Which compressor?

[PRR]

> despite the side chain forking off left of the LDR

But "after" the dominant 68k.

So yes: 96% feedback. The 4% kick is a hack to extend the compression when the LDR loses linearity.

Time "constants" (not constant at all) are entirely up to the LDR. The LA2 people worked with the LDR bakery and did a LOT of selection to get "nice" time response. I did a lot of playing-around in the 1980s, and even by then "good" LDRs were hard to find. I used the scheme for "rude" limiters to protect drivers against accidents and abuse.

When I heard the RNC1773 I stopped paling-around with obsolete chemistry. Yes, four separate channels of RNC is not cheap. I think "all" PA gear companies have a 4-channels compressor on similar technology, it's not obscure. If you convince THAT Corp you have factory and distribution, they can be eager to help.

Behringer MULTICOM PRO-XL MDX4600 V2
4-channel Compressor/Limiter/Gate
Rated 4/5 3 reviews
Item ID: MDX4600V2
$159.00

It's a cheap box and jacks but you can't touch it DIY for that price, even before you build and test guts.

[Fancy Lime]

If you like the way it sounds and feels, you were probably lucky in the vactrol lottery. In that case, there is nothing wrong with using the design as it is. Don't fix what isn't broken, no matter what some dude on the internet sais

However, in my experience feedback limiters do not respond well to long attack and release times. For this type of circuit you want your attack as fast as possible and your release just slow enough to avoid rippling on the lowest notes. Slow attack causes duck-and-swell and long decay causes pumping. Both rarely wanted. Seems that your vactrol happens to give you good values as is is, so no need to make it worse, imo.

There are a few things you should add, though. Add a 100 Ohm resistor between the vactrol LED and 1/2V and do the same for the indicator LED. This is called current sharing and makes it so that both LEDs light up even if they have different thresholds.

If you want to experiment with slower attack and decay, just place a cap from the top of the LEDs to ground. This is called the holding cap or timing cap. I would start with 22uF or even 100uF. If that has too little effect, try increasing that 56 Ohm resistor before the LEDs to 220 Ohm.

The buffer I was talking about would go between the holding cap and the LEDs. In the scheme described above, the holding cap discharges through the LEDs, which is not at all linear and cannot easily be controlled or changed. The buffer separates the discharging if the holding cap from the charge necessary to drive the LEDs, allowing to then use a (variable) resistor to set the discharging rate.

Andy

[Greenballs]

If you like the way it sounds and feels, you were probably lucky in the vactrol lottery. In that case, there is nothing wrong with using the design as it is. Don't fix what isn't broken, no matter what some dude on the internet sais

However, in my experience feedback limiters do not respond well to long attack and release times. For this type of circuit you want your attack as fast as possible and your release just slow enough to avoid rippling on the lowest notes. Slow attack causes duck-and-swell and long decay causes pumping. Both rarely wanted. Seems that your vactrol happens to give you good values as is is, so no need to make it worse, imo.

There are a few things you should add, though. Add a 100 Ohm resistor between the vactrol LED and 1/2V and do the same for the indicator LED. This is called current sharing and makes it so that both LEDs light up even if they have different thresholds.

If you want to experiment with slower attack and decay, just place a cap from the top of the LEDs to ground. This is called the holding cap or timing cap. I would start with 22uF or even 100uF. If that has too little effect, try increasing that 56 Ohm resistor before the LEDs to 220 Ohm.

The buffer I was talking about would go between the holding cap and the LEDs. In the scheme described above, the holding cap discharges through the LEDs, which is not at all linear and cannot easily be controlled or changed. The buffer separates the discharging if the holding cap from the charge necessary to drive the LEDs, allowing to then use a (variable) resistor to set the discharging rate.

Andy

Thanks for taking the time Andy, much appreciated. I'll learned some useful things there that I'll certainly be experimenting with over the next couple of days.

[Greenballs]

Behringer MULTICOM PRO-XL MDX4600 V2
4-channel Compressor/Limiter/Gate
Rated 4/5 3 reviews
Item ID: MDX4600V2
$159.00

It's a cheap box and jacks but you can't touch it DIY for that price, even before you build and test guts.

