DH pitch shifter, questions.

[Govmnt_Lacky]

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.

Last question Taylor.

If the FAKE ground (DC jack, Input sleeve, and Output sleeve) receives it's GND or V- when you insert the DC wall wart....... then,

When or how does the REAL ground actually receive it's GND or V-? On the PCB they are not directly connected.

Looking at the above link to the layout, on the bottom of the board, you can notice that there are 2 ground points. The FAKE ground is connected to the DC jack black wire (which according to you should also connect to the In/Out sleeves) and the REAL gnd is directly to the right of that.

So how does the REAL ground actually receive it's GND or V- referecne?

Thanks for helping with this.

[Taylor]

You're using the terms real and fake backwards from the way I am.

Let's instead say -V and "ground". -V is the black wire coming from the power jack. "Ground" is everything marked ground on that board.

You want to remove the input and output jack grounds from "ground" and instead connect them to -V.

[Govmnt_Lacky]

You're using the terms real and fake backwards from the way I am.

Let's instead say -V and "ground". -V is the black wire coming from the power jack. "Ground" is everything marked ground on that board.

You want to remove the input and output jack grounds from "ground" and instead connect them to -V.

Got it! 

Thanks for the help and I will post back with results when I can. 

[El Heisenberg]

Hmmm, i think ill try those clips now

[Taylor]

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.

The thing does act in a kind of gated way. I don't understand why you changed that resistor, or why you made it bigger, but if the LED's not lighting you're probably not getting any pitch shift signal.

[Govmnt_Lacky]

Well..... I followed Taylor's advice and resoldered the grounds accordingly.

The effect works.... KINDA!

I still have a massive volume drop when engaged. (when Volume knob is cranked it is about 1/2 of input) Also, I can tell the Vibrato is working however, I still do not know about the Robot momentary 

The Pitch shift is good however, it is VERY GATED as Taylor implied. Almost like there is signal clipping going on.

I wonder if the Volume drop is because the dual op amp is running at a lower voltage? (3.6VDC)

Bottom line is that this one is going back on the shelf. Not mad at it, but not board-worthy either 

[El Heisenberg]

Taylor i changed that resistor on breadboard and the thing worked mch much better. I told you to try it!

If youre gettig the gating thing than you should try it

The gating problem i had was playing too hard. With too much signal going the thing would go quiet and then fade back in. I would have to play softly to get rhe effect so i just put more resistance in there and it worked but with volume drop. So i added an LPB-1 after it and boxe it up. Gotta record a clip still... Its got the vibrato swirch but didnt wanna make room for robot since it never seems to work. Stuffed it in a 1590B.



The datasheet says that the robot mode is engaged when the thibg powers on and you have to turn it off first ()

Check it out

[sundgist]

I've played around with the HT8950A chip a few months ago(Same as the HT8950 just with different switching). It's still on my breadboard and I've slowly been adding bits to it. Not the highest fidelity by any stretch, but I've found it a lot of fun and came up with a lot of useful sounds (for me anyway).

I came across a velleman voice-changer kit at Maplin's a while back and thought I could mould that into something interesting. I found the input signal needed for it to sound good (especially robot mode) was quite a narrow range. Too hot and it would distort and sound awful, too quiet and it would gate and cut notes off early. At the moment I've got it pretty much following the Dean Hazlewinter schematic but with a different power supply and a few mods.

The extra bit I've been playing with was adding a SA571 (NE571/570 etc) stage before and after the HT8950A. Someone suggested this in a different thread (possibly different forum).

The signal path goes something like this.
Input - Opamp stage - SA571 stage(compres) - HT8950 - SA571 stage(expand) - Opamp stage(clean/FX blend) - Output
The first opamp splits and buffers the input. Can't remember if there is some signal gain there as well
The SA571 compresses the signal going into the HT8950 keeping it in that narrow range and does seem to make it less noisy too.
I'm probably spending more time on this than is warranted but I'm learning as I go and I like making interesting noises so it's all good. I'll post a schem once I'm happy with what I've got.

[El Heisenberg]

And some clips i hope!

here's a pic of mine. I removed the footswitch since I took this pedal off me board.

[Taylor]

I was thinking about this, and unless I'm crazy, the single battery way GGG shows this thing just doesn't make any sense. The audio is not biased to half the supply. What the heck?

I'm going to design a new circuit for this chip, because that one is just ridiculous.

[Govmnt_Lacky]

I was thinking about this, and unless I'm crazy, the single battery way GGG shows this thing just doesn't make any sense. The audio is not biased to half the supply. What the heck?

I'm going to design a new circuit for this chip, because that one is just ridiculous.

