LM317 as a ... ¿compressor?

[mac]

Well, the day is cloudy, rainy, no beach today. Perfect day to play with something like a LM317 voltage regulator. Call me insane but if vout = 1.25V(1 + R1/R2) is independent of vin it sounds like a limiter or compressor, at least for DC -- see National datasheet--. What about AC?
I set a TL071 as a simple amplifier and used the constant voltage at the output as vin = vcc/2 by coupling TL071 pin6 to LM317 vin. Instead of the standard 240R I used a pair of back to back diodes... because it looks more fancy . To adjust vout I used a 1M-B pot.
The tricky part was the output. I had to provide both a DC and AC load to the LM317. I set the pot to 1M and played with the last resistor R8 until the LM started to clip badly.
That's it. I can't really say if its a limiter, a compressor or simply LM317 distortion, but I like the way it ¿compress/limit/distort? the signal when I crank the master with a little gain.




mac

[MKB]

That's an insane idea!!  Insane enough to work...  at first blush I'd think it would work more as a limiter than compressor.  Also, the LM317's have a minimum current they like to see to regulate properly (although you're certainly not using it in such an application); you might consider trying a LM317L, it looks for a lower minimum current and might work better for this purpose.  It's in a TO92 package and is rated for 100mA max.

So, what happens if you run the signal into the ADJ pin and connect the IN to 9V, then use the OUT pin for the signal output?

[doug deeper]

i tried this with a 5v regulator just the other day, got nothing!
the adjustable version would work better i imagine (if its going to work)
it didnt color the sound at all in fact which is good news.

[lovric]

can you comment on:
www.diystompboxes.com/smfforum/index.php?topic=51719.msg402077#msg402077

and similar:
http://goofaman.free.fr/papa/rectifet/rectifet%20principe.JPG
pls!

[bancika]

Well, the day is cloudy, rainy, no beach today.

no beach today? I really feel sorry for you 

[petemoore]

• Guaranteed 1% output voltage tolerance (LM317A), >food for DIST.lol?
• Guaranteed max. 0.01%/V line regulation (LM317A) >DC I guess/
• Guaranteed max. 0.3% load regulation (LM117) >low load appl.
• Guaranteed 1.5A output current >probly won't burn up here
• Adjustable output down to 1.2V > here ya go?
• Current limit constant with temperature  > cool
• P+ Product Enhancement tested  > always a P+lus
• 80 dB ripple rejection > always nice to have some R 'n R on signal path
• Output is short-circuit protected > with BTB Diodes lets hope so

General Description


--------------------------------------------------------------------------------
The LM117 series of adjustable 3-terminal positive voltage regulators is capable of supplying in excess of 1.5A over a 1.2V to 37V output range.
  >I don't know what's up with this idea, insane or not, I decided to have some fun with it..below is the text I was looking at, making funny comments on [printed again..feel free to add comments], in case anyone else would like topic fodder for offering more tangible dialoge than what I can provide.
  I got the idea the 317 is for DC voltage regulation, and the chip 'consumes' 1.5 v and 2.5  operating voltage.
  working as an audio device processing AC, _/+ signal swings is interesting. perhaps someone will replicate and verify that is exactly what you have going on.

[markm]

Interesting indeed.

[Paul Perry (Frostwave)]

Like most "musical" applicaions here, the "magic" (if any) is happening when the device is used "out of spec".
That is to say, there isn't any data on what the LM317 does, when the input is less than a couple of volts above the output setting.
Plus it's going to be VERY asymmetric

Definitely worth experimenting with!

[petemoore]

  Is cool looking, anyone with 317 exp. will immediately notice the Vin is coming from an audio amp output...very 'strange' indeed.
  So..what's going on inside a regulator with a 'trickle' [compared to the 1.5a rating] of AC is introduced to the Vin, and the control is...controlled such as Mac has shown there
  Thanks Mac for showing the 317 all whacky like that !!
 

