dyna red distortion very low volume

[Steben]

as said: jFETs are finicky. You could indeed try lets say 10 different 2N5458's and they will self bias differently.

[Vivek]

What is the effect of the link from after the shunt diodes to the inverting pin of Opamp ?

Is this frequency dependent due to proper choice of cap in that link ?

Is this similar to what Sweet Honey and Klon do ? or is Klon feed forward and this one is feedback ?

[BJF]

Hi There,


Yes 2N5458 like most J Fets have a pretty wide tolerance and would typically bias in a self biasing buffer like this between 1V2 to 3V5, 2N5457 on the other hand typically form 0V3 to 1V7
It's basically same transistor presorted for IDSS range and ugs and marked 2N5457, 2N5458,2N5459 accordingly

Yes I fine sort my J fets so I I can stay within  a few 100mV's deviation in typical circuits.


As sourcefollower exact bias point might not be critical as long as the expected signal peaks are within dynamic range and as mentioned in previous post one can access a nonlinerity at the end ugs range

Filter at the output changes dynamically as impedance at the shunt diodes goes down when they conduct ........but assuming they do conduct then filter has one one hand a limit at 2KHz with a slope of -6dB / octave and
in addition to this treble control either at minimum changes ft to 700Hz or at any other setting creates a staircase filter. Now the 220pF adds to speed or angle of slope by returning frequencies high enough to invertinginput
and thus null them through negative feedback.

Yes the value  can be tuned within a span but care must be taken to avoid oscillation and so with bigger values a series resistor may be needed

It was a way of reducing a certain high frequency experienced when using Celestion Vintage 30 speakers.

The composite filters used in Honey Bee are more elaborate but they serve similar function reducing certain high frequency content. This really because the guy helping me with that design said time and time again as I tuned filters
" remove that frequency-it sounds like distortion box"

The idea of this type of filtering was shown to me by a fellow technician as he showed a technical approach to reduce and continously control slew rate in OP amps by external dominant roll off
Like this if you have an feedforward compensate LM 301 and after that connect a singel pole LP filter and then buffer that say with a LM 310 and from the output of LM 310  bring negative feedback to LM301 inverting input
This way a dominant pole is created within the feedback loop and can be continously controlled to set slewrates lower than that of the feedforward mode. The LM301 will remain stable as the new dominat roll off is introduced.

Now slew rate is defined as the accuracy with which the output can swing i .e. how close it can reproduce a perfect square wave

If the application is to create a waveform with less overtones than a squarewave from a square wave a slewrate limiting technique could be used

It follows that this would be similar to low pass filtering but in that only similar

However it does suggest possibilities of filtering techniques to control the production of overtones

There is a circuit using two diodes and three resistors and a selected J Fet that can be used to make a sine wave out of a square wave but only at a defined signal level and the function is that the J Fet cannot open nor close fast enough
combined with a nonlinerity at the edges of trasfer and thus it chops off the edges symetrically and at precise signal level and optimum selection of J Fet the input square wave becomes a fine sine
Unfortunately this is only possible within a narrow band and not at a wide dynamic range.

It would however give food for thoughts on possibilities of nonlinear amplifiers that could do this under a fairly wide dynamic range.



At your service
BJ
BJF Electronics
Sweden

[Vivek]

Dear BJ, I thank you for your kind reply

Here is an article on your pedal that I found last year when researching it

https://squidguitarinc.wordpress.com/2020/01/07/bjfe-honey-bee-od-origin-story/

[sopapo]

Hello, I just build this pedal, it sound very nice, i think its a real good vision of a hotroddedplexi-jcm800, and the response its very amp like (the hype its true¡¡) it does the low gain marshall thing very good and it clean wonderfully with the vol pot, something not so common in a pedal....kudos to Bjorn¡¡¡
I built this vero layout


I have doubts in the bias of the jfet, My readings are. g=0 s=1,84 d=8,64, its correct this biasing?? how can I to fix this biasing?? I notice that the level isnot quite on par with other pedals ,, I have it maxed. Its maybe because of the biasing???



Also I noticed that the 1 uf caps that are used in coupling are electrolytics and the 0,47uf in the filer too, there is a specific reason for this??
these values are generally used in poly-metallic film caps, at least until 1 uf, its un common to see in a pedal a 1uf electrolytic in signal path, its very affordable and generally better to use an mkt or similar...Maybe the electros add "something" to the sound or its better for other reasons??
Thanks for your time
Best regards

[antonis]

JFET is an output buffer so I'd place another 1M resistor between Gate and +E..
(10k Source resistor could be lowered, optionally..)

