Keeley Parallel Mixer

[gaussmarkov]

On http://www.muzique.com/home/gbook.htm I just came across a parallel mixer project by Robert Keeley: http://www.robertkeeley.com/manuals/Keeley_Parallel_Mixer.pdf
I was surprised to find no references to it here. Has anyone built this?  Apparently Robert Keeley Electronics is also going to supply them soon.

[amz-fx]

Although it was announced, I'm not sure he is coming out with it...  we spoke at length on Saturday and this product was not mentioned...  he has some extremely cool products planned for the near future inclduing some things you can't really d-i-y effectively

regards, Jack

[gaussmarkov]

Although it was announced, I'm not sure he is coming out with it...  we spoke at length on Saturday and this product was not mentioned...  he has some extremely cool products planned for the near future inclduing some things you can't really d-i-y effectively

regards, Jack

Thanks, Jack.  It did seem that the announcement was old.  

I have been interested in making such a pedal.  J. D. Sleep has talked about designing something like this on his site, under the project name The Paralleyzer.  For those interested see
http://generalguitargadgets.com/index.php?option=content&task=view&id=94&Itemid=122

Having only made an A/B box, I am puzzled by a few things on the schematic.   Will someone shed some light?

What does the coment "Open is high 2" mean for SW1?  

In RK's comments, he says that he chose the TL072 for the IC but the schematic shows the LF412.  Use either?[/list:u]

[B Tremblay]

Splitter-Blend

[gaussmarkov]

Splitter-Blend

terse.  :shock:  but a google search reveals ... the splitter-blend at runoffgroove.com:  http://www.runoffgroove.com/splitter-blend.html
and http://www.runoffgroove.com/sb2pcb.pdf.  thanks, man!

google also gives the bblender by sean maclennan http://www.seanm.ca/stomp/bblender.html who links to the buff'n blend signal mixing circuit by joel purkiss http://www.geocities.com/jrtookmyfalseteeth/

anything else? :?

[seanm]

What do you want the blender to do? My B. Blender was an attempt to make a general purpose tool to mix an effect with the dry signal. The splitter blender and Keeley's are for mixing two effects, although you could just short out one side for the dry.

The Paralleyzer looks more like a simplified mixer.

Really it all comes down to some buffers to prevent feedback and possibly add some gain and/or attenuation. You could go wild with tone controls and the like.

The more detail about what you want the blender for, the more help we can be.

[Mark Hammer]

While it is obviously useful, I suspect the reason why you've seen no mention of it here is because it simply combines a couple of things in one circuit that have been around forever.  I encourage all of you to make yourselves a splitter/mixer/invertor combo and experience the mysteries and joys of parallel processing.  Anderton introduced us to it more than 25 years ago, and I haven't looked back since.

[gaussmarkov]

What do you want the blender to do? My B. Blender was an attempt to make a general purpose tool to mix an effect with the dry signal. The splitter blender and Keeley's are for mixing two effects, although you could just short out one side for the dry.

The Paralleyzer looks more like a simplified mixer.

Really it all comes down to some buffers to prevent feedback and possibly add some gain and/or attenuation. You could go wild with tone controls and the like.

The more detail about what you want the blender for, the more help we can be.

hey sean!  thanks for dropping in.  i got interested in building the kind of tool that you designed when i thought about generalizing on voodoo lab's sparkle drive.  i suppose many others have had the same idea.  robert keeley certainly talks the same way in his description of his parallel mixer.

i wasn't after tone control or boost/attenuation.  just a dry/wet mix with bypass in a single stompbox.  seems like having buffers on both channels would add flexibility to the tool.  like mark hammer says, parallel is worth exploring.  i have been limited to effects in series.  boost/attenuation when the signal is mixed seems like a sensible mod, too.

did you ever make joel purkiss' buff'n blend?  how about your 90% b.blend?  thanks for your comments about the purpose of the buffers.  gives me some motivation to read up on buffers.  i haven't found anything on in jack orman's buffers stuff about preventing feedback.  i'll comb through this forum.

[gaussmarkov]

While it is obviously useful, I suspect the reason why you've seen no mention of it here is because it simply combines a couple of things in one circuit that have been around forever.  I encourage all of you to make yourselves a splitter/mixer/invertor combo and experience the mysteries and joys of parallel processing.  Anderton introduced us to it more than 25 years ago, and I haven't looked back since.

more than encouraged, i'm stoked!    lessee.  25 years.  that's a long time.  when was the first time you remember the queen visiting canada? :wink:

[gaussmarkov]

Splitter-Blend

oops.  i did not realize that was a link, even though i wondered how you got blue letters.  :oops:   i'll stop posting those long url's and use your example.

[seanm]

did you ever make joel purkiss' buff'n blend?  how about your 90% b.blend?  thanks for your comments about the purpose of the buffers.  gives me some motivation to read up on buffers.  i haven't found anything on in jack orman's buffers stuff about preventing feedback.  i'll comb through this forum.

