# lm317 regulator question for anodizing mild steel.

#### RIKRIK

Joined Oct 11, 2019
146
Okay so recently i watched a great video and read a paper on anodizing mild steel, Paper found here https://pdfs.semanticscholar.org/c532/91aeb43b2ff35b04d88ee973e8df58e42fca.pdf

The quick and dirty, by placing a piece of mild steel in a warm bath of either 50% KOH or 50% NaOH in 50% distilled water. then clipping that piece of mild steel to the positive of a low voltage , low amp power source, & the negative up to apiece of iron. A magnetite oxide layer then forms on the mild steel. Acting like a rust proof coating.

As the chemicals im using are relatively caustic , The experiments are done outside and the anodizing process takes about 5 minutes. So Something portable makes good sense to me.

my main issue is the set up is acting abit like a fuel cell so having a highly conductive solution plus having amp draw which is determined on how close the piece of iron and mild steel are apart. there needs to be some sort of current regulation. The circuit below took my interest.

My questions -
1) it says its a 6v charger but when the resistors R1 & R2 are worked out its 6.9v. Is there some voltage drop from the npn current regulator. 0.6v would bring it down to 6.3v
2) If that is the case could i change the voltage to just over 4v by just changing R1 500ohms - R2 1.5K - output 5V , Minus 0.6 = 4.4V
3) If i wanted to set the current to 250ma would R = 0.6/0.25 = 2.4ohms be correct for R3.

Any help would be appreciated.

Kind regards,
Rick

#### crutschow

Joined Mar 14, 2008
31,137
1) it says its a 6v charger but when the resistors R1 & R2 are worked out its 6.9v.
I calculate 6.98V also as the regulated output.
Is there some voltage drop from the npn current regulator. 0.6v would bring it down to 6.3v
No, the NPN only operates when the current limit is reached.
Don't know why they said it was 7V unless there is an error in their calculation.
2) If that is the case could i change the voltage to just over 4v by just changing R1 500ohms - R2 1.5K - output 5V , Minus 0.6 = 4.4V
The 0.6V only enters into it near the current limit.
For a low current output of 4V, use R1 = 240Ω(maximum value for R1) and R2 = 523Ω.
3) If i wanted to set the current to 250ma would R = 0.6/0.25 = 2.4ohms be correct for R3.
Yes.
But note that the short circuit limit of that circuit is the internal limit of the LM317 (about 1.5A) since the minimum output of the LM317 is 1.25V.

#### RIKRIK

Joined Oct 11, 2019
146
I calculate 6.98V also as the regulated output.
No, the NPN only operates when the current limit is reached.
Don't know why they said it was 7V unless there is an error in their calculation.
The 0.6V only enters into it near the current limit.
For a low current output of 4V, use R1 = 240Ω(maximum value for R1) and R2 = 523Ω.
Yes.
But note that the short circuit limit of that circuit is the internal limit of the LM317 (about 1.5A) since the minimum output of the LM317 is 1.25V.
Thankyou . i wonder if anyone else can solve the 6.98v . maybe its a charging voltage.

#### crutschow

Joined Mar 14, 2008
31,137
if anyone else can solve the 6.98v . maybe its a charging voltage.
Could be.
7V would be about the right voltage to charge a 6V lead-acid battery.

#### RIKRIK

Joined Oct 11, 2019
146
Well il do some testing but here is the basic idea, The zener voltage regulator is abit of a power hog, but i like a bright "i am on". will also probably also modify the circuit to make a 7.5v to 25v wide voltage input 1A phone charger. As most of the stuff i order comes in packs of 10 . Then put the led with a single resistor on the output side, With a 5.1v zener .

#### crutschow

Joined Mar 14, 2008
31,137
D1 is not a good way to the regulate the LED as you are trying to use a voltage regulator to control D2's current.
LEDs require current control, not voltage control, which is typically done with a series resistor.

If you want to make D2's current independent of the input voltage, then it's better to use a constant-current regulator, such as can be done with a couple of transistors.
An LTspice simulation of such a circuit is below;
As can be seen, the LED current is essentially constant as the voltage input goes from 7.5V to 15V.
The LED current is approximately 0.6V/R1, so you select the value of R1 to give the LED current you want.

The current is also essentially independent of the LED forward voltage drop.

Last edited:

#### sparky 1

Joined Nov 3, 2018
732
you can use a bigger voltage regulator the LM338 because current can be a factor and voltage supply to match.