[sopapo]

Thanks for your answer Antonis, so do you think the buffer have to be biased at vcc/2 right? I Will try this.
Bjorn, the designer Talk in this very same topic, that this buffer dont have to be biased this way, he say that its in a selfbiased configuration and in this Mode the bias is in the Range of 1'2 to 1'6 v that is enough for the jfet to work on buffer Mode if i have understanded right
Al so he talk about certain non linearities that add something tonthe sound, what do you think of all of this??
If i bias at 4'5v It Will change the sound ir its depende of the jfet??
Thanks for all

[antonis]

Thanks for your answer Antonis, so do you think the buffer have to be biased at vcc/2 right?

Not necessarily, but I'd feel more comfortable with a Source follower biased higher enough than half of signal amplitude together with a lower value Source resistor, for current sinking abillity..

As it is and with your particular measurements, JFET can easily swing up to 2.5V (for a signal of 700mV amplitude) but its downwards swinging ability is strongly restricted from its Source resistor value and next effect input impedance..
(it might result into negative waveform clipping, hence output lower volume/level - exept of "extra" asymmetrical distortion..)

[sopapo]

Ok, maybe this extra assymetrical distortion is what he refers as "non linearities"....

[PRR]

The output level is limited by the 1N4001 diodes. You can never get over 0.6V here.

To confirm my wild idea: snip those diodes. Now the limit is the 1.6V of the LEDs.

That may be LOUD (I assume nobody would use the Volume knob). Try 2 diodes each way (4 diodes total), still in the 1N400X family, for 1.2V limit.

[antonis]

Or, just for curiosity delight, configure JFET as CS amp..
(you can do it by replacing Drain jumper with a resistor and drag 1μF cap into Drain..)
Of course, output signal should exhibit phase reversal..

[BJF]

Hi there,


This answer is long over due but here goes

@Vivek

Yes that is a good article. It is now over 20 years since introduction. I have made a DIY kit in conjunction with moody sounds
For those interested in building by themselves. There is also a podcast that goes in deep detail at guitar geeks podcast
#301 Honey Bee 20 years that also goes into technical details


@Sopapo
Thank you for the kind words.
Your biasing will be just fine at source resting at 1V84. Max signal at gate will be 1V peak to peak square wave.
The diodes used 1N4007 has at the currents involved between 500 and 550mV forward voltage
Now the grid space or gate space is one half of a 2nd degree quadrant function and the AC signal must fit into this.
The grid Voltage to Drain Source current is not linear as described above but to the edges transfer bends off and this causes nonlinerity
The negative flank stretches farther than the positive so the vital point here is the resting for working point at Source.
Lowest limit is around 1V2 at Source. In the region of 1V2 and 1V6 there is a slight distortion caused by the gate to Source relation.
Actually this is so but to a lesser degree at resting points lower than 4V due to the 2 quadrant behaviour. So really lowest tolerable at 1V2.
Yes this depends heavily on the J Fet used. The point is to slightly shall we say sand off the egdes of the square wave.

This circuit is meant for battery operation and the buffer is also used to isolate the filter from volume control or volume control would alter the slope of the filter . The low current results in a drive capability of a few K Ohms but it is not required to drive such a load typically.
Voltage range is 5V to 15V and towards the lower and internal bias systems of the OP amp will collapse and cause additional distortion which some enjoy as added complexity. Yes the transistor can be wired as common source amplifier for increased output  ( absolute phase angle is of no real concern in series and mono) but this would affect Voltage range. The buffer remains fairly constant over the range.

With J Fet buffers the output drive capability is very much dependent on the cross current through the source load ( meaning the resting current)
One could use a source current of a couple of mA and bias the J Fet with a voltage divider at gate to have Source rest at half the supply voltage and in the process minimise difference between transistors used and distortion levels would be lowered.

I answered the about the Elytic capacitors in a Japanese publication the Effector book some ten years ago but that is in Japanese
The interviewer was concerned about the brand of capacitors.
Right so these are miniature electrolytics that are actually close to tantalums used in the 70s. They cannot be reversed polarized or they will become a totally different component as in part a zener diode. They can also process excellently AC voltage as long as they are DC polarized and AC voltage is below 1/10th of Working voltage and below the bias across the capacitor. These E lyrics are 50V types and AC voltage through them is about 4V max. These are considerably smaller than plastic type caps and both building height and circuit board real estate is important in the production size

Yes I have a soft spot for fine capacitors on an emotional level and yes I collect capacitors but practically from pure technical performance that is another matter. Also in this case there may be a tiny distortion contribution that is likeable.

Historically in guitar effects and musical electronics in general through 60s to 90s budget elytics were commonly used and as repairman I can say break down of those budget caps is most common. Those older elytics could however tolerate reverse bias sometimes up to 30% of Working Voltage which is an important consideration when replacing these and it can be necessary to replace with plastic types.

At your service
BJ
Bjorn Juhl
BJF Electronics
Sweden