One-Bottle Preamp

[ciminosound]

Heres' a chart I found on the web somewhere.
(I hate using things without giving credit!)

http://www.geocities.com/ciminosound/E180CCtable.jpg

(CUT AND PASTE LINK INTO BROWSER WINDOW! DAMN YAHELL!....)

Gerry

 

[pucho812]

Dave looks good. Any audio samples?

 

[NewYorkDave]

I'm planning to cut a song demo within the next couple of weeks, and I'm planning to use my new "fum ribbon" along with the one-bottle preamp for some of the tracks. So I can post that when it's ready. But if it ends up going on the back burner for whatever reason, I'll try to find time to just put up a mic and cut a simple sound sample or two.

 

[PRR]

Another way to make a one-tube preamp, one tube not one bottle, though it is a heck of a tube and doesn't come in a bottle.

GE 7768 microwave triode datasheet

At Vp=200V, I=24mA, you get Gm=50,000μmhos (yes, 0.05 Siemens), Rp=4,500Ω, which works out to an Amplifcation Factor of 225.

Yes, it it a triode, not a tetrode.

Yes, Mu is over 200 and the curves are as linear as many triodes, better than some.

One oddity is that Gm is so high they do not suggest self-bias, but a cathode resistor and +6V on the grid. Like you would do with a transistor, and for about the same reason.

With transformer loading into 10K, the voltage gain is over 130. 10K:600 is 4:1, so we have gain about 30 from grid to load.

Grid-Plate capacitance is about 2pFd. Miller that by 130, input capacitance is around 300pFd. High. Round up to 500pFd for strays and winding capacitance, the source impedance must be under 20K. Say 10K, with 200Ω primary, a 1:7 step-up.

1:7 times 130 times 4:1 is forward voltage gain of 200 or 46 dB.

Power consumption at the book conditions is 200V at 24mA or ~5Watts. I think you could work it around 120V and 15mA, ~2Watts. Heater demands 6.3V at 0.4A, 2.5Watts.

What is this 7768? It is a cylinder of ceramic rings and copper disks, about 1/2" diameter and 3/4" long. No glass at all: no glow. One source is asking $129 each. To this must be added a fairly beefy output transformer (10K and 24mA is a lot like a type'50 SET power-amp), and the input iron of your choice.

 

[fazeka]

Wow, PRR. That is freaky looking, sort of a 50s sci-fi look to it.

What kind of socket does that thing take? :shock:

Owensboro, KY eh? Home of current-day Richardson, though something tells me they don't make this one anymore!

 

[mikka]

I remember an oddball guitar amp from about 1980. Had one massive valve in it designed just for the amp. I think it was US made. Theory was that it would sound better...reducing the capacitance I guess. Can't remember if it had an OT. The sales guys were really excited. Then I went and asked them how long it would take to get a replacement if I blew the tube/dropped the amp and broke it. Genuine question .... I thought they would keep them in stock. Psssssssssssssssssssssssssssssssssssssssssshhhhhhhhhhhhhhh.....

 

[Guest]

From the 7768 datasheet:
"especially suited for use where unfavorable conditions of mechanical shock, mechanical vibration, and nuclear radiation are encountered."

sounds like it might be great for location gigs... :twisted:

 

[CJ]

I have been wanting to build a one bottle mic pre with one of these guys for years. Tons of gain, but I don't know how I would wire it. Grid leak or cathode bias.
Can anybody draw me a circuit as a starting point?

 

[pstamler]

PRR's design, with a 1:7 input transformer, would have a noise figure of 0.185dB with a 150-ohm microphone, assuming a perfect transformer and an unbypassed cathode resistor. In effect, the tube would add virtually no noise; all noise would come from the transformer and the mic itself. Pretty impressive! Not to mention the stares it would draw if you put a window on the preamp so people could look at it...

Peace,
Paul

 

[bcarso]

Umm 7768. Another example of where optimizing for one app leads to benefits in a totally different area. Sort of like the 2N4403 in sandland, designed as a switch with no thought given to amplifier use.

Another triode that is pretty interesting is the readily available 6C45. I don't have the precise characteristics at hand right now but I remember high gm, although not near as high plate resistance as the 7768 hence lower mu. The one thing I noticed about one I played with was a very high Q microphonic at about 10kHz.

I'll bet the microphonics on the 7768 are nice and low.

 

[BradAvenson]

[quote author="CJ"]I have been wanting to build a one bottle mic pre with one of these guys for years. Tons of gain, but I don't know how I would wire it. Grid leak or cathode bias.
Can anybody draw me a circuit as a starting point?

