Tube piezo pickup preamp?

[Gus]

What you need to do first is scope the waveform from the piezo pickup. Now you might want to build a voltage divider resistor string to not load down the piezo pickup when it operates into a >3 meg load OR just hang a scope probe on the piezo output not connected to anything else and hit the bars and note the peak to peak output voltage. Or use Scotts circuit and math before the scope

You might be generating more voltage than from a guitar etc so you will need to check

This is so you can DESIGN for the output voltage of the pickup. You will need to know the input and output headroom to not clip.

Using a 12au7
-4 grid for input headroom
300 VDC B+
3.3meg, 10meg, 22meg what ever sound good input resistance
lower voltage fil for more like a tube microphone input resistance
Looked at the 12au7 curves I found on the web I picked the resistor values 510K and 8.2K you might want to cap bypass the cathode resistor. 510K and 8.2K are a first try values under heating at 5.8VDC and lower plate current.

If the distance between the circuit and the tube amp with a 1meg input is short you should not need a CF. Now if you want to use a direct coupled CF you might need to raise the heater voltage above ground.

HV mosfet as a SF, be careful with the power up and the gate to source voltage breakdown voltage.

Scotts circuit being a SF and having a voltage divider should work very well. I do have a question about it why is there a input cap and not just two 4.7 megs top or middle node switched to the gate when using a piezo? Does that low of a gate voltage cause dielectric stress in the piezo element that changes the sound if direct coupled?

Miller cap will have an effect look at a km84 circuit note the 4pf cap drain to gate (and the pad cap) that works with the capsule cap and the open loop gain of the fet stage and the transformer ratio to set the gain.

The summed up piezo pickups cap divided by the plate to grid cap and the open loop gain of the tube circuit. If you have to much gain maybe a cap plate to grid to work it more like a charge amp.

 

[tv]

If you put a 510k plate resistor, then most mosfets will have too much a gate capacitance to be driven as followers, no?

what about active loads? (mainly just for fun... perhaps too much hassle for an instrument pre)

here: http://home.pacifier.com/~gpimm/self_bias.htm
main: http://home.pacifier.com/~gpimm/

 

[Gus]

http://www.zetex.com/3.0/product_portfolio.asp?pno=ZVN0545G&h=49#applinks

 

[Mbira]

OK. Sorry for the delay. The mbira is finished, and I just tested the peizos voltage. I get max 2.5v peak to peak. The average is lower than this, but that is the highest voltage for the loudest keys.

Here is a pic of the front of the mbira:

And the back:

Things are messy until I get the wiring decided. Here is a basic (and ugly) quick schematic I just threw together. The ugly thing on the top left is supposed to be the 1/4" jack that will go into the guitar amp:

I have them wired this way because I'm thinking about either doing fixed r-c networks to the specific piezos, or do tone controls for each. There is enough room on the back body that I could route a space for some pots.

What happens is that a few notes inevitably ring out a lot louder than others with theise piezos. I'd like to tame that somehow.

 

[Gus]

Thanks for the voltage measurement.

Now a question why three pickups? What I am thinking is maybe you could use one and move to different places on the back until you get a balanced sound. You might have done this.

If you want all three maybe three of the SF type FET circuits one to each pickup and then mixed together to a parametric EQ to find the fixed EQ you might want.

 

[beatpoet]

You could make your life simple and just use a nice 1/3 octave EQ to 'flatten' it out

 

[Mbira]

[quote author="Gus"]
Now a question why three pickups? What I am thinking is maybe you could use one and move to different places on the back until you get a balanced sound. You might have done this.[/quote]
Because I use one external pickup now that I can do just that with. It works well, but is fragile and expensive. Every mbira I have built has a slightly different hot spot for certain frequencies.

I'd like to use the three and eq them individually (with a fixed resistor and cap once I figure out the value?) slap on the coverplate and forget about it.

I think I will wire up a type of gibson pickup wiring scheme (volume and tone for each pickup) and see what happens. If I like the results, I'll slap it together.

I play many different mbiras all with different eq needs-that's why an external eq wouldn't work so well. Remember, these are just being plugged straight into a silver-face Super Reverb.

 

[Mbira]

Alright. I finished the mbira pickup. This is certainly the best I've everr made. I hardwired resistors set to balance each of the three pickups. It sounds great.

Now, I am back to wanting to build an external tube pre for gigs when I don't have the Super Reverb around.

 

[Mbira]

is there a specific ratio that I want for the output transformer into an XLR?

I'm thinking of building the first preamp stage of a fender Super (including tonestack) but then going to XLR output.

I'd like to even use a tube rectifier.

 

[Mbira]

Thinking about this more, should I just measure the peak to peak signal going into a super reverb circuit and measure the signal coming out of the tonestack and figure out the ratio and that is the ratio I'm looking for in a transformer?

Thanks,
Joel

 

[PRR]

> I just measure the peak to peak signal going into a super reverb circuit and measure the signal coming out of the tonestack and figure out the ratio and that is the ratio I'm looking for in a transformer?

