Author |
Message |
Thalassa
Joined: Jan 27, 2006 Posts: 95 Location: Spain
Audio files: 5
|
Posted: Wed Nov 25, 2009 4:13 am Post subject:
|
|
|
Just share my own experience building this filter because I also had problems with the resonance and input levels.
I've build this filter two times, the first time I have problems with the resonance, the levels etc. I was changing few times resistors , the JFET .... at the end I had nice filter sound but the PCB was a bit damaged due soldereing/ desoldering so I made another PCB and on the places that I would need to test components I installed pin sockets ( http://www.musikding.de/product_info.php/info/p243_20-Pin-inline-socket.html )
As Yves says on his site, probably you will need to check few JFETs to find the right one. I test the filter with more than ten 2N3819 to choose the one that give good resonance levels. With some of them the filter don't self-oscillate and with others resonance was limited , only two or three I have good results.
Hope that this could help you |
|
Back to top
|
|
|
arnoid
Joined: Aug 23, 2009 Posts: 57 Location: Belgium
|
Posted: Wed Nov 25, 2009 7:55 am Post subject:
|
|
|
I also had problems to get self-oscillations of the resonance on this module I tryed different JFET without result.
Changing 680K in 150K did the trick for me.
I love the sound of this filter thx for the design yves. |
|
Back to top
|
|
|
numbertalk
Joined: May 05, 2008 Posts: 992 Location: Austin, TX
Audio files: 5
|
Posted: Wed Nov 25, 2009 9:27 am Post subject:
|
|
|
Thanks for the tips. I changed that resistor to 150K and it self-oscillates, which makes me think it's working, no? I'd hate to go messing with the transistor if it's not going to help - if I needed to change the transistor it wouldn't oscillate, right? Maybe it is just a matter of getting the input level right?
Would it be the case with a 10Vp-p signal if not properly attenuated I would observe this behavior - of somewhat weak resonance until it just oscillates? I changed R8 back to its originally specified value and am just using the input attenuator now. When dialed in to what seems right, I get some pronounced resonance, but it's all at settings close to the point before it self-oscillates. Any setting of the resonance pot between 0-7 doesn't really add much, it's subtle - it only really seems to kick in and get noticeable and nasty closer to the point of oscillation - around 8-9. Does this sound right to you guys, or do you get a rich "expected" 4P LPF resonance sound at lower settings as well on yours? EDIT - I see this is a 3P filter, so does that maybe explain my confusion - is it normal to need to turn the resonance this high to really get the effect? |
|
Back to top
|
|
|
arnoid
Joined: Aug 23, 2009 Posts: 57 Location: Belgium
|
Posted: Thu Nov 26, 2009 1:00 am Post subject:
|
|
|
numbertalk wrote: |
Any setting of the resonance pot between 0-7 doesn't really add much, it's subtle - it only really seems to kick in and get noticeable and nasty closer to the point of oscillation - around 8-9. Does this sound right to you guys |
It's the same on my filter the self-oscillation kicks in at potsetting 8, I think this is normal ? Can someone confirm this |
|
Back to top
|
|
|
yusynth
Joined: Nov 24, 2005 Posts: 1314 Location: France
|
Posted: Thu Nov 26, 2009 1:18 am Post subject:
|
|
|
Hi numbertalk
resonance should grow linearly with the setting of the resonance pot until it starts oscillating around tick 9.
Are you using a 25K linear pot (if using ALPHA pots or japaneese pots it should read 25KB not 25KA) ? _________________ Yves |
|
Back to top
|
|
|
Thalassa
Joined: Jan 27, 2006 Posts: 95 Location: Spain
Audio files: 5
|
Posted: Thu Nov 26, 2009 3:32 am Post subject:
|
|
|
arnoid wrote: | numbertalk wrote: |
Any setting of the resonance pot between 0-7 doesn't really add much, it's subtle - it only really seems to kick in and get noticeable and nasty closer to the point of oscillation - around 8-9. Does this sound right to you guys |
It's the same on my filter the self-oscillation kicks in at potsetting 8, I think this is normal ? Can someone confirm this |
The same on mine, the resonance increases normally within 0-8 and after 8 the self-oscillation appears very very strong.
I don't know... I have like hate-love relation with this filter, I'm not very satisfied with accuracy of the building, it's very very hard to match diodes because the voltage drop drift a lot with temperature and you can have totally different lectures on the same diode just because you has touched it with your fingers or the room is cooler or warmer than the previous reading.
