DH pitch shifter, questions.

[El Heisenberg]

I was planning on using those flatter wider momentary buttons they have at radioshack.  I want flatter buttons since I'll be controlling the rotary switch with my foot. In the same way as the Maestro Parametric Filter. But the ones with the black and red tops in the link are flimsy. Really crappy.

I rehoused a keyboard in a box and used em to be able to stomp on certain letter keys. I also used em for a little bit when the switch on my electronic cigarette fried when I used it at 6v instead of 3.7v. The switch broke so I got a flatter one from radioshack. Now I can carry it in my pocket without the thing goin off. And I can vape at 7v. They have tons of different types of momentary switches to choose from.

[glops]

I like the video, as well.  Ordered the chip from Mammoth. Thanks!

Just looked at Mammoth, they only sell the HT8970 Echo chip. The one you need for the GGG Pitch Shifter is the HT8950.
[/quote]

Oh geez, wrong chip!  Thanks for pointing that out, I probably would have bboarded it and wondered why I wasn't getting sound.
guess i'll do something else with the 8970...

[Taylor]

OK, I built it...

I don't know what to think about this thing. It's weird, and it does a good job of digitally obliterating my sound, but I'm not sure I'd say there's anything like pitch shifting happening, and none of the switches seem to work the way you'd expect.

I wired it with a knob for vibrato speed (not shown in the GGG project) and momentary SPST switches for robot, vibrato, and TGD, which should cycle through the pitch shift amounts, I thought. TGD doesn't seem to work at all, and the other 2 do seem to work, but only when I hold the switch down (in other words, it seems to expect a latching switch, not momentary as the datasheet and GGG schem show).

Will have to ponder this more. But it's definitely fun for folks who like glitchy sounds, and I do.

Edit: TGD/TGU won't work for me, but switching pins 1-3 works.

BTW, glops, I bought more of these chips than I really need. If you can't get one easily I'll send you one for the cost of postage.

[El Heisenberg]

Wha--- you biult one just like that? In like less than 24 hours? Did you etch a board? Perf? Jeez.

[Taylor]

Perf. I bought the chips and the zeners months ago because I wanted to build it, then forgot about it. I had to dig around under my bed to find the chips.   I think I had some plan to do like 3 of them in parallel for nasty harmonies, which must be why I bought more chips than necessary, but the thing is too nasty - more than one at a time would just be noise.

Now that I sorted out the switch issue, I am getting pitch shifting, and it's great! In a really ugly way.

It also does a really nice ring modulator when you set the oscillator frequency to maximum. Very bell-like, like a sine wave ring mod.

After just looking at the other clip, I really want to get my robot function working... wonder what the deal is. That guy is using the HT8950A which is a different chip with different pinout. Wonder if that's the trick?

[glops]

Hey Taylor, thanks for the offer.  I realized that I can get them from ebay easily and I also realized that I need to finish a BUNCH of other stuff before doing
anything else.  For instance, I've had the Gristleizer PCB that I ordered from you that's waiting to get stuff when I return home from Texas.

[Taylor]

Eureka! I'm a total bozo.

This circuit has a weird power scheme. The Holtek chip needs unipolar 3.6v, but I guess the designer decided to run the opamps on +-3.6v, so you see the whole zener power setup which makes a faux-bipolar supply. Then you have virtual ground.

Now, the GGG schem and PCB both show all the normal grounds (the ground on your in/out jacks, the bypass switch, etc. connected to the fake ground, not real ground.

My first mistake was that I am so used to wiring pedals up the normal way that I wired all my grounds, real and fake, together because I wasn't paying attention (that's what happens when you build a pedal 24 hours after deciding to build it). I fixed this and wired everything according to the schem and PCB layout. Suddenly, my robot, vibrato, and TGD/TGU switches all worked (I could tell because I could hear the oscillator vibrato-ing, etc.) but I had no audio.

This makes sense, I think - if you have your virtual ground connected to the audio jack sleeves like GGG shows, and then you have any other pedals in the chain, whose jacks are connected to real ground, then you're shorting your virtual ground to fake ground, right? How can that work? (Sincere, not rhetorical, question).

So I then rewired all the grounds. In and out jack ground, volume knob grounds and bypass switch ground go to the ground tab on the power jack, and all the grounds for the momentary switches go to the virtual ground on the board. Now everything works perfectly.