In the data sheet you can find a circuit with the LM334 has nice adjust-ability good up to 10mA , however It can be found in circuits requiring more amperage such as metal anodizers it uses a power a transistor to increase the current and the LM338 Voltage supply to keep up. The LM334 has adjustable current 10uA - 10 mA and the power transistor will give off heat as a result of increased current. The LM338 has higher current capability than LM317.
The PWM has been popular it will reduce heat save energy however the DC output may require more circuitry.

Last edited:

#### crutschow

Joined Mar 14, 2008
31,137
Note that the LM317 will dissipate up to 3.5W so will need to be mounted on an appropriate heat-sink (thermal resistance no more than about 10°C/W), otherwise it will go into thermal shut-down.

#### Alec_t

Joined Sep 17, 2013
13,232
will also probably also modify the circuit to make a 7.5v to 25v wide voltage input 1A phone charger.
Assuming the phone is expecting a 5V input, the 3.5W dissipation limit of the LM317 that Crutschow mentions means that at 1A current there can only be 3.5V across the regulator, hence the input voltage needs to be less than 3.5V + 5V =8.5V to avoid thermal shut-down.

#### shortbus

Joined Sep 30, 2009
9,762
Before spending much time on this project, unless it's for a school type thing or for your own education, you might want to do some more digging into it. With just a quick look at the PDF, their charts mention they did it on C1010 steel, good luck finding that anywhere. While it is a known alloy it isn't common, most mild steel in the market place is C1018. seems funny they didn't use it instead, or it may be that anything with even a slightly higher carbon content(1018 vs 1010) doesn't work out well. But that's just a guess on my part.

Have a read of this link, where people that do metal finishing for a living have to say about anodizing steel. https://www.finishing.com/374/71.shtml Personally after spending my life time in metal working I never even heard of the term anodized steel. There are better faster and chemically ways of making mild steel rust resistant but even then it is never rust proof.

The big problem I can see with doing this electrically is the power source would need to be adjustable for each size or surface area of the steel, just like when you electro plate something. A 1" square would take far less than a 10" square of metal

#### crutschow

Joined Mar 14, 2008
31,137
the 3.5W dissipation limit of the LM317 that Crutschow mentions means that at 1A current there can only be 3.5V across the regulator, hence the input voltage needs to be less than 3.5V + 5V =8.5V to avoid thermal shut-down.
That 3.5W max. dissipation is at the 260mA output he wants, so I don't understand why you are referring to 1A current (?).

Thermal shutdown only occurs if the heat-sink is not sufficient to keep the junction temperature below the shutdown limit at 3.5W dissipation.

#### RIKRIK

Joined Oct 11, 2019
146
Personally after spending my life time in metal working I never even heard of the term anodized steel. There are better faster and chemically ways of making mild steel rust resistant but even then it is never rust proof.

The big problem I can see with doing this electrically is the power source would need to be adjustable for each size or surface area of the steel, just like when you electro plate something. A 1" square would take far less than a 10" square of metal
Heres also a video.
The pieces i am wanting to anodize are only 3.9" squares. I want to test the corrosion resistance.

And thank you everyone for the comments will use the two npns for the led

#### crutschow

Joined Mar 14, 2008
31,137
will use the two npns for the led
OK.
You can substitute 2N2222's for the 2N3904's, if you desire, with negligible change in its operation.

#### Alec_t

Joined Sep 17, 2013
13,232
That 3.5W max. dissipation is at the 260mA output he wants, so I don't understand why you are referring to 1A current (?).
In post #5 the TS states
will also probably also modify the circuit to make a 7.5v to 25v wide voltage input 1A phone charger.
It wasn't clear if the TS was aware that the modification would need to provide a bypass auxiliary semiconductor, as the LM317 alone wouldn't handle the power.

#### crutschow

Joined Mar 14, 2008
31,137
In post #5 the TS states

It wasn't clear if the TS was aware that the modification would need to provide a bypass auxiliary semiconductor, as the LM317 alone wouldn't handle the power.
Okay. I was just looking at the schematic and missed that.

#### shortbus

Joined Sep 30, 2009
9,762
The pieces i am wanting to anodize are only 3.9" squares. I want to test the corrosion resistance.
Did you read the PDF you linked completely, or watch the video to the end? The PDF was more about coloring the steel and so was the video. The PDF talked about a form of salt spray test I never heard of, Spray the salt water one time then wait! Most salt spray tests do it continuous for a certain period of time.

When bot sources of information actually talked of rust prevention/resistance from doing this, they say it is used as a pretreatment to some other process. And both mention the fact that a chemical process is better. Something like a phosphorus wash, and there is a well known product to do this *Ospho* http://www.ospho.com/directions.htm. If you think an electrically applied film of 4 nm = 1.5748E-7 inch, is actually going to prevent rust I wish you luck.