9 volts output with an LM7805

[Silvio55]

Hi, I need to build a 9 Volts PS for a pedal but only have 5 volts regulators. I knew that you can increase the voltage output of a regulator by placing a resistor between the ground leg and ground. But many schematics online show another resistor between the ground leg and the output leg, forming a voltage divider like this:



What is the purpose of R1 in that schematic? Wouldn't it just work with R2 alone? Thanks!

[iainpunk]

no, it won't work with just a single resistor.
what the 7805 does is create a 5v differential between the out and COM pin. the COM pin doesn't allow for much current flow. this leaves you with that the current through the 2 resistors is roughly the same.
putting a 5k resistor between out and the COM and a 4k resistor between COM and ground, you'll get a nice 9v source.
those values aren't easy to come by, so i suggest you use a 5k6 and a 5k trim-pot.

cheers

[antonis]

You can use ANY fixed voltage regulator as an adjustable one by simply setting respective Vref instead of 1.25 into LM317/350 respective formula Vout = 1.25 (1 + R2/R1)
(of course, output lower adjustment should be Vref..)

[PRR]

.... Wouldn't it just work with R2 alone?....

What size R2 would you use? (Do you know your E I and R?)

[antonis]

What size R2 would you use? (Do you know your E I and R?)

What is the value of GND pin current (I_Q).??

[Silvio55]

no, it won't work with just a single resistor.
what the 7805 does is create a 5v differential between the out and COM pin. the COM pin doesn't allow for much current flow. this leaves you with that the current through the 2 resistors is roughly the same.
putting a 5k resistor between out and the COM and a 4k resistor between COM and ground, you'll get a nice 9v source.
those values aren't easy to come by, so i suggest you use a 5k6 and a 5k trim-pot.

cheers

Thanks! I found this for R1 and R2 values:


They are lower but common values.

[Rob Strand]

no, it won't work with just a single resistor.
what the 7805 does is create a 5v differential between the out and COM pin. the COM pin doesn't allow for much current flow. this leaves you with that the current through the 2 resistors is roughly the same.
putting a 5k resistor between out and the COM and a 4k resistor between COM and ground, you'll get a nice 9v source.
those values aren't easy to come by, so i suggest you use a 5k6 and a 5k trim-pot.

cheers

Thanks! I found this for R1 and R2 values:


They are lower but common values.

The idea works.  The low resistor values help improve the regulation characteristics.

The current down the ground leg isn't small on the LM78L05, it has a wide tolerance, it varies with the input voltage and with the load.   Using low value resistors helps "water down" these effects.   Check out the IQ and delta IQ values in the datasheet.   The LM317 has a much lower IQ and very much less delta IQ as it has been designed with the intention of having R1, R2 present.
http://users.ece.utexas.edu/~valvano/Datasheets/LM78L05.pdf
https://www.ti.com/lit/ds/symlink/lm317.pdf

You should allow for a small increase in voltage due to IQ through R2.

The down side of low resistor values is you have to waste a lot of current down R1 and R2.   So there is a trade-off between regulator performance and current wastage.


You can put  zener in place of R2 which is a better trade-off.  (see on-line)

[PRR]

> You can put  zener in place of R2

Then you still want an "R1" because <7V Zeners are awful soft and the GND current of the 7805 is awful undefined.

To the point that the choice would be 50:50, except Zeners are typically MUCH more expensive than resistors. (Yes, I know: this crowd may have a bag under the bench of ninety-seven of those 3.9V Zeners from a phaser bias project.)

[Rob Strand]

PS, sorry accidentally edited your post but it's restored.

> You can put  zener in place of R2

Then you still want an "R1" because <7V Zeners are awful soft and the GND current of the 7805 is awful undefined.

To the point that the choice would be 50:50, except Zeners are typically MUCH more expensive than resistors. (Yes, I know: this crowd may have a bag under the bench of ninety-seven of those 3.9V Zeners from a phaser bias project.)

Sure, that was kind of implied.  It's yet another trade-off.

In either case it's about being conscious of the limitations and trade-offs.   The designer has to make sure it does what it is supposed to do.

(FWIW, the temperature coefficient is affected by both methods.  For audio you might not care but for something else it might.)

[ElectricDruid]

Another way to do the same thing is use a string of series diodes from the common pin down to ground to bump the output voltage up.

Put six or seven 1N4148s in there instead and see what you get.

Then again, the resistors is probably neater if you need to go all the way to 9V.

[Rob Strand]

Here's a 2xresistors vs zener + resistor.

The low bias design uses less than half the current but drives the zener with enough current to keep away from the low current region.   The low bias design uses a larger R1 value and wastes less current.  At that current a 4.3V zener would be competing with a chain of silicon diodes.   I've allowed for the zener voltage tolerance.  If you wanted to be nasty you could double the variation on the IQ part of the zener but as you can see the zener tolerance contribute more anyway.  Also even with the zener tolerance the resistor version is worse.



The take home message is LM317's and LM78L09's exist for a reason.