How to choose filter capacitors for power supply

Подписаться 167 тыс.

Просмотров 34 тыс.

50% 1

How to choose filter capacitors for power supply
Answering a viewers question on how to choose capacitor values for a power supply.
5 boards for about $22 in about 7 days www.pcbway.com
Yes!! I now have merch: teespring.com/stores/learnele... Electronics themed T-Shirts!
HEY YOU ABOUT TO POST SOMETHING STUPID:
Yes, I know my hands are large/fat. It's because of a heart condition I suffer from called Congestive Heart Failure. Here's a link to some information about CHF: cle.clinic/2TdS2Ux
One of the symptoms is that is causes fluid retention, mostly in the hands, feet/ankles, and face.
It will kill me one day. I'm doing ok right now, but someday...
--------------------------------------------------------------------------------------------------------------------
If you are shopping for electronic components, test gear or consumables please consider visiting my Amazon shop @ amazon.com/shop/learnelectronics
Please check out www.patreon.com/learnelectronics and pledge a dollar if you can. It will go a long way to keeping the channel alive. It costs a bit of money to buy all the items and produce these videos. You help is appreciated.
---------------------------------------------------------------------------------------------------------------------
Or....if you'd like to send a one-time donation you can use this link: paypal.me/learnelectronics
----------------------------------------------------------------------------------------------------------------------
Here is all the gear I currently use:
Oscilloscope 1: Rigol 1054Z amzn.to/2PRybtN
Oscilloscope 2: Siglent SDS1202X-E amzn.to/31Kk1yH
Power Supply 1: TekPower TP3005N amzn.to/2TeQnxI
Power Supply 2: RUIDENG DP50V5A amzn.to/2sNicSe
Signal Generator: JS2800 ban.ggood.vip/d9vk
DMM1: Owon18B amzn.to/2SbIGLs
DMM2: Aneng 8008 amzn.to/2sNQncI
DMM3: Aneng 8009 amzn.to/2S5qeEp
DMM4: Uni-T UT801 amzn.to/2S89SuL
Soldering Station1: Zeny (Yihua)862D+ amzn.to/2S49z4e
Soldering Station2: KSGR T12 amzn.to/2Wk7Dn9
Light Meter Smart Sensor AS803 amzn.to/2B2gyRg

Наука

Опубликовано:

13 июл 2024

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 75

@KissAnalog 4 года назад

Great video Paul. Here’s a simplified guide; use 8500 uF for each amp in load current for 1 V ripple (that’s peak to peak ripple for the equation not rms), then if you want to cut the ripple in half double the size of capacitor. If you want a fourth, so the 250 mV you talked about, then you multiply by 4 (to half the ripple twice). If load current is doubled, then multiply by 2. Example 1: 1 V ripple and 1 Amp, then 8500 uF Example 2: 1V ripple and 2 Amp (multiply by 2), so 17000 uF Example 3: 250 mV ripple (multiply by 4) and 1 Amp load, so 8500 uF x 4 = 34000 uF Example 4: 250 mV ripple (multiply by 4) and 2 Amp load, so 4500 uF x 4 x 2 = 67000 uF I hope that makes sense. I rounded up in most cases.

@SkullDraizer 4 года назад

I like this way of looking at it.

@abeditani8293 4 года назад

I think in my opinion that convert AC to DC with more 10 amp is not ideal for 50 or 60 Hertz . What if I want a pure DC 24v 25Amp with ripple 1% 240mv It will need a big capacitor and that capacitor will need a very very big current in the starting from 0 to 24v ( that may short the life cycle for the capacitor ) . I think the switching power supply is the best way for more than 8 amp

@trelligan42 4 года назад

@@abeditani8293 Use a resistor (low ohms, high watts) to limit the input rush. As you say, saves the capacitor from becoming a high-cost short (or open). To calculate the resistance, use the input voltage peak and max inrush current (from capacitor spec sheet); R = E / I. To calculate the wattage, use 5 time constants of the RC value, and assume you'll have a minute (depending on the project) before you'll need to turn it on again, worst case. I^2 * R * Tc / (60 seconds). That's a conservative value, and if your project is power-hungry you may need to incorporate thermal load calculations. This resistor adds a tiny bit of ripple reduction, giving you a larger margin if you don't count on it.

