Author |
Message |
steffensen
Joined: Jul 11, 2012 Posts: 103 Location: Sweden
|
Posted: Sat Jan 03, 2015 12:50 pm Post subject:
Saw-core Hardsync |
|
|
Been playing around with the hardsync on my Odyssey VCO clone. (Layout by Andy)
As i don't have the real thing to compare with, i have no idea if it's suppose to sound like it does. The issue i have with it, is that its pretty noisy in a bad way each time you have the 2 needed VCOs 6 notes apart. It sounds great when sweeping the pitch in between tho! It's like when the master VCO and slave meets in the middle of the reset-phase they interfere with eachother. It sounds abit like a PW control going over its maximum limit.
I've seen the kind of hardsync the oddy have in several other circuits, seems to be a common thing. So i wonder, what can i do to improve it?
I have narrowed the trig signal switching the reset-JFET on/off, but the sound didnt improve much. I have set different kinds of limits for trig signal, so the JFET doesnt completely shut off and such. No difference really.
Out of ideas at this point, anyone with a bit more insight on the matter out there?
Here is a pic of how it looks like when this happens. The large trig/spike you see follows the syncing input all thru the freq range but its only at this kind of setting i have problems. I have tried lowering it, but the slave-VCO wont reset then, so i guess its supposed to be like that.
(im real bad with words for these things, sorry if i sound too waffly)
Description: |
|
Filesize: |
1.58 MB |
Viewed: |
159 Time(s) |
This image has been reduced to fit the page. Click on it to enlarge. |
|
|
|
Back to top
|
|
|
piedwagtail
Joined: Apr 15, 2006 Posts: 297 Location: shoreditch
Audio files: 3
|
Posted: Sat Jan 03, 2015 9:59 pm Post subject:
|
|
|
I guess what you're hearing in this case is the input overload on the following opamp as that spike bites into the rail. The sync pulse and magnitude probably isn't the culprit.
Maybe a small current protection 220 ohm resistor inline and a 11 to 12ish Volt Zener to ground after the VCO, such that it conducts only in this exception.
Since the Odyssey VCO Ramp is 0-10V iirc, the zener should just kick in during this circumstance.
Robert |
|
Back to top
|
|
|
steffensen
Joined: Jul 11, 2012 Posts: 103 Location: Sweden
|
Posted: Sun Jan 04, 2015 3:55 am Post subject:
|
|
|
Thanks for the reply,
Im not sure thats it, as the spike is 2V under the rail-limit.
(im powering this with 12V)
The original Ramp waveform is indeed 0-10V, with the Syncing spike going from -5 to 15V, according to the service manual.
By powering it from 12V, i get a Ramp that is about 0-7.8V with a Spike from -4.6V to 12V, which seems to be expected.
(well, it should be up to 12V going by my simulations, but real life value is actually around 11V, dont think the Oddy goes up to 15V either))
I however just found this info from Wikipedia regarding Hardsync:
"This effect may be achieved by measuring the zero axis crossings of the master oscillator and retriggering the slave oscillator after every other crossing.
This form of oscillator sync is more common than soft sync, but is prone to generating aliasing in naive digital implementations."
That explains my issue a lot better than my rambling. Its just when the Master and Slave meets at the zero crossing that i get this "noise" i mentioned. So i guess thats the Aliasing they refer to. Which means this is also to be expected?
Edit: This "noise" is also bound to the stepping, its just this kind of stepping i would like to improve, to have the sync sweep a lot smoother if possible.
A kind gentleman is recording his Oddy sync for me today, so i have something to compare with. This might just be normal, i have no idea. |
|
Back to top
|
|
|
Ricko
Joined: Dec 25, 2007 Posts: 251 Location: Sydney, Australia
Audio files: 27
|
Posted: Sun Jan 04, 2015 7:13 pm Post subject:
|
|
|
Aliasing is an issue for digital or sampled systems. Your VCO is analog not digital.
Does the problem still occur when synching to a square or pulse master? |
|
Back to top
|
|
|
steffensen
Joined: Jul 11, 2012 Posts: 103 Location: Sweden
|
Posted: Mon Jan 05, 2015 5:13 am Post subject:
|
|
|
Well true. Bad word for the issue i guess, perhaps stepping and the noise that happens when that occurs is a better explanation.
The issue is more obvious when syncing a square actually, the saw not so much.
Dont think theres much more to do than keeping the trig-spike really short, will try and shorten it even further. |
|
Back to top
|
|
|
commathe
Joined: Jul 26, 2013 Posts: 153 Location: Beijing
Audio files: 5
|
Posted: Wed Jan 07, 2015 12:20 am Post subject:
|
|
|
2V under the rail would still make sense for it to be op-amp saturation. Few op-amps can make it to the rails, and almost all of those can only be used from 0-5V |
|
Back to top
|
|
|
steffensen
Joined: Jul 11, 2012 Posts: 103 Location: Sweden
|
Posted: Wed Jan 07, 2015 3:22 am Post subject:
|
|
|
Thats true. The square does go thru a op amp while the saw doesnt, might have something to do with the square sounding worse. Doesnt explain the "noise" with the saw though.
After a lot of googling, this kind of noise happening at the zero axis seems to be common. Will update with my results from a shorter sync-spike. Im hammered with a bad cold atm. |
|
Back to top
|
|
|
steffensen
Joined: Jul 11, 2012 Posts: 103 Location: Sweden
|
Posted: Tue Jan 13, 2015 7:58 am Post subject:
|
|
|
Just thought i'd update with some results here. Still having a bad cold, but managed to still my curiosity with some testing even so.
Not sure anyone is interested, but in case you are here is what i discovered.
I tried a short trig/pulse/spike to switch the JFET used for sync, and i really cant hear any difference between using a trig of 1usec to 6usec wide.
The original trig normally used (from the VCO-Core's Comparator) is about 1.5usec wide.
So i played with the gain of the trig. (original is having its spike going from -5v to +15V)
Changing these levels didnt cause much change in sound really, besides that it stops to work when going too far up from -5V. (the JFET obviously is open all the time then, so no wonder)
Also, the only difference when lowering the higher threshold of 15V is that it might not sync, jumps around a bit. When it syncs, the sound is the same as before.
I noticed Bergfotron (in the BVCO) sent a short trig in to the VCO-Core's comparator section instead of discharging thru a JFET. So i tried this instead, and even if it sounds great, that kind of noise between the stages i've been talking about still persists.
As i still havent found anyone willing to record the original Sync behaviour for me, i have nothing to compare with. This all might be just normal after all, but its was fun to play with and to still all my ideas for improving the sync behaviour.
Anyway, might still someone elses curiosity. Feel free to add any more ideas for testing if you want, im all out. |
|
Back to top
|
|
|
|