[mac]

Like most "musical" applicaions here, the "magic" (if any) is happening when the device is used "out of spec".
That is to say, there isn't any data on what the LM317 does, when the input is less than a couple of volts above the output setting.
Plus it's going to be VERY asymmetric
Definitely worth experimenting with!

Someone wrote here recently that JRC never expected to see a 4558 distorting in one of the most famous pedals ever. I was looking for a fet when I saw the LM317... "why not?"
The problem is that I'm playing by ear here. Just my intuition and what I've learnt in this forum. Calling Mark & RG!!!

So, what happens if you run the signal into the ADJ pin and connect the IN to 9V, then use the OUT pin for the signal output?

Interesting, I'll try it. I also thought of vin <--> vout ie, an expander.

• Output is short-circuit protected > with BTB Diodes lets hope so

Aren't guys at national smart?

no beach today? I really feel sorry for you

Thx for your condolence words, no beach today but happy hour starting 2am tonight 


mac

[Steben]

The input of the LM317 is about 4.5Volts, since there is no decoupling cap. So the LM works not that way out range.

[Paul Perry (Frostwave)]

The regulator may not be very far from its normal operating voltage input, but internally it is opearting in a mode TOTALLY UNLIKE what it is intended to.
So, for one thing, I would not guarntee that regulators from different manufacturers aill sound the same. As we have seen when overdriving op amps, of course.

[mac]

The input of the LM317 is about 4.5Volts, since there is no decoupling cap. So the LM works not that way out range.

The LM317 amplifies the signal by itself, no need for the opamp, but the problem is to keep a constant vin=vcc/2 while turning knobs from output to gnd, from adj to gnd, for adj to vout, etc. I used pots as voltage dividers at vin and vout (1M-B pots both, lug 1 to gnd, lug 2 to vin & vout, lug 3 to Vcc) and with the 1M-B pot from adj to gnd at its maximun value I set vin=vcc/2 and play with vout pot until I get distortion. Decreasing the adj pot I got cleaner tones but vin moved from vcc/2.
The opamp makes thing easier.

Also I tried vin to vcc, input to adj, and output to vout. I put pots everywhere. It works without the need of an opamp. It has the advantage that vin is fixed. I got some distortion when hiting hard but maybe a resitor at the inout may help. Need to experiment further.

The regulator may not be very far from its normal operating voltage input, but internally it is opearting in a mode TOTALLY UNLIKE what it is intended to.
So, for one thing, I would not guarntee that regulators from different manufacturers aill sound the same. As we have seen when overdriving op amps, of course.

BTW chip is from ST, LM317T


mac

[brett]

Hi
I wonder what would happen if you replaced those diodes with a cap and a pot?
This is because the output is referenced to the common leg (2).  The diodes give an on-off effect, whereas a cap and variable resistor might give a smoother effect.
You're some sort of mad genius for trying this.  Congratulations on getting it going.

[mac]

Hi
I wonder what would happen if you replaced those diodes with a cap and a pot?
This is because the output is referenced to the common leg (2).  The diodes give an on-off effect, whereas a cap and variable resistor might give a smoother effect.

That will make the ratio R1/R2 = 0 since R2 is infinty, so vout=1.25 to DC. But what will happen to AC? Going to try.
There is a 100k pot in my breadboard replacing the BTB diodes but finally I set it near 100-220R

You're some sort of mad genius for trying this.  Congratulations on getting it going.

[snap]

was it Joe D., who had a voltage regulator Fuzz about 6 years ago? 7805?

[Marcos - Munky]

was it Joe D., who had a voltage regulator Fuzz about 6 years ago? 7805?

It was Joe, the year is 2002, and the regulator is a LM317T.

[mac]

It was Joe, the year is 2002, and the regulator is a LM317T.

Do you have a link or schem?

I've been experimenting with the above suggestions but finally came back to the original idea.


mac

[Marcos - Munky]

I have the schematic, but I believe Joe doesn't want his schematics to be posted in another site but Joe's. But if you, or anybody, are interested in the schematic, send me an e-mail to diygescorp at yahoo dot com dot br.

[aron]

That is so clever. I need to try this one!