ornaments+crimes extended / polymorphic 4x16bit CV generator

From circuitbending to homebrew stompboxes & synths, keep the DIY spirit alive!

Moderators: lisa, luketeaford, Kent, Joe.

Post Reply
User avatar
mxmxmx
Super Deluxe Wiggler
Posts: 2243
Joined: Thu May 31, 2012 1:54 am
Location: berlin

Post by mxmxmx » Tue Apr 30, 2019 7:43 am

J_Behan wrote:Trying to calibrate at the mo.
On channel A, and I've got up to 5v done. But when I try to dial in 6v, the figure in the top right corner has maxed out, and I can't get higher than 4.6v on my meter.

Is this likely down to a build mistake?
yes.

but hard to say what/which without further information. when you start over using the "default" calibration values: what's the (uncalibrated) output voltage at the -3V step? are the default values incrementing as expected otherwise, ie is the distance between the (uncalibrated) steps ~ 1.0V?

User avatar
J_Behan
Common Wiggler
Posts: 53
Joined: Wed Jan 03, 2018 7:57 am
Location: Ireland
Contact:

Post by J_Behan » Tue Apr 30, 2019 8:20 am

mxmxmx wrote:
J_Behan wrote:Trying to calibrate at the mo.
On channel A, and I've got up to 5v done. But when I try to dial in 6v, the figure in the top right corner has maxed out, and I can't get higher than 4.6v on my meter.

Is this likely down to a build mistake?
yes.

but hard to say what/which without further information. when you start over using the "default" calibration values: what's the (uncalibrated) output voltage at the -3V step? are the default values incrementing as expected otherwise, ie is the distance between the (uncalibrated) steps ~ 1.0V?
Thanks, I had a feeling it might be.
On channel B, the calibrated value for -3v is -2.442v. If I recall correctly, it was around the 4v and 5v steps on channel A that I began to suspect something wasn't right. The initial values seemed way off, around 2v or so.

I have another working unit, which I calibrated without any problems, so I'll compare the two builds for mistakes. I probably put a wrong value resistor in there somewhere or something.

User avatar
mxmxmx
Super Deluxe Wiggler
Posts: 2243
Joined: Thu May 31, 2012 1:54 am
Location: berlin

Post by mxmxmx » Tue Apr 30, 2019 8:58 am

J_Behan wrote:
Thanks, I had a feeling it might be.
On channel B, the calibrated value for -3v is -2.442v. If I recall correctly, it was around the 4v and 5v steps on channel A that I began to suspect something wasn't right. The initial values seemed way off, around 2v or so.

I have another working unit, which I calibrated without any problems, so I'll compare the two builds for mistakes. I probably put a wrong value resistor in there somewhere or something.
i see. before you start looking for wrong value resistors, it'll definitely help to know if the steps are 1V (as they should be); and if the behaviour is identical across channels. in this case, it's probably just the offset, that's wrong. but if things start at -2.4v and max out at +4.6v, it's probably not that. could be that the gain is wrong, could be something else.

User avatar
J_Behan
Common Wiggler
Posts: 53
Joined: Wed Jan 03, 2018 7:57 am
Location: Ireland
Contact:

Post by J_Behan » Tue Apr 30, 2019 9:27 am

mxmxmx wrote:
i see. before you start looking for wrong value resistors, it'll definitely help to know if the steps are 1V (as they should be); and if the behaviour is identical across channels. in this case, it's probably just the offset, that's wrong. but if things start at -2.4v and max out at +4.6v, it's probably not that. could be that the gain is wrong, could be something else.
It's easy to calibrate up to 5v. The initial uncalibrated value for the 5v step is around 4v, and it feels like it takes an age to dial it up to 5v, but it gets there.
The initial uncalibrated value for the 6v step is 4.27v, and it's maxed out at 4.57v, I can't get that value any higher.

edit: same story for channel B. The 6v step maxes out at 4.5v.

The channel C -3v step starts out at -11.3v, so something clearly wrong there.

