THE ENGINEER'S THUMB... At last, a better compressor!

[Tubebass]

I used a 13600 in my build. Although I don't have it boxed yet, bench testing confirms that it does work fine.

[KazooMan]

Well, I am finally back with my comparison.  It sure took a while.  It all came down to a very sharp utility knife, a trip to the ER, eleven stiches in my hand, and waiting for it to heal enough so I could get the stupid bulky bandage off.  The hand will be just fine.  The woodworker's ego is still having a few issues.

I completed the build of the pcb version and used sockets for the chips so I could compare the NJM13600 chip that Steve sent to me with the surface-mount LM13700 I purchased from him at Smallbear.  This was my first attempt at a pedal in this size enclosure and that brought its own issues to bear.  I did incorporate the two protection leds to minimize distortion, I used a trimpot to replace the resistor to have the threshold capability, and I temporarily added a switch to control the treble boost.

First off, the 13600 IC did work as anticipated.  I tried the treble boost switch here and did not find a lot of difference.  As has been discussed, the threshold trimpot is not really that useful over just the release pot.  I really like the compression and especially the sustain.  However, there was quite a noticable amount of distortion at higher settings of the release control and the pedal reduced the output to the extent that I had to have the level pot at about 80+ % to equal the bypass level.

I tried to solder some leads to the legs of the surface-mount IC with the expected negative results.  I decided I had to make a mini pcb to mount the board and provide contacts for regular leads (not many leads are required).  My first attempt at this involved running leads from this board to the sockets of a millmax socket.  I built it, epoxied it all together and had my own chip ready to plug in.  It worked great, but it was too thick with the other socket already in place and I couldn't get the bottom of the box on

I made another pcb with the LM13700 and just bent the leads into the desired locations (using a socket for alignment) and then applied some epoxy.  The resulting "chip" worked just fine!

These are subjective "ears only" comparisons at this point.  The REAL LM13700 improved the level issue.  I know that this is true from AB comparisons with the pots held constant.  The distortion issue seems much better with the 13700.  I really need to explore this with my oscilloscope to see if my ears are fooling me or not.  I had removed the treble boost switch so I cannot report on that. 

I may have a go at modifying the pcb layout to incorporate the surface mounted chip.  This will not be trivial since all of the connections are mirror images and there are already several traces running under the chip. 

Here's a quickie photo of the Rube Goldberg IC chip in the pedal.  The picture gives some funny color cast to the solder joints.  It is not there in the real pedal.



Oh yeh... I should add that the reason the 47 uf cap is laying on its side is that some dummy cut the notch for the DC plug on the wrong side of the box.

[merlinb]

I built another one. This one also has a gain reduction meter of sorts, and a pregain stage with active treble boost/tone control, making it probably the quietest 9V compressor you can lay your hands on. It was also my first attempt at acrylic flow technique (aka swirl painting). Annoyingly, the lacquer seems to have reacted somehow and caused some grey speckling n the top right...

[Chugs]

Cool. How did you implement the treble boost/tone control?

[deadastronaut]

looks great merlin:


ok, i decided to give it another go on breadboard and its working now, with the leds on input too..

but its very distorted/broken up?...especially with the sustain right up....or even halfway....i'm guessing this isn't normal right?.

should i ground all unused pins of the 13700n?.

cheers rob.

[garcho]

should i ground all unused pins of the 13700n?

Based on my experience breadboarding and building this, it shouldn't be distorted, nor should you have to ground the inputs.

I believe certain circuit's inputs are grounded because they can oscillate or amplify other noises that end up in your signal somehow, not because they actually alter the other op amps or inverters or what-have-you within the same IC.

Do you have a few ICs to swap out? They'll fry out with too much current going to the amp bias input pins.

[deadastronaut]

cheers gary,

yep i have a couple i'll have a swap...

[midwayfair]

cheers gary,

yep i have a couple i'll have a swap...

