01V) to line level
-Tone shaping via James tone stack
-Tone shaping via active mid cut/boost a la ampeg svt
-Lightly compress and distort the signal for a bit of coloration
-Drive a power amp input
-Drive a long balanced cable length to sound board

I'm very open to constructive criticism on my design and advice on how you might accomplish these goals.

Thanks in advance!
 

Attachments

  • Bass Preamp.pdf
    77.7 KB · Views: 262
What is your B+ going to be?? Some design decisions depend on this.
The first 12AX7 will probably be biased around 1.5V. A lot of basses, even passives, put out peaks way above that. I measured peaks up 5V on one of my active basses. So you might consider a second padded input (-15dB for example), so you can leave the drive/distortion completely up to the second stage.
For reference, see pic below, just ignore the second stage and other and on:
inputstage.png

One other thing is the master volume after the 4th 12AX7. The 1M volume put is in parallel with the 30k resistor. Why not use a 50k pot instead and reduce the noise produced by the 1M piece of carbon and reduce the part count.

It also seems you amplify the signal a substancial amount a couple of times and the reducing it again by means of a voltage divider. Signal to noise ratio will suffer severely.
 
Last edited:
What is your B+ going to be?? Some design decisions depend on this.

I'm planning on 300V.

One other thing is the master volume after the 4th 12AX7. The 1M volume put is in parallel with the 30k resistor. Why not use a 50k pot instead and reduce the noise produced by the 1M piece of carbon and reduce the part count.

Good idea, I like it.

It also seems you amplify the signal a substancial amount a couple of times and the reducing it again by means of a voltage divider. Signal to noise ratio will suffer severely.

This is the area I'm most unsure about. How do I benefit from the massive gain of tube amplifiers while still maintaining line level output? I didn't want to give the dc coupled cathode follower a tiny signal, assuming I would not benefit as much from it's inherent compression and non-linear characteristics. Should I decrease the beginning gain stage, or only have one voltage divider, or both? Or completely redesign :confused:
 
To be frank :

1) it looks way too much like an Ampeg SVT preamp.
As in 95% alike.

2) it looks like you didn't want to straight copy an Ampeg SVT preamp, so, without *really* varying the design, you modified each stage, one by one, replacing it with either a home brewed stage or something copy and pasted from somewhere else.

Either way it does not "better" the original design but potentially worsens it, and adds errors.

As example:

a) the tone controls.
The treble boost and cut have been raised one octave relative to a classic Ampeg stack.

This will work properly in a Hi Fi preamp, where you have cymbals and such, reproduced with tweeters, but in a Bass amp with Bass speakers it will be perceived as "weak"

The Bass range has also been raised roughly 1 octave, it will be perceived as having too much midbass, maybe "muddy" or lacking depth.

The full adjustment range has also been reduced by some 12/15 dB, so it will be perceived as "doing little".

I suggest you download Duncan's "TSC" , the tone stack checker, and trace it both with original Ampeg values and with the new ones, you can save and compare memories.

Believe me that Ampeg values have been the benchmark for over 60 years for some reason.

3) the way the midrange control was implemented, it works almost as the original when attenuating, but on boost, instead of the original *very* noticeable 20dB, it will have very anemic 3 or 4 dB, almost nothing.

4) you copied wrong the balanced line driver, the top unity gain follower is fine, but the bottom stage is open loop (ay, gain 1000000 at low frequencies) and its output will probably be slammed against a power rail juust because of uncompensated offset.

Personally I would straight copy an SVT preamp (in fact I do exactly that on customer demand) , play with it, get to know it well, and then, based on that experience, start tweaking it to better some aspect where I think it can be done.

As of the line driver, it's almost there, but in the bottom Op Amp, connect the 10K resistor from the output to the -in , and ground the +in.

And for safety, add a 10K to 22K resistor from Master pot wiper to to +in, and from there reverse biased diodes to +15 and -15 rails, because with that tube preamp, if you set the master to 10 and play loud, you *may* surpass 15 V peak either way and damage the TL072 input.

EDIT: when posting I read:
the dc coupled cathode follower a tiny signal, assuming I would not benefit as much from it's inherent compression and non-linear characteristics.

Don't worry about that.
You will have no compression and very little (if any) non linearity in that stage, unless you hit it with a sledgehammer (signal wise) and at the same tine load it with a too low impedance signal .... and nothing of that will happen.
 
Last edited:
2) it looks like you didn't want to straight copy an Ampeg SVT preamp, so, without *really* varying the design, you modified each stage, one by one, replacing it with either a home brewed stage or something copy and pasted from somewhere else.