User avatar
mxmxmx
Super Deluxe Wiggler
Posts: 2243
Joined: Thu May 31, 2012 1:54 am
Location: berlin

Post by mxmxmx » Wed May 01, 2019 2:50 am

J_Behan wrote: The initial uncalibrated value for the 5v step is around 4v, and it feels like it takes an age to dial it up to 5v, but it gets there.
The initial uncalibrated value for the 6v step is 4.27v, and it's maxed out at 4.57v, I can't get that value any higher.
edit: same story for channel B. The 6v step maxes out at 4.5v.

The channel C -3v step starts out at -11.3v, so something clearly wrong there.
mmh, there's something clearly wrong with all those channels, i'm afraid. hard to say what though; it doesn't look like you've simply put some wrong value resistors somewhere, at least not consistently wrong...

have you checked the DAC output is ok? you can tap that at the one leg of the 24k9 resistors. when stepping through the default calibration points, you should see the DAC outputs decrementing in ~0.25V steps, starting at about ~2.3V. (i'm guessing you'll find that it is ok, but just to make sure)

User avatar
n0rd
Veteran Wiggler
Posts: 547
Joined: Thu Aug 02, 2012 6:01 am
Location: Down Under

Post by n0rd » Sat May 18, 2019 9:04 am

Meta-Q seq_mode no longer works correctly in v1.3.5.

Steps to reproduce:
Set s#1 to some scale and deselect all notes in mask except one. (Eg Octaves only).
Set s#2 to same scale and select different/additional notes in mask.

Set seq_mode to TR+1.
Set scale to s#1.

Apply varying voltage to CV1 and a clock/triggers to TR1 to sample.
You will get octaves output. (Since s#1 has only one note selected).

When triggering TR2, mask on LCD screen changes but pitch does not.
ie Out A still outputs only octaves.
Expected:
Out A should now output notes from s#2.

The same occurs when "CV aux" is set to "scl#". Mask visually changes on LCD screen but actual output does not change.

Manually changing "scale #" from O&C menu works ok.

Steve
:tardis:

User avatar
mxmxmx
Super Deluxe Wiggler
Posts: 2243
Joined: Thu May 31, 2012 1:54 am
Location: berlin

Post by mxmxmx » Sun May 19, 2019 1:20 am

n0rd wrote:Meta-Q seq_mode no longer works correctly in v1.3.5.
ups, can you try this? https://github.com/mxmxmx/O_C/tree/dev136

User avatar
n0rd
Veteran Wiggler
Posts: 547
Joined: Thu Aug 02, 2012 6:01 am
Location: Down Under

Post by n0rd » Sun May 19, 2019 4:35 am

mxmxmx wrote:
n0rd wrote:Meta-Q seq_mode no longer works correctly in v1.3.5.
ups, can you try this? https://github.com/mxmxmx/O_C/tree/dev136
Quick test - seems to work.
Might be an issue when changing "seq_mode" from "TR+3" to "-" and then back to "TR+1". Seems to skip one or two TR2 inputs when doing this.
:tardis:

User avatar
mxmxmx
Super Deluxe Wiggler
Posts: 2243
Joined: Thu May 31, 2012 1:54 am
Location: berlin

Post by mxmxmx » Sun May 19, 2019 5:16 am

n0rd wrote:
mxmxmx wrote:
n0rd wrote:Meta-Q seq_mode no longer works correctly in v1.3.5.
ups, can you try this? https://github.com/mxmxmx/O_C/tree/dev136
Quick test - seems to work.
Might be an issue when changing "seq_mode" from "TR+3" to "-" and then back to "TR+1". Seems to skip one or two TR2 inputs when doing this.
thanks, i've just pushed another commit (re TR2); not 100%, but think that might have been that.

User avatar
Setherian
Wiggling with Experience
Posts: 353
Joined: Wed Mar 09, 2011 5:16 pm
Location: Berlin, Greenland

Post by Setherian » Tue May 21, 2019 7:31 am

Can anyone sell me a regular Ornament n Crime panel in Europe (cheap! :)?

jimfowler
Veteran Wiggler
Posts: 562
Joined: Mon Feb 22, 2016 1:13 pm
Location: New York
Contact:

Post by jimfowler » Sat Jul 20, 2019 6:49 pm

So I've built a handful of these and recently sold my Euro version. The buyer messaged saying the quantizer output is at max no matter what the input is so I got the unit back and sure enough the quantizer is stuck high...sort of. It's actually pitching the output up by a handful of octaves. If you go into the settings of the program and adjust the scale (I think it was scale) you can get it to actually play pitches but the intervals are not at all right. Measuring the output (of all 4 channels) shows 11.xx volts, which is (to my knowledge) as high as it'll go (12v system and all). Weird thing to me is that the problem is isolated to the quantizer program. LFO, etc. all sweeps through the full voltage range.

