Very simple tube preamp

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

NewYorkDave

Well-known member
Joined
Jun 4, 2004
Messages
4,378
Location
New York (Hudson Valley)
Just an idea... It looks good on the f**king lying piece-of-shit simulator, but it's going to be a while before I have a chance to try it out in real life. I'm posting it here in case some DIYer with more free time than I have wants an easy new circuit with which to play around.

verysimplepreamp.gif
 
hi NewYorkDave,

is that shunt-regulated-push-pull thingy? I made a srpp pre-amp using 12au7. less gain, but runs into 2k4 instead of 10k. Sounds fantastic! I built the hf section of passive helios eq(@10kHz) and used srpp stage as gain make-up. Still using it today. Fabulous "air" on vocals. So tight! perhaps because of no feedback?

Blah!


Andy P
 
No, it's just a common-cathode stage direct-coupled to a cathode follower, with about 16dB of negative voltage feedback. There's nothing shunt-regulated or push-pull about it.

The gain figure quoted (40dB) is net gain, including the stepup of the input transformer. The gain of the tube circuit alone is 20dB with feedback, 36dB open-loop. A higher-ratio input transformer (say, 1:15) will give correspondingly higher net gain.
 
That looks simple enough even for me to understand and build. :wink:

The RCA BC2B is also a one-tube design. This reminds me of it for some reason (similar fixed gain, simple design, etc.)

Could you do a 600:600 output trafo for balanced out?
 
This thing won't be happy driving loads under 10K. A typical 600:600 transformer may have too low of a primary inductance, even if the load on the secondary is 10K or higher. Try it if you want, but measure your frequency response and distortion, especially at low frequencies. Or you could use a 10K:10K transformer and make sure you always run it into 10K or higher loads... or you could use a 10K:600 and get 12dB less net gain, which you could make up by using a higher ratio input transformer. I would not reduce the feedback to increase the gain. As it is, there's only 16dB of feedback; you generally aim to use at least 15-20dB.
 
Don't forget Jensen Transformers site.

www.jensentransformers.com

They have a simple pre too, cascode to SRPP. Sounds different to my ear than what Dave designed here. I want the BIG box of Crayola crayons please....

http://www.jensentransformers.com/as/as054.pdf

Good work Dave! :thumb:
 
Consul's mention of the BC-2B preamp got me thinking... so here's another variation of the cheap 'n cheerful single-tube preamp concept:

cheapbastardpreamp.gif


As you can see, it's vaguely similar in concept to the BC-2B, but using a cheap tube and a standard non-gapped output transformer. Net gain is about 40dB. The electronic gain is 30dB, and there's 16dB of feedback. The maximum output figure is into a 600-ohm load at 1.5% THD (well, 1.5% according to the f**king lying-assed simulator).

One day, I'll find the time to actually breadboard this crap and see how it really performs.
 
Hey Dave,

Regarding the first circuit, how do I know what the reflected impedance is at the primary without there being a grid resistor?

Thanks.
 
> what the reflected impedance is at the primary without there being a grid resistor?

Why do you care, as long as it is much-much higher than the microphone impedance?

It could be 5K or 50K. With 200Ω mikes, the difference is an inaudible 0.3dB. Even 1dB or 2dB of loading isn't a big deal.

Actually, at the low end it is the transformer inductance. In the midrange, it is iron-loss. In the treble it is winding capacitance, grid capacitance, and winding leakage inductance. If the transformer is any good (for the purpose), the transformer impedance will be much-much higher than rated impedance over the rated bandwidth, even including some reasonable amount of stray and grid capacitance.

Yes, we often find resistors and even capacitors on the transformer. If that turns you on, or gives a better graph, or sounds better, do it. But note how many classic designs used just a naked tranny and grid. Loading secondaries unnecessarily has been known since the 1930s, but did not become a fad until the late 1970s. And in a "very simple" amp, it goes against the concept.
 
Actually, at the low end it is the transformer inductance. In the midrange, it is iron-loss. In the treble it is winding capacitance, grid capacitance, and winding leakage inductance.

That's the to-day most compact description of audio transformer characteristics..

:grin:

Jakob E.
 
[quote author="PRR"]> Why do you care, as long as it is much-much higher than the microphone impedance?

