MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

Serge DUSG noticeably drifting
MUFF WIGGLER Forum Index -> Buchla, EMS & Serge  
Author Serge DUSG noticeably drifting
jersupereq
So I've been using my DUSGs and DTGs as a clocking source but while recording I noticed the tracks keep going out of phase. Is this an issue that can be fixed or should I consider getting a more reliable master clocking source?
cebec
If the Rise/Fall aren't exactly the same, you may notice some drifting, but it should be minute if you can eyeball it closely enough.

You could also try turning the Rise/Fall pots fully counterclockwise and patch something like a Scaling Buffer or ACPR offset into the VC IN of both to control them with a single knob.

The VC Clock and VC Timegen Osc are the most stable clocking modules, however.
cygmu
I think it's inevitable that the DUSG cycle rate drifts a bit with temperature, except perhaps on the Random Source versions. The rate is controlled by pairs of transistors, which are matched to provide a certain amount of stability, but the additional temperature compensation or stabilisation you would find in a VCO is not present in the DUSG, other than the Random Source version (I think -- never actually seen one but someone told me there was temperature compensation in there).

Are yours STS, Random Source, or CGS DIY ones? If DIY, you might be able to increase stability by matching the transistors better.
tmeade
Can you hear it drift at audio rate? I had one that wavered slightly and changing the pot did the trick.
Navs
jersupereq wrote:
... while recording I noticed the tracks keep going out of phase.


What do you mean by this? Can you describe the set up for us? e.g. do you have two clocks that are running, or one clock that is driving multiple slaves or are you recording multiple passes of the same free-running clock?
jersupereq
I believe I had the USG's rise and fall set differently which could be the cause for the noticeable drift at sub audio range. I'm getting a gator so maybe the VC Timegen Clock will make a better master clock for my system.

It's was a newer STS USG

I haven't noticed it at audio rate, but I could just be missing it.

So from 1 USG set to loop at a sub audio rate, I use the trigger out to play a percussive sound, then a second USG to divide the first by two to play a second sound. I recorded this into logic. I then did a second recording of this patch with some variations on the sounds being played but not the same tempo, they drift in and out of being in sync.

I'll try some of these tricks and report back, Thanks!
Navs
jersupereq wrote:
So from 1 USG set to loop at a sub audio rate, I use the trigger out to play a percussive sound, then a second USG to divide the first by two to play a second sound. I recorded this into logic. I then did a second recording of this patch with some variations on the sounds being played but not the same tempo, they drift in and out of being in sync.


Why you would expect them to be in sync, especially given the two recordings feature different tempi?
jersupereq
Navs wrote:
jersupereq wrote:
So from 1 USG set to loop at a sub audio rate, I use the trigger out to play a percussive sound, then a second USG to divide the first by two to play a second sound. I recorded this into logic. I then did a second recording of this patch with some variations on the sounds being played but not the same tempo, they drift in and out of being in sync.


Why you would expect them to be in sync, especially given the two recordings feature different tempi?


I meant the same tempo. The clocking DUSG was never altered during the two recordings.
wackelpeter
If i understand it right you're clocking your second slope by putting the end out pulse from the first slope into the trig in, correct?
In this case the second slope will only trigger again when the cycle is completed... means if the slope is slightly longer as the slope of the first one it's end out pulse will vary...

Or do i misunderstood you and also the timing of your clocking slope alters and drifts?
johnnywoods
What sort of case/bus board do you use? I was having a lot of drift issues with my serge, and it turned out my busboard was picking up some interference. I moved it to a different location in the case, and everything cleared up.
Navs
jersupereq wrote:
Navs wrote:
jersupereq wrote:
So from 1 USG set to loop at a sub audio rate, I use the trigger out to play a percussive sound, then a second USG to divide the first by two to play a second sound. I recorded this into logic. I then did a second recording of this patch with some variations on the sounds being played but not the same tempo, they drift in and out of being in sync.


Why you would expect them to be in sync, especially given the two recordings feature different tempi?


I meant the same tempo. The clocking DUSG was never altered during the two recordings.


OK. But even given the same tempo, they will phase unless you sync them somehow. It's the same if you record a fixed-frequency VCO twice ... or for that matter a drum machine. Or am I still misunderstanding? hmmm.....
jersupereq
Navs wrote:
jersupereq wrote:
Navs wrote:
jersupereq wrote:
So from 1 USG set to loop at a sub audio rate, I use the trigger out to play a percussive sound, then a second USG to divide the first by two to play a second sound. I recorded this into logic. I then did a second recording of this patch with some variations on the sounds being played but not the same tempo, they drift in and out of being in sync.


