transistor recommendation for input buffer

[PRR]

> if it seems like I have a major mistake in the diagram

Diagram is fine.



Circuit action is clearly wrong one way, and dubious another.

hFE (Ie/Ib) appears to be barely over 100, while spec-sheet gives much higher values in this range of current.

The collector-base voltage is very close to 7V.

Guess: you have E and C reversed.

A BJT will "work" backward. However the BE junction is processed for high gain at the expense of breakdown voltage, the CB junction is processed the other way. Reverse-connection hFE may be very much lower than normal-way hFE. And the BE junction voltage breakdown is typically 7V (the 4.5V on the sheet covers slop).

The low gain alone might "work". But if it is stuck in breakdown, this isn't a transistor, this is a Zener, a very solid voltage-holder. It must be just a hair into breakdown, else the input impedance would be (0.1uFd plus) 3.3K or the ~~100r of a Zener.

As an academic tool, yank the base out of circuit and see if "E" stays very near 1.1V-- that would mean it is (and was) in breakdown not transistor-action.

For practical debuggery: check-check-check the pinout, then go ahead and pirouttte the transistor anyway, see where it comes to.

Assuming hFE=600 and nothing brokedown: base=3.3V emitter=2.7V

[fuzzy645]

Thanks....that was it, I had the pinout backwards.  FYI - when I took the base out first I did still get 1.1V as you guessed.

OK, so now it works fine, but the next thing I noticed is that it sounds awfully bright in front of the easy drive.   When I A-B it with and without the buffer out front the sound without the buffer is more "pure" to my ears. 

Thanks   

[waltk]

Diagram is fine...

Paul,
Your notes on the diagram, and explanation of the issue was one of the most clear and concise explanations I've ever seen.  Thanks for being here and answering questions like this.

[R.G.]

OK, so now it works fine, but the next thing I noticed is that it sounds awfully bright in front of the easy drive.   When I A-B it with and without the buffer out front the sound without the buffer is more "pure" to my ears.   

Human hearing happens in the brain, not in the ears.  

A good buffer prevents loading of the guitar, which prevents a certain loss of treble, other things being equal. This makes the guitar sound brighter, because less of the guitar's native treble is lost.  Some people think this is "clearer" sounding. Some people don't like it because they like the slightly "turned-down" treble without the buffer.

Unless there is some actual distortion happening - which you'd likely hear as distortion - you're hearing the increased treble, which you're not used to. That may sound clearer, or may sound less pure.

It is possible that there is something else going on, but more treble is one of the things a buffer is supposed to do.

You do say "in front of the easy drive". I don't have the easy drive schemo in my head, but if it's an overdrive/distortion, more treble can make it sound harsher. This is a quirk of how clipping distortions work. There is a description of this effect in "The Technology of the Tube Screamer".

[fuzzy645]

...more treble can make it sound harsher. This is a quirk of how clipping distortions work. There is a description of this effect in "The Technology of the Tube Screamer".

Thanks.

That is exactly what is happening.  The tone of the easy drive with the buffer in front of it is harsher, particularly because I am A-B-ing this with and without the buffer.  It sounds like the buffer is doing what its supposed to do.  The buffer definitely is increasing the input impedance for sure and that is good news   I will check out your article about the Tube Screamer.

Will the choice of transistor, or any other component help in this situation, or perhaps increasing the value of the 10 uf bypass cap??

Thanks again!

[fuzzy645]

Diagram is fine...

Paul,
Your notes on the diagram, and explanation of the issue was one of the most clear and concise explanations I've ever seen.  Thanks for being here and answering questions like this.

Agreed 100%.  Thanks to both Paul and RG for very clear, informative, detailed and very helpful answers!!!

[R.G.]

Will the choice of transistor, or any other component help in this situation, or perhaps increasing the value of the 10 uf bypass cap??

One thing buffers do is to make your circuit independent of what the input guitar/cable/etc. do to the signal. If you have a buffer and want less treble, you can always lower the treble after the buffer. One simple way is with a single-R, single C treble cut. You put the R in series after the buffer, and the C to ground, and take the signal into your easy driver (or whatever) at the junction of the R and C. The treble starts to roll off at a frequency of F = 1/ (2*3.14*R*C) where R and C are expressed in ohms and farads.