Make sure you aren't using something besides a TL072. Most any BJT-based op amp will distort like crazy. It has to be a FET-based OP amp.

[deadastronaut]

ok swapped the 13700n..

and have 2x tl072's...

i'm using 2n3906,  /  2x5mm red leds.../ 1n4148's

i took the 13700n off my mutron breadboard so i know its fine.

still distorting...


edit:  when i strum the guitar hard the sounds dips to nothing, then swells back up...

[midwayfair]

Rob: Post pics and voltages.

[Neurowork]

Hi,

First of, many thanks to Merlinb for this great design.
I was going to build the 2 knobs version when he posted the 5 knobs version with the bright switch so I figure ... let's give it a try.

I made a neat 2 layers prototype board layout I'm sending to the fab today.



It is 2" wide which means it'll fit in a 1590B.
I'll have 2 spare boards if anyone's interested.

Can't wait to populate and rock this baby.

Cheers.

[mth5044]

^Awesome layout!

@Merlin - what chip/circuit did you use for the LED meter? Looks SO COOL

[TheWinterSnow]

I am curious about a small mod to the design.  Since I don't have much knowledge in OTAs but was wondering if the circuit could be adapted to a +/-9v supply by swapping the original ground to -9v, the original Vref to ground and well 9v to 9v.  If not, are there simple modifications to the circuit that could make it compatible.

[samhay]

I am curious about a small mod to the design.  Since I don't have much knowledge in OTAs but was wondering if the circuit could be adapted to a +/-9v supply by swapping the original ground to -9v, the original Vref to ground and well 9v to 9v.  If not, are there simple modifications to the circuit that could make it compatible.

I was wondering the same thing recently - http://www.diystompboxes.com/smfforum/index.php?topic=102280.msg906530#msg906530
In summary, it is not as trivial as it sounds.

[stevie1556]

Neurowork - is your design for a 3PDT switch layout or for the millennium bypass as per the original? If it's for a 3PDT I'll be interested in a board if there is still one free, just let me know the cost with postage.

[Neurowork]

3PDT
I have a big stack of 3PDTs in my stash so I ditched the millennium bypass.
I have one board left. PM me if you're interrested.

[merlinb]

Neurowork - is your design for a 3PDT switch layout or for the millennium bypass as per the original? If it's for a 3PDT I'll be interested in a board if there is still one free, just let me know the cost with postage.

You do realise that you can still use a 3PDT even if the board has Millennium Bypass on it, right? I don't want anyone to think that providing the facility for Millennium bypass somehow prevents you from using a 3PDT if you prefer. Nothing ever stops you from using 3PDT, whatever the circuit.

[Neurowork]

I do absolutely. I just removed the millennium bypass to lower the part count

Thanks again for the great schematics merlinb! My board are still in the fab but I can't wait.

Cheers!

Edit : Just realized the previous message was not for my benefit. Got to put the soldering iron down and hit the bed !!

[Roger Martin]

@ Neurowork
Maybe a pic of the copper below layout can do me a lot of help 

Thank you

[MrStab]

just wanna say i built the 5-knob version of this a coupla weeks back, and now i hardly jam without it. you've no doubt heard it a million times, but awesome work, Merlin! funnily enough, a coupla days after making it, i was asked to fix (what i think was) a Ross clone from the early 80's - crackly pot - and the ET just felt that little bit more transparent and was much more tweakable.

at the moment it's in an old mains transformer block of all things, strangled by electrical tape, but i plan to transfer it to a more sturdy enclosure at some point. the only issues i have are some weird clicking when bypassed that varies with the ratio control (although i attribute this to me jamming the input, output, DPDT and separate Millennium board [used Sabrotone layout] untidily into to small a space), and it starts to squeal when Release is at 0. i'm not too bothered by that, though. just hope i don't ruin it when i change enclosure because i would really love to use this thing live.

any ideas on the range of each setting in milliseconds etc., that might help with future labelling (or just adjustment generally)?