@davidluther3408 4 года назад

WHY DID YOU START WITH 4500uf?HOW DID YOU GET 66000uf?

@KissAnalog 4 года назад

david luther Thanks David, I’ve been on cold medicine - so I guess I’ll use that excuse;)

@ViniciusCZanini 4 года назад

Nice video!!! Very informative please make more videos like this... They could tedious for some people but they’re really important too!

@Taran72 4 года назад

WOW, Awesome video! I really liked it because you explain things clearly and straight to the point. And the information is really useful if you want to build something. love your channel! :)

@TheBinarydeity 3 года назад

awesome video, i like the technical vids you do allot.

@bblod4896 4 года назад

Great subject, thank you very much. Stay healthy Paul.

@jacopocanova7583 4 года назад

can't get 510mv out of the formula. 1=0,0047f x (dV/(1/120) ??

@NickBR57 3 года назад

Looks like nonsense maths to me. Seems to me you get 1.76 volt ripple from this

@zachdavis5834 2 года назад

Yeah totally agreed. The math on this video is absolute garbage. And assuming “1 over the freq”?? Why?? For a current value? Makes absolutely no fucking sense

@d614gakadoug9 8 месяцев назад

@@zachdavis5834 If you use the entire half-cycle period (8.3 ms at 60 Hz) you are assuming that the time the cap spends being charged from the transformer is zero. That overestimates the amount of capacitance required for a particular ripple voltage. The discharge from the peak voltage is constant current because of the regulator. If you have 5 volts out at 1 ampere, the current into the regulator is also 1 ampere and would be taken to be constant for purposes of calculation. As the capacitance is made larger and larger the waveform becomes more and more like a sawtooth with rapid rise and linear fall. How long the charging current flows is difficult to estimate accurately because it depends a lot on the source impedance (transformer). His 510 mV is nonsense.

@Taylor_26GE93 2 года назад

The Diode on the first schematic (0V to -15) is the wrong way around. 0V is 15V greater than the -15, would forward bias the diode etc...

@AdamTheAd-vanc3d Месяц назад

Yeah thanks for pointing that out nearly caught me, perked so both should be pointing in the dame direction , yes Anode out.

@AdamTheAd-vanc3d 4 года назад

Thanks for the video. Very useful and demystifying to limited knowledge on power supplies. Any chance you can do a follow up vid (if you haven't already covered this), how to select the correct caps for voltage regulators based on say the different voltage you may require. ( Did that make sense lol). Thanks again.

@willyrivero470 4 года назад

I'll stick with your approximation. It was the bug of my troubles and never quite used that formula because of that. Finding the right filter can be tricky. Don't use too much, and don't use too less. Thanks for the video. I think I'll watch it a few times. Take care. Thumbs up.

@opalprestonshirley1700 4 года назад

Always enjoy the math. Stay safe. Thanks Paul.

@Jose-tw9bl 2 года назад

super intresting! thank you! if i could give some ideas for future videos, i would also be interested in choosing capacitors for reactive power compensation, on the primary of a transformer😁

@calpardo 3 года назад

you deserve more than a like :) (Just subscribed)

@dalenassar9152 2 месяца назад

Since we are dealing with relatively large values here, the differential equation "I = C dV/dT" may be written as IT=CV. Rearranging: I/V = C/T and plugging in V= 1 volt and I=1 AMP, we get ONE volt per amp = C/T. Plugging in (1/120)th of a second for 'T' gives C=8333uF. This tells us: THE RIPPLE (UNDER LOAD) IS REDUCED BY ONE VOLT PER AMP FOR EVERY 8,333uF OF ADDITIONAL FILTER CAPACITANCE. And, everything follows this 'proportion'. For example, (1) adding only HALF of 8333uF or 4166uF; you get a 2v drop per amp. (2) using a half wave rectifier (a single diode) will require twice as much 'C' for the same droop as a FWB.....etc, etc... Hope this is useful. --daLe

@michaelpadovani9566 4 года назад

This brings back memories when my dad told me as a kid that rectified AC is not a straight line (this was shocking to me as I was confused how electronics could tolerate such a thing). Anyway this is actually a complicated subject, may want to do a follow up video. Different types of capacitors have different amount of ripple even if the capacitance value is the same. ESR and especially ESL come into play b/c an ideal capacitor does not exist in real life.