[/quote]

I'm not concerned from a performance standpoint -- I'm just trying to figure out how to calculate this stuff, or at least have an intuitive sense of what's going on, which I now have thanks to your eloquent four-liner.

[quote author="PRR"]> Actually, at the low end it is the transformer inductance. In the midrange, it is iron-loss. In the treble it is winding capacitance, grid capacitance, and winding leakage inductance. If the transformer is any good (for the purpose), the transformer impedance will be much-much higher than rated impedance over the rated bandwidth, even including some reasonable amount of stray and grid capacitance.
[/quote]

Thanks PRR!
 
I have proto'd someting similar to both those schems, and I think you will find the RC coupling of the feedback in the 2nd schem can produce a low freq response peak, depends on the values. You can twiddle the cap values to decrease it.

There was a similar schem used on a popular Dynaco preamp, don't remember the number, but I do recall an article in Glass Audio about fixes for this issue.

In the first schem you could take the feedback from the output side of C3 and eliminate C2, it will work fine.
 
[quote author="Scott S."]In the first schem you could take the feedback from the output side of C3 and eliminate C2, it will work fine.[/quote]

Hi Scott, and welcome to the Lab. What you describe will work fine if the load is known, fixed and non-reactive. If the load varies, or is reactive, then the frequency and phase shift characteristics of the feedback loop become variable, which is generally not a good thing. (On the other hand, it is common to include a reactive component--the output transformer--in the feedback loop in power amp designs, but the designer is usually working with a particular transformer, with known characteristics, in mind).

As drawn, the 1uF coupling cap and 100K feedback resistor are flat in the audio band; but like you say, you could use a smaller coupling cap to reduce the feedback at low frequencies and give the amplifier a low-frequency rise. One must be careful of excessive phase shift, of course. Even though it's kinda "cheating", stuff like that is sometimes done to compensate for, say, a cheap input transformer with poor bass response. And of course, it's normal practice in equalized amplifiers such as RIAA preamps.
 
I can see two issues with the design. First, there's no termination on the input transformer, and most transformers these days are designed to be used with a specific termination. On the Jensen JT-115K-E, for example, it's 150k. If you don't use the termination, the tranny's likely to ring.

Second, your signal-to-noise is going to suffer. You're adding about 9k to the impedance feeding the tube's grid, without any corresponding voltage gain. So instead of an equivalent source Z of about 15k with a 150 ohm microphone, it's now about 24k, and your signal-to-noise will be decreased by about 1.8dB. (Actually somewhat less, because there'll be a bit of noise from the tube, but not much.)

Peace,
Paul
 
Actually, what I had in mind was a cheap 'n cheerful way to make use of old 600/150:50K input transformers. These were generally designed to run into an unterminated grid. It's well-known (or should be) that the termination needs of different model transformers are not the same across the board, so it's implicit that a constructor would terminate as needed.

Don't forget that the OSI of the tube is rather high. I don't expect much, if any, degradation going from 50K to 59K source Z.

Who would use a $60-$100 Jensen on a $20 preamp, anyway? :wink:
 
[quote author="NewYorkDave"]Actually, what I had in mind was a cheap 'n cheerful way to make use of old 600/150:50K input transformers. These were generally designed to run into an unterminated grid. It's well-known (or should be) that the termination needs of different model transformers are not the same across the board, so it's implicit that a constructor would terminate as needed.

Don't forget that the OSI of the tube is rather high. I don't expect much, if any, degradation going from 50K to 59K source Z.

Who would use a $60-$100 Jensen on a $20 preamp, anyway? :wink:[/quote]

Well, I would, but that's me. With a 50k-secondary, you're right; the additional 9k would make little difference.

The original design, though, showed a 1:10 transformer, and on that there'd definitely be a difference.

Peace,
Paul
 
Points well taken. I should have marked the transformer as "600CT:50K" or "150/150:50K" instead of "1:10" to avoid confusion. The 1:10 refers to the ratio of the full primary winding to the secondary. With the usual 200-ohm (nominal) mic, you'd only use half the primary winding (or, in the case of dual primaries, both in parallel instead of in series).

Of the two schematics I posted in this thread, I like the second one much better anyway.

It's good to see you here, by the way. I enjoyed your magazine series on building a small mixer (although I never managed to get all the parts of the series).
 

Latest posts

Back
Top