The only other time I've seen something similar is on my 2O_C and it was found during calibration and solved by reflowing the DAC chip.

Tried re-installing firmware and re-calibrating from scratch. Any ideas? This is beyond my skillset.

- Jim

pld
Ultra Wiggler
Posts: 937
Joined: Thu Mar 05, 2015 5:15 am
Location: Germany

Post by pld » Sun Jul 21, 2019 1:23 am

So I'll pull this part over as well...
jimfowler wrote:so when you're calibrating the DAC volt per octave it'll generally be way high or way low (something like + or - 11.00v) and the thing calibrates successfully - e.g. 1v = 1.000, 2v = 2.000, etc. But this is the voltage range of the quantizer. It's as if the DAC is scaling the output very high (as in: 5-6 octaves) and it eventually tops out at the max voltage the DAC can get to

and yes, the fault was not there, then intermittent and now consistent. consistent is good...at least easier to fix.
So you're saying that the initial output voltages during calibration are far off, but can be reached? What are the initial values at -3, -2, -1...?

I'm not sure what you mean by "voltage range of the quantizer" -- the other apps might not use all the calibration points (except 0V) but can't override the hardware. The DAC doesn't do any scaling. It mostly reads like the output stages are just built wrong or there's simply a short. I'd start by checking the 2.5V and the voltage divider (2x 47K) 1.25 V. Then check all the resistors at the 2172s (220/24.9/100K). If in doubt, hi-res image of board.

electricanada
Super Deluxe Wiggler
Posts: 1187
Joined: Wed Nov 29, 2017 12:26 am
Location: Norfolk, VA

Post by electricanada » Wed Aug 21, 2019 5:13 pm

Does the closed-loop calibration work with Hz/V oscillators? I have a Metasonix RK-4 in mind.
Eléctrica (electric) Nāda (the yoga of sound).

pld
Ultra Wiggler
Posts: 937
Joined: Thu Mar 05, 2015 5:15 am
Location: Germany

Post by pld » Thu Aug 22, 2019 1:09 am

electricanada wrote:Does the closed-loop calibration work with Hz/V oscillators? I have a Metasonix RK-4 in mind.
Unlikely, but maybe I'm missing something. Even if the octaves somehow calibrate, the internal calculations are still linear.

User avatar
aragorn23
Common Wiggler
Posts: 216
Joined: Wed Jul 26, 2017 6:02 am
Location: Johannesburg, South Africa
Contact:

Post by aragorn23 » Wed Aug 28, 2019 10:41 am

Hey all, is the following OLED suitable for an o_c build?

https://www.robotics.org.za/W14242?sear ... 3&quot%3B=

pld
Ultra Wiggler
Posts: 937
Joined: Thu Mar 05, 2015 5:15 am
Location: Germany

Post by pld » Thu Aug 29, 2019 12:39 am

aragorn23 wrote:Hey all, is the following OLED suitable for an o_c build?

https://www.robotics.org.za/W14242?sear ... 3&quot%3B=
Looks like the pinout is incorrect.

User avatar
Waz
Wiggling with Scurvy
Posts: 2745
Joined: Mon Mar 02, 2009 1:36 pm
Location: Detroit

Post by Waz » Sat Oct 12, 2019 1:27 pm

I am trying to update to the newest firmware and keep getting a message that the Hex is too large. I've went through this entire thread about this topic and can't seem to solve the issue. I am using a known data usb cable, so that's not the issue. Any suggestions?

*Edit - Found the issue. The latest pre-release was too large. The older version (1.3.5) flashed perfectly fine. I'm on a teensy 3.2 btw.

Bug maybe?