As an example, you can make it start rolling off at 2kHz by choosing R = 22K and then computing C = 1/(2*3.14*22000*2000) = 3nF. Or if you pick R = 2.2K, C = 30nF.

This gives you GREAT flexibility in getting the amount and kind of treble rolloff you want, and it not being dependent on what happened before the buffer.

[artifus]

found a nice online calc recently: http://sim.okawa-denshi.jp/en/CRlowkeisan.htm

*edit*

i'd like to echo the sentiments expressed above and also offer a huge thank you to both r.g. and prr whose posts i always read regardless of subject as i find them to always be so thoroughly and concisely informative, educational and illuminating, not to mention  how generous they are with their time and energy.

i appreciate everybody here and there are many others who deserve an honourable mention (y'all know who you are) but those two deserve some kind of award or something. thanks guys!

and seasons greetings to all x

[fuzzy645]

Will the choice of transistor, or any other component help in this situation, or perhaps increasing the value of the 10 uf bypass cap??

One thing buffers do is to make your circuit independent of what the input guitar/cable/etc. do to the signal. If you have a buffer and want less treble, you can always lower the treble after the buffer. One simple way is with a single-R, single C treble cut. You put the R in series after the buffer, and the C to ground, and take the signal into your easy driver (or whatever) at the junction of the R and C. The treble starts to roll off at a frequency of F = 1/ (2*3.14*R*C) where R and C are expressed in ohms and farads.

As an example, you can make it start rolling off at 2kHz by choosing R = 22K and then computing C = 1/(2*3.14*22000*2000) = 3nF. Or if you pick R = 2.2K, C = 30nF.

This gives you GREAT flexibility in getting the amount and kind of treble rolloff you want, and it not being dependent on what happened before the buffer.

Thanks RG.  So I can do something like this below then?  I will experiment with different values for R and C and have a listen. 

Also, I'm wondering if I should play around with the value for the input cap on the easy drive too? I have tried both .047 uf and .1 uf for the input cap on the easy drive. The .047 uf  has sounded better to my ears than the .1, but that was before the buffer was there.  Now I'm thinking with the buffer it might be best to move that back to a larger value to help with the bass response.  Thoughts??

Thanks again!

[PRR]

> it sounds awfully bright in front of the easy drive.

By conventional design standards, the Easy Drive is very F***ed-Up. It was of course f***ed-up in a way which works with guitar... but maybe/probably not for other sources.



The input impedance in the linear zone (under a few dozen mV) in the middle of the audio band is about 30K. Worked against a naked pickup with 5H inductance, treble is reduced significantly (to avoid high/harsh intermodulations).

Guitar generally has a Volume pot which masks some of the pickup inductance.

Over a few dozen mV the EZDrive overloads and input impedance is different, mostly lower. And of course that's the range of most interest (why have an overdrive if you are not over-driving it?). And this action is pretty crazy.

IMHO.... put 100K series pot between the buffer and Drive. You don't need an added hi-cut cap, the Drive has that (but not the resistor: Joe depended on the pickup). Zero resistance gives what you have: excessive bright and probably irritating after a while. Full 100K wil be dull and boring. Somewhere between may be nice.

[fuzzy645]

> it sounds awfully bright in front of the easy drive.

By conventional design standards, the Easy Drive is very F***ed-Up. It was of course f***ed-up in a way which works with guitar... but maybe/probably not for other sources.



The input impedance in the linear zone (under a few dozen mV) in the middle of the audio band is about 30K. Worked against a naked pickup with 5H inductance, treble is reduced significantly (to avoid high/harsh intermodulations).

Guitar generally has a Volume pot which masks some of the pickup inductance.

Over a few dozen mV the EZDrive overloads and input impedance is different, mostly lower. And of course that's the range of most interest (why have an overdrive if you are not over-driving it?). And this action is pretty crazy.

IMHO.... put 100K series pot between the buffer and Drive. You don't need an added hi-cut cap, the Drive has that (but not the resistor: Joe depended on the pickup). Zero resistance gives what you have: excessive bright and probably irritating after a while. Full 100K wil be dull and boring. Somewhere between may be nice.

