Scott Hampton Class A Opto Comp with schemo and parts list!

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

matta

Well-known member
Joined
Jan 11, 2005
Messages
1,640
Location
Cape Town, South Africa
Hey Guys,

Some of you may or may not be aware that Scott Hampton contributed a basic, yet functional Class A Opto Comp design at the Tape Op Con this year. It was discussed briefly in this thread: http://www.groupdiy.com/index.php?topic=8364

These boards were only available to the conference goers but being the kind and generous guy he is he has now posted the schemo, parts list (including suppliers) on his site. Best of all this comp only cost $20 in parts to build! A little more if you mod it to have variable Attack and Release.

OPT19.jpg


OPT17.jpg


I asked him for his PCB layout and he has now posted that as well. He has given me permission to share it with you but by his own admission the board layout is a little unclear for making a home etch but maybe someone here would like to retrace it and share it here?

From what I have heard this is not a comp that will 'blow' you away but is great for having a basic, no frills Opto comp for everyday studio use, clean and quite, plus his designs are really cool and creative as a lot of us know.

At $20 I'd be willing to give it a try, and it is simple design, maybe a great projects for DIY newbies?

He has even set up a photo gallery of photos, build info and tips and even some troubleshooting tips.

Anyone willing to retrace the board?

Thanks again to Scott and his willingness to give back to the DIY community.

I'll try start working on making a PDF or Xls BOM and post it when I am done.

Here is the link to all the relevant info: http://www.hamptone.com/tapeop2005.htm

Cheers

Matt
 
I would really like to build one of these - I need a really small compressor for my live rig - I can probably make a stereo version of this for $50 or so - the only little stereo comp that price is the nanocompressor by alesis - basically a 3630 in a little box. Thanks for getting the ball rolling!
 
Nice and simple little compressor.
Have you noticed the fact that the sidechain can work as feedforward or feedback. This is an option I want to implement into Fred's optocomp with a three position switch: FF, FB and FB/FF

chrissugar
 
Hey Chrissugar to implement the FF/FB thing on the Forssell opto all you need to do is toy with the values of the restors before and after the sidechain send. The one before the send is the FF resistor, and the one after is the the FB resistor. By playing with the values you can blend between full feedback, full feedforward, or somewhere in between (which is how most folks are implementing it). Make R43 0 and you have a feedforward design, make R44 0 for pure fedback. Would be cool to put a 5k pot in there and be able to fully sweep between feedforward and feedback. Then again, I think 2 or 3 settings out to do.

Sorry to hijack this thread a bit, but I've been doing a fair amount of work on the Forssell opto these days.

Cheers,

Kris
 
Very nice little one for doing a multi-channel insert compressor!
And maybe perfect for a discrete voice channel in 1 HE chassis...
Maybe of use to tame some hot synth signals i use to feed in my mackie LM-3204...
It's so SMALL and simple!
How does it sound really? What can it be compared to?

:grin:

Martin

(btw looking for a low cost comp? alesis 3630 powered up like in buttachunks thread...it's sooo sweet now, i can't get it! my best friend got big ears not hearing the any of the artifacts and click/pops anymore that were so dominant before! Really sounds like a good pro box now, so he mentioned i should do a custom frontpanel for it so nobody should ever think this is a budget compressor but only trust them ears...i think i will use it for summing/mastering purposes and give it balanced connects, no joke...)
 
Just for my own jollies, I marked-up the schematic with the parts values so I could see what was going on.

Marked-up schematic

Comments not meant as criticism:

The attack/release times are similar; myself I like them very different. The jumpers or trace-cuts are there to add pots (outside the original time/cost) budget. A further hack for extended experimentation would be to put loops of heavy wire in all the C and R pads there, so you can solder-in various values without burning and lifting the PCB pads.

The rectifier is (as mentioned) half-wave. I don't see a way to make it full-wave at this price. Using half-wave with some solo musical instruments/voices means you get different limiting levels depending which polarity the signal happens to be. Since this is an "effect" limiter and not a Protection limiter, you will trim it to taste and any asymmetrical level effects are meaningless.

You could make it full-wave with a second U5 stage, fed from U5 with a resistor divider to make the voltage gain unity, and adding another set of C7 R15 D1, to rectify the other side of the signal. If someone does a new layout, you might leave pads for that. Similar to how U2 feeds U3. The divider will have to be trimmed to cancel the gain of the second U5, since I don't know what that gain is (about 17) (hmmm, I suppose you could wire U5a with two resistors as a unity-gain no-trim inverter).

The actual threshold seems to be the silicon drop in D1, so is slightly temperature sensitive. Again, in an effect limiter that is just meaningless. Your harpsichord and drums etc will be going out of tune much faster than your limiting levels will change.

R1 and R2 work the LDR in a low-Z situation; my little limiters seemed to like a higher Z here. I note that jumper RAT allows you to insert more resistance; to get full effect replace R2 with 100K or so. However Scott is socking the LED with current, so it might not limit well with bigger R1 R2 unless many other changes are made.

The feedforward connection is interesting. "Flat" limiting range is less than with feedback, but some 20dB of near-flat limiting is possible feedforward. This may be why the LDR works low-Z and with substantial idle bias (R21): speed is improved and you get off the low-light knee of the LDR. I never had real speed problem using a similar V/I converter with feedback, because the V/I will beat the LDR down hard; in feedforward there is no servo effect to overdrive the LDR. In feedback, you might try R21 at 10K or even infinity.

In some situations, the device feeding the limiter may leak DC. Bad form, but stuff happens. This will thump the LDR, and upset bias on U2 and U5. You could add a 2uFd non-polarized (film) capacitor at the input; that costs as much as everything else.
 