User avatar
mxmxmx
Super Deluxe Wiggler
Posts: 2243
Joined: Thu May 31, 2012 1:54 am
Location: berlin

Post by mxmxmx » Sun Oct 13, 2019 3:08 am

Waz wrote: Bug maybe?
mmh, -ish; looks as if there's a mismatch between what the IDE things is "too large", and what teensyduino loader app thinks is "too large".

should work now: https://github.com/mxmxmx/O_C/releases/tag/1.3.6

also, if anyone was bored ... superficially it looks as if compiling with the -Os option (= Tools > Optimize > Smallest Code) results in working code; that would bring down the memory usage to 65% (rather than 98%), we could bring back in a lot of preset scales at least, but needs some testing.

User avatar
SphericalSound
Wiggling with Experience
Posts: 300
Joined: Thu Feb 01, 2018 9:23 am
Location: Spain

Post by SphericalSound » Sun Oct 13, 2019 7:27 pm

So the teensy 3.2 is getting more and more difficult and expensive to get. Meanwhile 4.0 is at 18USD. Any chance of getting a revision of the hardware to accomodate T4.0? (Plus a lot fresh RAM and CPU!!)

:deadbanana:

pld
Ultra Wiggler
Posts: 937
Joined: Thu Mar 05, 2015 5:15 am
Location: Germany

Post by pld » Mon Oct 14, 2019 12:14 am

MapacheRaper wrote:So the teensy 3.2 is getting more and more difficult and expensive to get. Meanwhile 4.0 is at 18USD. Any chance of getting a revision of the hardware to accomodate T4.0? (Plus a lot fresh RAM and CPU!!)

:deadbanana:
Seems unlikely, see e.g. here and here.
I might be swayed by a really compelling use case but as-is, it's not a super interesting endeavour...

User avatar
mamonu
90s ambient miscreant!!
Posts: 657
Joined: Tue Jul 20, 2010 4:08 pm
Location: Surrey, UK

Post by mamonu » Mon Oct 28, 2019 2:13 pm

can the O&C work on +-15V? ...easily?

I have seen a Buchla version and a MOTM version
somewhere but not many detals on how

pld
Ultra Wiggler
Posts: 937
Joined: Thu Mar 05, 2015 5:15 am
Location: Germany

Post by pld » Wed Oct 30, 2019 2:22 pm

mamonu wrote:can the O&C work on +-15V? ...easily?
Off-hand two things that might be affected are the +5V regulator which would have to dissipate a bit more heat, and possibly the load resistor for the LM4040.
The schematic for the 2OC and hOC from Northern Light should be available.

User avatar
Sikkboy
Common Wiggler
Posts: 107
Joined: Wed May 15, 2013 10:12 am

Re: ornaments+crimes extended / polymorphic 4x16bit CV generator

Post by Sikkboy » Thu Jan 09, 2020 3:49 pm

Any way of getting the v1.3.6 hex file flipped from github?
---
4/9

User avatar
Spastic_Colon
Learning to Wiggle
Posts: 44
Joined: Mon Nov 06, 2017 2:11 pm
Location: London

Re: ornaments+crimes extended / polymorphic 4x16bit CV generator

Post by Spastic_Colon » Thu Jan 30, 2020 11:17 am

Sooo, something’s not right with my build, or I’m
Misunderstanding elements of the calibration process. I’m struggling seeing pet 3 through, having had no issues with calibrating the outputs.


3. calibrate CV inputs:

the next page says ADC CV1 / ADC value at 0V: with no signal connected to the inputs, twist the right encoder so that the value shown is as close to 0 as possible.


Is this the value on my multimeter, or am I just ‘zeroing’ the figure on the O_C? If it’s the MM value, I’m getting no reading. Im a real DIY novice.

User avatar
Altitude909
Super Deluxe Wiggler
Posts: 3394
Joined: Wed Aug 24, 2011 5:13 pm
Location: Meesheegan

Re: ornaments+crimes extended / polymorphic 4x16bit CV generator

Post by Altitude909 » Thu Jan 30, 2020 12:00 pm

No, you are just zeroing the value on the front panel.. Thats the ADC trim so its done with nothing (aka "0") plugged. It should be in the ball park (-20-20), if its way out of wack, you have soldering issues on the input circuitry
Last edited by Altitude909 on Thu Jan 30, 2020 3:03 pm, edited 1 time in total.

Post Reply

Return to “Music Tech DIY”