I tried your suggestion of the 100K pot and it really worked quite well!!!  I put the 100K pot between the output of the buffer and the input of the easy drive.  Like you said, rotated one way it was bright (as if the pot wasn't there) and rotated the other way was way too dark.  I found the sweet spot in the middle and measured 26K pot resistance at that point.  I suppose a fixed 25K resistor in that spot would do the trick instead.

....oh, and I forgot to mention that I measured the NEW input impedance (as per discussion above about measurements) and although the method might not be 100% perfect, I now came up with 536K.   A far cry from the 7K I was getting earlier.

Great stuff!!! 

Thanks for all your helpl

[tempus]

You asked earlier if there was any way to make a buffer without relying on high value resistors. I made this:



and it seems to work fine. I don't know what the input Z would be, but I'm assuming it'd be the JFET resistance, which is crazy high. The 22K is there to limit potential current discharges when you plug in.

As a side note, this design seemed unreliable with my setup as I sometimes couldn't get a signal. I later added a 10M resistor from the 22K to ground and that seemed to fix the problem. It turns out it didn't, and I later tracked down the intermittent problem (bad jack on my pedalboard). I left the 10M in (I couldn't hear any increase in noise with it anyway) mostly because I've been too busy (read: lazy) to take it out.


Woops...

Forgot to add in the 9v through a 1.5K resistor to the source.

[fuzzy645]

Tempus -  thanks for sharing.

RG / PRR -  I tried the buffer tonight again with a guitar with an extremely hot pickup and the buffer went into distortion. As an experiment, I then tried replacing that 47K resistor to ground in the voltage divider section with a 50K pot so I can try varying resistance from 50K downwards, and the distortion went away completely when the pot was turned down to the 7K-8K range, even when playing hard.   I will stress that a guitar with lower output single coil pickups did not exhibit this same problem, and the 47K worked perfectly fine.

In general, are there any "hidden" disadvantages in lowering this resistor that I am not aware of?  I know that if both resistors are the same value it will cut the 9V in 1/2 to 4.5 V, but with the 2nd resistor to ground lowered to 7.5K (for example) that would bring the voltage down to only around 1.2V.   

The good news is I measured the input impedance both before and after this change (via the imperfect but still useful method) and the impedance measurements came out the same.

Thanks!

[PRR]

> replacing that 47K resistor to ground in the voltage divider...to the 7K-8K range

Doesn't make sense yet. For max headroom you want the emitter up above 4.5V, not down low.

You never reported-back with your as-working voltages. I'm guessing R.G.'s values set the emitter near 2.0V-2.5V, which may be less than max, but really should be ample for any un-crazy guitar.

[fuzzy645]

> replacing that 47K resistor to ground in the voltage divider...to the 7K-8K range

Doesn't make sense yet. For max headroom you want the emitter up above 4.5V, not down low.

You never reported-back with your as-working voltages. I'm guessing R.G.'s values set the emitter near 2.0V-2.5V, which may be less than max, but really should be ample for any un-crazy guitar.

Thanks for the info about headroom.  For a more accurate test/comparison, I swapped the 50K pot out for a switch to toggle between a fixed 10K vs. fixed 47K and then I recorded all voltages:

As original with the 47K resistor here are the voltages:
* battery 8.9V
* collector 8.9V
* base 2.9V
* emitter 2.5V
* at the junction of the voltage divider 4.4V

With the 10K resistor instead
* battery 8.9V
* collector 8.9V
* base 1.15V
* emitter .65V
* at the junction of the voltage divider 1.6V

I'm almost wondering if the problem is that this particular pickup is way over the top too hot (a DiMarzio X2N).  When I used a different guitar with a typical 6K strat pickup, and another with a fairly hot P90 pickup (in the 10K range) all was good as originally planned with that 47K resistor to ground at the voltage divider.  Its just that darn DiMarzio X2N sent the buffer into distortion unless I lowered the voltage into the base by using a smaller resistor to ground at the voltage divider. 

My fear is that voicing it for this "over the top" X2N might sacrifice tone for my other guitar which have more normal pickups.