How to get more upper octave from a fender blender...

[chicago_mike]

Hey guys,

I have a Fender Blender reissue that I use for bass. SO im already going to try different cap values and replace the ceramics with wimas or even some solens I have lying around.

But I would also like a little more upper octave effect. Can I do this by adding more ge diodes in the clipping stage?

[Mark Hammer]

No.

The octave is generated by that 3rd transistor stage that has an 8k2 resistor on the emitter and collector.  That configuration is referred to as a "phase splitter" insomuch as it generates equal amplitude (but opposite phase) outputs at the emitter and collector.  Those two equal outputs pass through the cap and diodes immediately after them.  The diodes block one half cycle of the signal, and pass the other.  When those two remaining halves are recombined, in equal measure, the result is a doubling of frequency, or at least what is heard as a doubling of frequency.  Nothing is actually doubled.

Technically, the 8k2 resistors should produce equal amplitude complementary outputs, but of course as 5% tolerance resistors they might not.  Both of those diodes should have the same forward voltage, but of course there will be some variation there too.  This is why you'll see things like JC Maillet's "nulling" mod for the Green Ringer. The GR uses the identical split/shave/recombine approach to producing octave-up, but permits the insertion of an adjustable resistance in one half of what gets recombined to produce/fine-tune an equal balance and more robust octave up.

What can be done in the case of the Blender?  I suppose one thing you can do is to measure the forward voltage of those two series diodes and assure they are within range of each other.  YMMV, bt I find I can easily get up to 60mv difference in forward voltage between Si diodes of the same batch (e.g., when you buy a bunch of 50 on one of those paper spools).  Ideally, you want them to have the same forward voltage, so aim for withn 5mv of each other.  As well, you might want to consider a fixed-resistor plus trimpot on the emitter leg of the transistor to achieve equal-amplitude emitter and collector outputs.  Indeed, Bernie Hutchins (demi-god of synthesizers) wrote in Electronotes over 30 years ago that it is sometimes preferable to have the emitter resistor juuuuuuuust a shade lower than the collector resistor for equal-amplitude inverted outputs.  So, a 6k8 fixed resistor and 5k or 2k trimmer might not be such a bad idea.

So what role does that clipping pair nearest the output play.  First, they impose some dynamic constraints such that the signal stays somewhat compressed.  Second, they keep the harmonic content cranked up so that whether you get a discernible octave or not, things will sound "busy" and forceful.  Adding more diodes will raise the ceiling on the signal.  With a higher clipping threshold, you'll have much more output level to be sure, but less harmonic content.  That might not be such a bad thing for bass, but I guess you'll have to try it and see.  My experience with removing such diodes entirely (i.e., eliminating/lifting that path to ground) is that you will still have audible distortion even without them, although you'll get noticeably more with them.  The resulting tone just might work well for bass.  I guess the only thing to do is try it and see.  Alternatively, you could try a compromise and use a 2+2 diode complement.

[chicago_mike]

Thats some good info to chew on. THanks Mark!

[amptramp]

This page can be a gold mine (or a land mine):

http://www.falstad.com/fourier/e-fullrect.html

This shows the harmonic content of a fullwave rectified sine wave.  If you drag the cursor over the "knobs" on either of the bottom two lines which indicate the harmonic content and phase, you will see the harmonic superimposed on the rectified sinewave at the correct amplitude.  Any imbalance will allow the fundamental to get through, so Mark's suggestion to add a balance control is a good idea.  Note that the output impedance of a transistor is much higher at the collector than the emitter, so an inverting and non-inverting op amp may provide outputs that are easier to balance and low output impedance, so that would no longer be an issue.

The previous page shows halfwave rectification which has a pronounced fundamental.  Combine fullwave with a small amount of halfwave and you will get what you have with an imbalance.

[johngreene]

When those two remaining halves are recombined, in equal measure, the result is a doubling of frequency, or at least what is heard as a doubling of frequency.  Nothing is actually doubled.

Correct, it is actually squared.

Technically, the 8k2 resistors should produce equal amplitude complementary outputs, but of course as 5% tolerance resistors they might not.  Both of those diodes should have the same forward voltage, but of course there will be some variation there too.  This is why you'll see things like JC Maillet's "nulling" mod for the Green Ringer. The GR uses the identical split/shave/recombine approach to producing octave-up, but permits the insertion of an adjustable resistance in one half of what gets recombined to produce/fine-tune an equal balance and more robust octave up.

which is why double-balanced mixers use matched diode rings.

What can be done in the case of the Blender?  I suppose one thing you can do is to measure the forward voltage of those two series diodes and assure they are within range of each other.  YMMV, bt I find I can easily get up to 60mv difference in forward voltage between Si diodes of the same batch (e.g., when you buy a bunch of 50 on one of those paper spools).  Ideally, you want them to have the same forward voltage, so aim for withn 5mv of each other.  As well, you might want to consider a fixed-resistor plus trimpot on the emitter leg of the transistor to achieve equal-amplitude emitter and collector outputs.  Indeed, Bernie Hutchins (demi-god of synthesizers) wrote in Electronotes over 30 years ago that it is sometimes preferable to have the emitter resistor juuuuuuuust a shade lower than the collector resistor for equal-amplitude inverted outputs.  So, a 6k8 fixed resistor and 5k or 2k trimmer might not be such a bad idea.

You should be able to compensate for the mis-match in diodes with the trimpot.

So what role does that clipping pair nearest the output play.  First, they impose some dynamic constraints such that the signal stays somewhat compressed.  Second, they keep the harmonic content cranked up so that whether you get a discernible octave or not, things will sound "busy" and forceful.  Adding more diodes will raise the ceiling on the signal.  With a higher clipping threshold, you'll have much more output level to be sure, but less harmonic content.  That might not be such a bad thing for bass, but I guess you'll have to try it and see.  My experience with removing such diodes entirely (i.e., eliminating/lifting that path to ground) is that you will still have audible distortion even without them, although you'll get noticeably more with them.  The resulting tone just might work well for bass.  I guess the only thing to do is try it and see.  Alternatively, you could try a compromise and use a 2+2 diode complement.

Because those diodes are clipping an octave signal, it will generate many even order harmonics. So you are not just getting 'an' octave, but pretty much every octave on up. This is very close to how comb generators are designed. If you primarily want the second octave, remove those clipping diodes. Then you have something more similar to an Octavia.