My-very-first-gadget (tm)

[PRR]

ladewd> 12VAC plug in wall transformer, then used an old 220VAC to 9V transformer and wired it backwards

How hot does that run?

Transformers have a voltage ratio, and your math is OK. But transformers also have a maximum voltage (for a given frequency). Feed them more voltage, and the iron saturates, large current flows, winding gets hot. There is some headroom above the rated voltage of a power transformer, but headroom is expensive so they only give you enough to cover likely line variations. Maybe 20%. You are running 33% over rated voltage ("9V" winding on 12V source). If both transformers are working cool, or just warm, OK. But if they run hot, you know why and what to do.

NewYorkDave> Increasing the feedback does reduce the output impedance, but there is an irreducible minimum load that a tube can drive to a given level regardless of the amount of feedback. This limit is imposed by the current-handling capability of the tube.

In any feedback amp, feedback changes the small-signal impedance but NOT the large-signal impedance. A 12AU7 is about a 6K source and should be loaded >12K. Wiring it as a cathode-follower gives 300Ω small-signal impedance; wrapping it with a 741 can give a 1Ω small-signal impedance, but the large-signal impedance is -still- 6K, best loaded >12K.

Working a small triode into 600Ω raises THD, though this can be "fixed" with NFB. But it is still a 6K large-signal source: load it in 600Ω and you get about 1/10 the voltage and power that you could get with an optimum load resistance. Since we sometimes need very-small power, a few mW, and they don't make very-small triodes (or not cheaper than the 1-watt 12A_7 series), this is often acceptable.

> 8.8K in series with a 100K pot. At minimum gain--with the pot shorted out--that low value of -fb resistor will load down the plate, but ...it can still do about 12VRMS...

Plausible. You are demanding about 2mA in the feedback net, which is no big load for that tube.

Note that open-loop without feedback and a more optimum load, that supply voltage would allow 40V RMS output at 5% THD. You have enough NFB going to keep the THD low up to the clip-point.

> A reverse-log pot would be preferable

Take another coffee-break, Dave. You know better. Consider an op-amp inverter with 1K fixed input resistor and 100K variable feedback resistor. Straight-audio taper gives approximately constant dB per knob-mark. Your non-inverting connection will bend at the low-gain end, and lack of ample NFB will bend the top, but mostly gain is proportional to feedback resistor and you want log-taper resistor to get log-taper control.

It is a bit tricky to set up, but you can run a PARAM sweep. Give the resistor a value-list like 0, 5K, 10K, 55K, 100K to see the gains for a 10%-taper pot at 5 settings 0∠, 68∠, 135∠, 203∠, 270∠ (where ∠ is the degree symbol that this board does not seem to support).

pstamler> also generates distortion with a lot of high harmonics

SPICE tube models I have seen just don't have enough terms to account for high-order non-linearity. If they did, they might run too slow for easy use. I would not say lying-ass: we give it a simplified description, it gives simplified answers. It isn't clear that an ideal tube will have large high-order THD (except grid-clipping) but real tubes with imperfectly wound grids sure do have high-order nonlinearities.

 

[NewYorkDave]

Yes, a coffee break is in order (Mmmmm... coffee!). I typed "reverse-log" as a knee jerk reaction to the notion of varying a feedback voltage, and didn't realize my mistake until Paul's reply made me actually stop and think about it, which is something we all fail to do sometimes, I guess :green:

 

[7string]

Hello everyone. I think I'm going to give Dave's design a try as well. I'll have to figure out the pad add. I plan on using this one after some of my balanced line level guitar gear and before the sound card inputs. I won't need alot of gain, but I will probably be hitting the input a little hot. THank you for the design, Dave. I think it will be a great first project.
Pat

 

[xldave]

I am also interested in building the pre as a first project. Can anyone point me towards a good power transformer thats has 240 volt primary to use with this circuit as I will have to order one.

Cheers
Dave

 

[ladewd]

PRR,

Thanks for the advice. Fortunately, the xformer runs relatively cool. I've burned in this preamp for a couple of days and it seems pretty bullet proof.

