Voltage Regulator Orientation

[canman]

Hi guys,

Just built up a Rebote 2.5 Delay, very cool delay...it's no wonder it is so highly regarded in the DIY community.

On to my question...I checked the pinout of the 78L05 regulator and installed it as per what I thought was correct.  I fired up the pedal and I got delay, but the regulator started to get really hot.  So I flipped it around to see what would happen, didn't seem to affect the delay effect.  

What kind of damage do you think I did by flipping the regulator around?  It didn't get hot in the flipped position so as of now, it's still flipped.  I figured those things aren't supposed to get that hot so it must be correct in the flipped position.

While it was in the "hot" position, I did notice that sometimes the delay would get farty for like a second and then get back to normal.  Is this probably caused by what I described above?  I also noticed that when the delay is after fuzz/distortion, the delays are a lot quieter then when the signal is clean.  Is that normal?

Also, while I'm at it, how would I go about changing the filtering of the delays?  I'd like them to be a little more crisp/clear.  They're a little muddy right now and I know if I roll off less high end, there will be more potential for noise, but I'd like to experiment a little bit.

Sorry for so many lame questions...I'm not very good at all of this yet..!

Thanks!

[tombaker]

A little more information is always helpful. Checkout the debugging thread.

Just built up a Rebote 2.5 Delay

Is it vero/etched pcb/perf, what's the layout? We'll only know what "flipping" the pinout has done to the circuit if we know what layout you've used.

On to my question...I checked the pinout of the 78L05 regulator and installed it as per what I thought was correct.

What datasheet/casing are you going with? This one? http://www.ti.com/lit/ds/symlink/lm78l05.pdf

While it was in the "hot" position, I did notice that sometimes the delay would get farty for like a second and then get back to normal.  Is this probably caused by what I described above?  I also noticed that when the delay is after fuzz/distortion, the delays are a lot quieter then when the signal is clean.  Is that normal?

Are you using a battery or a wall wart power supply? What are the voltages at the source and at the pinout of the 78L05?

Also, while I'm at it, how would I go about changing the filtering of the delays?  I'd like them to be a little more crisp/clear.  They're a little muddy right now and I know if I roll off less high end, there will be more potential for noise, but I'd like to experiment a little bit.

Did you check the Rebote thread?

[deadastronaut]

with the flat side of the 78l05 reg facing you , you should have 9v going in the right hand leg..

middle to ground...and the left leg putting out 5v...


if you want louder repeats swap the 33k at the top of ic2 for a 22k...

[duck_arse]

the regulator will only work as a regulator when it is installed the right way round. if you can measure +9V or so at the input pin, and +5V at the output pin, with respect to the ground terminal, then you can assume it is the right way around and working correctly. it will be less costly to just look the datasheet, put it in the right way the first time.

[canman]

Ah, sorry my bad, didn't mean for this thread to sound like debugging...the circuit works fine, I was just curious as to what *could* happen, generally speaking, if a voltage regulator is in backwards.

This is the layout I used:



I did check the datasheet, but when I installed it as per the datasheet (mentioned in the first post) it got really hot so I got somewhat confused.  This wouldn't be the first time the pinout from Tayda confused me.

I'll check the voltages of the pins and see what I find, thanks for all the help.  I also wasn't aware there was a Rebote thread...

[R.G.]

What *could* happen is anything between no lasting effect and absolute failure, and anything in the middle.

The simplest way to debug whether your regulator escapade is causing the audible issues is to replace it. They're about $0.50.

As a comment on the process of building things, I always have at least one spare of each kind of component. If I need 12 transistors, I order 13 or 14. Well, actually for resistors and transistors, I order by 100's as the price for 100 is often only 10-20% more than a dozen. But caps and ICs, I make sure I have one spare for the one I will undoubtedly kill by putting in backwards. This doubles the expense if I need only one, but is trivial if there are several needed on a project. Then for the cost of storage, I have most of the spares ... free... next time.

[peterg]

If you still are running into heat issues you could swap out the 78L05 for an LM7805. It can take more heat and has the same pin spacing. I've used this one for a few PT2399 based circuit because of heat issues with the 78L05.

[canman]

Luckily I ordered two...not because the first one went bad on me, but because somehow it fell out of the socket and now I can't find it, hah.  I typically order a couple extra of every part as well, luckily it worked out for me this time.  Probably not a bad idea to have a few of these regulators on hand anyways...

Thanks for the LM7805 tip...I'll check the voltages on the regulator in a little while and see if I'm getting proper voltages.  If so and I'm still getting heat issues I'll pick up a couple of the LM's.

EDIT:  Voltages on the regulator look good.  I guess that would be the problem with the regulator getting too hot when I had it in backwards.

[R.G.]

As a specific note on installing ICs backwards:

Modern IC construction has all of the transistors making up the IC "printed" on top of a slab of purified/treated/processed silicon. The transistors are isolated from this slab by reverse-biased diode junctions by doping the first layer above the slab with the reverse polarity of the slab doping, so any voltage on the features above it can revers bias the complete "diode" area to the base slab.

It's like there's a reverse biased diode from every circuit feature in the IC to the base. The base is often used as the the "V-" source for the entire IC. When you insert any normal IC backwards, the substrate/base slab diode is forward biased, and conducts all it can within the limits of the voltage and current available.