@ku4nz 4 года назад

and I always thought you just used what you had in the pile...just kidding..good job as usual.

@jeshy1958 4 года назад

I am immune compromised so I am in seclusion for the time being. Your heart problems may put you in the same situation. Stay far away from people. It may sound harsh but no one knows the incubation period for this bug!!!!! Your telling us to be safe, YOU be safe!!!!!!! And by the way, could you explain the formula a little more in depth, I didn't quite get it! Good luck and be safe!!!

@pulesjet 4 года назад

The Flu is killing 100 time more people every day. This corona is just a not so minor wave.

@trelligan42 4 года назад

The formula is actually calculus, so the explanation/derivation is 'beyond the scope'. (Man, I'm coming up with *lots* of puns today.)

@zachdavis5834 2 года назад

Hey man. Really confused how you just got from current in a cap formula to the 1/120 Hz. First thing the units don’t match…one is in seconds, other is volts per second. Would love some clarification on that lol.

@AnalogDude_ Месяц назад

When do you know you need to use inductors in psu?

@keithking1985 4 года назад

just read Deb Y comment there Paul, look after your self until this madness has passed. we need ya here..

@pulesjet 4 года назад

I always used the brute force method. Just keep adding uF until I get what I needed.

@trelligan42 4 года назад

Also known as 'cut and try', a well regarded engineering technique.

@MrWasim100100 4 года назад

nice video

@lobdsk 4 года назад

So how do I order one?

@viswanathancr349 3 года назад

good videog

@tolerbearALTII 4 года назад

why aren't other capacitors used in a power supply other than the electrolytic capacitor? Thanks.

@peternuninba1941 2 года назад

I have capacitors connected on a board all is sixteen in numbers, how can I use it to store electricity

@chickenfoundation9323 Год назад

So is a filter cap just a regular cap with very high capacitance or is it specifically made for filtering

@d614gakadoug9 8 месяцев назад

It depends on the lifetime you want. Electrolytic capacitors have a "wear-out" mechanism. A good capacitor for this would be specified for high ripple current at 120 Hz. That goes with low equivalent series resistance (ESR). Where it becomes hard to compare is with lifetime and temperature specifications. It used to be common for run of the mill electrolytic caps to be rated for 1000 hours at 85 °C at the rated ripple current (and often would be no ripple current or ESR spec published). Now we have lots of caps rated at 105 °C and with lifetime ratings of 2000 to 10 000 hours. For each 10 °C reduction in operating temperature you approximately double the life of the cap. If the ripple current is reduced the life will increase. Low frequency ripple current rating is often much less than the high frequency ripple current rating for the same part. It's all rather messy. Major manufacturers will specify their parts well and often include graphs that help. At one time large capacitors with screw terminals could be had in "computer grade." These are pretty much nonexistent now because computers all use switchmode power supplies.

@mikeoliver3254 4 года назад

Math ugh, just playing great video Paul.

@JohnnyG1956 4 года назад

I've often wondered the same thing. I have a question about the waveform of the rectifier. If the rectifier is just removing the negative part of the wave, wouldn't the time spent at 0 volts be one half of a wave? In my mind, I am picturing a halfwave hump above 0, a halfwave space and another halfwave hump above 0. This leads me to another question. If you were purchasing just one scope, would you get a 2 channel or a 4 channel? I will use the scope to troubleshoot electronic devices of all kinds and to design breadboard circuits for self-education.

@ron.owensby 4 года назад

The four diodes configured that way is a "Full Wave Rectifier" if only two diodes were used it would be a "Half Wave Rectifier" and would appear to the O'scope as you described. The full Wave actually flips the negative half of the sine wave as he describes in the video. I would think a two channel scope would be sufficient for your needs, but I don;t know your needs or how much you want to spend.

@trelligan42 4 года назад

In the various (and varied) shops I worked in, we often had a 4-channel scope. It was always on the shelf because we almost never needed the capability. In a few cases, I did breadboard a switch so I could alternate between test points.

@bjl1000 4 года назад

Funny. I always precede my caps with a resistor.

