String pluck detector??

[markusw]

and....there's your VCA!!!   U could sub a dual OTA like the 13700/13600 easily.

That's for the next module 

I've got an interesting issue at the moment. I replaced the two 4049 inverter stages in the pdf schem with a 4093.
Interestingly, at first it didn't work. OK this isn't really astonishing. After searching for wiring bugs I started checking voltages. Once I connected the DMM to the output of the 555 it started working again, like before I replaced the chip. If I disconnect the DMM it stops working.
I tried to replace the DMM with resistors (10k to 10M) from the NE555's output to either gnd or V+  but it didn't help. Notably, it keeps working as long as the DMM is connected even if I add a 10k to V+ or gnd.
It's really funny because performance is exactly the same as before I replaced the 4049 with a 4093, it just needs the DMM connected to the output of the NE555.

What kind of voodoo is that? 

Any ideas would be highly appreciated because I need my DMM for something else. 



Markus

[toneman]

Assuming both of the two 4093's inputs are tied together, U have just made a schmidt-trigger inverter.
Not really necessary for this application.....but.....
So... U used only TWO sections of the quad NAND gate? to make TWO inverters to replace the TWO 4049 sections??
whew!  almost as long-winded as Mr Hammer   
Did U remember to tie ALL unused CMOS inputs to gnd or pos???   
JUST the inputs..  ......leave unused outputs floating.
lmk

T

[markusw]

Did U remember to tie ALL unused CMOS inputs to gnd or pos???   
JUST the inputs..    ......leave unused outputs floating.

Yup, the inputs are tied to gnd. Thanks for the tip! 

12 h later it still only works if the DMM is connected to the NE555's out?

Any other ideas?

Markus

[markusw]

I found it. It doesn't work if I tap off the CV1 directly from the NE555's out. If I have two 4093 inverters in series connected to the NE555's out  and take CV1 and CV2  from their outputs it works. CV1 is then taken from the second inverter.
Any ideas why it doesn't work to CV1 directly from the 555??

Markus

[markusw]

Here's another update:

http://www.aronnelson.com/DIYFiles/up/PLL_AMS100_PD_NE555_current_schem_24-03-07.pdf

Changing the 555 to retriggerable was definitely a good modification. Even if you play faster than sampling time no notes are missed.
I also added a 4015  shift register plus a couple of resistors for sine approximation. With a simpe passive lp filter distortion is around 0.1%.

It's already real fun playing with it. 

Next I will try to add the programmable divider from the EMM synth....



swt, I had a look at the Electrax LFO. Maybe decoupling of the LFO opamp power rail might help?? I hope that others with more experience will chime in.



Markus

[MR COFFEE]

Very cool Markus!!!

Can't wiat to hear the sound clips. How many steps per cycle in your 4015 sine converter?

Keep us posted.

[markusw]

How many steps per cycle in your 4015 sine converter?

Should be 16 steps. It divides by 16.

One Q: I tried to tap off a square wave at 1/16 of the frequency from one of the 4015's outputs but it seems that I can't get a pure square wave from them. I suppose due to the mixing resistors it's a mess. Any ideas how to get the square wave at the same freq as the sine from the 4015 without using a buffer? Also would like to avoid another divide-by-16 chip.....

BTW, the programmable divider from the EMM synth is already working  Just have to tweak the 4046 VCO timing cap to adapt to the higher divider.......

Can't wiat to hear the sound clips.

I will try to da a sample of the sine approximation vs. dry guitar signal  First I will have to tweak the PLL to run with a 128 to 192 divider......


Regards,

Markus

[toneman]

none of the outputs of any shift register (4014, 4015 or 4017) have SQwave outs.
it's just the nature of the beast.
To get a 50/50 square wave U need a divide-by-2 each time.
the 4015 is 2 4bit shift registers.
the 4014 is a single 8bit shift register.
Several web sites show the output counts of this shift-registers.

I'd say, use one of the divided clocks to generate a sine using a 4040/R2R technique.
check your email
afn
T

[markusw]

Hey T,

thanks for your help!!

Obviously I had a wiring bug yesterday evening. Now it works. I get a 50/50 square wave (at the same freq as the sine approximation)  from every output of the 4015.
If I take it from the 4th output it's in phase with the sine.

In the meantime I added a 4040 between the 4046's VCO out and the 4015 sine generator to select the various octaves.
At the moment there are about 15 ICs on breadboard. So the PCB(s) won't fit into a 1590BB 

BTW, I intensively checked whether there is any advantage of having a 4093 stage between the 311 guitar signal squarer and the PLL's in, and to be honest, it really seems the 311 is sufficient. At least in my hand I do not get an improved performance if the 4093 is added.
Nevertheless, I probably will leave it in because there is one stage of the 4093 left.