[/quote]

http://www.pmillett.com/tubedata/HB-3/Miscellaneous%20Types/407-A.PDF
http://www.pmillett.com/tubedata/HB-3/Receiving-Type%20Industrial%20Tubes/5842_417A.PDF
http://www.pmillett.com/tubedata/HB-3/Miscellaneous%20Types/408-A.PDF
I bet these datasheets would lead you in the right direction.

You gotta love the TUBE DATASHEET LOCATOR

 

[PRR]

> the tube would add virtually no noise

This is perfectly possible with "common" tubes and a little more step-up. The 2dB noise figure of a good tube amp is a lot about transformer losses.

> an unbypassed cathode resistor

Maybe just telephone company caution, but they seem to suggest a large grid bias voltage and a large cathode resistor. Might be less necessary with a plate resistor; but in microwave and in transformer-output there is no 1/Mu feedback from the plate resistor drop.

Also to get the large voltage gain with low output impedance, we need to bypass the cathode resistor (or at least most of it).

> optimizing for one app leads to benefits in a totally different area.

A "benefit" of dubious merit. Two-triode amps, like NYDave's, give more gain/feedback with much cheaper triodes that come two to a bottle and are mass-produced, and far less supply current to transform and filter.

The B+ feeding is non-trivial. A 12AV7 amp eats say 7mA, the 7768 around 28mA. Four times the power. And the 12AV7 amp has around 10dB-20dB negative feedback which will reduce power supply ripple; the 7768 amp has less than 4dB ripple rejection. So you need 24dB more filtering on 12dB more current, 36dB or like 60 times as much capacitance! Even with the low price of Asian Al caps, this gets to be awkward for more than a couple channels.

And with little to no feedback, the 7768 amp won't be the cleanest tube-amp on the planet, though it may have only small simple errors that don't offend the ear.

> the stares it would draw if you put a window on the preamp so people could look at it...

What stares? The 7768 is about as exciting to look at as a ceramic transmitting capacitor. Small dark ceramic lump.

> draw me a circuit as a starting point?

Tube mike input transformer to 417A grid. 5K,25mADC,+20dBm:600 output transformer to +150V. 60Ω cathode resistor bypassed with 470uFd.

 

[PRR]

> Can anybody draw me a {417A} circuit as a starting point?

I took the tube-book conditions almost literally, just dropping B+ to 120V; these may not be best, but actually might be a good starting place for mike-amp use (if cash is no problem). 417a/5842 data

Voltage gain, as shown, is a trace over 10. Input capacitance is low enough to use ordinary 1:10 grid transformer for 200Ω mike input, so total gain is 100 or 40dB, minus a dB or two for transformer losses.

Frequency response is all up to the transformers: the tube is flat from DC to many-MHz.

Output winding is grounded only for E-Z simulation: you would normally leave it floating.

Distortion at various levels, using a fairly decent (not 3/2-power) 417A model, runs about like this:

1V peak, 0.7V RMS, -1dBm
TOTAL HARMONIC DISTORTION = 0.1%

10V peak, 7VRMS, +19dBm
HARMONIC___NORMALIZED
____NO______COMPONENT
____ 1______1.000E+00
____ 2______7.100E-03
____ 3______1.326E-03
____ 4______2.605E-04
____ 5______3.403E-04
TOTAL HARMONIC DISTORTION = 0.72%

19V peak, 14VRMS, +25dBm
TOTAL HARMONIC DISTORTION = 3%

33V peak, +29dBm
TOTAL HARMONIC DISTORTION = 8%

None of these numbers are to be trusted. The detailed 1V numbers look like numeric garbage: I did not crank-up SPICE's resolution, and I doubt the model is accurate enough to warrant it. The high-output conditions will surely be very bent and about this much, but the exact distribution of harmonic strengths depend on bias, load, and model-errors.

The 10V numbers are probably not far wrong. Note that is is 0.7% 2nd and 0.13% 3rd, a sweet drop. Higher harmonics seem to be dropping nicely, though I doubt the model has enough kinks in the equation to be even semi-correct past the 3rd or 4th. (The 5th probably should not be "higher" than the 4th, but both are in the numeric noise.)

Distortion would be lower with a higher load impedance; but gain and power would drop too. That could be interesting, except the hefty supply current (17mA) already requires a large and well designed lump of iron, and trying to get 10K or 15K impedance at that current, with good bass response, would be tough.

Output impedance is about 200Ω. Gain and power would rise slightly if we "matched" to the plate (about a 2K primary) but distortion would rise and it would be sensitive to loading. Keeping Zout about 200 means a modest 3dB rise from 600Ω to unloaded.

