How do I Fine tune a 78L05 ?

[Dimbulb]

Does anyone have experience adjusting a TO-92 78L05 ?

I would like to change the output 4.995 V to 5.000 V

at national semi:

http://www.national.com/ds/LM/LM78L05.pdf

under Typical Applications the second example is given.
with a formula

VOUT = 5V + (5V/R1 + IQ) R2
5V/R1 > 3 IQ, load regulation (Lr) ≈ [(R1 + R2)/R1] (Lr of LM78L05)

.. should I modify this equation?
.. which resistors are good for R1 and R2

lets say for example 10 mA
but 50 mA would be cool too

I want precision and stability

Tnx

[Chris Smith]

First, they are fixed and not made to adjust.

Second, the amounts you mention may not consider the VOM you are using.

Are you sure its that accurate?

Third, you can alter the ground pin with a resistor to ground, and the Regulator will alter its out put.

Your stability here depends on the quality of the resistor.

[rshayes]

The circuit that you referred to can only raise the voltage above the point that the individual part regulates at. According to the data sheet, the LM78L05 alone may produce an output of as much as 5.25 volts. You can't guarentee that the part can be adjusted below this level in that circuit.

I would suggest an adjustable part, such as a low power version of the LM317. The output voltage can be set with a divider made with metal film 1% resistors, which will help stability, and the adjustment range can be set to just enough to compensate for the part tolerance and provide the range that you want.

You might consider using one of the shunt regulator chips similar to the TL431. These are also available in TO-92 packages, and the initial tolerance can be as low as 1 or 2 percent. These may also be more stable than the series regulators. They will not supply a large amount of output current, however.

If you need an accurate and stable 5 volt reference, you might look at some of the voltage references from Analog Devices. These parts are usually not very adjustable, but they are trimmed to very tight tolerances. A voltage divider on the output could provide the adjustment range you want if the impedance of the divider is tolerable.

[Robert Reed]

Dimbulb
Apparently you are trying to trim a 5 volt source to less than 5 mv of change.This will mean that the circuits stability will have to be greater than 0.01% (less than 100 PPM drift). Unfortunately, with off the shelf components and even better than 1% resistors, the circuit probably will not have the stability to remain at the desired voltage setting. It would almost have to be temperature controlled in its own oven and with very restricted load ranges to accomplish this.

[Craig Kendrick Sellen]

Dear DIMBULB;
I calulate on my calulator it's only 5mV differance. How crrital is it? You can put a Schottky barrier diode between ground and ground of 78L05 but thats only 100mV differance. Or make a custom regulator by using a PN2222A transistor and a pot. say 10K and at least a 1K in series with the base. You will have to do a custom trim job. It also may be your meter is off a little?

[Dimbulb]

all good ideas but staying with the original problem for now and
yes my meter can do 5 places.


If I use a 10K ohm for R1
.1 ohm resistor for R2

and modify the formula

4.995+(4.995/10000+.05)*.1 = 5.00004995

theoretically a small fractional resistor in this senario should trim the 78L05 using the data sheet instuctions

give me some feedback

[Chris Smith]

The fixed VR requires no resistors.

Play with your ground lead if you want to tweak it voltage.

But the value of the added resistor will be the out come of your stability.

[Robert Reed]

Dimbulb
You can tweak it to any voltage you desire, but what good will it do if it won't stay there?

[Chris Smith]

He will find at the heart of most VRs is a precision Zener reference that only keeps values at a given.

EVERY thing has a plus and minus.

He can play, but like Robert said, what’s the difference if it drifts with the out side temperature well within its norms.

If your looking for DEAD accuracy, then super cool it and maintain that temp, then the VR will stay put within your params.

[positronicle]

--Edited by Positronicle--

[haklesup]

Are you looking for a 5V source of a particular current output and dead on accuracy or are you looking to use the 78L05 as a precision voltage reference you can calibrate.

If you want a precision voltage reference, I suggest you look at one of the many IC devices designed for that purpose. Designing with the 78L05 is kind of like only using the LM741 as your only op amp. There aren't that many in the TO-92 package but the search tools on Digikey, Mouser etc. make cross referencing and searching by spec pretty easy.

Just build it and tell us what you find out. Its only a couple bucks in parts.

Cooling and regulating the temp will result in improved stability but not necessarily accuracy.

A precision supply can be made by using a precision reference and buffering that through either a transistor circuit or monolithic op amp with high current output and unity gain. With feedback from a temp compensation circuit you can maintain good stability. Each decimal point you want will cost exponentially more than the last though.

[Chris Smith]

I think you meant using an oven and"heat".

The exact opposite is true.

[positronicle]

--Edited by Positronicle--

[Chris Smith]

That’s nice, were not dealing with crystals, were dealing with the transistor.

HEAT, and the lack of it stabilizes most things unless their atomic structure need the heat factor to work properly. [Crystals]

Transistors, semiconductors, all of the above work best at a loss of heat, not a gain.

Crystals in general don’t dissipate extra energy or heat. [Negligible]

Yes, you can preset the oven and place your device in it hoping it will remain the same as you inject more energy into it, but the reality is as you add in more energy into a transistor the temperature will rise and will throw off the semi conductors values.

Crystals are different, they require some what of a room temperature in order to operate, thus when you control its temperature, and it does not unilaterally increase its temperature, it remains fairly constant and thus accurate.

Not the same for the transistor, it needs to dissipate energy, and any heat is already energy that you don’t need to deal with, thus extract it long before it becomes a problem.

[positronicle]

--Edited by Positronicle--