I never made the buff'n blend, been too busy with distortion effects

I breadboarded the 90% blend just to make sure it works. It basically assumes that the effect is fully buffered or that the effect does not like a buffered input. The b. blender guarantees a high input Z and a low output Z, the 90% has the input and output Z of the effect, which may not be a bad thing. The effect may want to have a low Z input to overload the pickups on purpose.

You need the buffer in the 90% blend or you will get feedback since you effectively have the output directly connected to the input. I don't think Jack ever mentions this, but it is a common use for buffers.

I breadboarded the sparkle drive and really like how the clean knob works. I am thinking of buying one (I like to support the small effects makers when I can).  A blender could easily replace the clean knob on the sparkle drive, but all you have done is removed the dual-ganged pot (warning: you still need 100k to +4.5 off pin 5 of the 4558) and now have an unused opamp on the 5532.

Unless you wanted to *always* use the effect with the blender and never put effects in series. For example, in the sparkle drive you could drop the input buffer, drop the 5532 completely. Add gain to the output buffer from the return and you have a working pedal. I test designs all the time like this. Warning though, I think some of the sparkle of the sparkle drive comes from the 5532. You might want to build the blender with this chip.

There are disadvantages to generic blenders, like the b. blender. If the output of the effect is too hot, the output buffers can distort. This could be fixed with something like the AMZ superbuffer or by running the b. blender off 18V (it can handle it).

A second disadvantage is that if the effect has a lot of phase shift (not a reversal, that can be corrected) the blended result can sound like two seperate streams rather than one coherent stream.

[MartyMart]

The splitter blender works great on a chorus/flanger pedal, I always find
them too "wet" and this lets you "blend in" an amout of "dry" signal
A fabulous idea which also works well with distortion, but that wasn't my
reason for building one !

Kudos once again to ROG

Marty.

[Mark Hammer]

lessee.  25 years.  that's a long time.  when was the first time you remember the queen visiting canada? :wink:

Funny story, there.  Some time in the very early 60's, her Majesty came to our city, and I went with my parents downtown to Parliament Hill to view the procession.  A rather impatient (and not especially tall) lad, I started moving through the crowd, trying to get a better view.  I eventually found myself behind the official viewing stand that had been erected by the gates near the entrance on Wellington Street.  Young enough to be thrilled by such events I began waving madly as the open coach approached and the ever present crown and Admiral cap could be seen.  It suddenly dawned on me that I was waving through the space between someone's legs up on the stand.  I looked up and it was our Prime Minister, Lester Pearson!

Yes folks, I get to say I greeted Queen Elizabeth from between Lester Pearson's legs!! :lol:  :lol:   Then there was the time we were moving between apartments in Victoria, and I arrived at an intersection in the car, chock full of boxes, at precisely the moment that Prince Charles and Princess Diana (in the everpresent pillbox hat and polka dot dress) came gliding by in their limo on the way to the governor general's residence and we exchanged waves.  Um, just so we're clear, I wasn't between the legs of anyone famous at that moment.  I mean, it was an intersection for heaven's sake, and a small car.

All true.  Of course, the days of a person of ANY age being able to elbow their way up behind a viewing stand of dignitaries (and of parents remaining calm when their kid has momentarily disappeared in a crowd) are pretty much long gone.

[gaussmarkov]

lessee.  25 years.  that's a long time.  when was the first time you remember the queen visiting canada? :wink:

Funny story, there.  ... I looked up and it was our Prime Minister, Lester Pearson!

Lester Pearson!  I remember him, too, though we were never that close.  :twisted:  That was back in the glory days of the Leafs!

In case anyone has a moment, I am still interested in the answer to an earlier question:  

What does the coment "Open is high 2" mean for SW1?[/list:u]

[Mark Hammer]

What a difference a font makes! :lol:

It says "Open is high Z", not "high 2".  Easy mistake to make, and I would have made it too if not for the conspicuousness of its location.  "Z" is frequently used as a symbol for impedance, as in Lo-Z and Hi-Z mics.  I don't know what the origins of the symbol are.

The idea is that if the signal input comes from a low impedance source, you can skip the input cap by shorting it via the switch.  If the source is high impedance, open the switch to run the signal through the cap.  Craig Anderton used to use two input jacks in some of his projects, one with cap and one without.  This version takes up less space.

[gaussmarkov]

did you ever make joel purkiss' buff'n blend?  how about your 90% b.blend?  thanks for your comments about the purpose of the buffers.  gives me some motivation to read up on buffers.  i haven't found anything on in jack orman's buffers stuff about preventing feedback.  i'll comb through this forum.

I breadboarded the 90% blend just to make sure it works. It basically assumes that the effect is fully buffered or that the effect does not like a buffered input. The b. blender guarantees a high input Z and a low output Z, the 90% has the input and output Z of the effect, which may not be a bad thing. The effect may want to have a low Z input to overload the pickups on purpose.