I'm not so sure if I will keep this filter on my system I love the sound that you get filtering noises but maybe is not enough reason to keep it. In comparative with the ARP filter I don't have the same excitation using this one...... maybe the VCS3 filter it's not for me |
|
Back to top
|
|
|
numbertalk
Joined: May 05, 2008 Posts: 992 Location: Austin, TX
Audio files: 5
|
Posted: Thu Nov 26, 2009 10:00 am Post subject:
|
|
|
yusynth wrote: | Hi numbertalk
resonance should grow linearly with the setting of the resonance pot until it starts oscillating around tick 9.
Are you using a 25K linear pot (if using ALPHA pots or japaneese pots it should read 25KB not 25KA) ? |
Hi Yves. Yep 25K linear pot. When I get home from traveling this weekend I plan to try lowering that resistor value even more - is there a value I need to be careful not to go below (it's 150K right now). Might just hook a pot up there to see if I can find a resistance value there to improve the resonance range.
I've had problems matching the diodes too. No diode ever measures the same across multiple measurements and the cv bleedthru is pretty bad. |
|
Back to top
|
|
|
yusynth
Joined: Nov 24, 2005 Posts: 1314 Location: France
|
Posted: Thu Nov 26, 2009 10:44 am Post subject:
|
|
|
If you have noticeable bleedthrough try to change the CA3046 _________________ Yves |
|
Back to top
|
|
|
numbertalk
Joined: May 05, 2008 Posts: 992 Location: Austin, TX
Audio files: 5
|
Posted: Thu Nov 26, 2009 11:59 am Post subject:
|
|
|
I'll try that thanks
so no danger going too low with that resistor? |
|
Back to top
|
|
|
yusynth
Joined: Nov 24, 2005 Posts: 1314 Location: France
|
Posted: Thu Nov 26, 2009 4:28 pm Post subject:
|
|
|
numbertalk wrote: | I'll try that thanks
so no danger going too low with that resistor? |
I would not go lower than 50K _________________ Yves |
|
Back to top
|
|
|
numbertalk
Joined: May 05, 2008 Posts: 992 Location: Austin, TX
Audio files: 5
|
Posted: Sat Nov 28, 2009 6:37 pm Post subject:
|
|
|
Thanks Yves. Just got home and quickly experimented with a 120K resistor. Still same thing basically - oscillates a little earlier but still basically very subtle resonance between 0-8, then at 8 it gets super dirty, which is great, I guess I'm just looking for more in between. Might try a lower resistance here, might just live with it.
Anyway, I also tried swapping in a few different 3046s, all with the same terrible CV bleedthrough. Is it possibly the batch of 3046s I have or more/equally likely my diodes aren't matched well enough?
Thanks. |
|
Back to top
|
|
|
yusynth
Joined: Nov 24, 2005 Posts: 1314 Location: France
|
Posted: Sun Nov 29, 2009 8:22 am Post subject:
|
|
|
Bleedthrough is due to misbalance between the left and right side of the ladder, that's why it is best to have the diodes matched by pairs. The same stands for the CA3046. I experienced the same problem on the prototype of my minimoog filter "clone" and it was solved y using the right brand of CA3046. One must avoid Malay brands and should preferably use true Intersil CA3046. _________________ Yves |
|
Back to top
|
|
|
numbertalk
Joined: May 05, 2008 Posts: 992 Location: Austin, TX
Audio files: 5
|
Posted: Sun Nov 29, 2009 8:43 am Post subject:
|
|
|
So the diodes can be matched in pairs and don't *all* need to be matched to each other? And just to be clear about how to determine which are the pairs, D1 & D2 are a pair, D3 & D4, etc...? Still, I've tried this is different rooms of my house, let them settle, not touched them and I measure a diode once, go back in 5 minutes to measure the same one again and get a totally different measurement.
I have LM3046 chips. I'll see if Scott at Bridechamer has the Intersils and grab one there.
