are caps required for a voltage regulator?
pico
Posts: 29
in Propeller 1
Hey guys,
When powering a propeller chip from a 3.3v regulator, are capacitors required for normal operation?
If I leave them out, can the propeller still function perfectly?
Thanks!
When powering a propeller chip from a 3.3v regulator, are capacitors required for normal operation?
If I leave them out, can the propeller still function perfectly?
Thanks!
Comments
The Prop runs a lot more reliably if it has a couple 1uF caps nearby.
So, if for some reason you can't use them in your circuit, try without to see if your circuit is stable.
-Phil
Definately required, and don't just increase the cap value either!
And the input and output caps are required right at the regulator pins!
And just like the prop requires its' caps right at its' pins!
Just because it works today, does not guarantee it will work tomorrow !!!!!
Yes. All of the above is good advice. It's no different to any other engineering. Basically, low-ESR capacitors physically located against the regulator gives the power regulation a sturdy structure upon which everything else downstream depends on.
A wobbly picnic table works only as long as you are treating it as such.
Modern LDO regulators certainly DO need caps, on both IN and OUT for stability, and just to keep novices on their toes, many have ESR specs to worry about too...
As always, read the data sheet, for the part you are using !
-Phil
Both Ceramics and Tantalums are known for always having low ESR - more ideal capacitor. The notable difference between them is achievable capacitance. Ceramics used to be limited to sub-uF whereas tantilums had no problems with 100 uF or more.
Aside from that, they each have their own problems though. Tantilums have a weakness to stress - They fail easy, usually explosively - Can be problematic in power supply circuits. Ceramics have very loose operational tolerances - Their capacitance varies with everything - Not usually an issue in power supply circuits.
After trying to repair some notebooks' motherboards, whose main defects were related to their batteries' charging and source-switching power distribution circuits, and becaming frustrated with the net results of my efforts, at the point to put my own skills in doubt, I decided to gather more information about ceramic caps, because I was almost sure that they were the main culprits of the bad results I've initially achieved.
As usual, relying on information provided by others, with similar experiences, has saved me many labour days, and a lot of frustration too.
https://pdfserv.maximintegrated.com/en/an/TUT5527.pdf
Hope it helps a bit.
Henrique
-Phil
Sure it was, for me too!
Before I found it, and other ones with similar contents, I was always unsure of specifying and using ceramics, at least in the uF range.
I'm still uncertain if voltage regulator designers take such effects into account, when they write their datasheets, or not.
When the circuit allows enough tolerance, I usualy oversize their values (and dimensions too, sometimes), to stay on the safe side of the fence.
As for power supplies, a rule of thumb that seems to ever work, is switching several loads, between the minimum and maximum designed limits, while observing both current and voltage waveforms, with an osciloscope.
If they have a trend to oscillate, or even fail, as the output load changes, I must know it, way before any customer calls me, complaining.
Henrique
I only use X7R for bypass caps, but had no idea the smaller 0603 were not the same specs as 0805. I guess we are fortunate that the prop is 3V3 so it's a lower voltage
Even so, there is the 5V side of the regulator.
High frequency inductance of the capacitors is small, on the order of 0.5 nH, but circuit traces add a lot. That is why the bypass capacitor needs to be close to the regulator. And the Propeller switches fast, needs and instant supply of current, high frequencies to bypass, so additional bypass capacitors need to be placed right next to the propeller pins. All those capacitors taken together form a bypass network, a beast indeed.
An aside: Avoid low-voltage high-K ceramics for audio coupling. Voltage dependency leads to harmonic distortion at low frequencies where the capacitor's reactance is a significant.
When verifying this characteristic, look for the term "load life" at the datasheets.
Examples from comercialy available MLCCs:
X7R ageing = 12.5% (eight years expected life)
Y5F ageing = 30.0% (THREE YEARS EXPECTED LIFE)
The above term is not to be confused with "life under load conditions", wich it surelly appears to be, in our technical minds, but it is not.
In fact, It is an intrinsic characteristic of the chemical/physical compounds used in the construction of capacitor's ceramic dielectric layers.
And it starts counting from the very moment the capacitor leaves its production line.
How to overcome that apparently terrible limitation (either eight years or three years seems to be too low an useful life aniway)... Temperature cycling all devices (shelf, stock, mounted or not onto PCBs) above the Curie Point of the particular ceramic compound; e.g. - X7R >= 125ºC - recommended = 150ºC, for enough time (two hours IIRC).
Sources (AVX is only an example, every vendor should have this data available):
https://alliedelec.com/m/d/8d6d41767ef490fe339db59d8653fb2f.pdf
sphere.bc.ca/test/production-parts/avx-cer-x7r.pdf
Autoritative reading, for study addicted ones (like me,
Electroceramics - Second Edition
Materials Properties Applications
A. J. Moulson and J. M. Herbert
John Wiley & Sons Ltd
Henrique
-Phil
http://www.ti.com/jp/lit/an/slva115/slva115.pdf
The book I've cited in my last post (electroceramics) has almost all about it, but it's huge (> 550 pages) and sure, it's also a dense reading, as the subject it deals with.
I've just scratched it, at the surface, but it seems such a steep learning curve, even for my 62 y.o. seasoned brain.
In the meantime, i've also found some insightful explanations of the phenomena involved in aging.
One of them is:
https://johansondielectrics.com/ceramic-capacitor-aging-made-simple
IIRC AVX also has a good video about that subject, though I couldn't find it, at the moment.
Henrique
This shows why advice that says "the bypass capacitor needs to be close to the regulator" is not always the best advice.
You can meet the ESR specs, with longer traces to the larger CAP (which is often better placed closer to the load, not regulator).
eg a 10 mm loop, of 10 mil trace calcs at 38 mOhms