THE ENGINEER'S THUMB... At last, a better compressor! [documentation]
- HEAD
- Resistor Ronker
Good points! But I realy have to admit that for guitar purposes it does its job pretty well so far. What I particular like about his design is how you are able to controll the attack time. This makes the thing at least to me pretty versitile. From hardly noticable to pretty funky.
- rocklander
- Old Solderhand
Information
- Posts: 2726
- Joined: 18 Apr 2008, 11:33
- my favorite amplifier: my jansen bassman 50
- Completed builds: rebote 2.5; supreaux; odie; heartthrob tremolo; ross phaser; dr. boogey; thor; baja black toast; slow gear attack, rebote, tri-vibe, small clone, little angel, magnus modulus, echo base, hex fuzz, big muff, 22/7.
- Location: Rotorua, New Zealand
- Has thanked: 1406 times
- Been thanked: 231 times
- Contact:
failed for me too.. could be an anti NZ thing? suspect local DNS servers don't know about it or something (haven't looked further into it). hidemyass.com gets to the page, but the project files don't appear (at least links to them). viewed the page souce and found the links so I can email them to you if you want.jstbrowsin wrote:um that seems to be a dead link
world's greatest tautologist ...in the world
Ronsonic wrote:...the lower the stakes the more vicious the combat.
atreidesheir wrote:He should be punched in the vagina.
- pz
- Solder Soldier
Or maybe i'll just post the files in a separate thread in the Ready-To-Build Effect Projects section?rocklander wrote:failed for me too.. could be an anti NZ thing? suspect local DNS servers don't know about it or something (haven't looked further into it). hidemyass.com gets to the page, but the project files don't appear (at least links to them). viewed the page souce and found the links so I can email them to you if you want.jstbrowsin wrote:um that seems to be a dead link
I think Merlin's thread is enough derailed now
Not sure why the page doesn't work where you are. Probably a DNS problem.
Piotr
- pedalgrinder
- Cap Cooler
http://www.hexeguitar.com/diy-tadek_e
Heres the link to it. I've seen it heaps of times cool compressor just not my taste. But hey most guitarists will cream over it.
Heres the link to it. I've seen it heaps of times cool compressor just not my taste. But hey most guitarists will cream over it.
What's the best thing about fat chicks and scooters? There both fun to ride around until your mates find out!
- pedalgrinder
- Cap Cooler
heres what i have got from that link enjoy
i can get to it not sure why others can't maybe some sort of filter between countrys
i can get to it not sure why others can't maybe some sort of filter between countrys
What's the best thing about fat chicks and scooters? There both fun to ride around until your mates find out!
You can control the release time to whatever you want- see the vero layout.pedalgrinder wrote: Hey merlinb i reckon you've done great job. I just reckon your release could be a longer time constant and your attack could use a seperate buffer.
Why does it need a buffer? The attack time is already independent of the release time.
- pz
- Solder Soldier
I think this is not a right place for these files. Let's not make any more mess in this thread, it's about Merlins compressor. (thanks for another cool design!)pedalgrinder wrote:heres what i have got from that link enjoy
i can get to it not sure why others can't maybe some sort of filter between countrys
I've set up a separate thread with all the project files for those, who cannot access the page: https://www.freestompboxes.org/viewtopic ... 13&t=17486
- pedalgrinder
- Cap Cooler
can a moderator shift them into a correct place for me i was only trying to help so all could see them. In regards to merlinb comments your using the feedback path of that op amp on the release i see. Correct me if iam wrong what i was suggesting which i should explain myself better is one op amp and cap time constant for each control. As you can see in the compressor schematic i posted it shows a sidechain like that to help explain what iam on about. Sometimes i don't explain myself properly. Iam not having a go at your compressor merlinb iam just suggesting that maybe this might help. i've found have a time constant for the attack seperate op amp then feeding into another op amp which seperates the controls essentially i see it as a form of a buffer into another larger cap on that that one that controls the release. Thats just my thoughts. I think you've done a great job you've got my support for sure.
