Filthy BJT! We hates it forever! -- A mini rant

[jafo]

It's weird; JFETs never gave me any trouble without having to bother with mu-amps or higher-values Rs and Rd, but BJTs? Nope. I can screw up even a simple NFB boost (say, an Electra without the clipping diodes) nearly every time. A Bazz Fuss? I've built it 8 times and I can't get it to work at all -- even with a functioning minimal BJT circuit using only Rc (cap into B, E to ground, C with a resistor to power and a cap after for output (and that's when I can get one to work at all!)), adding a diode of any type or orientation makes it stop working... and for that matter, removing the diode doesn't restore functionality. Very frustrating! Should I just stick to JFETs, or should I pray to some dark god (Morgoth, perhaps, as befitting the title)? Yeesh. How do you all do it?

[PRR]

> cap into B, E to ground, C with a resistor to power and a cap after for output

You HAVE to have current to the Base.

Very old GE parts leaked enough  to kind-sorta work with no other part.

Silicon parts will just sit there dead to the world until you bleed some current to Base.

Are you out of known-good plans?

[anotherjim]

Well, there has to be base bias current flowing from base to emitter (in the right direction depending on NPN or PNP). Your minimal circuit seems to lack this.
Breadboard one of these basic boost/fuzz BJT circuits, show us the schematic, and we'll get it working.
BTW, some grab-bag small BJT's may have current gain at the low end around 40 to 60 and they won't cut it in many audio circuits. Try with 2n5088 which have plenty more gain.

[fowl]

I remember Blue Thunder.

[Steben]

It's weird; JFETs never gave me any trouble without having to bother with mu-amps or higher-values Rs and Rd, but BJTs? Nope. I can screw up even a simple NFB boost (say, an Electra without the clipping diodes) nearly every time. A Bazz Fuss? I've built it 8 times and I can't get it to work at all -- even with a functioning minimal BJT circuit using only Rc (cap into B, E to ground, C with a resistor to power and a cap after for output (and that's when I can get one to work at all!)), adding a diode of any type or orientation makes it stop working... and for that matter, removing the diode doesn't restore functionality. Very frustrating! Should I just stick to JFETs, or should I pray to some dark god (Morgoth, perhaps, as befitting the title)? Yeesh. How do you all do it?

Bias bias bias.
Once you are there you might just starting sticking up your nose to many jFETs...
Double check your transistor pins as well ...

[amptramp]

The one thing designers like about the BJT is that the emitter - base voltage in operation is always going to match within a few millivolts.  You can design a circuit and go into production with the knowledge that Vbe is always going to be the same.  If you look at the input and output buffers in Boss equipment, you might be able to say they could have done what they needed to do with a JFET but source voltage would have been all over the place, so they used BJT's even with the input impedance penalty.

You would be lucky to get JFET Vgs to be within ±50% of a single rated value.  I have a Sencore signal generator that uses a JFET source follower and the effect of a high Vgs was to fry the source resistor - it was supposed to be 390 ohms but overheating raised it to 2195 ohms measured.  But the JFET was within its rated Vgs.  If you have a good JFET design and can afford the Vgs variation, it is a great choice - there are a lot of Tillman amps out there that run well with J102 devices but these are getting scarce for a reason - it is hard to keep the Vgs at the tight limits of the J102 spec.

When I need JFET characteristics like high input impedance, I use a JFET op amp like the TL07X series.  In most cases, the offset voltage difference of ±13 mV at the input over the full temperature range is not going to cause a problem and the two input devices are subject to the same processing so they are close to a match.  Even so, I have a design right now where I need to use the zero adjust capability of a TL071 to ensure acceptable circuit operation.

[antonis]

If you want a working stage , use FETs..
If you want a working stage in the way you wish to, use BJTs..

[jafo]

> cap into B, E to ground, C with a resistor to power and a cap after for output

You HAVE to have current to the Base.

Very old GE parts leaked enough  to kind-sorta work with no other part.