Not even close.

Gitar pickup peak levels are 20mV to 500mV.

12AX7 has gain of 50, 1V to 25V out.

Fender BMT tonestack has very up/down loss (that's the point), but may be approximated as 20:1. 50mV to 1V out. Discount the high end because you have (should have) a Volume pot here. 50mV-100mV nominal off the tonestack.

The output impedance of the stack is critical. It is even more up/down, which means any loading will cause suckage. The only good load is a vacuum tube grid. Over 500K, more is best.

But leave-off for a bit. The other half of the puzzle is: what levels do you expect on this "XLR"? It could be Mike or Line. If "mike", the next question is: what's yer other mikes miking? Ribbon on harpsichord may not peak 2mV. Condenser on guitar amp (or large/hard percussion) may be over 1V. You picked your mike inputs, and set your trims, by what you mike. When everything was dynamic, I figgered 10mV to 100mV for a stage act, though I have sneak-reinforced small choir at 1mV and tried to capture mega-percussion at over 1V. But let's take 50mV as not too atypical.

OK, paste-up time. We have a 50mV source which wants >500K load. We have a 50mV input which, if on a long line, may look like 200 ohms at the top of the band, generally 2K midband.

There is no passive solution.

You do have reserve of gain in the mixer. If you assume it cranks to 5mV sensitivity, then a 50mV source could be transformed 10:1 voltage, 100:1 impedance. A 200K:2K transformer works on paper. However few trannies hold 200K down into the bass, and that's where the tonestack Z is highest. Bass will suck. This may matter for this purpose, I dunno what low-tones can come off this instrument. Additionally a 200K:2K transformer feeding a long line is likely to cut the highs quite severely. This may not be a disaster. However the problems are so great that there are few general-purpose transformers for this duty.

One thing to try, because it is a Standard Part, is a passive direct-box transformer.

Since you get two per bottle, another approach is to use the other triode just like Fender does. Change the 100K plate resistor to 22K, change the 1K5 cathode resistor to 470R||25uFd. Use 0.22u cap to couple to 10K:150 transformer.

 

[Mbira]

To narrow it down:

I love how the instruments sounds going through my Super Reverb and out of the speakers.

Lot's of gigs, I have to use a DI box straight into the board. I'm stuck with a thinner sound and general lack of "majik".

I wanna make something with toobs like a toob DI box because I think the tubes are what make the mbira sound so good...

Since you get two per bottle, another approach is to use the other triode just like Fender does. Change the 100K plate resistor to 22K, change the 1K5 cathode resistor to 470R||25uFd. Use 0.22u cap to couple to 10K:150 transformer.

uhhh....like this? (warning-this will probably kill you because I'm pobably VERY wrong!) Do you mean the first stage cathode resistor and cap for the change?

 

[PRR]

> this will probably kill you

Nope, still breathing.

V1b wants a cathode resistor (the plan you stole shares this with another triode). I'm guessing 470R, but some experimentation would be good.

I don't think a bottle rectifier gains you anything. There is no "sag" such as large class AB-amp fans rave about. First because class A runs a steady current idle or whacked; second because the available 100mA+ rectifiers won't begin to sag with this mere 4mA load.

410V supply will be very clean, also costly. Start with 300VAC to make 420V DC on a 20uFd 450V cap, then go 10K-20uFd-10K-20uFd-10K-20uFd filter to get the hum down. You will land in the 250V to 350V range. I won't compute it, because your next step is to tack 5K or 20K resistors in place of the 10K, and see how that affects sound.

A key point (I think) of the Fender Input Stage is how it runs from very low THD to significant THD as the plucker leans into his pick and as notes decay. Woodwinds have strong timbre change with breath effort; thin steel strings on solid boards don't; the preamp curvature adds "feeling". If we fed a weak guitar to a preamp run on 600V, we'd never get into the feeling zone. If we feed a strong guitar into a preamp on 150V, it gets "too ripe" going on "blatttt". (Which is an effect also.) The 410V shown comes from an existing HIGH-power output stage, and also where Leo and Forest wanted to position this model that year. A fine choice. You might like another voltage range better for your instrument/pickup and music.

I have a feeling the output stage should be 12AT7. Can work at the input too, though it may want a lower supply voltage. But now we are deviated rather far from the Classic, and I can understand why you might not want to muck with success.

 

[Mbira]

I've got an old stancor PT here that'll give me around 350V. I'll probably use that because I need this pre alot more than I need some Deluxe clone sometime in the future.

sooo.....?

Inevitably there seems to be one really hard part to get. I'm hoping I have something here that will work an an OT.

 

[Scodiddly]

I wonder if maybe it's not any tube magic but rather the sound of the amp speakers that you like in this situation. Maybe a research step would be to fool around with some EQ into a PA speaker, to see if you can get closer to the amp speaker sound. Cut the very top, midrange boost, etc. Perhaps you could borrow one of those little guitar wonder boxes with amp modelling just to fool around with different speaker emulations?