Wiring a ceramic capacitor

Hi all:),
first I wish to express my gratitude to the creators of Arduino and all his supporters from this forum. I'm really enjoying this new "world", and the info I get from this forum is awesome.

I have read that when using a current stabilizer like the L7808, it is better to put in the output a small ceramic capacitor (that has no polarity), which scope is to filter out interferences produced by the integrated itself.
The question is: in the picture below, are both the wiring doing the same implementation of the schema above?

I know that this question maybe is ridiculous, and I swear I tried without success to get an answer from the web. Personally I'm more inclined to answer positively to the question, but before printing the PCB I thought was better to ask to you.
Thanks

Hi, Welcome.

To answer your question. In theory the one on the left is better. The reason is that wire or clad length reduces the the ability of a capacitor to filter very high frequencies.

In practicality the two layouts likely make no difference.

If I were doing the layout, I would put the capacitor just to the left of the regulator. That would make the clad between the two devices very short, so slightly better high frequency filtering.

Also look at the 7808 spec sheet. I believe there should be a capacitor before the regulator. This will likely be of higher capacitance that the one you have shown. It might be a polarized capacitor.

John

I said it was better to ask! thank you very much JohnRob for the answer, that is NO!
So I will use the first wiring because I cannot really put closer the capacitor to the L7808.
Regarding the capacitor in input, please tell me if I'm wrong again :slight_smile:
I intentionally removed it because I thought that it was an overkill considerating that the current comes from a stabilized power supply and the 8v output are furthermore stabilized by the Arduino Nano current stabilizer.
Is it still necessary?

I would add a capacitor at the input of the 7808, even is only a 0.1µf ceramic. The 7808 is an analog amplifier that requires low input impedance for stability. You might get away without it but if you can possibly place it on the PCB you would be best to do so.

BTW adding Karma is a good way to show appreciation.

John

The idea is to out the cap AS CLOSE AS POSSIBLE to the linear regulator (which for reasons unknown, you are calling a "current stabilizer". I have a feeling Google Translate is to blame.)

JohnRob you have now more Karma than ever :wink:

Yes, I meant a linear regulator, not a google error but mine!
The PCB space is quite limited, 8x13 cm for an irrigation system, including a PS 24V 2A (8.5x5.8 cm).
The L7808 with the cooler, will be plugged in a three pins Dupont socket, this to make it easily replaceable. So I realize now that the capacitor could be moved immediately close and below the integrated (there is enough height).

But now I wonder if temperatures are a problem: the PCB is fitted in a waterproof box, so everything is perfectly sealed, no breeze will mitigate the temperature of the TIP120+L7808 dropping tens of volts in August!
I want to create something that lasts 10 years! A small capacitor near an IC at 100 degrees is at risk? What can happen it the capacitor, in that position, explodes and has a short circuit?

Thank you very much for your support. The project when completed and tested will be shared for free, including of course the software that it is the part where I'm much more skilled.

If you use a chip capacitor, it can be soldered directly across the wires of the regulator.

Not only do you need a ceramic capacitor on the output but you also need one in the input and a larger polarised capacitor on the output. If you do not do then then the regulator might very well oscillate.

You need to look at the data sheet for the manufacture you use, these are different for each manufacturer. It tells you the minimum capacitor value to use.

no breeze will mitigate the temperature of the TIP120+L7808 dropping tens of volts in August!

I don’t know how the TIP120 in involved, However the 7808 dropping 24-5=19 volts (a guess) times the current draw of 5 ma (another guess) means you must get rid of 100 mw. If I’m even close you might not be in too bad a shape. However you need to use the actual numbers to be sure.

Almost forgot.

The L7808 with the cooler, will be plugged in a three pins Dupont socket

.

Dupont sockets are the most unreliable connection there is! (maybe I shouldn’t be so harsh but they are very bad)

If you want 10 years, you will have to solder the regulator in.

JohnRob:
I don't know how the TIP120 in involved, However the 7808 dropping 24-5=19 volts (a guess)

The 7808 drops to 8V, not 5V. That'd be the 7805.

If the OP decides to use an Uno the minimum current draw of the thing is more like 50 mA, giving 50 * (24 - 8) = 800 mW waste heat. That's a lot for a closed copntainer. Then we also don't know what else the OP has connected (or indeed even which exact Arduino model is in play). A packed 9x13 cm PCB implies there's a lot more to this project than just an Arduino.

The proper way to power an Arduino off a 24V supply is by using a 24-5V buck converter (~90% efficient, vs. the 20% of a linear regulator), connected to the Vcc or 5V pin, not using the Vin pin or barrel jack. That way you can easily draw 1-2A of current without having problems with heat.

