Author Topic: THE ENGINEER'S THUMB... At last, a better compressor! (Read 292498 times)

batelfixx · « **Reply #60 on:** August 25, 2012, 05:30:54 AM »

Hi guys,

midwayfair has me inspirited.
I ´ve made a PCB. http://forum.musikding.de/cpg/thumbnails.php?album=619
The numeration is like merlinb.
Tell me please if you find a mistake! It´s one of my first!

Not verified!

Bastel

merlinb · « **Reply #61 on:** August 25, 2012, 09:20:31 AM »

Quote from: batelfixx on August 25, 2012, 05:30:54 AM

I ´ve made a PCB. http://forum.musikding.de/cpg/displayimage.php?album=lastup&cat=1&pos=0

I think that's the wrong link?

batelfixx · « **Reply #62 on:** August 25, 2012, 02:09:55 PM »

Upps,

this is the old brain!
this are the pictures

By batelfixx at 2012-08-25

By batelfixx at 2012-08-26

By batelfixx at 2012-08-26
I hope now it is right - the preview will not go.
Modified on 2012-08-26
Bastel

garcho · « **Reply #63 on:** August 25, 2012, 03:38:35 PM »

nice job on the pcb! but the link is asking for a password

midwayfair · « **Reply #64 on:** August 25, 2012, 07:23:33 PM »

Hm. batelfixx, will that trace from R14 to the connection to Pin4 on IC3 print properly? You might have better luck putting R14 horizontally (keep the same top hole attacked to Pin 1 of IC2) and just connecting the other trace to the pads for R16/18. There's plenty of room between them and the PNP transistor pads. I don't etch, but I know that some people who do might not have a printer that can manage a good enough resolution for that trace.

Anyway, good job on it. It's by no means and easy circuit to do a layout for despite not having a particularly high parts count (took me four tries to get mine done).

PRR · « **Reply #65 on:** August 25, 2012, 09:33:52 PM »

> will that trace from R14 to the connection to Pin4 on IC3 print properly?

Military/Industrial PCBs have been running traces between DIP pins since 1970 or before.

I think any modern PCB house can handle it fine.

It's too-tight for my Sharpie-mask techniques.... but I can always jumper it with scrap wire.

As you say, there's other ways to route this, and prolonged brain-pounding (or auto-route) could maybe neaten it a bit. But nothing looks "wrong" to me, and I've seen very sloppy layouts sound great.

batelfixx · « **Reply #66 on:** August 26, 2012, 03:56:04 AM »

ok, midwayfair,
now the PCB more better to etch.
You can it use by copy and paste and print without scaling.

By batelfixx at 2012-08-26

and new

By batelfixx at 2012-08-26

I have no web-space to upload the pdf - sorry!
Bastel

batelfixx · « **Reply #67 on:** August 27, 2012, 11:33:23 AM »

... and now the link -2. approach

http://dl.dropbox.com/u/100857549/Engineers%20thumb%20compressor.pdf

(Thanks, Jon)
Bastel

Kesh · « **Reply #68 on:** August 31, 2012, 07:03:01 AM »

Quote from: merlinb on April 21, 2012, 10:17:37 AM

I would dearly love to know what the jack sockets are that can be put in line with the foot switch. mine are always too big, even with dpdt

merlinb · « **Reply #69 on:** August 31, 2012, 09:18:51 AM »

Quote from: Kesh on August 31, 2012, 07:03:01 AM

I would dearly love to know what the jack sockets are that can be put in line with the foot switch. mine are always too big, even with dpdt

I think they're all basically the same. They usually have a plastic bump on the end which I cut off with a sharp knife! There's a photo of what I mean here:
http://valvewizard.co.uk/makingpedals.pdf

mth5044 · « **Reply #70 on:** August 31, 2012, 04:43:57 PM »

Quote from: Kesh on August 31, 2012, 07:03:01 AM

I would dearly love to know what the jack sockets are that can be put in line with the foot switch. mine are always too big, even with dpdt

Keep in mind, that is also a DPDT, not a 3PDT. It isn't as wide.

Kesh · « **Reply #71 on:** August 31, 2012, 05:06:23 PM »

Quote from: mth5044 on August 31, 2012, 04:43:57 PM

Keep in mind, that is also a DPDT, not a 3PDT. It isn't as wide.

yes, i can count to six

mth5044 · « **Reply #72 on:** August 31, 2012, 05:48:29 PM »

Yikes, sorry for my concern.

midwayfair · « **Reply #73 on:** November 11, 2012, 11:40:21 AM »

This weekend I learned that NE5532 won't work in EITHER the audio or power section. Although I don't understand why it won't work, I did find that 4558, LM358 and OPA2134PA (and TL062, obviously, though it's noisier in the audio path) DO work* as subs. What's going on here? I can't remember ever running into a situation where a 5532 wasn't a workable sub for a TL072.