Yep, even the Drawmer DL441 is only £250 second hand these days. I think I might struggle to match even that price. Still, I'm enjoying the process of breadboarding different types of compressor and learning about how they function (with help from the good folks here), so its not a complete waste of time and effort! 

[Fancy Lime]

One general observation with hardware these days is that DIY has long stopped being the "economy option". All the standard stuff like studio compressors, mic preamps, headphone amps etc. can be had for much cheaper brand new from a dozen or so companies than you could ever make it yourself, even if you value your working hours at zero. And these things that are cheap new are very cheap used because people feel the need to upgrade from Behringer to something more reputable as soon as they can, whether it sounds better or not. From a purely economic standpoint, DIY only makes sense now if you want something extraordinary that is so far out of the mainstream that the cheapo companies don't bother with it. So my advice on the "you need a quad compressor" front would also be to get a used cheap one. As long as it works at all, It'll do just fine for your stated purpose of getting drums and bass and other stuff into a computer. Complicated compression (if that is what you want) is best done in the DAW these days anyway. To scratch the DIY itch, it is still fun to create your own "secret sauce" compressor for a particular purpose. One goal of the FET compressor I linked earlier in the thread, was to accentuate the nice fat overdrive of this type of control element in extreme settings and also to allow to dial in some rippling to taste. Both things that commercial designs try to minimize because you rarely want them. But I happen to like them a lot on the niche application of fretless finger style funk bass lines. Hot, fat and dirty, like all the best street food

Andy

[PRR]

.... in my experience feedback limiters do not respond well to long attack and release times. For this type of circuit you want your attack as fast as possible and your release just slow enough to avoid rippling on the lowest notes. Slow attack causes duck-and-swell and long decay causes pumping.

IMHO, this applies just as much to feed-Forward topology. If you have a rapid attack instrument (guitar, piano) and slow-attack limiter (your tiny demon was out carousing all last night), it's gonna splatt.

Even so, I wish you would consult for my cable TV company. duck and SWELL indeed. Not even good for business: frequently the name of the product is swallowed-up by an overblown first transient. Though this is more about over-compressing pre-compressed crap than about time factors. 

[Fancy Lime]

True, the same problems apply to feed forward designs. I just don't have any hands-on experience with those. There are ways of tackling these problems, like clipping the side chain, adaptive attack speed, etc. But all the ones I know or tried are hacky and finicky and not well suited for production circuits.

[rankot]

Can the Bajaman LA2A style optical compressor be modded to allow control of the attack and release of the effect either by means of a switch, between preset fast/slow modes, or a potentiometer? Or, if anyone has the time, can someone give me a brief analysis of the attached schematic please so I can try and figure it out myself? Thanks.

I've built this one some time ago and never made it work as intended. It always distorts the sound. Tried different vactrols, both rolled or factory made, they made some changes, but newer succeeded to remove distortion. NSL32, LCR-0202 and VTL5C10 (original NOS) work the best, but non of them is perfect.

[rankot]

The LA Light is a bit of a blast from the past...
https://www.diystompboxes.com/smfforum/index.php?topic=30106.0

The circuit is quite simple and it was designed specifically with vocals in mind, while also capable of doing a stand-up job with instruments as well. The links in the post linked above are broken (posted before my .com was ripped off), but they are updated on this page...
http://moosapotamus.net/fatNpretty.html

For recording, I would go with Johan's original and simpler design, LA-Light, rather than my modded stompbox version. The transistors are J201. Looks like synthcube's smallbear parts shop may have some.
https://synthcube.com/cart/index.php?route=product/search&search=j201&description=true

[/memories]

I've built Fat'n'pretty few years ago and never made it work at all. No idea why. I didn't use that isolated DC converter, used LT1054 negative voltage generator, maybe that was the problem? No idea.