Thanks.
yusynth wrote: | Bleedthrough is due to misbalance between the left and right side of the ladder, that's why it is best to have the diodes matched by pairs. The same stands for the CA3046. I experienced the same problem on the prototype of my minimoog filter "clone" and it was solved y using the right brand of CA3046. One must avoid Malay brands and should preferably use true Intersil CA3046. |
|
|
Back to top
|
|
|
yusynth
Joined: Nov 24, 2005 Posts: 1314 Location: France
|
Posted: Sun Nov 29, 2009 10:05 am Post subject:
|
|
|
Quote: | And just to be clear about how to determine which are the pairs, D1 & D2 are a pair, D3 & D4, etc...? |
The measure must be brief (a few seconds) in the same room. It is depending on temperature and if the temperature of your room is not steady you will see a change in the measured value. Another important point is not to handle the diodes with your fingers because it warms them ! _________________ Yves |
|
Back to top
|
|
|
numbertalk
Joined: May 05, 2008 Posts: 992 Location: Austin, TX
Audio files: 5
|
Posted: Sun Nov 29, 2009 6:23 pm Post subject:
|
|
|
Thanks Yves. I painstakingly replaced all the diodes and don't think I'm capable of getting them any closer - did it all within the same room, within minutes of each other. Still fluctuation between measurements but it's the best I could do. Still terrible bleedthrough. I'll see if Bridechamber has Intersil 3046 chips and see if replacing with that helps at all and report back. |
|
Back to top
|
|
|
numbertalk
Joined: May 05, 2008 Posts: 992 Location: Austin, TX
Audio files: 5
|
Posted: Mon Nov 30, 2009 10:55 am Post subject:
|
|
|
Hi Yves.
Heard from Scott and the 3046s he has are RCA - would these be good enough or do I need to hunt down an Intersil chip?
Thanks. |
|
Back to top
|
|
|
yusynth
Joined: Nov 24, 2005 Posts: 1314 Location: France
|
Posted: Mon Nov 30, 2009 12:17 pm Post subject:
|
|
|
Yes it should do, importantly is not to use cheap asian subtitutes _________________ Yves |
|
Back to top
|
|
|
numbertalk
Joined: May 05, 2008 Posts: 992 Location: Austin, TX
Audio files: 5
|
Posted: Mon Nov 30, 2009 1:26 pm Post subject:
|
|
|
yusynth wrote: | Yes it should do, importantly is not to use cheap asian subtitutes |
Thanks. At $1 a pop I'll give it a shot, but realized that I have LM3046, so those *should* be made by National, so still a little nervous it won't solve the issue. |
|
Back to top
|
|
|
yusynth
Joined: Nov 24, 2005 Posts: 1314 Location: France
|
Posted: Mon Nov 30, 2009 1:30 pm Post subject:
|
|
|
By the way, how much of a bleedthrough do you get because you can't expect from the EMS design to avoid it completely ? _________________ Yves |
|
Back to top
|
|
|
numbertalk
Joined: May 05, 2008 Posts: 992 Location: Austin, TX
Audio files: 5
|
Posted: Mon Nov 30, 2009 1:34 pm Post subject:
|
|
|
yusynth wrote: | By the way, how much of a bleedthrough do you get because you can't expect from the EMS design to avoid it completely ? |
It's pretty bad - with no audio input, if I feed a LFO into the CV input I get the LFO signal really loud at the filter output. Is this to be expected? It seems excessive to me. I can upload an audio sample later if that would help. |
|
Back to top
|
|
|
Sebo
Joined: Apr 27, 2007 Posts: 564 Location: Argentina
|
Posted: Wed Dec 02, 2009 8:28 am Post subject:
|
|
|
yusynth wrote: | I just had the opportunity to read the service manual of the JEN SX2000 |
Hi Yves:
Where can I find the SX-2000 service manual? A friend of mine wants to add
CV/Gate to it.
Thanks. _________________ Sebo
---------------------------------------
My Music:
https://www.facebook.com/cosaquitos/ |
|
Back to top
|
|
|
yusynth
Joined: Nov 24, 2005 Posts: 1314 Location: France
|
Posted: Wed Dec 02, 2009 10:31 am Post subject:
|
|
|
Hi Sebo
PM me I will send it to you. _________________ Yves |
|
Back to top
|
|
|
Sebo
Joined: Apr 27, 2007 Posts: 564 Location: Argentina
|
Posted: Thu May 13, 2010 1:52 pm Post subject:
|
|
|
Hi Yves and all:
I just finished the EMS VCF and I want to share my experiences.