I wonder if the other people who had problems with switches not working made the same mistake I did? Also, eniacmike, you mention getting opamp distortion - I had the same problem, but I think it's gone now. I think wiring the grounds the way GGG shows basically clips your opamp rails to 3.6v, but the gain stages are designed for twice that, so you get clipping.

[Skruffyhound]

Bookmarked

[El Heisenberg]

mwuahahaha! So how is it NOW??

Ya did it on perf? Grar! I guess I'll have to do mine on perf too.


That diode switching is gunna be a pain!

Taylor is yours boxed up? What did you put it in?

What does the TGD/TGU do?

Also, you're just using 9v right? Using two 9v batteries would fix that right? Do you think it'd really improve the sound of the holtek chip or just the opamp?


Why a 4558?? I don't have any. I have quick access to TL082s. Guess I could find an NTE replacement at Fry's Electronics. BUt is it that big of a deal? Or did he just use a 4558 cos he figured all us DIYers would have a ton of em from TS builds.

[Taylor]

It's awesome. It does pitch shifting, but also you can turn on both vibrato and robot and then turn the oscillator way down get all kinds of bloopy arpeggios happening.

As I said earlier, I added the vibrato control pot - just swap the resistor between the vibrato pin and OSC1 pin with a pot. 250k would probably be ideal, but all I had was 100k. You should probably add a small series resistor like you see for the oscillator pot. The vibrato pot is pretty cool, it seems not to change the speed of vibrato, but actually the depth. The vibrato LFO is square-like.

I didn't mess with the diode switching. I just added a TGU momentary. This is on one of the top right pins of the chip, pin 16 I think, but check the datasheet. This cycles upward through the pitch shift and robot modes. It seemed the simplest way to get preset control, and  could see it being useful to be able to cycle through pitch shifts with a foot switch like you could on the older versions of the Whammy.

I'm using a 9v power supply. I don't ever do batteries. I don't think there's anything you could do to make the Holtek chip sound better, and you can't run it on anything higher than 3.6v I think.

I used a TL072. I use them for all dual opamp duties, never notice a difference.

[Morocotopo]

Taylor, you mean this thing actually does decent pitch shifting? I threw it into the "PCB graveyard" drawer a couple of years back, because I couldn´t get a decent sound from it. Time to rescue it?

[Taylor]

Pitch shifting? Yes.

Decent? In the ear of the beholder. Probably no.

As I said above, I've got a real, high quality pitch shifter I made using the Spin FV-1 DSP chip. That thing sounds like a Whammy, and I can program it with whatever presets, features, and controls I want.

What the DH pitch shifter does is glitchy, low fidelity sounds, including pitch shifting. If that doesn't sound fun to you, it's probably not worth your time. It's definitely in the "noise toy" category. I happen to really be into that because I like weird avant garde music. But if you wanted something like a Whammy, this doesn't do it.

So it depends what you're looking for. There are some people who will hear "glitchy lo fi arpeggiating robot pitch shifter" and start drooling. If that's not you, then I'd pass this up.

[Morocotopo]

OK, I know this isn´t a squeaky clean thing, just wondering if, since I´ve already built it, it might be worthwhile to try to get it to perform as intended. Seems to be just a little rewiring. Maybe I´ll try it. And yes, I like lo-fi sometimes! In this case, beauty is in the ear of the beholder, hehehe...

Cool find.

Cheers!

P.D: Any thread here about your Spin pitch shifter to look at? Sounds like an interesting thing.

[El Heisenberg]

Taylor, what did you box it up in? Room to step on all those momentary switches???



I'm also interested in the FV-1 pitch shifter! I dunno nothin bout DSP tho...but think it's time to stop hoping I can squeeze this sorta stuff outta cheap chips.

[Taylor]

All the info for the FV-1 is at spinsemi.com. If you guys get into it check out the forum there. It's quiet, but Frank, who co-designed the chip, hangs out there and helps, and I will also help with your code. The best place to start is to read the knowledge base section - that will get you pretty far, and modifying the code examples that are up there will be a good way to get into it.

I built it in a box I had lying around which has unusual dimensions. It's a BUD box, about 6" x 3" x 2". I put the 4 stomps (bypass, TGU, vibrato, robot) in the 4 corners and the knobs in the center. It was pretty "whatever" because I was doing it really quickly. But since I don't gig, ergonomics don't matter much to me.

[Taylor]

Responding to a PM, thought I'd post this here to consolidate info...