Either way it does not "better" the original design but potentially worsens it, and adds errors.

The only thing that I took from ampeg was the mid control method...I haven't seen any other maker do an active one like this. Everything else was following the book Designing Tube Preamps for Guitar and Bass by Merlin Blencowe. When I compare this to the 1972 svt pre, I don't see a ton of similarities. Some things are bound to look similar, as they are in almost all guitar/bass applications, but I wasn't trying to rip off the ampeg svt.

Good catch on the op amp section...I was kinda rushing at that point.

I'll have to take another look at the James tone stack. I did use the Duncan Tone Stack calculator and it seemed to have a good looking response to me.

I don't quite follow why my implementation of the mid control doesn't match the same gain capabilities as the svt. Can you explain that?

Thanks for your feedback, I am excited to wire this up.
 
The only thing that I took from ampeg was the mid control method...I haven't seen any other maker do an active one like this. Everything else was following the book Designing Tube Preamps for Guitar and Bass by Merlin Blencowe.
1) please take what I answer with no offense (I'm sure you didn't, but it's better keeping things as transparent as possible) :)

2) you also took the block diagram, to call it some way, but replacing what's inside each block with a functional equivalent, that's what I mentioned.

The end result is still very recognizable and amounts to about the same.
Didn't know where you had pulled stuff from, and if it's Merlin's book, fine, it's a good resource.
Never read his book but saw his contributions in various Forums and he's clearly a serious guy :)

When I compare this to the 1972 svt pre, I don't see a ton of similarities. Some things are bound to look similar, as they are in almost all guitar/bass applications, but I wasn't trying to rip off the ampeg svt.

Well, I already told you what I see, but to clear doubts, "an Ampeg SVT preamp" architecture (which is recognizably very different to, say, a Fender or a Marshall, to name 2 "big ones") can be described , in a *very* simplified form, as follows:

"a roughly 50X gain stage > passive fixed EQ (lossy, around -10X) > 1M volume pot > gain stage to recover Passive EQ losses (around 10X) > James tone control (lossy, around -10X) > active (3 tubes) midrange gain block. Gain=8X (56K/8K) > cathode follower > output."

*IF* you pull the lossy fixed EQ (Ultra Hi/Lo) *and* the gain stage which was added only to compensate that loss, the preamp stays exactly the same as before, simply "you didn't push any of those buttons".:)

"a roughly 50X gain stage > 1M volume pot > (you will need here some kind of buffer to replace the 10X gain stage) James tone control (lossy, around -10X) > active (3 tubes) midrange gain block. Gain=8X (56K/8K) > cathode follower > output."

The preamp won't sound, say, Fenderish or Marshallish but still *very* Ampegish .... because that's what it is.:)

If you wish, in other moment I can split Fender and Marshall typical preamps in blocks so you see both are very different to this one (and between them).
Now I don't want to mud what we are talking about.

Now to yours:
"a roughly 50X gain stage > 1M volume pot > (unity gain buffer ) James tone control (lossy, around -10X) > active (1 tubes) midrange gain block. Gain=(let's calculate it when analyzing this midrange control)[/] > cathode follower > output.[/I]"
As you see, it's very much the same.
So much so that on purpose I copied and pasted the Ampeg description .... and it still fits yours :eek:

What do I mean by that?:

I look at circuits as functional blocks, "what they do" , which by the way is excellent for quick troubleshooting.

Don't care that much for "parts" alone, even less so about brands, and talks about cap or resistor colours, composition or vintage make me sick , I *do* care (very much) about how systems (which are a combination of parts in a certain way) behave, measure and sound.

So if a gain block was made with a tube which carries a certain name, and now is replaced by another with a different name but which works roughly the same, clips at about the same level, has about the same gain, etc., for me it's functionally the same.
Good catch on the op amp section...I was kinda rushing at that point.
Fine, always happens. :D
When I post some schematic, I also ask others to double check.:eek:

I'll have to take another look at the James tone stack. I did use the Duncan Tone Stack calculator and it seemed to have a good looking response to me.
Sure it looked good.
In fact it's fine, but more useful in Hi Fi (or PA or Keyboard) amps like in a Bass one.

FWIW, Polytone amps use "Hi FI preamp" type tone controls , as I said way too much midbass and weak treble ... maybe that explains why those amps are loved by the Jazz cats and their archtop fat body guitars, using *fat* strings ;)

I don't quite follow why my implementation of the mid control doesn't match the same gain capabilities as the svt. Can you explain that?
Yes, of course.

Ampeg made a very high gain block using 3 tubes, then applied heavy feedback to lower that gain to only 8X, but they have a lot of gain in reserve if needed.

