Low voltages, low freqs suffers?

[stonerbox]

I'm working on a fuzz [five stages] that has some hum/noise issues. I'm staving the collector at the fourth stage and I'm getting great results, except for one little thing. The fullness of the lower freqs seems to suffer.

I know that tubes driven by low voltages often has dogdy bass response. Could the same thing happen when lowering the voltage on the collector of a transistor or am I just imagining things?

[GibsonGM]

Well...I hadn't considered this before, but it's an interesting question, Stoner.   For lack of concrete understanding of this particular issue, I would put out as a possibility that as you alter the collector voltage, you're moving the transistor's operating point, and affecting bias.  That could change inter-element capacitances (to what degree I don't know), and maybe that's affecting bass response?

Just a guess, I don't have any real way to back that up.   Maybe you could "sim" that in LTSpice, see what you get for an AC analysis?

[R.G.]

No, it is unlikely that this is an issue. I suspect it may be due to a "starved" stage clipping harder and putting in more high harmonics, sounding more trebly. But the bass is very unlikely to be cut.

Bass in audio is very, very close to DC compared to the upper audio ranges. While there is a change in the input impedance of a bipolar with changes to the collector current (see "Shockley Resistance") it would take an odd circuit to make this have much of an effect on bass response.

Changes in internal capacitance will affect highs much more than lows, and in modern silicon devices, the capacitances are so low that changes in "treble" will be in the RF region in most cases. This was not necessarily true for old germaniums but with more than half a century of semiconductor processing under our belt, devices are much better (in the technical sense).

I would suggest rethinking your coupling caps and emitter resistors (if they are bipolars, which I infer).

[stonerbox]

GibsonGM Don't know why I didn't think of that. Thank you, I will surely keep that in mind.

R.G. Yes, they are biopolar. Thanks for the in depth explanation. I guess I interpreted the "tightness" that the miss-biasing can brings to the plate, if overdone, as a loss and/or added tightness in the bass response.

I'd like to retain the slightly loose, full and open style of the fuzz but without the hum and noise, of course. And as you suggest I better go back and check for alternatives on coupling caps and emitter resistors. Also, for the first stage I use a SC21685 [S 290-460 Hfe] and the two following are SC21685 [Q 160-260Hfe] and I believe they in their self are a little noisy. Better check those noise figures. Again, thanks R.G.

[GibsonGM]

No thanks for me, Stoner, R.G. has it nailed - my answer was only a guess, his is based on a lifetime of studying this stuff and being able to think on the fly about things like this!    Good analysis - a change in clipping characteristics. 
Poly or Mylar on the way?

[stonerbox]

GibsonGM I meant thanks for the sim idea, a thanks which I believe you are entitled to. I'm using Poly but I'm confused what you meant with clipping characteristics?

[GibsonGM]

I was commenting about what R.G. said, which was:
"I suspect it may be due to a "starved" stage clipping harder and putting in more high harmonics, sounding more trebly. But the bass is very unlikely to be cut."

Your last (starved) stage may be clipping harshly, and adding harmonics that make it appear that you have lost bass, when in fact, you may have ADDED those higher frequencies without really affecting the bass response of the stage.   So, the clipping characteristics of the stage may have changed when you starved it...   

I take that to mean that you may be able to increase the bass response of that stage by trying other (larger) coupling/emitter bypass caps....

[stonerbox]

Ok, I see what you're saying now. It's late over here (Sweden) so I'm a bit slow.

[GibsonGM]

Perfectly understandable!  Let us know what you find out - now I am curious if this was the cause of the problem...

[R.G.]

Me too. I do a lot of theorizing. It's always good to hear what Mother Nature thinks of my ideas.   

[amptramp]

Me too. I do a lot of theorizing. It's always good to hear what Mother Nature thinks of my ideas.   

Mother Nature is a bitch.

[stonerbox]

Me too. I do a lot of theorizing. It's always good to hear what Mother Nature thinks of my ideas.    

Nearly necroposting but here we go!

I did some tests and the lows and lower mids response does actually differ, at least from what I can gather. I did a simple spectrum analyzer test and this is what I got.

- Q3 collector - "normal" bias. -


Note: A smooth drop from around 120Hz down to 30Hz.

- Q3 collector - chocked (Gating) -

Note: Notice the gaps in this one, especially around 60-80hz. The 100-130Hz area appears to be a lot flatter too.

[FiveseveN]

Were you using white/pink noise as your test signal?

[R.G.]

Interesting, but to come up with more than we've already said (with any possible hope of being right, at least  ) we'd have to see the schematic.

There are always special conditions. As one college prof noted, "The real universe we live in is just a special case, albeit an important one."

[stonerbox]

Were you using white/pink noise as your test signal?

I used a short riff on repeat.

Interesting, but to come up with more than we've already said (with any possible hope of being right, at least   ) we'd have to see the schematic.

There are always special conditions. As one college prof noted, "The real universe we live in is just a special case, albeit an important one."

The schematic is far from finished and I would be embarrassed to show it to you guys right now.. However this experiment got me intrigued so I'll build some generic transistor stage/s and analyze the frequency responses. I'll post  the results and schematic when it's done.

Did another test on the fuzz feeding it with a sine wave sweep (16Hz-22Khz).

"Standard Bias"


"Chocked Collector"

[mac]

Are the bias resistors too different in both cases?

mac

[amptramp]

If you look at a woofer when it is operating, it has to have a longer throw to provide more perceived audio output than a midrange speaker.  The throw is proportional to the voltage, so yes, the bass will suffer if you don't have enough headroom.  I used to witness testing of avionics that included constant g-force vibration from 5 Hz to 2000 Hz.  Getting the typical values required almost a foot of displacement at 5 Hz declining to almost invisible travel at 2000 Hz.  BTW getting a pair of 5 KW amplifiers and the vibration test drivers would make an awesome stereo system, capable of being heard for miles.  I must check the industrial surplus sales...

[stonerbox]

Are the bias resistors too different in both cases?

mac

I'm not sure I interpreted you question right but I'll try to answer anyway. The biasing was done on one and the same transistor (Q3) in the circuit, both times. First with a 10k resistor then with a 10k potentiometer set to right where the noise and hum gets cut off.

If you look at a woofer when it is operating, it has to have a longer throw to provide more perceived audio output than a midrange speaker.  The throw is proportional to the voltage, so yes, the bass will suffer if you don't have enough headroom.  I used to witness testing of avionics that included constant g-force vibration from 5 Hz to 2000 Hz.  Getting the typical values required almost a foot of displacement at 5 Hz declining to almost invisible travel at 2000 Hz.  BTW getting a pair of 5 KW amplifiers and the vibration test drivers would make an awesome stereo system, capable of being heard for miles.  I must check the industrial surplus sales...

Yeah, this is what I was thinking too. Whether or not this applies to a transistor stage and how much in this particular circuit we'll just have to see, but I'm pretty positive.  

[mac]

I'm not sure I interpreted you question right but I'll try to answer anyway.

I asked because if, for example, you have to lower the collector resistor considerably to get a different collector voltage/current, the corner freq of that resistor with the decoupling cap decreases. *Maybe* that's what you are hearing.

mac

[R.G.]

Interesting, but to come up with more than we've already said (with any possible hope of being right, at least  icon_biggrin ) we'd have to see the schematic.