First I used a 680K for R41, and the resonance pot did almost nothing, then
I changed R41 to 220K and I get some resonance but not self oscillation, also
I measured the voltage at the gate of the FET and was swinging from 0 to
-3V.
Then I replaced R41 for a 100KA trimpot an a 100K resistor in series (that
was what I have at hand), and tweaked the resistance to get self oscillation.
I have to say that the behavior of the resonance pot is as reported by other
builders. Tweaking the trimpot enhaced a bit the response, I found that with
a resistance of 162K (100k+62K from the trimpot) the resonance pot have
the best response I can get. If the resistance is smaller the oscillation
kicks earlier and the response is worse. With larger resistance the response
is better but I can't get full oscillation.
Also noted that in opposition to other filters I build (Minimoog, Steiner, WASP),
this one self oscillate from sub sonic frequencies but stop oscillating at
about 9KHz. I that normal?
I didn't calibrated properly yet, I have a gig tonight tomorrow will do it
and take some pictures. _________________ Sebo
---------------------------------------
My Music:
https://www.facebook.com/cosaquitos/ |
|
Back to top
|
|
|
Sebo
Joined: Apr 27, 2007 Posts: 564 Location: Argentina
|
Posted: Sat May 29, 2010 8:33 am Post subject:
|
|
|
Hi:
I've calibrated the filter, it sounds great without resonance, very clean, a
lot different of my other filters. With high resonance the sound is nice too
but when start to self oscillate the filter do some weird things, the cutoff
frequency becomes unstable, but when the resonance reaches full self oscillation
the behavior gets back to normal. Also is strange that the self oscillation disappear
at high frequency (beyond 9KHz).
I don't know if all this is normal.
One thing that really surprise me was that the filter tracks 1V/Oct very well,
far better than the Minimoog VCF. Is perfectly in tune by 2 octaves, but out
of that range the deviation is minimal, less than half a semitone per octave.
A funny thing is that if I modulate the cutoff with a LFO and I move the
cutoff frequency, the LFO frequency compresses and expands itself, it's
hard to explain...
Here are some pics of the VCF.
Description: |
|
Filesize: |
168.88 KB |
Viewed: |
503 Time(s) |
This image has been reduced to fit the page. Click on it to enlarge. |
|
Description: |
|
Filesize: |
174.55 KB |
Viewed: |
673 Time(s) |
This image has been reduced to fit the page. Click on it to enlarge. |
|
_________________ Sebo
---------------------------------------
My Music:
https://www.facebook.com/cosaquitos/ |
|
Back to top
|
|
|
yusynth
Joined: Nov 24, 2005 Posts: 1314 Location: France
|
Posted: Sat May 29, 2010 9:29 am Post subject:
|
|
|
Sebo wrote: | Hi:
I've calibrated the filter, it sounds great without resonance, very clean, a
lot different of my other filters. With high resonance the sound is nice too
but when start to self oscillate the filter do some weird things, the cutoff
frequency becomes unstable, but when the resonance reaches full self oscillation
the behavior gets back to normal. Also is strange that the self oscillation disappear
at high frequency (beyond 9KHz).
I don't know if all this is normal.
One thing that really surprise me was that the filter tracks 1V/Oct very well,
far better than the Minimoog VCF. Is perfectly in tune by 2 octaves, but out
of that range the deviation is minimal, less than half a semitone per octave.
A funny thing is that if I modulate the cutoff with a LFO and I move the
cutoff frequency, the LFO frequency compresses and expands itself, it's
hard to explain...
Here are some pics of the VCF. |
I Sebo
As a matter of fact the behaviour of this file is depending very much of the input level. If you feed it with yusynth VCOs, you have to set the input level to mark 3 (or maximum 4) to get the EMS sound, if you go beyond that mark the filters starts sounding like a Moog filter up to mark 6 where the filter starts to saturate. If you are using levels above mark 6 you will have the behaviour you described close to auto-oscillation. Somehow the autoscillation starts to interfere with the input signal giving harsh sounds.
I could have limited the input level of this filter to the "EMS range" that is under mark 4, but I thought it is nice also to go beyound and have larger palette of sounds with the same filter.
This is a very general remark concerning all the original filters (Moog style, ARP style and EMS style) they were designed to work with a given input level. Therefore, when incorporated as a single module in a modular one must play the input levels to obtain the desired behaviour _________________ Yves |
|
Back to top
|
|
|
|