In tube amps, losing gain is *easy* (that explains passive tone and gain controls everywhere) but boosting gain is not, because to begin with there's very little, if any, "untapped " gain available.

To some (many?) this will sound like an insult, for me it's the highest compliment: that Ampeg gain stage is almost as good as an Op Amp :eek:

Its gain can be accurately predicted (notice I stated 8X while all others were "around something").
Gain is 1 + R35/(R29+R30)= 1 + 56K/8K=8X

Ampeg states that the midrange control has "+/-20dB range", at the resonant frequency, of course.

Let's see:

To get 20dB boost (10X gain) we need to parallel (R29+R30) (8K) with something 1/10 its value, say, around 800 ohms .

What do we have?= R34 (470 ohms) + the DC resistance of the coil.
Will those add up to roughly 800 ohms?
I bet they do ;) .... or quite close :p *

Now to the attenuation.
The attenuator is made out of R37 (47K) and R36 (6K8) .

We need something in parallel with R36 with impedance, at resonance, around 1/10 that or around 680 ohms.

Well, we have R39= 620 ohms (surprise :D) plus the DC resistance of the inductor, which I estimated before in around 100 to 200 ohms. Or not very far from that.

As you see, Ampeg design was *incredibly* accurate, clearly made by very good and experienced Engineers.
Not "something drawn on a napkin and tuned by throwing random parts at it"

Now you'll understand the main problem in yours: you have a single gain stage, run already *almost* full gain, there's *very* little extra.

So it will work properly (like in most classic tube circuits) attenuating, you put 470 ohms in parallel with 4K7, fine, similar to what Ampeg achieves.

But on boost, even if you skip the series resistor, you are putting the inductor DCR in parallel with the 1K5 cathode resistor of a 12AX7 .... you have at most 3 or 4 dB available to boost .... very far from the original 20dB.

So I took 1000 words to repeat what I had already said in 20 or 30, oh well :p

Hope it was somewhat useful ;)

Thanks for your feedback, I am excited to wire this up.

You're welcome :)

It will work, no doubt, but if measurements (or perceived sound) don't come up to what you expect, don't worry.

And please don't rebuild it using $300 capacitors made out of Platinum , Egyptian mummy papyrus and Moby Dick white whale oil, because that's not the point.

Magic is irrelevant, the point rests on hard work and knowing what you do.

Good luck :D
 
And just a thought. You have this tube preamp going, and then one lonesome TL072 driving a balanced out. That is going to require a whole extra split power supply just for the one chip. Maybe consider just biting the bullet and buying a nice transformer, which your tubes could easily drive.
 
1) please take what I answer with no offense (I'm sure you didn't, but it's better keeping things as transparent as possible) :)

No offense taken at all...I put my schematic up for critique. I admit I'm a noob, and expected more of a beating than I actually took. I appreciate your comments.

2) you also took the block diagram, to call it some way, but replacing what's inside each block with a functional equivalent, that's what I mentioned.

I'll briefly describe my design methodology. Working backwards:
I wanted to end with a line level output
I wanted to buffer the overall sound at the end and control it with a master gain
I wanted the distortion/compression character after the eq, thus the DC coupled cathode follower
The eq section comes after the initial gain stage for maximum effect and minimum noise

To be honest...this is one of about 5 topologies that I drew on a napkin (sticky notes actually).

Don't care that much for "parts" alone, even less so about brands, and talks about cap or resistor colours, composition or vintage make me sick , I *do* care (very much) about how systems (which are a combination of parts in a certain way) behave, measure and sound.

This is very much how I would like to operate. I'm more of an engineer than a cork sniffer.

But on boost, even if you skip the series resistor, you are putting the inductor DCR in parallel with the 1K5 cathode resistor of a 12AX7 .... you have at most 3 or 4 dB available to boost .... very far from the original 20dB.

I follow everything up to this point...I thought that I'm getting a pass band gain of about 30 at this part...and then a gain of ~50 for the resonant LC circuit when the mids knob is set to max (resistance of 0 ohms from the cathode). The mids knob shifts the circuit from an unbypassed grounded cathode to a bypassed grounded cathode with the LC circuit determining what frequencies get the extra boost from the bypassed arrangement. Curious to see if you follow my logic here or if I'm mistaken. :cool:
 
And just a thought. You have this tube preamp going, and then one lonesome TL072 driving a balanced out. That is going to require a whole extra split power supply just for the one chip. Maybe consider just biting the bullet and buying a nice transformer, which your tubes could easily drive.