You need the buffer in the 90% blend or you will get feedback since you effectively have the output directly connected to the input. I don't think Jack ever mentions this, but it is a common use for buffers.

thanks for all of this, sean.  can you (or anyone else) recommend some reading?  there was a texas instruments doc ref'd on your original b.blend thread.  but that looks like it's a little over my head.  :cry:  I just figured out that Z is a kind of impedance in trying to understand your comment. :shock: Jack Orman's Basic Buffers has been helpful, and every time I read it I realize how much I missed the previous times.  :?  (B Tremblay: note the cool link in my post. )

I breadboarded the sparkle drive and really like how the clean knob works. I am thinking of buying one (I like to support the small effects makers when I can).  A blender could easily replace the clean knob on the sparkle drive, but all you have done is removed the dual-ganged pot (warning: you still need 100k to +4.5 off pin 5 of the 4558) and now have an unused opamp on the 5532.

Unless you wanted to *always* use the effect with the blender and never put effects in series. For example, in the sparkle drive you could drop the input buffer, drop the 5532 completely. Add gain to the output buffer from the return and you have a working pedal. I test designs all the time like this. Warning though, I think some of the sparkle of the sparkle drive comes from the 5532. You might want to build the blender with this chip.

no.  i don't want to dedicate the blender to an effect.  i want it as a general tool.  

huh.  the "dual-ganged" pot on the sparkle drive.  i was wondering about that, too.  i've seen a schem and scratched my head about that part of it.  there were no words to go on but now i know what that is.  thanks!

There are disadvantages to generic blenders, like the b. blender. If the output of the effect is too hot, the output buffers can distort. This could be fixed with something like the AMZ superbuffer or by running the b. blender off 18V (it can handle it).

yeah, i was going to ask about this because rk makes the same comment and opted for 18v in his design.  i guess one could just plug in that part of his design into the b.blender?

[seanm]

yeah, i was going to ask about this because rk makes the same comment and opted for 18v in his design.  i guess one could just plug in that part of his design into the b.blender?

You could. Or you could just put the two batteries in series to replace the 9v battery. The TL072 handles 18v no problem. The divider for Vr (that's the two resistors and the cap in the bottom left) will produce the correct +9 for the reference.

Make sure any electrolytic caps you use are *at least* 25V if you want to use 18V. 50V caps would be even better.

I notice there is a small bug in my schematic. I always run a 100uF cap between + and gnd. I would make sure this one was 50V just in case.

[gaussmarkov]

I have learned much more about op amps and transistors in the last few days.   Great stuff.  Now I understand a lot more of R.G. Keen's analysis of the tubescreamer.  I feel like I'm making some progress.  I hope three more questions will fetch answers.  These could be new threads but I'll try them here first.

So, as I understand it right now, a unity gain op amp (or "buffer" or "voltage follower") doesn't amplify at all--it's all about impedance.  For a high impedance input like a guitar pickup, the signal picks up fewer "errors" if it goes from a high impedance input to a low impedance output than it would going into a low impedance input and out a low impedance output.  (Mark Hammer mentioned in another thread that low impedance inputs are a good thing for the traditional fuzz box.  But let's put that aside for now.)

First question: What is the effect of running low impedance into high impedance?  After all, when we put a series of stompboxes together this will often happen.  Like two tubescreamers in series.  I guess this is this the topic of cascading buffers?

Second question: Why is there a transistor in front of the polarity switch in the Splitter-Blend?  I don't see this component in the polarity switch mod for the B.Blender.

Third question: In these blenders, Sean and B Tremblay are using ordinary pots to mix the dry and wet signals.  As Sean pointed out, the Sparkle Drive uses a dual-ganged pot.  How would you choose between these options?

Thanks to all who respond.  "High 5" to Mark for explaining "high 2"!

[seanm]

Second question: Why is there a transistor in front of the polarity switch in the Splitter-Blend?  I don't see this component in the polarity switch mod for the B.Blender.

I think it is a simple buffer to make sure the input impedance is always high. After looking at my mod again, I am suspicious that when the switch is closed, the input impedance is very low. I have not tested this mod extensively since I have never needed it in practice.

[seanm]

Second question: Why is there a transistor in front of the polarity switch in the Splitter-Blend?  I don't see this component in the polarity switch mod for the B.Blender.

I think it is a simple buffer to make sure the input impedance is always high. After looking at my mod again, I am suspicious that when the switch is closed, the input impedance is very low. I have not tested this mod extensively since I have never needed it in practice.

Ok, you *need* the buffer. So I have removed the polarity reverser from my web site. My initial tests worked due to a flaw in the testing.

This discussion is great since it is making me think more about the designs. Which leads me to take a stab at the third question.

The Sparkle Drive seems to have different resistor values on the dry and wet paths before the final output stage. Both variable resistors are 100k. I think we can ignore the buffer since it is just there to stop feedback and is unity gain. So the dry path goes through a fixed 13.3k resistor and the wet path through a fixed 100k.

I am guessing here, but I would guess that the two different values compensate for the higher gain of the distort/tone stages over the dry stage.

With the blender designs, we have no idea what is going to be placed in the send/return loop (in fact it could be nothing), so a simple blend pot is simpler.

While I love my blender, a well designed blend added to an effect can actually work better since you have more control over where and how you place the blend.