In the case of voltage regulators, the positive side regulators have the substrate/base connected to the most negative voltage available, that being ground, and all the works float above it on the operating voltage. So by reversing it, you set up the output and input pins to conduct to the "ground" terminal backwards. It got hot. Since it's now working, it seems like the currents it was provided were low enough not to burn out things on the chip. For cases where someone is using, say, a car battery for a voltage source, there is enough current available to literally vaporize the silicon die or encapsulant and make the plastic case explode. I ... um... know this for a fact.    

As another note, this is why negative regulators like the 7905 as opposed to the positive 7805 have a different pinout. They have the most negative voltage attached to the die (and the metal heat sink plate) too, but that happens to be the input voltage for a negative-side regulator. So the heat sink tab has to be connected to the input voltage, not ground.

[canman]

Wow, thanks for that explanation! (Even though much of it is well over my head)

So what *could* have happened is that more than 5v could have gone to the PT2399?  I'm looking at the datasheet right now, and it doesn't specifically say what voltages are safe for operation, but the echo application circuit shows 5v...does that mean if more than 5v got into the PT2399, it could have fried it?

Since we're now on the topic of voltages...couldn't you insert a voltage divide before pin 1 of the PT2399 (if you're running the rest of the circuit at 9v) instead of using a voltage regulator?

[R.G.]

So what *could* have happened is that more than 5v could have gone to the PT2399?  I'm looking at the datasheet right now, and it doesn't specifically say what voltages are safe for operation, but the echo application circuit shows 5v...does that mean if more than 5v got into the PT2399, it could have fried it?

Well, "could" should have stars and elipses before and after it and should be bold, italic and underlined. Digital stuff (the PT2399 is) generally doesn't do half-works. If it got too much current, it would probably die. The PT2399 has protection diodes to +V and ground on all inputs. What it has on the V+ input is unknown, but might be a few diode junctions to the substrate (ground or V- terminal) and again you get into one of those imponderables where damage could be anything from "let it cool down" to complete death. But if it makes sound at all, it's probably OK.

All of the range of damage depends on how big the overvoltage current was and what it overheated, just like reversing the power supplies may cause damage depending on how much reverse current flowed. It's kind of unknowable unless you had the meters on it when the overcurrent happened.

Since we're now on the topic of voltages...couldn't you insert a voltage divide before pin 1 of the PT2399 (if you're running the rest of the circuit at 9v) instead of using a voltage regulator?

You could. In fact, if you knew the current into the PT2399 very well, you could just insert a resistor that was carefully calculated to drop exactly the right amount of voltage. The problem is that (1) you don't in general know the current, (2) the current changes and (3) a voltage divider is very, very wasteful of power - like 10X the power used must be wasted in the resistors to have some guarantee that the changes in current won't let the voltage go out of bounds.

Voltage dividers sound good in theory, but they're near-useless in practice in real circuits as power supply stuff. They're good for very low power things like reference voltages and bias voltages.

[canman]

Very cool stuff here, thanks for giving of your knowledge!  I really appreciate it.  I have sooooo much to learn..

[Fp-www.Tonepad.com]

the current demands of the 2399 change with frequency, a lot more current is drawn at short repeats than at long ones.

[R.G.]

the current demands of the 2399 change with frequency, a lot more current is drawn at short repeats than at long ones.

Good point. I suspected that was the case, although I have not measured it. I know you already know this, but the following pontificating is for the audience who's learning.

Practically all modern large scale integrated circuits use some variety of CMOS circuits inside. CMOS dissipates essentially zero power when its clock rate is zero, but when it's clocked, all the input capacitances and interconnecting wire capacitances have to be filled up if they change. To fill the capacitors, the drivers that charge the inputs and interconnect lines dissipate power, and it's a fixed amount of energy for each change of logic level. So the faster you clock the circuit, the more times per second those little caps get charged/discharged, and the more little packets of current are used per second.

That is, in general, the faster you clock CMOS, the more current it takes. For short delays on a fixed-length delay line (and the PT2399 is), you have to clock it very fast, and that needs more current.

[canman]

You guys are awesome, thanks so much for helping me in the learning process!

I need to order some books and do some serious reading, it seems...

[duck_arse]

on the topic of pt2399's, is there a datasheet other than the 5 pages from princeton anyone knows of? maybe something with more than just pin connections?

[canman]

if you want louder repeats swap the 33k at the top of ic2 for a 22k...

Sorry, just saw this post...if I replace the 33k with a 22k, wouldn't that make the repeats louder across the board?  Right now my repeats are pretty quiet, but only when I have a distortion pedal running before the delay.  When the signal is clean, the repeats seem to be fine.  Would swapping out this resistor fix that or would it make the repeats louder for the clean signal as well?

[deadastronaut]

no it wouldn't sort the distortion issue...but will raise the level of repeats..

not sure why that would be, never had that problem...hmmmm...

have you tried a different distortion just out of curiosity?..

[canman]

Yeah, I've tried a few distortions and a fuzz...same thing happens with all of them!

I know next to nothing about circuitry, but could the problem be somehow that with a distorted signal hitting the delay, something in the delay circuit doesn't like the distortion and filters it out or something?

I know it was mentioned above that the PT2399 doesn't do "half-works" and that it either works or it doesn't, but maybe this could be the result of the voltage regulator being in backwards?

[bluebunny]

on the topic of pt2399's, is there a datasheet other than the 5 pages from princeton anyone knows of? maybe something with more than just pin connections?

There's Merlin's PT2399 notes here.