The tardis - modded blue box

[Bastis]

Hey guys!

I'm working on a setup of pedals for saxophone. The blue box worked really well with the sax, but I wanted to be able to play in 4ths or 5ths, so I decided to add the CD4017 rocktave mod to the blue box. I wanted to add a clean blend pot as well, and it's working pretty well. I couldn't fit it all on my breadboard, so I soldered the blue box to a vero with a couple of modifications and connected it to my mod on the breadboard. It worked pretty well, but the square wave that I tapped from the blue box to send to the cd4017 oscillates (or it might be some kind of feedback issue) quite easily when I connect the saxophone mic. Using a noise gate helped it a bit, but since I have no oscillation problems with the blue box it should be fixable.

I play through a wireless system with line level output.

I've changed a few things in the schematic from the version I have on my vero board, and I figured I'd run it by you guys before trying more stuff on my own. I've been building pedals for like 6 months, and had very little knowledge prior to that, so keep that in mind when lecturing me on the mistakes I've made.

I've added buffers before splitting signals and it seems like the signal needed to be buffered before the blend pots to not interfere with each other, I might have added more buffers than needed though. The LP filter on the top right hand is to improve tracking.

A few questions:
1. For some reason the buffer before the CD4017 chip wouldn't work as a non-inverting buffer, but since I needed one after as well, I figured I could use inverting ones there. However the signal wouldn't go through IC3b without the 10k resistor... no idea why, something to do with impedance?

2. The little red box is where the original blue box part of the circuit starts, I could probably remove the two resistors and maybe the capacitor as well?

3. Can I skip the IC3a buffer? Is the signal already buffered from the Q1 transistor?

4. Can any other buffers be omitted?

5. Anything else I've missed that should be changed or can be omitted, or do you guys think I'm on the right track?

I hope I havn't made any typos in the schematic. I've looked over it a few times to make sure, but you never know!

Edit: I found a mistake... the blend pots should be wired differently... For each Blend 1a and Blend 2a, the signal should go into lug 1, out from lug 2 and lug 3 to ground. Blend 1b and 2b is the opposite, signal goes into 3, out from lug 2 and 1 goes to ground. Just a schematic error, I didn't wire em that way on the breadboard. Oh, and I labeled Blend 2a Blend 1a... I'll make a new version shortly!

[Mark Hammer]

1) Are you Bunk Gardner? or Charles Lloyd?  I well remember ads for sax players using the Conn Multi-Vider in the late 60's.

2) I salute your diligence in pursuing this.

3) If the 4017 Rocktave mod does 4ths, it probably does them  waaaaayyyyy below the fundamental.  Low enough that they may have little musical value unless you play soprano sax.

4) The original Rocktave used the 570/571 as a sort of noise gate, or rather, a downward expander.  The original purpose was to gate out the divided signal before it got too sputter.  Since you are feeding it a reed instrument with relatively constant level, rather than a guitar that falls below threshold as notes decay, sputteriness is much less of a risk and does not require as much protection.  The Blue Box uses the transistors just ahead of the Blend pot as its noise gates.  I will note that the risk of sputter from the 4017 is the same as from the 4013, so you might wish to implement something similar for the octave+5th than is in place for the two octaves down.

5) Strikes me that maybe what you really want is a simple output mixer stage with level controls for each of the individual signals you want to select from.  So, a level pot for clean, another for 1-octave down, for 2 octaves down, for an octave and a 5th, and any other divisions you intend to make use of.  I'm guessing about your needs, but I think that might be less complicated than having some individual level controls, blend controls, and master-level controls.

[ElectricDruid]

1. For some reason the buffer before the CD4017 chip wouldn't work as a non-inverting buffer, but since I needed one after as well, I figured I could use inverting ones there. However the signal wouldn't go through IC3b without the 10k resistor... no idea why, something to do with impedance?

IC3b isn't currently wired up as a buffer of any type. It's close to an inverting op-amp configuration, but without the input resistor (like you have it) it's actually an I-to-V convertor - turns a current output into a voltage.
Without the 10K resistor..I don't know what it'd be, so I'm not surprised it didn't work.

2. The little red box is where the original blue box part of the circuit starts, I could probably remove the two resistors and maybe the capacitor as well?

Yes, probably. You need to think about what bias voltage each stage is expecting to see, and what voltage you're feeding it. The first buffer stage has a capacitor so that only AC signals can pass (a "DC blocker") and has a resistor to Vref, since that's the bias level we want - all our signals should swing about aound the middle voltage (4.5V if we're using a 9V supply). But after that, our signal is *already organised* to swing about around 4.5V, so we don't need to use AC coupling between the stages necessarily, and we don't need to use resistors to Vref. It looks to me like you've often had to put a buffer and a Vref resistor in because you had a capacitor which was blocking the *correct* DC bias that you'd established earlier in the circuit.