Taylor,

1) What did you mean by "fake ground?"
2) Could you explain, in a little more detail, exactly how you grounded the whole circuit?

I guess I fell into the "traps" you referred to however, I do not understand your explanation about your WORKING grounding scheme.

I also had a large volume drop with the pedal engaged. Did you experience the same thing with the bad ground scheme?

I appreciate your help as this is the ONLY box I have that still remains on the sidelines.

Ok, let's look at the schematic. There are two different power circuits we can use, outlined in dotted line boxes on the right. I used the upper one, which uses a single 9v power supply. What this does is use a pair of zeners to create a 3.3v positive rail, a fake ground, labeled ground, and a -3.3v negative rail (negative relative to the fake ground). This is not how we normally do biasing in effects these days, honestly I'm not sure why it was done this way, other than that maybe the designer comes from a synth background.

The HT8950 runs on 3.3v, but IMO it would have been much more intuitive, at least for those of us used to our stompbox world, to run it on 3.3v and run the opamps on 9v and bias the audio to the center of the supply. But they did it this way instead. This would be fine, except that we're used to calling the center of our power supply "vref" or "vbias" or just "vr", and calling the negative voltage rail ground, rather than "-v".

This leads to the central problem that I suspect everyone who had problems faced: the ground tab of your power supply is not the ground of the circuit! The "ground" of the circuit is 3.3v positive relative to the ground of your power jack. We're used to connecting all of our grounds together, but you cannot ground the ground tab of your power jack to the parts of the circuit  labeled ground - if you do this, you're shorting the negative rail to the center of the supply.

A major problem crops up because other pedals' input/output jack ground will be connected to the negative rail of the power supply, so then you will be shorting the pitch shifter's "ground" to its negative supply rail. So you need to wire the grounds like this:

In and out jack ground, volume knob grounds and bypass switch ground go to the ground tab on the power jack, and all the grounds for the momentary switches go to the virtual ground on the board. Now everything works perfectly.

It's basically really confusing the way they've done it, and if you follow their wiring scheme I believe you will be boned. Maybe at some point I'll design a new circuit using this chip, which uses a normal power supply scheme and is less confusing.

Hope that helps clear it up, I haven't messed with this much since I built it so I hope it was fresh enough in my mind that I made no mistakes in explaining it.

[El Heisenberg]

Ive been meaning to record mind and post here.


I was excited about the effect for about a day after i built it.

Im afraid mine is wrong or this is just how bad it is.

Oh yea i remember i had a problem with gating. I would play and its like a gate would close with the chips LED lighting up. I switched the resistor feeding the chip from 10k to 120k. This stopped my problem, and givvese a sound but not sure its what i shouls get.

The LED doesnt light up anymore. Cos the signal aint strong enough wheb it hits the chip. Still have a switch so i can toggle it back to stock.

[Taylor]

Taylor,

Thanks for helping but I am still a bit confused  
Correct me if I am wrong but I should tie the following grounds together:

REAL GND) Input sleeve, Output sleeve, 3PDT, Volume pot, Effect Mix pot, Dry Mix pot, and DC jack.

FAKE GND) 7PST switch, Robot momentary, and Vibrato momentary.

Also, where do these connect together with relationship to the circuit board?
Sorry, but if you could simplify it a bit more that would help. (Like A+B+C get tied at Point X on the board)

You've got it right. The fake ground and real ground don't connect, that's the thing. You could actually connect the pots and the 3pdt ground to the board as shown in the GGG layout, it would work the same. The input and output jacks NEED to connect to the power jack, and everything else can ground to the "ground" on the PCB.

[Govmnt_Lacky]

You've got it right. The fake ground and real ground don't connect, that's the thing. You could actually connect the pots and the 3pdt ground to the board as shown in the GGG layout, it would work the same. The input and output jacks NEED to connect to the power jack, and everything else can ground to the "ground" on the PCB.

OK. So going by the following layout:

http://www.generalguitargadgets.com/diagrams/dh_pitch_shifter_lo.gif

And, what you have said before AND what you are saying above...

All I have to do is disconnect the Input and Output sleeve grounds from where they attach to the board in the layout and reconnect them to the DC jack ground?

That is all?

Seems TOO easy 

[Taylor]

I think so. Keep in mind I didn't use the PCB, I built straight from the schem onto perf. But looking at the layout, if that's what you used, then that's probably the solution. Give it a try and see, anyway.