7805 7807 & 7808 voltage regulator question

[FUZZZZzzzz]

When I look at the datasheet for positive voltage regulators like 7805 7807 7808 etc. I always see examples with caps from the input voltage to grnd and the output voltage to ground, but everyone uses different cap sizes. How do you guys deal with this for guitar pedals (lets say 9 or 12 v 1a).

Also, Looking at this picture. Does this mean I cannot use the 7808 from a 9 volt supply as the Minimum Input Voltage = Vout + 2.5 = 8 + 2.5 = 10,5 Volt? or am I missing something?

[FUZZZZzzzz]

found some cool reading as well:

http://tronixstuff.com/wp-content/uploads/2010/06/fairchild-lm7805c.pdf

[BetterOffShred]

When I look at the datasheet for positive voltage regulators like 7805 7807 7808 etc. I always see examples with caps from the input voltage to grnd and the output voltage to ground, but everyone uses different cap sizes. How do you guys deal with this for guitar pedals (lets say 9 or 12 v 1a).

Also, Looking at this picture. Does this mean I cannot use the 7808 from a 9 volt supply as the Minimum Input Voltage = Vout + 2.5 = 8 + 2.5 = 10,5 Volt? or am I missing something?

It's been my experience as limited as it is, that you want to supply about 3 more volts to the regulator than you're expecting to get out.   There are some variables of course.   

[anotherjim]

Sadly it's true. With input V too low, output is input -2.5V.

If I wanted a little lower lower than 9v, I'd just drop it as close as I can with a series diode or two.
A zener (8v2) can do it. Needs a series resistor but there would only be 1v drop on it.
There are LDO (low dropout) regulators than can run closer to Vin.
If you really wanted a properly regulated 8v with a 78xx, then you'd probably have to use a DC/DC converter to produce say 12v from the 9v first.

[ElectricDruid]

Also, Looking at this picture. Does this mean I cannot use the 7808 from a 9 volt supply as the Minimum Input Voltage = Vout + 2.5 = 8 + 2.5 = 10,5 Volt? or am I missing something?

No, you're not missing anything. You're spot on.

If you want a regulated voltage that's closer to the input voltage, you have to look for "low drop-out regulators" aka "LDO regulators".

As others have said, there are other ways, and that may not always be necessary.

HTH,
Tom

[amptramp]

These old school regulators had to drive a large NPN transistor which was usually connected as a darlington internally (meaning a drop of 1.4 volts) but additional voltage was required to drive the transistor from a differential stage.  This added up to as much as three volts drop from the supply just to get it to turn on.  The low dropout (LDO) regulators use a PNP transistor for a positive regulator meaning the collector could go down to less than a diode drop from the supply rail.  There would be an additional drop for the current limiting in both cases, but it could be small.

[R.G.]

A.T. is right - there exist low drop out (LDO) regulators for exactly the reason you're wondering about.

However, (you knew that was coming, didn't you?     ) LDOs have issues with capactive loading making them unstable, because the PNP pass device actually has a variable voltage gain with changing currents and loadings, and so certain capacitive loads make them unstable. The datasheets for these (... usually) contain the magic limits of how much capacitance is too much and how much is too little.

[Rob Strand]

anotherjim already mentioned the LDO's.

The datasheets for these (... usually) contain the magic limits of how much capacitance is too much and how much is too little.

I've seen LDOs which specify a *minimum* ESR for the output cap (for obvious reasons).
Think about how you would meet that requirement when you have bypass caps all around the board.

[R.G.]

anotherjim already mentioned the LDO's.

Ooops. Missed that. Sorry.

The datasheets for these (... usually) contain the magic limits of how much capacitance is too much and how much is too little.

I've seen LDOs which specify a *minimum* ESR for the output cap (for obvious reasons).
Think about how you would meet that requirement when you have bypass caps all around the board.

The only good way is with a resistor, of course.

[Rob Strand]

The only good way is with a resistor, of course.

Exactly!  You wouldn't even consider doing that in the old days.

Another thing I've discovered about the new era regulators is some of the low quiescent current regulators have *unbelievably poor* line regulation - to the point where you would only run it off the output of another regulator.

[R.G.]

After reading some LDO datasheets, with critical capacitances and ESR requirements, I came to a simple internal guideline: don't use LDOs unless you simply can't do anything else. I've been tempted to use them several times, and got as far as prototypes in one case. I went back and designed them out every time.

[Rob Strand]

I came to a simple internal guideline: don't use LDOs unless you simply can't do anything else.

Not a bad policy.  Plenty of gotchas lurking around these days.

[rankot]

You can always regulate with a circuit like this one:



Regulated voltage is equal: V_zener-V_{BE_junction_droupout}+V_f(D1)

D1 can be omitted, or replaced with Schottky type, or you can combine more diodes to achieve desired output voltage. Maybe not as nicely regulated as with 780x or similar, but voltage drop can be really small - up to 1V.

[FUZZZZzzzz]

Wow! a lot of useful information. Thank you very much! So, if I decided to work with lets say a 7806. How do I determine the value of both in and output cap to grnd? If I want to use with a 9v 1a power supply (one spot or similar)

LDOs is a whole new learning curve for me. Lot of stuff to read.

[rankot]

For 78/79xx familty, input capacitor is 330n (or a little bit more, but not less), output capacitor i 100n.

[thomasha]

since we're talking about regulators...

if I use a regulator for tube heaters in series, something like 440mA at 12v for some russian subminiatures,
how do I avoid the current protection?

When I connect the pedal to the regulated supply the initial resistance of the filament is very low, and the regulator sees a short and stops working.
If I turn everything on at the same time, with the regulator's capacitor uncharged, it works. In front of the regulator my supply also has a 1000uF capacitor for a 15v output, so it takes a lot of time to it to discharge it and for the thing to start working again.

There was a trick using another capacitor to overcome this, but I can't find the source.
Or is the only solution using a resistor, like 1M5, to discharge the first cap faster?

[rankot]

Why using a regulator for heater at all? You can drop voltage with resistor.

[thomasha]

that's part of the wall wart.

I have similar problems with other high current devices, as for battery charging.

[amptramp]

You can use an LM317 as a current regulator by connecting a resistor from the output to the adjust pin and taking the output off the adjust pin if you want to control heater current.  The current is the resistor value divided by the 1.25 volt difference between the adjust and output pins.  The advantage of controlling current is that there is no turn-on surge from a cold heater - the current remains constant, so the warmup may be slightly longer but much less stressful.  At a constant voltage, heater current usually starts at about four times the steady-state current.

[antonis]

So, if I decided to work with lets say a 7806. How do I determine the value of both in and output cap to grnd? If I want to use with a 9v 1a power supply (one spot or similar)

IN cap is mainly set for eliminating remaining voltage ripple due to long distance from PS filter & OUT cap for improving stability and transient responce..

If you realy want to mess with those caps you have to take in mind your specific regulator line & load regulation (which vary in accordance to junction temperature  ) and maximum load current to calculate their values for regulation to be set more closer to your taste (if it differs from manufacturer's taste..)

Or just throw 300nF & 100nF to In & Out and sleep in piece...

P.S.
A 1A current should be mandatory for heat-sinking..