gml opamps
- Thread starter rafafredd
- Start date
Help Support GroupDIY Audio Forum:
SSLtech
Well-known member
There are a goodly number of switched-capacitor filter systems and dedicated ICs specifically for that purpose.
There's be little or no point in PWM-controlling an LED intensity to me, because you're just going a very long way round the houses to do something that can be done without PWM... there's never been a problem modulating an LED's brightness with a 0-10V DC signal for example... and you still end up with an opto-compressor either way... hardly seems worth the effort. :sad:
The big deal about PWM is that you use the FET (or transistor, or phototransistor...whatever) as a switch, so you're avoiding the nonlinear distortion. The FET has to be fairly fast, and you trade noise, level and distortion limitations for a prescribed bandwidth limitation right up front... but that's the whole thing to me: different approaches yielding different results, therefore different sounding product at the end of it all.
Keith
There's be little or no point in PWM-controlling an LED intensity to me, because you're just going a very long way round the houses to do something that can be done without PWM... there's never been a problem modulating an LED's brightness with a 0-10V DC signal for example... and you still end up with an opto-compressor either way... hardly seems worth the effort. :sad:
The big deal about PWM is that you use the FET (or transistor, or phototransistor...whatever) as a switch, so you're avoiding the nonlinear distortion. The FET has to be fairly fast, and you trade noise, level and distortion limitations for a prescribed bandwidth limitation right up front... but that's the whole thing to me: different approaches yielding different results, therefore different sounding product at the end of it all.
Keith
Mark Burnley
Well-known member
[quote author="Gus"]
There are some problems with 555s when it switches it turns on hard and can make an almost inpossable to remove power glitch. I believe some of the newer versions might take care of that problem.[/quote]
Yep,
CMOS 555's are the 7555. They have minimal "crowbarring" of the power rails, and have a very low quiescent current draw. Current increases with increasing f though.
Mark
There are some problems with 555s when it switches it turns on hard and can make an almost inpossable to remove power glitch. I believe some of the newer versions might take care of that problem.[/quote]
Yep,
CMOS 555's are the 7555. They have minimal "crowbarring" of the power rails, and have a very low quiescent current draw. Current increases with increasing f though.
Mark
NewYorkDave
Well-known member
I agree. This is a very interesting project, and it deserves its own thread.
Mark Burnley
Well-known member
Keith,
Sounds like the gauntlet has been thrown down...
:wink:
(BTW, spent the afternoon in Keith's when I shoulda been getting on with stuff :guinness: :green: )
Mark
Sounds like the gauntlet has been thrown down...
:wink:
(BTW, spent the afternoon in Keith's when I shoulda been getting on with stuff :guinness: :green: )
Mark
SSLtech
Well-known member
Gus,
I suppose it is in a fashion. As soon as I can get someone to host the file, I'll start a new thread explaining everything.
...anybody with 100k web space to host the file?
Keith
I suppose it is in a fashion. As soon as I can get someone to host the file, I'll start a new thread explaining everything.
...anybody with 100k web space to host the file?
Keith
Kev
Well-known member
I can do that
SSLtech
Well-known member
Email(s) sent.
:green:
Keef :thumb:
:green:
Keef :thumb:
Kev
Well-known member
done
it's in two places
my place
http://www.celestial.com.au/~rosswood/data/CompressorSchematic.pdf
and the old place
http://recording.org/users/kev/CompressorSchematic.pdf
looks like a PA comp
is that a double high pass filter I see at the input ?
The use of 5534 at the output could be subbed for 5532 or LM833 to save opamps and simplify the PCB.
LM319 + 555 ... I don't get this yet
Leave me alone .. I've had a long day at work and I'm tired !! :sad:
it's in two places
my place
http://www.celestial.com.au/~rosswood/data/CompressorSchematic.pdf
and the old place
http://recording.org/users/kev/CompressorSchematic.pdf
looks like a PA comp
is that a double high pass filter I see at the input ?
The use of 5534 at the output could be subbed for 5532 or LM833 to save opamps and simplify the PCB.
LM319 + 555 ... I don't get this yet
Leave me alone .. I've had a long day at work and I'm tired !! :sad:
Interesting, to say the least!
What is the gain/attenuation range of this VCA? Could it be large enough to incooporate an expander/gate sidechain as well?
Jakob E.
What is the gain/attenuation range of this VCA? Could it be large enough to incooporate an expander/gate sidechain as well?
Jakob E.
SSLtech
Well-known member
Jakob,
I tested it down to -40dB or so. -I never thought about adding a Gate, and sonce the compressor is in feed-back, the gate woud interfere with the compressor's smooth operation without more complexity.
Jakob, if you'd like to surgically split this thread, I'd be very happy to add a circuit description or something, but at the moment it's sort of hijacked another thread, so I'm reluctant.
Keith
I tested it down to -40dB or so. -I never thought about adding a Gate, and sonce the compressor is in feed-back, the gate woud interfere with the compressor's smooth operation without more complexity.
Jakob, if you'd like to surgically split this thread, I'd be very happy to add a circuit description or something, but at the moment it's sort of hijacked another thread, so I'm reluctant.
Keith
SSLtech
Well-known member
-A-ha!