Silicon parts will just sit there dead to the world until you bleed some current to Base.

Are you out of known-good plans?

Yep, that's one of the things I've tried, sometimes with larger C-B resistors, sometimes with diodes; using 2n2222, 2n3904, even BC337, correctly oriented, spun 180° (emitter to ground, emitter to +V). The only thing I've ever had success with was Earthscum's Super Simple Anti-Sparkle Drive, which uses a reversed 2n2222, but even then I had to unreverse it to work, and even then it was too muffled. Schematic (because it no longer exists on the forum page) here:



I'll have to try germanium.

[jafo]

I remember Blue Thunder.

Is THAT where it's from? A buddy started calling me that a few centuries ago, and wouldn't tell me why, but it stuck.

[jafo]

Well, there has to be base bias current flowing from base to emitter (in the right direction depending on NPN or PNP). Your minimal circuit seems to lack this.
Breadboard one of these basic boost/fuzz BJT circuits, show us the schematic, and we'll get it working.
BTW, some grab-bag small BJT's may have current gain at the low end around 40 to 60 and they won't cut it in many audio circuits. Try with 2n5088 which have plenty more gain.

Hmm, I've tried "Rc (cap into B, E to ground, C with a resistor to power and a cap after for output"; what did I miss? Anyway, good idea to use a breadboard and 2n5088; saving my pennies!

[GibsonGM]

"JAFHO" - "just another first-hand observer"   

Did you forget that feedback/bias resistor?  The 100k from collector to base...

[anotherjim]

You also have to give voltages when you need help. An NPN amplifier base has to be about 0.7v higher than the emitter or it no worky. The collector voltage has to sit higher than both.

Have you got a multimeter with transistor test sockets?

[jafo]

If you want a working stage , use FETs..
If you want a working stage in the way you wish to, use BJTs..

Heh. My next question will be how to bootstrap to raise input impedance!

[antonis]

Heh. My next question will be how to bootstrap to raise input impedance!

Input impedance of any value should be totally useless for a non-working CE amp..

[jafo]

Weird; now the very simple things are working -- probably because I'm being more careful.

I also think I may have found a problem in my "design": I've been bypassing the feedback resistor/pot with the diode, instead of replacing the resistor (or putting them in series). Something about putting them in parallel feels wrong, but I can't figure out why... probably just because nobody seems to do that. Electricity follows the path of least resistance, so it seems like in this configuration it would go through the (silicon) diode if > 0.6V, and through the resistor if < 0.6V, right?

[Steben]

Weird; now the very simple things are working -- probably because I'm being more careful.

I also think I may have found a problem in my "design": I've been bypassing the feedback resistor/pot with the diode, instead of replacing the resistor (or putting them in series). Something about putting them in parallel feels wrong, but I can't figure out why... probably just because nobody seems to do that. Electricity follows the path of least resistance, so it seems like in this configuration it would go through the (silicon) diode if > 0.6V, and through the resistor if < 0.6V, right?

uhm.... of course. The bazz fuss does not have them in parallel, does it?
Why hate something if you do not build the schematic?

Just some tips:

1. try to make a schematic of what you really have built
2. try to post voltages at most points, especially the 3 pins of transistors
3. if you are not an electronic master, do not throw any component in with the demand of getting a fantastic result. You are dealing with 9V probably (right?  )  so it won't that dangerous, but doing these things will if you go higher.
4. stop hating components

[amptramp]

There used to be some questionable designs that used a reverse-biased diode in place of the feedback resistor from collector to base.  The idea there was to have the diode leakage current track the collector to base transistor leakage current with its rising value with respect to rising temperature.  And in some cases, no bias device was used for the base - everything was biased by leakage current and the signal came from an electrolytic capacitor that had its own leakage as well.  You might get away with it using germanium.  Silicon requires you to design the stage for very low leakage current.

[jafo]

Thanks, everybody. Things seem to be working well now; electronics is fun again!