Funny you mention that...I originally had in mind a Jensen 10k:600 output transformer, and in another thread it was mentioned I could go with op amps. Either way I'm giving up real estate for a transformer.
 
.I thought that I'm getting a pass band gain of about 30 at this part...and then a gain of ~50 for the resonant LC circuit when the mids knob is set to max (resistance of 0 ohms from the cathode). The mids knob shifts the circuit from an unbypassed grounded cathode to a bypassed grounded cathode with the LC circuit determining what frequencies get the extra boost from the bypassed arrangement. Curious to see if you follow my logic here or if I'm mistaken.

Your logic is good, so far you are seeing what I'm pointing at, but you are not SEEING it . :eek:

You calculated the gains exactly (or picked them from Merlin's book or a datasheey, it's the same, they are correct), you went *almost* all the way and now need to go the extra inch:

Yes, the stage gain is 30 with the pot on "5" (on a 0/10 scale) and rises to 50 when on "10" ....

But "50" is not 20 times more than "30"

It's merely 1.5X 30 , or between 3 or 4 dB more.

*Barely* audible , while 20dB gain is *screaming* .

Hope this makes sense, otherwise refresh dB scales, ear response, etc.
Read it and at some point it will "click" in place.

And thanks for your kind words :)
 
BIG thumbs up to LMFahey!

Regarding the opamp output and extra split supply. If you decide to keep this arrangement, you might consider adding an FX loop and buffered tuner output, both opamp driven off course, to really start making good use of the extra PSU section.
 
Your logic is good, so far you are seeing what I'm pointing at, but you are not SEEING it . :eek:

You calculated the gains exactly (or picked them from Merlin's book or a datasheey, it's the same, they are correct), you went *almost* all the way and now need to go the extra inch:

Yes, the stage gain is 30 with the pot on "5" (on a 0/10 scale) and rises to 50 when on "10" ....

But "50" is not 20 times more than "30"

It's merely 1.5X 30 , or between 3 or 4 dB more.

*Barely* audible , while 20dB gain is *screaming* .

Hope this makes sense, otherwise refresh dB scales, ear response, etc.
Read it and at some point it will "click" in place.

And thanks for your kind words :)

Dang, I get it now. I was making the mistake of using the input to the triode as the bench mark...i.e. 50x gain would give me a whopping 34 dB of eq boost (who would ever use that!?) but I should have been using the output as the bench mark.

Back to the drawing boards :eek:
 
Curious to hear what you all think on separate transformers for each supply. I often see in the hifi realm that different transformers for the B+ and heaters is king, I just wonder how much overkill that really is. And if I'm doing an opamp output, following the above reasoning would mean three transformers which is a bit ridiculous for such a humble little project.

On the topic of power supplies, is it a good idea to use a high voltage regulator for the B+, or am going overboard again?
 
Think about it. Why does one regulate a power supply? One wants a steady voltage under changing load. Now look at your circuit. It is all just a string of class A stages. That means the current is fairly steady. So what's to regulate?

Also, it's a guitar amp. It is not a precision circuit.

In hifi, the goal is to sound perfectly neutral. add no coloration. SO if the thing works, it ought to sound about like any other. SO in the hifi world, hype is king. They need to have things to claim, like separate transformers. They need to take some tiny aspect of electronics, then exaggerate it way beyond reality, so they can tell you all the things they did to eliminate it. Can you imagine a guitar amp boasting it is ruler flat all the way to 30kHz?

So, yes, I think separate transformers is overkill. Show me one commercial tube preamp that needed more than one power transformer to work well.
 
Attached is my first design for a bass guitar tube preamp. The goals of this design are:

-Boost a passive bass signal (~0.01V) to line level

You might want to double check that number, unless you're using much different pickups than I'm familiar with.

Can you tell us a bit about what styles of music you will be using this for? You've already seen my take in the other thread you linked, but I have no interest in loud rock music at all and made many of my choices based on getting nice "bedroom tone" and building something quiet enough for my particular recording needs. ;)
 
You might want to double check that number, unless you're using much different pickups than I'm familiar with.

Can you tell us a bit about what styles of music you will be using this for? You've already seen my take in the other thread you linked, but I have no interest in loud rock music at all and made many of my choices based on getting nice "bedroom tone" and building something quiet enough for my particular recording needs. ;)

It's difficult for me to measure my passive signal without an oscilloscope. My audio interface is indicating -35dB consistently, and -30dB on the peaks. My active bass is way hotter...anywhere from -24 to -5 dB.

The style of music will vary...nothing too heavy, but not lounge music, either. Styles include jazz, blues, and bordering into pop rock.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.