CA

 

[Gus]

Dave add a IRF840 SF direct coupled with a 15V zener gate to source to protect the gate. Adjust the Source r for the current you want . Its still a 1 tube circuit.

Or maybe better yet a 840 with a CC in the source leg.

 

[NewYorkDave]

You like those MOSFET source followers, don't ya? :wink:

I don't do those hybrid combinations, myself. It just never felt right to me... I guess it always makes me think of those cheesy "tube" preamps with SS front- and back-ends, or those awful hybrid circuits I used to encounter all the time in commercially-made guitar amps. When I design with tubes, I generally put myself in the mindset that I'm working in 1955. Sticking in a transistor or IC to compensate for something that's not easy to make the tube do on its own feels like cheating to me. But that's my own personal hang-up.

 

[Gus]

I have not installed one in a guitar amp or tube pre yet but untill clipping EF SF and CF circuit tend to be clean. I posted it as an idea if the fil supply did not have enought current also you don't have to float a fil supply.

I forgot to add the gate will want a few hundred ohm resistor in series.

I am going to maybe help install one in a preamp this week.

 

[NewYorkDave]

Watch that series resistor--gate capacitance in a MOSFET tends to be pretty high, and also nonlinear. Of course, you're trying to isolate the driving source from that capacitance by including the resistor, but if it's larger than it really needs to be, your frequency response will suffer.

You could always buffer the MOSFET gate with a JFET source-follower... Now it's getting really complicated. :razz:

 

[clintrubber]

tmbg wrote:

Also I don't have any decent audio transformers

Sorry if I missed an eventual earlier alike suggestion, but wouldn't the cheap Edcor-TXs be suited for the 10k:600 ? (WSM10k/600)

 

[Gus]

Good point about the gate cap

I have built guitar overdrives using power mosfets. If you keep it under 5K it sounds ok for guitar. Some power mosfets can have 3,000 pf of gate cap IIRC, so for a preamp I would guess 220 to 1k.

 

[PRR]

> How about a one-tube preamp?

Too many parts! Aside from quibbling that the one bottle is two tubes, you have 3 or 4 transformers, two separate supplies, high voltage.

One transistor mike preamp:

Genuinely useful around a studio.

Pretty darn safe around a newbie.

Gain of the amp alone, into 10K load:
from low-Z source, is about 34dB hi-gain, 20dB lo-gain.
from 10K source, about 28dB hi-gain, 14dB lo-gain.

With the low-price Radio Shack, AudioTechnica, or similar XLR-1/4" inline transformer shown, total gain is about 45dB hi-gain, 31dB lo-gain.

Gains are 6dB higher if you have an infinite load, 18dB lower in 600Ω load (NOT recommended!).

At 5V RMS output, hi-gain gives about 2% THD, low-gain gives about 0.5% THD. At both settings you get mostly 2nd but plenty of 3rd under it. When it clips, it is asymmetrical: it will go sweet-fuzz rather than harsh-buzz. Distortion reduces at lower levels, of course.

Bass response in hi-gain is -3dB at 20Hz, better in lo-gain.

Treble response is -3dB at 16KHz with 1,000pFd load, so you want to use less than 30 feet 10 meters of cable on the output. Keep it near the console or recorder, not on the far side of a big barn. Remember that the lo-Z mike line can be any reasonable length.

This will work very well from four 9V batteries. In a typical operation, battery cost will be less than coffee/beer expense. But I've shown a dead-simple wall-power supply too.

I picked 36V-40V partly to get some "oomph" out of the one transistor, but also because most Phantom mikes can eat the same juice. Using the RS/AT/etc $10 transformer, gently take out the XLR end, tack one end of two 4K7 1/4W 5% resistors to the XLR pins that have transformer leads. Solder the other ends together with a length of thin insulated wire. Drill a hole in the side of the transformer case to bring this wire out. Connect to "Phantom" or the top of the four 9V batteries.