*Of the OP amps that DO work, it seems to be really, really sensitive to the op amp, more than anything else I've run into. It vastly changes the compression character and decay. TL07x is clearly the absolute best and most natural sound, but it's weird that it's such a drastic change with some swaps.

Anyway, just thought I'd pass this on in case anyone else is in a bind with their chips and is looking for substitutes: Basically ... don't substitute in this circuit.

garcho · « **Reply #74 on:** November 11, 2012, 11:57:40 AM »

Quote

I can't remember ever running into a situation where a 5532 wasn't a workable sub for a TL072.

It's irresponsible of me to say this without recalling the specifics, but I've found a number of circuits where this is the case. Most recently, a BP filter I've been noodling with.

Kesh · « **Reply #75 on:** November 11, 2012, 12:29:26 PM »

Quote from: midwayfair on November 11, 2012, 11:40:21 AM

This weekend I learned that NE5532 won't work in EITHER the audio or power section. Although I don't understand why it won't work, I did find that 4558, LM358 and OPA2134PA (and TL062, obviously, though it's noisier in the audio path) DO work* as subs. What's going on here? I can't remember ever running into a situation where a 5532 wasn't a workable sub for a TL072.

*Of the OP amps that DO work, it seems to be really, really sensitive to the op amp, more than anything else I've run into. It vastly changes the compression character and decay. TL07x is clearly the absolute best and most natural sound, but it's weird that it's such a drastic change with some swaps.

Anyway, just thought I'd pass this on in case anyone else is in a bind with their chips and is looking for substitutes: Basically ... don't substitute in this circuit.

5532 is not the most stable of internally stabilised op amps, and also has very low input impedance, even for a bjt op-amp, and the 1M VRef resistors are way too high for it. As the biasing 1M resistor off U1b is also a threshold control, I can imagine the op-amp's input impedance is going to strongly affect the sound. Stick to fet input op-amps like TL072

midwayfair · « **Reply #76 on:** November 11, 2012, 12:38:11 PM »

Quote from: Kesh on November 11, 2012, 12:29:26 PM

Quote from: midwayfair on November 11, 2012, 11:40:21 AM
This weekend I learned that NE5532 won't work in EITHER the audio or power section. Although I don't understand why it won't work, I did find that 4558, LM358 and OPA2134PA (and TL062, obviously, though it's noisier in the audio path) DO work* as subs. What's going on here? I can't remember ever running into a situation where a 5532 wasn't a workable sub for a TL072.

*Of the OP amps that DO work, it seems to be really, really sensitive to the op amp, more than anything else I've run into. It vastly changes the compression character and decay. TL07x is clearly the absolute best and most natural sound, but it's weird that it's such a drastic change with some swaps.

Anyway, just thought I'd pass this on in case anyone else is in a bind with their chips and is looking for substitutes: Basically ... don't substitute in this circuit.
5532 is not the most stable of internally stabilised op amps, and also has very low input impedance, even for a bjt op-amp, and the 1M VRef resistors are way too high for it. As the biasing 1M resistor off U1b is also a threshold control, I can imagine the op-amp's input impedance is going to strongly affect the sound. Stick to fet input op-amps like TL072

Yeah, I decided to do some light reading this morning after my post and didn't realize that the 5532 was BJT instead of FET (and obviously I can't just swap out a FET for a BJT all the time in a transistor circuit). Oh well. I'll just use a BB for now and swap it out for a 072 after my next parts order.

garcho · « **Reply #77 on:** November 11, 2012, 12:40:03 PM »

According to the TI datasheet, the input resistance in a TLO7x is 10,000,000,000,000Ω, where as the input resistance of the NE5532 is 300,000Ω. Also, total supply current (for each amp) in a TL07x is 1.4 mA, in the NE5532 it's 8mA. The unity gain bandwidth of a TL07x is 3MHz, the NE5532 is 10MHz. Outside of input impedance, I don't have a clue if these things are connected to why it won't work in this case, but I noticed them glancing over the datasheets.

merlinb · « **Reply #78 on:** November 11, 2012, 02:53:00 PM »

The input impedance of the 5532 will be the main problem. It is too small compared with the release resistor which will cause the peak detector to go too far negative and overcompress. It will also cause the caps to discharge quite quickly, reducing the release time. The offset voltages will probably be quite large too, which may cause additional problems.
The 5532 isn't always better!

sugonidamaso · « **Reply #79 on:** January 27, 2013, 06:09:52 AM »

Is it okay if I use 2n7000 instead of BS170? Looking forward for your 5 knob version. Thanks in advance!

News:

Author Topic: THE ENGINEER'S THUMB... At last, a better compressor! (Read 292498 times)