Yes, the age old debate of this cap versus that cap.
But I want to get a real world use case from those that are actually designing and implementing new designs. My design is automotive based so there will be the typical noise sources, transient voltage and current spikes, etc.. But for the sake of this conversation we are only discussing capacitors and usage in new circuit designs.
I had originally purchased some 220uF 16v electro caps, and .15uF ceramics to use for line filter ahead of Vin feed for the Arduino as well as a LM 7805 on the starter input signal.
Now according to the LM7805 it recommends a 0.22uF cap on the input side and in the schematic shows a non polarity cap, so assuming not an electro cap.
I am getting ready to place another Digikey order and have some 0.22uF electro's in my cart but I am also looking at just ditching the electros for ceramics.
I know that different caps have different filter capabilities, but in this case everything is analog 9.5v to 14.5v dc inputs.
So electro vs ceramic, should I ditch the tin cans or do they have their place in a automotive dc filter circuits?
These two capacitors are different by a factor of 1000. Standard ceramic capacitors aren't available for a 22µF value. I would use both in your circuit. The ceramic capacitor has a much lower capacitance but is much faster in response, the electrolytic capacitor will provide enough power if requested.
References please ! You are not in a pub, you are among nerds here.
When you say that the LM7805 recommends a 22µF capacitor, can you give a link to that datasheet. There are many LM7805 datasheets.
Every decent datasheet mentions what the requirements are for those capacitors.
When you talk about a capacitor at Digikey, can you give a link to it.
I have seen blown and leaking capacitors in every possible way. Ceramic capacitors don't have that problem, although I came across a ceramic capacitor with a short circuit.
At least Panasonic makes very good electrolytic capacitors.
It is partially application dependent and circuit dependent. In your case you need to use the ALUM cap as it meets the regulator requirement, the Ceramic cap does not. As how much voltage your circuit will see, if it is connected to the battery there is the 24V battery jump and reverse battery. You also need to consider load dump. Since it is automotive based is it first, second or third tier.
I am accepting both answers as true, thus Electrolytic Capacitors will be implemented as originally designed. Thank You both.
Gil, you bring up good point in regards to the Automotive circuit variables.
I don't see a 22µF capacitor in the datasheet.
They don't mention the ESR for the capacitor. Perhaps they assume that any ceramic capacitor should be good enough.
If you have a project with wires and motors and a lot of everything, then a 100...1000 µF capacitor for the power will help to make the circuit more reliable. Electrolytic capacitors are not always bad.
The datasheet says 0.22uF. That would be ceramic or some other non-electrolytic type. I don't think I've ever seen an electrolytic less than 1uF.
There is a strange characteristic of electrolytic capacitors that they "don't act like capacitors" at very-high frequencies and they may actually become (slightly) inductive. In applications like this where the capacitor is recommended for "stability" (to prevent oscillation) an electrolytic may not work because it will likely oscillate at high frequencies where the capacitor is less effective.
Most analog and digital IC's need "bypass capacitors", again for stability, so you'll often see one or more ceramic capacitors near every IC on the board.
Occasionally you'll see an electrolytic capacitor for the required higher capacitance plus a ceramic capacitor in parallel to "take over" at high frequencies.
Capacitors alone aren't great for noise suppression because the high current available from the car's electrical system can charge the capacitor very quickly and the heavy current load means that the capacitor can discharge (into everything in the car) very quickly.
The regulator itself provides a lot of noise suppression (because it's trying to hold 5V).
With an inductor you can make an L-C filter and the inductor "resists" current surges/changes. A regular R-C filter doesn't usually work as a power supply filter because of the voltage drop across the resistor.
The LM78xx is a 40-50 year old design by National Semi. Even back in '78 National stated: Input by-passing is needed only if the regulator is located far from the filter capacitor of the power supply. What size cap and what "far"meant was left up to the designer.
As far as I know, the origin of the 0.22uF or 0.33uF cap is unknown. If my power supply is located 6 feet from the LM78 is 0.22uF good enough?
Well, they are NOW. I've got a big pile of 100uF ceramic caps, even.
I don't think they existed when the LM7805 was designed, though, so if the datasheet says 22uF, it's almost certainly electrolytic. (OTOH, apparently the datasheet says 0.22uF, which would be ceramic. Probably a disk, if you want historical accuracy!)
The lm7805 is pretty primitive and bulletproof. The original Arduino Serial used a 7805 with 100uF and 0.1uF caps in parallel on the input side.
The datasheet does read .22uF, not 22uF, huge difference. I keep making rookie mistakes, thanks for not roasting me on this one, yikes. I chalk this one up to my new glasses that are great outside, but not ideal sitting in front of the computer.
I got point two two ceramic caps, TDK brand, I will keep the 22uF electro caps and I also threw some 100uF electro caps in for good measure.
Sorry @ramjeepee If you are mixing up uF, nF and pF capacitors I'd not be confident using any system you designed.
In an automotive environment you need to consider not just the capacitance. Every "capacitor" is a component that has capacitance but also resistance and inductance. For reliability you would also need to take into account the ripple current.
Hello John,
Thank you for your contribution to the conversation. In regards the '22pf' capacitor I referenced, that was an error in parts selection.
In regards to your comment regarding mixed capacitors in a system, if you look at something like a engine control unit (ECU) you will find mixed capacitors throughout such a design.
Here is a Bill of Materials for a very popular Engine Control Management system that has been installed in millions of vehicles.
So having mixed capacitance within a circuit design does not indicate flaws.
Now as to your confidence in my system design?
I am not an electrical engineer, I am obviously a noob trying to learn, so I wouldn't have confidence in my design either!
I wouldn't be asking this questions on this forum if I had confidence in my design.
Again, thanks for your contribution to the conversation.
As I mentioned before, the original National data sheet does not provide a value for an input bypass capacitor. How and why that value was chosen is a mystery.
My initial statement (which I need to go back and correct) was supposed to indicate a .22uF capacitor, but regardless that doesn't change the point your making.
In this use case, there is no 'power supply' other then the raw DC system voltage in an automotive environment. As you mention the .22uF is quite arbitrary and ultimately would be up to a systems designer to spec out what is needed based on the particular system at hand.
So with that said I will set it up with a series of protective diodes and a couple of capacitors. It is on the starter motor circuit which includes the starter relay and starter motor so I am sure there will be some noise on the circuit. I will just do my best to 'clean' up the noise on the way into the LM7805 and throw it into service. Its a one off device so its not likely to have any terrible consequences if this one transistor fails.
Back when the LM78 was invented, high voltage (more than 12V) ceramic caps were rare and expensive, a designer would have probably used an electrolytic cap if required.
Electrolytic caps are more robust when it comes to absorbing line transients. It's probably a good idea to have an electrolytic on the input to the regulator. Not that you need a large value but larger values have lower ESRs. So maybe 47uf, 100uf
Sorry if my comments seem negative, I'm tryingto warn you of problems but phrasing it badly.
So having mixed capacitance within a circuit design does not indicate flaws
No - but you have to have the experience to know which to choose- not just values, but tolerances, types (eg polyester, ceramic, electrolytic, tantalum, polarised / non-polarised ...) impedance, voltage and power ratings.
You might think that is the case. Sadly it doesnt always follow as not everyone conforms to the standards (of which to add to the confusion there are many)
Each has its place and the two are very different.
Automotive design covers a huge range of electronics, from very high voltage, current and frequencies to very sensitive components.
Perhaps if you were to give us an idea of what you intend we could give more positive advice.