[Rob Strand]

I've built this one some time ago and never made it work as intended. It always distorts the sound. Tried different vactrols, both rolled or factory made, they made some changes, but newer succeeded to remove distortion. NSL32, LCR-0202 and VTL5C10 (original NOS) work the best, but non of them is perfect.

The half-wave rectification doesn't help but designs like this rely 100% on the opto time constants to reduce distortion.   Without redesigning the whole thing it's probably worth playing with caps across the opto LED maybe 10uF upto 100uF,  just enough to shave off the distortion.   Not ideal but maybe enough to push it over the line.

[rankot]

The half-wave rectification doesn't help but designs like this rely 100% on the opto time constants to reduce distortion.   Without redesigning the whole thing it's probably worth playing with caps across the opto LED maybe 10uF upto 100uF,  just enough to shave off the distortion.   Not ideal but maybe enough to push it over the line.

Thanks Rob, I'll try that. I've built https://stompville.co.uk/?p=50 Clean Squeeze compressor and it has that parallel cap across the opto LED, but I forgot about that option. 

[Rob Strand]

Thanks Rob, I'll try that. I've built https://stompville.co.uk/?p=50 Clean Squeeze compressor and it has that parallel cap across the opto LED, but I forgot about that option.

You can only try it to see if it can be saved.   The distortion comes from the release time being too short.  However the parallel cap affects both the attack and release times.    What happens is by the time the cap is big enough to fix the release time the attack time becomes too long.  So it's a careful balancing act tweaking the parallel cap.   For feedback type compressors it not as bad as it seems since the release time gets stretched compared to the attack time by nature of the feedback.

The four diodes around the opto LED on the John Hollis compressor (from your link) will give you full-wave rectification, that helps reduce distortion too.  Which is useful.  However you can't use the same trick on the BAJA compressor since it adds a DC bias using the two small diodes.   To get full-wave and the DC bias on the BAJA you need to redesign that part of the circuit - which will make it more complicated.

[bowanderror]

Can the Bajaman LA2A style optical compressor be modded to allow control of the attack and release of the effect either by means of a switch, between preset fast/slow modes, or a potentiometer? Or, if anyone has the time, can someone give me a brief analysis of the attached schematic please so I can try and figure it out myself? Thanks.

I've built this one some time ago and never made it work as intended. It always distorts the sound. Tried different vactrols, both rolled or factory made, they made some changes, but newer succeeded to remove distortion. NSL32, LCR-0202 and VTL5C10 (original NOS) work the best, but non of them is perfect.

The Baja Optical Limiter is by far my favorite compressor, and it's been on my board for close to 10 years now. I didn't test any legit vactrols, but I had to try a few combos of LDRs (with different off resistances) and LED colors to get the right one. I think I ended up with a red LED & a quite low off resistance LDR.

Definitely worth troubleshooting if you still have it.

[Greenballs]

The Baja Optical Limiter is by far my favorite compressor, and it's been on my board for close to 10 years now. I didn't test any legit vactrols, but I had to try a few combos of LDRs (with different off resistances) and LED colors to get the right one. I think I ended up with a red LED & a quite low off resistance LDR.

Definitely worth troubleshooting if you still have it.

I'm just about to start buildng this today, with a couple of the minor mods suggested by Fancy Lime. Fingers crossed I won't have any issues with it.

I've also had AMZ's Q&D 2 compressor on the breadboard and was very impressed with it's range and versatility, not to mention the HUGE amount of output it has on tap, so I'll definitely be building that too, maybe with the addition of a switch (between different cap values) to change the decay time of the VCA.

Just out of interest, is there any way gain reduction and output meters could be hooked up? Schematic attached.

[Greenballs]

If you like the way it sounds and feels, you were probably lucky in the vactrol lottery. In that case, there is nothing wrong with using the design as it is. Don't fix what isn't broken, no matter what some dude on the internet sais

However, in my experience feedback limiters do not respond well to long attack and release times. For this type of circuit you want your attack as fast as possible and your release just slow enough to avoid rippling on the lowest notes. Slow attack causes duck-and-swell and long decay causes pumping. Both rarely wanted. Seems that your vactrol happens to give you good values as is is, so no need to make it worse, imo.

