LFO Output - Buffer

[Unlikekurt]

Hello all.
I'm working on trying to fine tune an LFO for a project I'm working on.  I'm quite happy with the LFO itself, the output is quite robust, the waveform is very smooth and the range of adjustment can be dialed into what I need with the right RC arrangement. 
The issue I'm having is that at the output of the buffer there is too much signal loss and the distortion is also somewhat undesirable; albeit useable.
I drew up a quick schematic.  "NUMBER ONE" shows the buffer as it stands.  "NUMBER TWO" shows something I threw together as a quick trial to see if I could get better results if I decoupled before the emitter follower to see if I could get a better bias and also not dropping signal level through that 100K off of the output of the oscillator.  The resistor selections were pretty arbitrary for the divider.
That being said, does anyone have some advice for a better approach?
Thanks

[antonis]

and also not dropping signal level through that 100K off of the output of the oscillator.

On No.1, LFO out total load is 4k7//(100k + h_FE x 4k7)
On No2,   //          //      //    4k7//(100k//100k//h_FE x 4k7)
(ignoring intrinsic Emitter resistor for buffers and small-signal output resistance for LFO..)

Despite Base resistor value, No.2 always loads LFO Collector much more than No.1..
(due to voltage divider biasing resistors appearence in parallel with Emitter reflected reistance..)

Try a lower enough Base resistor value (or even direct coupling) for No.1..
(as it is, you should have about 1V signal voltage drop for buffer BJT h_FE=100, 500mV for hFE=200, etc..)

Or proceed with an FET Source follower..

[bool]

Delete the R5/100k in front of your buffer? Or drop the value to something like 1k.

[PRR]

What is the buffer driving? If a low-Z load needing high level, none of these give much buffering. (The emitter resistor on Q2 is no lower than the collector resistor on Q1.)

[diffeq]

That being said, does anyone have some advice for a better approach?

Rail-to-rail op-amp buffer? Has the same footprint as 2 resistors, 1 cap, 1 BJT combo.

[Eb7+9]

... use circuit #1 and get rid of R5

crude ac analysis gives the following:

4k7 collector load of osc Q1 sees 470k assuming Q2 beta 100, in which case Zout becomes 47 ohms

2n3904 for Q2 should surpass those estimates, but that's ac analysis

circuit operates in large-signal mode so those estimates aren't static, still this is probably the best you can do using two bipolar devices

[garcho]

+1 for op amp buffer, low parts count, no math

[antonis]

What is the buffer driving? If a low-Z load needing high level, none of these give much buffering. (The emitter resistor on Q2 is no lower than the collector resistor on Q1.)

@Unlikekurt: take the above into account.. 

Q2 Emitter resistor is reflected on Q1 Collector h_FE +1 times its actuall resistance..
A Q1 directly drive of low Z value load should considerably lower Q1 stage gain..
(resulting into a not so smooth oscillator waveform or even into a just amplified ramp..)

Now,  buffer is called to do the current drive dirty job but with the limitation of load Z value marginally equal (for an ideal Vcc/2 Emitter bias) or higher than Emitter resistor value..

So, depending on your specific load current need, even a single op-amp buffer might be inadequate..

[Eb7+9]

see if this works ...

[Unlikekurt]

Thanks to all who chimes in.  I no longer believe the buffer to be my issue.
I'm finding that as I get toward the upper limit of the oscillator the amplitude drops off (as one would expect).  So I'm adjusting the resistor feeding the base of q2 to try to compensate.  However then when the frequency is decreased I end up clipping the rail.
I'm thinking I need to have a wider range of adjustment than this configuration will allow.  I hope to achieve 2-7.5hz or so.
Would adding another shunt resistor in the rc network open up the range?
Also, I'm very much considering switching to an opamp.  One dual and I'd get the oscillator and the buffer.
That said I'm still going to try out Eb's work up to see how it improves.

[PRR]

To get wide range from a CRCRCR network you need to vary all three resistors.

3-gang pots are rare. We vary just one. We can only get (at best) the third-root of the permissible change of resistance. Often the circuit will not tolerate 10:1 change of R. This leads to a 2.15:1 range of frequency. You are asking more than twice that. A 2-gang pot is worth a try. 

[Rob Strand]

I suspect the problem with the circuit is the low R4 value.   You probably need to use a higher R4 value then using smaller cap to bring down the frequency to what it was before.

Also your R4 value goes to a short, you really need to limit the minimum value.

EDIT:
Forgot to mention, with the buffer you might find connecting your C3 to the output of the buffer also helps.

[antonis]

Forgot to mention, with the buffer you might find connecting your C3 to the output of the buffer also helps.

Nice catch, Rob..!! 
(bootstarp ramp generator..)

[Eb7+9]

SPICE suggests sustained and stable 1.5Hz to 8.1Hz speed range using the following single-pot circuit and values ...






which I loosely derived from the following Fender Custom Shop Vibrolux Tremolo oscillator:




www.lynx.net/tremoloCircuits.html



you may have to use film caps for the smaller ones

and, to be clear you’d be replacing Rspeed by a 250k pot and 2k2 resistor wired in series

of course, you’ll be the first to try this out
good luck

[amz-fx]

Fixed link:

http://lynx.net/~jc/tremoloCircuits.html

regards, Jack

[Unlikekurt]

Eb!  Wow!
This is getting tested out this week.
I'll report back.
Thanks.

[Unlikekurt]

Have it up on the breadboard right now.
The output level and frequency are pretty unstable.  At higher rates it'll swing between 3.5Hz and 7Hz every couple of cycles.
Level is also quite low.
I'm feeding the scope only at this point.

[Eb7+9]

Sorry I'm on the road and don't have access to any of my electronics stuffy atm ... I'd try swapping the active devices by other 2n3904's first (their Beta values certainly would play a role here) ... the sims were a little finicky to get going ... so, measure your resistors and caps if you can and make sure they are all as close as possible to the sim values ... it may take a bit of fudging to get the real version going - but it n principle there is potential there ... while I'm here I'll state that the circuit was particularly sensitive to R4 value (1 Meg) ... I'd try playing with that value as well

Best of luck!

Ps. Thx to Jack for fixing that link ...

[Unlikekurt]

Went back and checked the components on the breadboard, wasn't able to get that circuit to oscillate.
I've since moved onto a CMOS oscillator, I believe from anotherjim, in this thread:
https://www.diystompboxes.com/smfforum/index.php?topic=107962.0

It's providing pretty good results.  I was able to dial in a little under 2Hz to 6.5hz with a 1M potentiometer.
There is a loss of level as the rate is decreased however.