[quote author="DrFrankencopter"]The one before the send is the FF resistor, and the one after is the the FB resistor. By playing with the values you can blend between full feedback, full feedforward, or somewhere in between (which is how most folks are implementing it). Make R43 0 and you have a feedforward design, make R44 0 for pure fedback. [/quote]

Exactly. This is how I want to do it. Fred already has the right value for the intermediate FB/FF. The rest is simple.

[quote author="DrFrankencopter"]Would be cool to put a 5k pot in there and be able to fully sweep between feedforward and feedback. Then again, I think 2 or 3 settings out to do.
[/quote]

A pot would be to much in my opinion. A three position switch would cover all the needs.

chrissugar
 
There is one problem with using an LDR compressor in feed-forward mode: it will always be at an infinity: 1 ratio.

Because it is a linear device (so much current in for so much change in opto impedance), there is no way to scale the sidechain voltage to effect different G/R slopes. You can in feedback mode because the relationship of the sidechain voltage to gain is non-linear, and with enough gain reduction is almost log in nature.
 
There is one problem with using an LDR compressor in feed-forward mode: it will always be at an infinity: 1 ratio.

Because it is a linear device (so much current in for so much change in opto impedance), there is no way to scale the sidechain voltage to effect different G/R slopes. You can in feedback mode because the relationship of the sidechain voltage to gain is non-linear, and with enough gain reduction is almost log in nature.

The VT5C1 opto is not linear with respect to current and resistance, it's logarithmic. Also, if you check out Fred Forssell's sidechain, you can get different ratios by putting a series current limiting resistor in line with the vactrol LED. This just drives the opto less hard for a given input...brilliantly simple, and simply brilliant!

What really changes between feedforward and feedback is attack and release times.

Cheers,

Kris
 
umm

just a suggestion...

its not a good idea to advertise your paypal id like that

I know its not your p/w but your providing half the game by putting your id out like that...

plus there are search bots that scan through and find email address that are listed as you have yours.. thats how you will end up on spam lists...

seriously.. I wouldnt do that

my address I expect to receive spam.. and I do.. but its not my personal address. and its not my paypal id
 
its never too late...

:)

I just learned a long time ago that a lot of people strike at the right opportunity.. just dont give them that opportunity
 
[quote author="DrFrankencopter"]
There is one problem with using an LDR compressor in feed-forward mode: it will always be at an infinity: 1 ratio.

Because it is a linear device (so much current in for so much change in opto impedance), there is no way to scale the sidechain voltage to effect different G/R slopes. You can in feedback mode because the relationship of the sidechain voltage to gain is non-linear, and with enough gain reduction is almost log in nature.

The VT5C1 opto is not linear with respect to current and resistance, it's logarithmic. Also, if you check out Fred Forssell's sidechain, you can get different ratios by putting a series current limiting resistor in line with the vactrol LED. This just drives the opto less hard for a given input...brilliantly simple, and simply brilliant!

What really changes between feedforward and feedback is attack and release times.

Cheers,

Kris[/quote]

Kris,

According to the data sheet, it is neither a linear transfer curve, nor log/antilog. Considering FF connection;

Utilizing the part values from the site, with a series resitor of 4.7kohms (lets say 5k for simplicities' sake), you achieve -6dB of gain when 1ma is put through the LED (5k in the cell).

Increase the input to the side chain (labeled "log" amp, but I don't see any log circuitry there) by 10dB, and now you'll have roughly 3ma of current through the LED. According to the graph, you will now have 1kohm in the cell, giving you a total of -14dB of gain. A change of 8dB.

With another 10dB increase in sidechain voltage, you'll have 10ma of current through the LED, with 300ohms resistance in the cell, giving you a total of -24dB of gain.

This turns out being a "soft Knee" compressor in FF mode, with roughly a 5:1 ratio for the first 20dB of gain reduction, and above that, it's infinity to one. You can change R1 all day, and all it will do is move the rotation point of gain reduction vs. LED input current. This resistor does not change current in the LED. Voltage across C6, feeding the V to I converter, U6 does that. If you change the gain anywhere from the compression pot R12 through U6, all this will do is change the threshold of onset of compression.

In FF connection there is no way to alter the ultimate ratio. Just the threshold, and rotation point from low ratio to infinity to one. In FB mode, that's another story. I've played with opto couplers and linear OTA amps for a long time, and because of the nature of the beast (linear transfer as opposed to log/antilog) you can't alter the ratio of compression in FF.

Having said that, this is a clever circuit.
 
damm I guess I walked into that on like a brick wall

I see you know him oy you have dug around his site

oh man no fair he has there names listed on the web site

shoot :shock:

Wilebee
 
Hey Folks,

As promised here is the BOM/Parts list with suppliers, parts, designation
and prices. All prices are based on quantities of 100 (sorry) so it will be
a little more expense for smaller quantities (unless there is a group buy).

HAMPTONE TAPEOPTO COMPRESSOR REV1 BOM XLS

Hope it proves useful. I?m looking forward to building it. Thanks to
everyone who has shared their insight. It sounds like a very modable
projects from what has been stated here.

I wouldn?t mind building it stock just to hear how it sounds. I?m under no
illusions that it is not going to be the best comp in the world, but for
$20 in parts I?m sure it could be a fun tool to have lying around. It is
also a great newbie project, akin to building something like Stomp Box.

Ptown how is the trace coming on?

Did anyone go to Tape Op and build one of these and are using it in there
studio? I would love to hear your comments, thoughts and remarks.

Cheers

Matt
 
[quote author="wilebee"]damm I guess I walked into that on like a brick wall

I see you know him oy you have dug around his site

oh man no fair he has there names listed on the web site

shoot :shock:

Wilebee[/quote]

Hahaha, so when should I expect payment? :wink:
 

Latest posts

Back
Top