There are a few things you should add, though. Add a 100 Ohm resistor between the vactrol LED and 1/2V and do the same for the indicator LED. This is called current sharing and makes it so that both LEDs light up even if they have different thresholds.

If you want to experiment with slower attack and decay, just place a cap from the top of the LEDs to ground. This is called the holding cap or timing cap. I would start with 22uF or even 100uF. If that has too little effect, try increasing that 56 Ohm resistor before the LEDs to 220 Ohm.

The buffer I was talking about would go between the holding cap and the LEDs. In the scheme described above, the holding cap discharges through the LEDs, which is not at all linear and cannot easily be controlled or changed. The buffer separates the discharging if the holding cap from the charge necessary to drive the LEDs, allowing to then use a (variable) resistor to set the discharging rate.

Andy

@FancyLime

Hi Andy,

Can I just check that the attached schematic (simplified to show just the relevant parts) is what you were describing regarding the buffer/holding cap in your previous post?

I have the Bajaman compressor built and it works but there is a little bit of background distortion, which I'm trying to figure out ways of eliminating. Am I right in thinking that the correct value holding cap should help to reduce the distortion?

[rankot]

I think that loading opamp output with a capacitor is a bad idea. Put it across LEDs.

[FSFX]

I think that loading opamp output with a capacitor is a bad idea. Put it across LEDs.

Would it not be better to make U2.2 an integrator by having the capacitor in its feedback path or am I missing something here?

[Rob Strand]

I think that loading opamp output with a capacitor is a bad idea. Put it across LEDs.

It's not that bad as the opamp drives the cap via a resistor.  On that design you are pumping a fair bit of current through the LEDs anyway.

FWIW, I had some thoughts later.  You might have more success placing a rectifier further back in the signal chain.   At least that lets you keep a faster attack and you can still set the time constants.   It also removes the need for large caps.

Here's the idea.  You could use a Schottky diode.


Please don't think of this as a working circuit.  It's only showing the general idea of shoe-horning a way to tweak the time-constants into the circuit.   It's still a half-wave rectifier.

There's an assumption that the distortion is caused by the time constants, while highly likely, it's not yet proven.

[rankot]

It's not that bad as the opamp drives the cap via a resistor.  On that design you are pumping a fair bit of current through the LEDs anyway.

FWIW, I had some thoughts later.  You might have more success placing a rectifier further back in the signal chain.   At least that lets you keep a faster attack and you can still set the time constants.   It also removes the need for large caps.

Here's the idea.  You could use a Schottky diode.


Please don't think of this as a working circuit.  It's only showing the general idea of shoe-horning a way to tweak the time-constants into the circuit.   It's still a half-wave rectifier.

There's an assumption that the distortion is caused by the time constants, while highly likely, it's not yet proven.

It seems to be working, at least in LTSPICE. Adding 330n across R11 on your schematic will further smooth the envelope curve, but it will also slower the attack a little. I will test this circuit later today, because I have one almost complete unit on the desk.

Also, would it be possible to adjust compression ratio with R15 as a potentiometer?

[Rob Strand]

It seems to be working, at least in LTSPICE. Adding 330n across R11 on your schematic will further smooth the envelope curve, but it will also slower the attack a little. I will test this circuit later today, because I have one almost complete unit on the desk.

If you need to extend the release time to reduce the distortion you can use a larger cap and reduce the 1k in series with the diode to keep the attack the same.   The threshold pot is fixed at 10k so you can play with the resistance to modify the release time.   Another angle is to change the 10k pot to a 100k pot - that will let you get longer release times with smaller caps and avoid small resistances on the output of the opamp.

Also, would it be possible to adjust compression ratio with R15 as a potentiometer?

In use it might give that impression but what it is really doing is limiting the maximum attenuation.   If the compressor sits at full attenuation it stops compressing at higher inputs and you get less (incremental) compression.    I haven't analysed it in detail and sometimes the details screw you over on these circuits.