What's the best thing about fat chicks and scooters? There both fun to ride around until your mates find out!
I don't think that would make any difference in this case. The attack and release resistors are already independent in my circuit, so each can be freely tweaked or replaced with pots for full control over both, with no interaction.pedalgrinder wrote: i've found have a time constant for the attack seperate op amp then feeding into another op amp which seperates the controls essentially
- pedalgrinder
- Cap Cooler
yeah i see what your saying merlinb but your were both talking about achieving the same thing using two different principles. Which approach is better well how long is a piece of string. I understand why you've done what you've done and i breadboarded your schematic sounds great i just played with it with another approach and it could be taste rather than an electronic aspect. But hey your design still sound sensational iam not knocking it just some friendly chat about it.
What's the best thing about fat chicks and scooters? There both fun to ride around until your mates find out!
- Seiche
- Old Solderhand
wtf did i just read? no idea about past grudges but the way people swooped in and started bashing was really fucked up IMHO and totally not in the FSB/DIY spirit. What happened to constructive criticism? And saying this doesn't work when you haven't built or simulated it or values are "wrong" is a little weird to say the least.
seriously, has anyone else built this? So far we have merlinb obviously who says it sounds great and HEAD who says it doesn't.
maybe you made a mistake somewhere when building/breadboarding it?? no offense, but as we all know this happens all the time.HEAD wrote:Hmm, by all means this compressor doesn't work or sound good on a guitar i.m.h.o. I built it and tried it and didn't like. Everyone is free to make his own decision...
seriously, has anyone else built this? So far we have merlinb obviously who says it sounds great and HEAD who says it doesn't.
- mictester
- Old Solderhand
Information
My quickly breadboarded version performed pretty much as I expected. The compressor reacts quickly to the initial note attack, but overshoots and you get a "swell" effect as the note begins to decay. It's an interesting effect in its own right, but not really a "clean sustainer" which is what we want a compressor to be. Tweaking the attack and decay times (halving both to start with) made the swell effect less obvious, and began to get towards normal compression sustain. With a sidechain take-off point AFTER the gain cell, the results were not dissimilar to the Morley pedal!
With a bit more tweaking, this could get very close to good, but would be radically improved with a delay line included in the audio path to overcome the overshoot issue.
The VCA as drawn is OK, but can be significantly improved by the use of the generously provided linearising diodes in the OTA, and by some recalculation of the bias values. The datasheet values are quite carefully chosen for optimum operation (the manufacturers really like to show their ICs off in the best possible circumstances), so these make for a good starting point.
There is an issue with loop stability when the level sensing is taken from before the gain cell, and this is why the vast majority of manufacturers take their rectifier inputs from the output of the VCA. This also confers other advantages, including minimising the parts required by providing a low impedance source to feed both the rectifier and the output stage.
If you look at the Dynacomp, the Black Finger, the Orange Squeezer or any other classical compressor design, you'll see the same topology. Electro Harmonix and Dan Armstrong shunt the audio to ground to provide variable level into a fixed-gain amplifier (though EH give you a gain control for that amp), and MXR use the current-variable gain of the OTA in the audio path. Neither approach is particularly good - shunting audio means that you have to make up level somewhere else, thereby introducing noise, and OTAs in the audio path are always nasty, causing distortion artifacts of their own. Using the OTA in the feedback loop of an op-amp (like here and in the Morley / Tel-ray circuits) is definitely preferred, but reduces the range of control available.
I find that two gain cells of this type cascaded, driven by a delay line and controlled by a "look ahead" feed-forward level-sensing rectifier gives unparalleled clean sustain with very low noise and distortion. The trick is to not to ask for too much gain through the op-amps in the gain control cells. Even so, I get a very controlled 12:1 compression ratio at maximum setting, and this sounds very "squashed"!
With a bit more tweaking, this could get very close to good, but would be radically improved with a delay line included in the audio path to overcome the overshoot issue.