I agree  

I think the op amp needs to run at AND be biased based on supply voltage. Meaning, the op amp should be running off of 9V and biased to 4.5V. The Holtek chips needs the V+ regulated down to 3.3V.

EDIT: I guess the only question is whether or not the HT8950 will be able to handle the signal output of the op amps.  

EDIT2: Is there any reason why the layout shows the supply voltage going to Pin 11 of the HT8950? According to the datasheet, that is for testing purposes only 

[slacker]

The audio is not biased to half the supply. What the heck?

Yeah it is, the supply is +3.6 volts, ground, -3.6 volts. Pin 8 of the opamp is getting +3.6 volts, pin 4 is getting -3.6volts and the non inverting inputs are connected to ground, so the audio is biased around ground, which is the middle of the supply.

I agree with you though, a more conventional single supply would be better.

[Taylor]

EDIT: I guess the only question is whether or not the HT8950 will be able to handle the signal output of the op amps. 

EDIT2: Is there any reason why the layout shows the supply voltage going to Pin 11 of the HT8950? According to the datasheet, that is for testing purposes only 

As long as we don't boost the signal much with the opamp stage in front, there shouldn't be any particular clipping issue (any more than there already is anyway). LEDs in the feedback loop might be a good idea to keep the 8950 from clipping.

The datasheet app circuit shows pin 11 connected that way, so DH used that verbatim.

The audio is not biased to half the supply. What the heck?

Yeah it is, the supply is +3.6 volts, ground, -3.6 volts. Pin 8 of the opamp is getting +3.6 volts, pin 4 is getting -3.6volts and the non inverting inputs are connected to ground, so the audio is biased around ground, which is the middle of the supply.

I agree with you though, a more conventional single supply would be better.

I must be misunderstanding the function of the zeners. Thanks. I will simulate this so I understand it better.

[slacker]

The zeners have a voltage drop of 3.6 volts, so the top of the zener will be 3.6 volts higher than the bottom, the resistors limit the current through them.
The mid point of the Zeners is tied to ground so Z1 holds V+ (the one in the square box) 3.6 volts above ground, Z2 holds ground 3.6 volts above V-. If you take ground to be 0 volts, then that gives you +3.6 volts and -3.6 volts.

Hope that makes sense.

[sundgist]

I'm assuming it's been done this way due to the holtek chip having a max voltage of 4 volts (ish, can't remember exactly). So the opamp runs off -+ 3.6V and the holtek wired between the centre 0V and +3.6V.

I've found the holtek chip likes to have a signal input of around 1V peak to peak for the best sound to avoid excessive distortion above this (this is not clean at the best of times) and to avoid gating while the signal is dropping. There shouldn't be any issue with the opamp distorting as there is no reason for it to amplify the signal so much.

I've used a 9V supply and biased the opamp halfway and used a 3.6v zener to limit the voltage to the 8950. The first half of the sa571 i'm using is set up as automatic gain control taken from the app notes which keeps the signal into the 8950 at just around 1Vpp. The second half is set up as an expander after the 8950 which is effectively acting as a gate to take out any noise amplified by the first half. I've moved the first opamp from the start of the signal chain to before the 8950 as a buffer, sounds and looks better on a scope. I'm really just moving building blocks around at the moment so haven't looked at it too closely. I'm slowly drawing up a schem, that will probably explain things better.

I'll try and post some soundclips once I'm able to but it doesn't really sound much different from the ones I've heard on youtube, try searching for "Death by Audio - Robot" which is basically the same thing.
Mine kind of 'feels' different due to the compression. It makes it easier to just plug in and play without having to fiddle with the input level or think too much about your playing technique. The clean blend at the end makes it a bit more flexible in the noisemaking sense, the pitch shift intervals don't track evenly across a wide range which makes for some quite discordant sounds.

[bram]

Dear all,
yup, I'm yet another person struggling with this weird power scheme of the dh pitch shifter.

So, if we call the power supply rails +9V and ground and the pseudo bipolar supply with the zeners V+, V bias and V-.

I have the HT8950a connected to V+ and V bias (so it's working on +3,6 V relative to V bias, checked it with the voltmeter). The momentary open switch TGU/TGD/ROB and VIB are connected to V bias, as well as V ref (with cap C5).

The Tl072 is connected to V+ and V- and biased to V bias (checked with the volt meter, 3,8V ; 0 V - 3,7V are the readings).
The in/out jack grounds are tied to ground, as well as the lug of the volume pot.

I checked all soldering, it seems ok to me. Now when I put power to the pedal I hear loud squealing and a very silent signal.
Could you help me out? What did I do wrong?