I saw a few glaring errors on the sketch.
There was nothing to buffer the meter in "level" mode, the rectifier won't work as drawn, the inverting followers to the 'chopper' stage are going to produce gain-lift in the "off" position.
These have been changed now, though I don't know when Kev will have time to update the hosting.
Oh, I also removed the cross-coupled output stage... I figured it's probably easier to use a DRV134 or SSM2142 or something... for now there's a straight inverter as per the Gyraf SSL outputs.
Keith
I saw a few glaring errors on the sketch.
There was nothing to buffer the meter in "level" mode, the rectifier won't work as drawn, the inverting followers to the 'chopper' stage are going to produce gain-lift in the "off" position.
These have been changed now, though I don't know when Kev will have time to update the hosting.
Oh, I also removed the cross-coupled output stage... I figured it's probably easier to use a DRV134 or SSM2142 or something... for now there's a straight inverter as per the Gyraf SSL outputs.
Keith
Learner
Well-known member
Hey keith,
Why don't you start a new thread to explain the concept of your design first? I think that's what we are all dyin to find out, even though there are some errors in your current layout the changes to be made aren't goin to change the concept and structure anyway.
:guinness:
Why don't you start a new thread to explain the concept of your design first? I think that's what we are all dyin to find out, even though there are some errors in your current layout the changes to be made aren't goin to change the concept and structure anyway.
:guinness:
tk@halmi
Well-known member
Wow, Keith! This IS something different. It does deserve its own thread.
Tamas
Tamas
Keith,
I was thinking of maybe a 82E10-type of sidechain - based on a RMS-converter, and feed-forward. I really like that type on drums, and the associated gate/expander is one of the most useable around. But it would take some minor conversion to change this pwm-circuit into one with the needed log-control-voltage type.
Sadly enough, it is not within my power to change thread content. My powers are - for now - merely destructive...
Jakob E.
I was thinking of maybe a 82E10-type of sidechain - based on a RMS-converter, and feed-forward. I really like that type on drums, and the associated gate/expander is one of the most useable around. But it would take some minor conversion to change this pwm-circuit into one with the needed log-control-voltage type.
Sadly enough, it is not within my power to change thread content. My powers are - for now - merely destructive...
Jakob E.
[quote author="gyraf"]Keith,
I was thinking of maybe a 82E10-type of sidechain - based on a RMS-converter, and feed-forward.
[...]
Jakob E.[/quote]
Just wanted to be supportive here. I've found that the idea of doing dynamics in a feed-forward architecture is the key to stabilizing a device's sound across uses and at all levels. Feedback topologies are themselves inherently unstable (overshoot, oscillation & etc).
The challenge with doing RMS detection and RMS-controlled VCA's is that it's very expensive (talking about parts here) to get right. On the other hand, it's the only way to do log (and audio after all is logarithmic) control.
George
I was thinking of maybe a 82E10-type of sidechain - based on a RMS-converter, and feed-forward.
[...]
Jakob E.[/quote]
Just wanted to be supportive here. I've found that the idea of doing dynamics in a feed-forward architecture is the key to stabilizing a device's sound across uses and at all levels. Feedback topologies are themselves inherently unstable (overshoot, oscillation & etc).
The challenge with doing RMS detection and RMS-controlled VCA's is that it's very expensive (talking about parts here) to get right. On the other hand, it's the only way to do log (and audio after all is logarithmic) control.
George
Learner
Well-known member
WoW!~
GM was here AGAIN!!!!!! :shock:
AND OFFERING HIS VALUEBLE EXPERTISE!!!!!!!!! :thumb:
Would love to see him drop by more often, if his schedule permits.
Thanks George! :guinness:
GM was here AGAIN!!!!!! :shock:
AND OFFERING HIS VALUEBLE EXPERTISE!!!!!!!!! :thumb:
Would love to see him drop by more often, if his schedule permits.
Thanks George! :guinness:
DrFrankencopter
Well-known member
Hi George, nice to see you back!
I've always liked feedforward topologies from an analytical aspect, but in practice, the feedback topologies seem to end up sounding 'better'. I realize that with a feedback arrangement you lose accuracy (time constants are a function of ratio), and gain a bit more distortion, but the result seems to be smoother than that with a feedforward setup.
Maybe it's just a matter of getting the feedforward right. I really like Mark McQ's (FMR RNC) approach with the digital sidechain...I think that's a great way to do feedforward (provided the sampling group delay doesn't interfere with the ability to get fast attack times). With DSP coming down in price and increasing in power it might be a great way to keep costs down too.
Cheers,
Kris
I've always liked feedforward topologies from an analytical aspect, but in practice, the feedback topologies seem to end up sounding 'better'. I realize that with a feedback arrangement you lose accuracy (time constants are a function of ratio), and gain a bit more distortion, but the result seems to be smoother than that with a feedforward setup.
Maybe it's just a matter of getting the feedforward right. I really like Mark McQ's (FMR RNC) approach with the digital sidechain...I think that's a great way to do feedforward (provided the sampling group delay doesn't interfere with the ability to get fast attack times). With DSP coming down in price and increasing in power it might be a great way to keep costs down too.
Cheers,
Kris
Similar threads