 

[CJ]

Cool! Thanks guys!

Looks like the 417, with all that transconductance, would be the best tube to try first. Also has more reasonable 6.3 heater. I guess they ised the 20/40 volt heaters to insure adequate voltage on a tube that might be miles away from it's pwr supply. (sitting on a telephone pole somewhere)

I like the even harmonics!

Easy to wire up a one tube pre, thats for sure.

I should have some big iron laying around here someplace.

Frame grid construction. Should be extremely quiet.

 

[dale116dot7]

[quote author="CJ"]Cool! Thanks guys!
I should have some big iron laying around here someplace.
[/quote]

CJ, you have some iron that hasn't met a hacksaw?

 

[CJ]

My apartment makes creaking noises all the time. I don't know if it's termites or iron.
Good thing I'm upstairs! :razz:

 

[PRR]

> I guess they ised the 20/40 volt heaters to insure adequate voltage on a tube that might be miles away from it's pwr supply. (sitting on a telephone pole somewhere)

Possibly.

But telephone systems developed before widespread power systems, and were reliable when power systems weren't. Telephone systems often have their own power, and the customary power was 48V battery.

 

[ToobieSnack]

Dave thanks for this post. nice stuff and great specs.

I'd like to build one of these
i keep re-visiting this post and seem to absorb a little more each time.
i think this would be and excellent sample circuit/discussion for the newbie meta or electronics 101. thanks for all the info ....guru's

I have a bunch of stupid questions but i think I will try to figure it out as I go along/learn more.
some cool topics for dicussion in the newbie/101 meta ,with this circuit. i think, could be:
the transformer/transformerless output and
actaully this whole lil section is interesting

With transformer loading into 10K, the voltage gain is over 130. 10K:600 is 4:1, so we have gain about 30 from grid to load.

Grid-Plate capacitance is about 2pFd. Miller that by 130, input capacitance is around 300pFd. High. Round up to 500pFd for strays and winding capacitance, the source impedance must be under 20K. Say 10K, with 200? primary, a 1:7 step-up.

1:7 times 130 times 4:1 is forward voltage gain of 200 or 46 dB.

and just a newbie/quickie ...question...or 2 (or hey... a newbie/snack... lol)
what does Gm stand for? and what is a zobel?
info much appreciated
thanks
ts

 

[bcarso]

Gm stands for transconductance---how much does the (plate) current change for a small change in (grid) voltage.

The eponymous Zobel network dates back to ancient phone company days, and is usually a series R-C network placed at the ~output of something to damp overshoots in step responses and quell oscillations. The French (bless their hearts) call them Boucherot cells...

 

[CJ]

OK, after five years of staring at these WeCo miniatures, I finally found the hour it took to wire up as a micpre with the 417A.

It sounds great! What blew me away most was that with a 1:30 input transformer, the gain was about the same as a V72 or API/Melcor. And this is single triode.

It seems to have a lot of clear high end without the cathode by-passed, which might be useful, so I put in a switch for the cap. Vocals would sw the cap in, cymbals or field work (this would make a great tube/field preamp) sw the cap out.

Noise is great, and went way down when I dropped the B+, but the gain of course went down also. Which gave me an idea for gain control on a fixed gain preamp like the V72, as you would not have to disturb the circuit. Once in a while the tube acted up with a little hash, which is probably why it cost me a buck at the surplus. I have a few more I can try. I should have grabbed a whole shhopping bag of these, as they go for silly money on evilbay.

There is some high frequency ringing when the tube is tapped, way above what most tubes ring at. So I will shock mount the socket and maybe add a tube cooler.

Now I know what Aunt Gertrude was really listening to when she complained about a bad phone line. It was tube hash!

Bias sat at about a buck and a half with 250 B+.
Went to 75 cents with 125 B+.

Thanks to Bill Y for the DuKane, not a bad transformer at all. Bi-filar, and a nice mu can.

This pre was built from junk pile stuff, so it did not cost anything. I can't wait to record some stuff with it. Here is the schematic. Pardon the large jpg, stiill trying to figure out this new software:

This would make a great beginners tube micpre. One tube, one input x-former, two resistors, two caps, an xlr and a 1/4 inch output jack. I am, as always using an external pwr supply.

Let me know if you want some pics. Not much to see with that parts count.

I have plenty of rack left, so I think I will do another, only trying PRR's circuit, with the iron on the output instead. With 20 ma un-balanced dc, it's gotta sound cool!


cj