3. Can I skip the IC3a buffer? Is the signal already buffered from the Q1 transistor?

See above. Yes, probably.

4. Can any other buffers be omitted?

IC5b and the 1M to Vref can both go. They do nothing apart from buffer an already-buffered signal from the buffer IC 5a.

5. Anything else I've missed that should be changed or can be omitted, or do you guys think I'm on the right track?

I agree about Mark's point about using a mixer to mix all the output signals together. Yes, you're on the right track. Did anyone mention how long the track is?!?

Good luck with the next steps!

Tom

[Bastis]

Thanks for the feedback guys! I only have time to reply Mark now, so I'll get back to you later, ED!

I am sorry to say that I am none of those guys. I did go to a Charles Lloyd concert a few years ago.

The octave + 5th down is pretty nice on the alto. Combined with one octave down from the blue box (I forgot to and the 1/2 octave switch to the schematics) I get to play in 5ths, and it's not too low frequencies to be useful.

Edit: Oh! I just reread your message Mark and I misunderstood your point on 4ths. What I meant was that with the BB two octave down I get a 4th between the notes, and with the 1/2 octave switch set to 1 octave below I get a 5th between the notes!

About the transistor noise gates, is it the Q1 transistor you are referring to?

And you're probably right, a mixer would be a better solution!

[Mark Hammer]

Q2 and Q3 form crude gates for the two principal signals in the basic Blue Box.  They are turned on when the envelope signal coming from D2 is hot enough, and shut down when it isn't.  I gather that increasing the duration/time-constants of that envelope will result in the gates remaining open past the point where sputteriness in the divider stages kicks in.  I also gather that usability of the frequency divisions can be optimized by playing with the different parameters affecting those gates, such as pre-rectifier gain, rectifier diode type, cap value, etc.

[WorkBench]

Does it sound all wibbly wobbly?      Sonic sun glasses.

[Bastis]

It doesn't sound very wibbly wobbly luckily! 

I love these threads. I get help with my build but more importantly for I understand the basics more and more!

I drew a new version of the circuit with some modifications. If I misunderstood your recommendations please point it out!

1. I removed a couple of (probably) unnecessary buffers.

2. I removed some connections to vRef, maybe too many? Should I add a resistor to ground before IC1a?

3. I added a noise gate transistor to the 4017 signal. That transistor should work as a buffer too, right, so I have all buffered signals before the mixer volume pots?

4. I set up a mixer that I hope would work. Do I need the 10k resistors after the pots? When you guys (ED for instance) helped me design my band pass filter pedal (http://www.diystompboxes.com/smfforum/index.php?topic=112581.msg1047300#msg1047300) I put 10k pots at the mixing section, but that circuit used switches instead of pots, so perhaps they arn't needed here?
Edit: After some thinking I realize that the 10k resistors shouldn't be needed... right? When the signal is turned all the way down the pot will act as a 100k resistor to ground... and the only point of the 10k resistors would be to... heck I don't know. I see no point of the resistors anyway!

5. ... Did I change anything else? Oh, yeah, I added the 1/2 octave switch. I'll work on giving the 1 and 2 octave down seperate volume pots. Any tips on doing that? I should probably buffer the signal going from pin 13 on the 4013 chip, then send it to pin 3 and into a transistor noise gate like the other ones? Or would that octave 1 volume pot interfere with the signal going back to the 4013 chip?

By the way, I sometimes struggle to find the right words writing about this stuff. Partly because this is complicated for me and partly because English isn't my first language. I hope you understand what I mean anyway!

Anyway, here's the circuit as it looks at the moment:

[Bastis]

Sorry about the spam!

I squeezed in my idea on how to implement the octave 1 and octave 2 volume pots. Since I had a hard time explaining it I figured I'd just post this version of the schematics. Is this a good way of doing it?

(Of course the Q5 is a 2N3904 like the other ones...)

[Bastis]

I just breadboarded this thing, and it turns out IC3a wasn't needed and I had to put a capacitor and before IC3b. I get a very low signal 1 oct down signal, the rest are fine. I suspect I just made some mistake on the breadboard, I'll get back to it later and double check everything!

Edit: Damn it, it turns out it wasn't a mistake on the breadboard that caused the 1oct down signal to be weak. It seems like it was only signal bleeding through the transistor gate power line that I heard... No matter what I try I can't seem to fix this one.

Like I said I had to put a capacitor and before IC3b and add a resistor to vRef. When I try to split the signal after that buffer the signal runs fine to IC2a, but tapping the signal through Q5 doesn't work... I tried to add another opamp with a feedback resistor to get the volume up on that thing, but for some reason it still wouldn't go through the Q5 transistor. I tried using another transistor but that didn't work either... With the 1M resistor in front of the transistor the signal is barely audible, but that isn't an issue in any of the other cases... Can anyone help me out with this one?