Markus

[StephenGiles]

Could you possibly post a picture of your breadboard as is - I must see this!

[markusw]

Could you possibly post a picture of your breadboard as is - I must see this!

Sure, I will take a pic later. To be honest, looking at this mess it's hard to believe that it works 
So it will for sure be OK for a good laugh 
Due to it's 3D wiring I do not yet have an idea how large the PCB(s) will be.
Markus

[markusw]

Here is a pic of the current mess



The most left board contains the AMS pluck detector (the 3 top chips) and an orange squeezer.
The second from the left contains parts of the pluck detector, the NE555, the 4093, the 4016, the LF398 and a LM311 (at the moment used as output buffer).
The 3rd one has the input gain stage, Vref opamps, the guitar signal squarer LM311, the 4046 and the two 4526's for the programmable divider.
The 4th one contains the input lp filter, an additional opamp stage (currently used as output buffer) and mainly some stuff from previous experiments (i.e. an adaptive Schmitt trigger and a 4013 divider) that is not in use at the moment.
The most right board finally contains the 4040 divider and the 4015 sine generator.

Markus

[slacker]

That's amazing 
Looks pretty neat and tidy to me.

[StephenGiles]

Amazing indeed, I've been there many times and often something that worked one day would not the next - as happened with the ADA Flanger, so I ripped it all out and started again!

[markusw]

That's amazing 
Looks pretty neat and tidy to me.

Looks pretty neat and tidy to me.

I just hope that it will look a bit more tidy when on PCB.  To be honest, I myself wonder quite often that it works.

BTW, I like tose tiny trim pots. They fit nicely onto two or three rows.

I've been there many times and often something that worked one day would not the next - as happened with the ADA Flanger, so I ripped it all out and started again!

Sometimes it's really pretty weird, it stops working....you do nothing, maybe unconnect it from the ps....then after a while power it up again and everything's perfect again.
Currently, since the 555 issue was solved it seems pretty stable...no DMM needed to keep it running   

so I ripped it all out and started again!

I think I would quit the project before wiring up a second time   

Markus

[toneman]

WOW!      

U gonna go surfacemount?


T

[markusw]

WOW!      

U gonna go surfacemount?


T

Hopefully not 

Markus

[markusw]

Another update 

Yesterday eve I added a simple gain control on breadboard to add an envelope to the VCO signal. I just wanted to give it a try with a LED/LDR combo    It worked out astonishingly well.
Basically, I tapped off the signal from the NE571 rectifier (part of the AMS-100 pluck detector) and run it through a non-inverting opamp stage with slight gain. Subsequent to the opamp stage there is a simple R/C low pass consisting of a 20k pot and a 10µ to gnd. Then there is another opamp buffer stage. From it's out there is a trimpot (current limiter) followed by an LED to gnd. The LED is part of a home-made LED/LDR combo. The LDR is wired in series with the output of the sine generator (or the square wave) followed by a 10k to gnd (like a volume pot). It needed just a little tweaking of the gain stage and the current limiting pot to null the signal when in idle state.
Interestingly, the LED/LDR combo seems to smoothen any ripple. So for fast attack the R/C low pass isn't even necessary. With slow attack (pot set to 5-10k) it sounds a bit like a reverse tape effect.
With the sine or square wave set to e.g. 3rd or 5th or suboctave it sounds pretty cool.
Will probably give it a try with a LM13700 VCA to compare it to the LED/LDR combo but actually I would like to stay with the LED/LDR combo....I simply like to use them.

Will try to do some samples this eve. Also the schem urgently needs an update ;-)

Markus

[markusw]

Here are two samples with the LED/LDR gain control. Two are sine, once with one octave down, second one 5th up. Third clip is again 1 octave down, this time square wave.
One channel dry again..

http://www.aronnelson.com/DIYFiles/up/PLL_sine_-1_octave_30-03-07.mp3

http://www.aronnelson.com/DIYFiles/up/PLL_sine_5th_30-03-07.mp3

http://www.aronnelson.com/DIYFiles/up/PLL_square_-1_octave_30-03-07.mp3

Regards,

Markus

[markusw]

As per suggestion from toneman (thanks!  ) I adapted the levels between left and right channel (wet and dry). Otherwise the samples are the same.
Didn't mention before. It was done with my cheap Strat copy directly into the PC. Also there is no tone control so far....

http://www.aronnelson.com/DIYFiles/up/PLL_sine_-1_octave_30-03-07_3.mp3

http://www.aronnelson.com/DIYFiles/up/PLL_sine_5th_30-03-07_4.mp3

http://www.aronnelson.com/DIYFiles/up/PLL_square_-1_octave_30-03-07_2.mp3

Regards,

Markus