JohnRob:
Dupont sockets are the most unreliable connection there is! (maybe I shouldn't be so harsh but they are very bad)

I have made some one-off projects using DuPont sockets and wires. I then used a hot glue-gun to fix the wires into the connectors.

There is a worse connection, much worse, but really off-topic here. That would be aluminum wiring in your home.

Thank you again for all your answer, I find very useful to listen to different points of views.
Datasheets of electronic components have become my favorite reading recently, but this doesn't mean that I have understood well all I have read :slight_smile:
On the PCB there is an Arduino Nano, an RTC DS3231 module, a TIP120, an L7808 and a few components and connectors (to 3 buttons, to an OLED screen, to the PS and to the solenoid valve.).
I will realize two versions of the project, one with 12V2A PS and one with a 24V2A PS.
24V is the only way I know to output 50W from a space of 8.5x5.8cm. I do really want to use a power supply already assembled and enclosed.

The system will be up and running for a maximum of 20 consecutive minutes (normally 5 are enough) and then it will have hours to cool down before the next watering.
The "brain" of the system (Nano+RTC+Screen) should not drain more than 150mW, so the L7808 must dissipate (24-8)*0.15=2.4Watt. Is it correct? Can a small passive cooler easily dissipate this amount?

I will immediately dig the suggestion from wvmarle of using a buck converter. I didn't know about this big difference with efficiency.

Regarding the stability of components plugged on Dupont sockets: Nano and the RTC are very firm, the TO220 not so much but they seem to be quite stable for the scope. Of course, I cannot play football with the box :slight_smile:
I will proceed with the Duponts in the first version because I wish to replace easily the components.

Hi Grumpy_Mike, regarding the larger polarised capacitor on the output, on the L7800
SERIES specs I cannot find this need. Is it really needed, can the interferences affect the Arduino computation capabilities?

Let's face it, no self-respecting technician or engineer would be caught dead using Dupont sockets for anything that needed to be reliable because the only reason anyone would use them for anything that was not temporary is that they can't solder. (don't know how, don't have the equipment etc.).
Dupont sockets ARE rated for 100mA but I don't think that's the issue here. The question is WHY would you use them if you are not solder-challenged and the project is NOT a temporary low current breadboard circuit ?

Thanks raschemmel (and JohnRob) for you considerations about the Dupont. It seemed to me a practical idea to have them replaceable without dismantling the enclosure/pcb, but now you convinced me to abandon this solution.
I tried to find a dedicated socket for this scope, but they exists for simply testing the IC and they are rated max 1A.

Hi Grumpy_Mike, regarding the larger polarised capacitor on the output, on the L7800
SERIES specs I cannot find this need. Is it really needed,

I once had to scrap a whole days production of set top boxes because some one in purchasing changed the manufacturer of the voltage regulator we used. The new regulator needed a larger minimum capacitor for stable operation. And you are asking about not using one at all!

Hi, i will refactor the PCB taking in consideration your advice.
I will solder the TIP120 directly to the board.
Regarding the l7808, I've investigated a bit the buck step down converters, and I think they win the comparison because because they are very cheap, small and efficient.

Sorry if insist with these Dupont, but now maybe the situation is slightly different: what if i solder pins to one of this buck (like in the picture below) and then i plug it in a 4 pins Dupont socket?
The 5V pin of Ardino (also plugged on a Dupont) will face the same current of the buck output pin. Since the buck is very efficient (97%) I suppose that also the input pin of the buck will manage more or less the same amount of current. Am I completely wrong in wanting to use these Duponts, at least when it is technically possible?

I think OP is referring to a large cap often seen on the output - in conjunction with the smaller ceramic cap that keeps the regulator itself stable.

Whether you need one of those and how big depends on the overall circuit. It won't hurt.

Ipposnif:
. Am I completely wrong in wanting to use these Duponts, at least when it is technically possible?

I don't see the point. Properly selected components just don't break.

Buck converters have lower input current than output current. 1A 5V output will have just over 200 mA 24V input.

Efficiency of 97% will only be reached at specific load and very low voltage drop, it's best case. 85-90% is more realistic a number.

You're in serious need of being weened from Dupont connectors. You are chained to them. They are designed for hobbyists and people wanting to perform some temporary circuit for some reason but have no intention of any long term use. Are you violating that guideline ?

raschemmel:
Let's face it, no self-respecting technician or engineer would be caught dead using Dupont sockets for anything that needed to be reliable because the only reason anyone would use them for anything that was not temporary is that they can't solder. (don't know how, don't have the equipment etc.).

I just inserted a case where DuPonts were a solution to a specific project. I had less than 6 hours to go from a breadboard to a finished project for a one-time use.