[ElectricDruid]

The mixer arrangement is wrong still. I haven't checked the rest, but that leaps out.

The input signals need to go to the top of the pots, not the centre(wiper). The input 10K resistors need to go to an op-amp wired as an inverting mixer like this:

http://colomar.com/Shavano/opamp-s3.gif

Currently you have them wired to the positive input like a non-inverting amp.

HTH,
Tom

[Bastis]

Awesome, thanks Tom! Is this how you suggested?

I don't quite grasp why it matters if the amp is inverting or non-inverting in this case... Is the signal inverted somewhere in the circuit, is that why?

Should the trim pots be wired like the volume pots, with the signal going into the top, and out from the wiper?

The big mystery to me for the time being is the "mystery box" that I added. I somehow need to split the signal between IC2b and IC2a, send one to a volume pot and one back into IC2a... Any suggestions?

Edit: Forgot to add the ciruit...

[ElectricDruid]

Awesome, thanks Tom! Is this how you suggested?

Yep, that's it.

I don't quite grasp why it matters if the amp is inverting or non-inverting in this case... Is the signal inverted somewhere in the circuit, is that why?

No, it's because if you build it with a non-inverting amp, essentially you've got a passive resistor mixer followed by a non-inverting amp. There's no advantage from the op-amp. The inverting arrangement is a proper active mixer.

Should the trim pots be wired like the volume pots, with the signal going into the top, and out from the wiper?

Well, I'm not sure. I was about to ask you about those. What do you want them to do? They're to trim the signals to the mixer pots so that everything is similarly loud? In which case, you should wire them as "variable resistors" rather than as "potential dividers" e.g. only two connections needed. The usual way is to go in at one end, the wire the wiper to the other end and take the output from that. You could just move your grounded ends and connect them to the wiper -same thing. That gives you a variable resistance above the mixer pot, effectively limiting its maximum setting to less than 100%.

The big mystery to me for the time being is the "mystery box" that I added. I somehow need to split the signal between IC2b and IC2a, send one to a volume pot and one back into IC2a... Any suggestions?

Sorry, I don't understand that part. I don't get what all the transistors+56K+1M parts are doing. I don't know the original circuit. Maybe someone else can help?

Tom

[slacker]

Sorry, I don't understand that part. I don't get what all the transistors+56K+1M parts are doing. I don't know the original circuit. Maybe someone else can help?

Tom

The transistors are being used as sort of VCAs, The collectors are "powered" by an envelope follower made out of D2, D3 and associated bits, and the a square wave version of the signal drives the base. So you end up with the volume envelope of the original signal chopped up into what ever frequency the square wave is at.

Bastis, I think the design can be simplified massively, instead of having a separate transistor stage for each output you can just take all the signals directly form the various ICs to their mixer pots. Then just send the output of the final mixer to one of the transistor stages, this will give you one VCA for all the signals instead of having one for each one.

[Bastis]

Thanks, Tom! That makes sense, even though I don't fully understand how and why some things work. I'm getting better though! I'll give the variable resistors a go!

That's a very good point, Slacker, there should be no point in keeping all the transistors. I'll redesign it.

I'll see how my issue with tapping the signal between IC2b and IC2a with the new design and see what is going on!

[Bastis]

I looked at the Rocktave solution for tapping the signal between IC2b and IC2a, and there was no magic to it, just lead the signal to a pot and mix it with the rest. I'll give that a go. Here's the schematics, I'll breadboard it next week! At first I sent all the mixed signals through the VCA, but then I realized that it probably has a bad effect on the dry signal, so I mixed em together afterwards in another inverting opamp. Do you guys think this will work?

[ElectricDruid]

You've got 100K pots feeding 10K input resistors.

Does that drag the input down and screw up the response of the pots? I can never remember whether it's supposed to be 10K into 100K or 100K into 10K works best. Seems like 10K into a larger impedance would be better to me now.

Tom

[Bastis]

I breadboarded the thing today. I didn't have 5 10k pots so I used mostly 100Ks, and I didn't have the trimpots needed either, but up to and including the IC3a it worked like a charm! Probably even better with 10k pots. However, that darned VCA gate won't work with the signal coming from the mixer. If I feed it a signal from the CD4013 it works. I didn't try the other sources, but it should work just as fine. When I wire it up like the schematics though I don't get a signal through. It's completely quiet when I probe the base or collector of Q2. Should I look for a different gate solution or should this be fixable? Could it have something to do with impedance? Or the signal being inverted compared to the signal coming out out IC1a?

[Bastis]

I finally got this thing working. Putting the VCA gate after the mixer didn't work for some reason, so I went back to having 4 different VCAs, one for each signal (except the dry signal of course) before the mixer. 10k pots worked fine. I soldered it to a vero board and it works, I'll finish the boxing tomorrow probably!