Why you would expect them to be in sync, especially given the two recordings feature different tempi?


I meant the same tempo. The clocking DUSG was never altered during the two recordings.


OK. But even given the same tempo, they will phase unless you sync them somehow. It's the same if you record a fixed-frequency VCO twice ... or for that matter a drum machine. Or am I still misunderstanding? hmmm.....


They used the same DUSG as the clocking source without modifying it between takes and continued to go in and out of phase. One overtaking the other.

Update: this issue seems substantially less noticeable when the rise and fall were set the same. So I guess that was the issue. Makes sense why the TGO only puts out a pure triangle wave.
Vsyevolod
It is the nature of analog to have various degrees of fluctuation. Yes, you can minimize the fluctuation but never get rid of it. I would just record a click track and feed that back into your modules when recording successive tracks. Bam, instant sync forever.

Stephen




.
nordlead
jersupereq wrote:
Makes sense why the TGO only puts out a pure triangle wave.


Actually that's a (reverse) sawtooth (Rise at max and Fall variable) - it appears that provides the best DSG tracking, whereby "best" is really relative - the (STS) TGO only covers a bit more than one octave. However, that setting is also recommended to get the (R*S) DSG mk2 to track (over 4+ octaves).

As Stephen said, without using a quartz, analog will always have variations in pitch / frequency - so even when recording the same DSG (or PCO) and layering it, to have no phasing / beating effect, the DSG would have to have *perfect* pitch, i.e. no variation at all - something that's typically not even desired (check out J├╝rgen Haible's thoughts on his Living VCOs page re vintage oscillators - the LVCOs are designed to actually use that "beating" against each other ...)
Jari Jokinen
I have a quick question somewhat on topic. (I sold my DUSG, but came to my senses and just bought a new one.)

If my Synthcube MU format DUSG is set to self cycle, frequency depends on which output(s) are tapped.
1. Bipolar output only: Let's say this gives reference frequency.
2. Unipolar output only: Frequency is a quarter tone lower.
3. Bipolar output and Unipolar output: Frequency is a semitone lower.

I wonder, if this is normal?
Navs
edit: answer to Jari's question, not the OP:

I don't know if it is normal but I have an idea of why it might be possible.

Have a look at the circuit at Elby's site:

http://www.elby-designs.com/webtek/cgs/serge/cgs75/cgs75_vcs.html

Both the 'AC' and normal output are back connected to the 100k trimmer. If I remember correctly, this sets the threshold for the comparator that enables looping. I don't know the MU design, but on the old Bananalogue and newish Doepfer builds, this trimmer can be used to set the maximum frequency. It also has the effect of changing the output amplitude, which stands to reason. Tim's description might help: http://www.timstinchcombe.co.uk/index.php?pge=vcs

So, my suspicion might be that that the voltage level is dropping and that is what is causing the difference in frequency. Is there a way you can check the change with a different load/ destination module? Send it to a buffer first to see if that makes a difference? Measure the frequency of the End Out by comparison? The point is, it shouldn't make a difference which output you use.

I can hook up my VCSs at some stage to compare. One last thought, tho: how are you measuring the change in frequency, by ear? Could it be perceived? As you can see from the circuit, in the versions that stick to that design, the 'AC' output is only crudely achieved, with a cap and resistor, not biasing ...
wavecircle
I notice in my recordings that when I put them into Ableton the clock does get very very slightly quicker over time (it makes warping a PITA). I think locking it to a MIDI clock is probably the best way to work with the serge in the context of a project.
Jari Jokinen
Indeed, there is no difference in pitch when the outputs are routed to Hordijk Mini Matrix which is buffered. I previously tried many destinations but not the Matrix.
cygmu
Navs wrote:

Have a look at the circuit at Elby's site:

http://www.elby-designs.com/webtek/cgs/serge/cgs75/cgs75_vcs.html

Both the 'AC' and normal output are back connected to the 100k trimmer.


Good spot, Navs.

This has reminded me: a friend pointed out a few months back that the STS PCBs differ from the CGS ones at exactly this spot. On the STS ones, the connection to that trimmer comes directly from the op amp output, i.e. from the other side of the 100R resistor. This explains why that is a good idea, I think. I wonder why Ken moved it? Stability perhaps?
MUFF WIGGLER Forum Index -> Buchla, EMS & Serge  
Page 1 of 1
Powered by phpBB © phpBB Group