The VCA as drawn is OK, but can be significantly improved by the use of the generously provided linearising diodes in the OTA, and by some recalculation of the bias values. The datasheet values are quite carefully chosen for optimum operation (the manufacturers really like to show their ICs off in the best possible circumstances), so these make for a good starting point.
There is an issue with loop stability when the level sensing is taken from before the gain cell, and this is why the vast majority of manufacturers take their rectifier inputs from the output of the VCA. This also confers other advantages, including minimising the parts required by providing a low impedance source to feed both the rectifier and the output stage.
If you look at the Dynacomp, the Black Finger, the Orange Squeezer or any other classical compressor design, you'll see the same topology. Electro Harmonix and Dan Armstrong shunt the audio to ground to provide variable level into a fixed-gain amplifier (though EH give you a gain control for that amp), and MXR use the current-variable gain of the OTA in the audio path. Neither approach is particularly good - shunting audio means that you have to make up level somewhere else, thereby introducing noise, and OTAs in the audio path are always nasty, causing distortion artifacts of their own. Using the OTA in the feedback loop of an op-amp (like here and in the Morley / Tel-ray circuits) is definitely preferred, but reduces the range of control available.
I find that two gain cells of this type cascaded, driven by a delay line and controlled by a "look ahead" feed-forward level-sensing rectifier gives unparalleled clean sustain with very low noise and distortion. The trick is to not to ask for too much gain through the op-amps in the gain control cells. Even so, I get a very controlled 12:1 compression ratio at maximum setting, and this sounds very "squashed"!
"Why is it humming?" "Because it doesn't know the words!"
*sigh* You really don't have the foggiest idea what you're talking about, do you.mictester wrote: The VCA as drawn is OK, but can be significantly improved by the use of the generously provided linearising diodes in the OTA, and by some recalculation of the bias values. The datasheet values are quite carefully chosen for optimum operation (the manufacturers really like to show their ICs off in the best possible circumstances), so these make for a good starting point.
There is an issue with loop stability when the level sensing is taken from before the gain cell, and this is why the vast majority of manufacturers take their rectifier inputs from the output of the VCA. This also confers other advantages, including minimising the parts required by providing a low impedance source to feed both the rectifier and the output stage.
If you look at the Dynacomp, the Black Finger, the Orange Squeezer or any other classical compressor design, you'll see the same topology.
Using the OTA in the feedback loop of an op-amp (like here and in the Morley / Tel-ray circuits) is definitely preferred, but reduces the range of control available.
- tube-exorcist
- Resistor Ronker
Please explain: where can I find the "unstable loop" in a feed forward compressor ?mictester wrote: There is an issue with loop stability when the level sensing is taken from before the gain cell, ....
BTW, which delay-line was used in the limiters mentioned by you to delay the audio signal ? A special BBD, less noisy and with more headroom than the common ones ? Or was it a 48-bit digital delay ? I mean the delay line must have at least 30db more headroom than the limiter otherwise the signal will be clipped by the delay line, right ?
"I've noticed there's an inverse relationship between cost of gear and talent. If you need the most expensive gear to get decent tones, then you suck as a player."
- hbo
- Breadboard Brother
Information
I'm going to give this one a try and thought I'd share my vero layout with all the suggested mods added. (Not yet verified as working when I write this).
- Greg
- Old Solderhand
Guys.. can we keep this thread on track and about The Engineer's Thumb please.
It's becoming a huge mess and has been cleaned up several times.
No personal attacks or banter (or vagina pics).. take them to the Cafe or Member's Corner.
Thanks
It's becoming a huge mess and has been cleaned up several times.
No personal attacks or banter (or vagina pics).. take them to the Cafe or Member's Corner.
Thanks
culturejam wrote: We are equal opportunity exposure artists.
- hbo
- Breadboard Brother
Information
Not to bring back a crazy thread, but I just had to comment on this. Finished building this (based on the layout I posted on the previous page) with all the bells and whistles, and while the result may look a bit untidy it sounds really, really great. This is probably my favorite compressor at this point, so a big thank you Merlin!