Matching Transistors for Valley People MicPre

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Your probably right. The front end is more important than the opamp.

What's wrong with the 3906? Noise?

TL072 won't drive 600 ohms, right?

I can post the rest of the schematic, but it is not proven (to me at least) so .......

Valley_People_Mic_Pre.jpg


Could I use a couple of zeners in place of that middle transistor in the current source?


Jeez, anybody ever get carpel tunnel from these darn mouses? Or is it mice?

here's a 1.8 mb pdf hype sheet
 
> Could I use a couple of zeners in place of that middle transistor in the current source?

Sure. You could even replace all that current-source crap with two 7K resistors; many good mikeamps do. But I thought you had thousands of transistors?

I looked up that 2N4355. It isn't claimed to be a very low noise part. It is claimed to pass 1 Amp well, so it is bigger than a '3906. It, or a specific run of it, may have tested well in VP's labs; or it may just have been cheap enough to sort for match and excess noise.

A 5532 is not a drop-in in this plan: note the huge resistors into the first TL07x. That needs a FET-input or a redesign. OTOH, the second TL07x has a very easy job, and a lot of chips could replace it.

No, a TL072 sure won't drive one end of a 600Ω load well; neither would the Valley. This was the dawn of the idea that you could do pro-quality audio without going down to 600Ω every stage. IIRC, this design was mostly used inside a box or small system, not where it would be asked to drive long lines. I forget what VP did for long-line drivers.

I do like the VP mike pream. It gives differential input and balanced output with just two opamps; many designs (like the Green) use four opamps. It also loses the DC without any huge capacitor. I don't know why it isn't used more often.
 
Not enough bias current for that first opamp, I see. Needs FET front.

Would stacked ceramics make beter coupling caps going into that first opamp?

Who makes a low inductance 0.68?
 
I have not read the entire thread so this may have been discussed already. IMHO, there may be a drawing error on the second opamp. The inverting and non-inverting inputs appear to be swapped. The non-inverting input should be grounded.
 
> Who makes a low inductance 0.68?

Why?

It would take a hell of a lot of inductance to bottleneck the 3 Meg load.

In fact I don't see any real problem here. With 3Megs and caps of that magnitude, the cap is comparatively dead-short through the audio band, no matter how bad it is.

I have not seen a microphonic ceramic. If there is such a thing as a piezo microphonic ceramic, that would be very bad news. But 0.6uFd(?) is not a value for good-quality ceramic. Use film-caps in your choice of flavor.

Actually, any but a worse-case 5532 would "work". 1uA bias current times 3Meg means the inputs will idle around -3V. Bias current is usually less. Offset current could bite, maybe 0.3V output offset, which really should not be present on a line-out. You could scale by 30: 100K input resistors and 22uFd caps, which need to be low-leakage (or Tantalum for that 1980 sound).
 
Thanks PRR!

Yes, the opamp pins are numbered wrong. I forgot to correct that. :oops:

This version did oscillate the first time I tried it, so maybe the polarity is wrong.
 
A little OT: I was asking about low inductance caps. I took apart one of thee old Siemens caps that I scored at a surplus store and noticed a weird construction. Instead of winding a bunch of foil in a roll, this guy uses individual rectangles of foil that are somehow spliced together. It's kind of like the pages of a paperback novel, only bound at both ends.I am wondering why they built them this way. Lower inductance, maybe?

siemens_cap.jpg


siemens_cap1.jpg
 
The op amp pins are numbered correctly but the polarity marks are wrong on both. If you wired them up based on the polarity symbols and not the pin numbers the thing would just latchup---as some call it, a zero-freq oscillator ;).

The cap you describe is a so-called stacked film, and yes, it typically has lower inductance than some other constructions.

PRR wrote: "I have not seen a microphonic ceramic. If there is such a thing as a piezo microphonic ceramic, that would be very bad news."

With the move to SMD there are some very piezoelectric caps out there now, for the so-called high-K dielectrics. One wouldn't let these get near a pro design but I have had to use them here and there in budget stuff.

If you get hold of a JBL On Tour, a battery-powered microsystem for MP3 players like the iPod, you will notice that after a couple minutes or so of zero input signal it goes into a sleep mode indicated by a flashing green LED. It wakes up for a mV or so of input signal, but you can just bang it a bit to turn it back on as well.
 
OK, I fixed the polarity on the opamps shown on schematic-page 2.

OPA2134 has same pinout as other duals, so I will lay out board for dual opamp. App note likes both chips on same piece of epoxy.

I noticed the output of this thing comes straight off the output pins of the opamps. No blocking caps needed? I guess thay are banking on low offset voltages.
 
This is about what the VP will look like when it's stuffed. Have a few corrections on the board to do, as well as some R and D for resistor values.

Transistors will be thermally coupled after the tweaking is done. Draws about 12 ma with the 8 transistor array.

vp+1.jpg


Off to the river. See you monday.
:thumb:
 
Thanks, Thomas!
Going to do a Rev 2 board tonite, bigger pads in some places, trace corrections, etc. I am also going to flip one row of transistors in the array so that the flat sides will face each other which will help thermal coupling a bit.
cj
 
I was measuring some 2N4403s and the OnSemi parts have pretty good Hfe. Most of them were around 200. Digikey has them for little money.
The Fairchild ones have less gain based on a sample of 25 transistors each brand.
 
Sounds good.

Hey, what is the difference between using 2 transistors per side in the array as opposed to 4? Do I get more gain or just less input impedance.

Noise?

Thanks!
cj
 

Latest posts

Back
Top