@Lann91 3 года назад

... and the search continous for the "how to decide the right value for a capacitor"

@modelking51 4 года назад

Can you run just house current into a power supply with out a transformer

@trelligan42 4 года назад

Yes, but then you'll have a potential potential (possible voltage :) between the case of your device and that wet floor you're standing on barefoot, or that fridge you grab a soda out of. A life is worth a *_lot_* more than the weight and cost of a transformer. There is a design criterion called 'double insulation'. Basically, a transformerless device would have a grounded metal case, then plastic outside that so you don't come into contact with a hot or neutral, even if insulation failed somewhere. But they're always making things foolproof - then finding a bigger fool. So if you put a butter knife inside a toaster, unplug the darn thing first.

@gfodale 4 года назад

The transformer is part of the power supply. In this case a step down transformer, meaning the secondary (output) puts out lower voltage than the primary (input). Otherwise you would be playing with DC voltages of 120 or more volts depending on your region.

@vs68s39 4 года назад

Max DC input for 78xx 79xx is 30/40VDC...Use

@curtstacy779 4 года назад

That is what the transformer does it was just not shown in the drawing. it would be to the left of this circuit. the transformer will step down the voltage to the level you need. remember you will have a drop in voltage after the rectifier because of the diodes also, about .6 to 1.2 volts. refer to the datasheet of the diode used for that number.

@vs68s39 4 года назад

@@curtstacy779 Tx

@ozzymandius666 4 года назад

Not a bad approximation, better than the fundamental principle of engineering: sin(x) = x.

@trelligan42 4 года назад

...for 'small enough' x.

@user-ww2lc1yo9c Год назад

Something here does not make sense, You wrote 5V, what purpose does that serve? Did you calculate C using formula, then why is it not on the left hand side? What exactly does 1A mean since when capacitor discharges, the current does not have a constant rate of change, it is an exponential decay. There are so many things that you did not clarify in this video.

@d614gakadoug9 8 месяцев назад

I am at a loss to figure out how he applied that equation and came up with 510 mV of ripple with a 1 ampere load and a 4700 µF capacitor. I get 1.77 V of ripple. In any case, using that method overestimates the capacitance required. It assumes the capacitor is continuously being discharged for the entire half cycle at constant current. The constant current part is OK - the regulator fixes the voltage across the load, thereby fixing the current. The larger the capacitor the more accurate the calculation becomes because the time during which charging of the cap is occurring is reduced (since charging only occurs when the instantaneous voltage out of the rectifier is greater than the cap voltage). He said that V in dV/dt is RMS. That is nonsense. dV/dt means "the differential of voltage with respect to time" or commonly "rate of change of voltage."

@abeditani8293 4 года назад

Help me about this matter . I have transformer 33v AC 50hrz 20amp . I put bridge 30 amp with capacitor 10mF 50v . But in some how the voltage get up to 52v because the capacitors detected the top voltage and it not stable . In 20 amp load the voltage get down to 30v and the current get up to 30amp in AC section !! There too much losses and too much heat and too much extra current in the ac section more than in DC section

@abeditani8293 4 года назад

@@andymouse I was looking in Google for formal between ac and DC power . I found that ( Vac÷74×100=Vdc ) I make some tests rectifier ac to DC with capacitor but with out loads and it was correct 100%

@rodney1297 2 года назад

Your second diode is backwards

@cory8791 2 года назад

Can you have to big of capacitors ?

@learnelectronics 2 года назад

Not really

@cory8791 2 года назад

@@learnelectronics what's your thoughts about converting a battery charger into a power supply for my cb amps would I just need to build a filter to clean up the ripping?

@d614gakadoug9 8 месяцев назад

Increasing the capacitance will increase the ratio of RMS current from the transformer to DC output current. Loss in the transformer windings due to their resistance is proportional to the square of the RMS current. Transformers ratings are for RMS currents because of this. Exact values are hard to calculate for real transformers. Typically you might get something like 0.6 ampes DC for each RMS ampere (i.e. if you need 1 A DC out you would need a transformer with a secondary rating of 1.67 A). With more filter capacitance you might need a transformer with a somewhat higher RMS current rating, perhaps 1.8 A RMS rating for 1 A DC output. Reputable transformer manufactures will usually have some ap notes for guidance. With really big caps you must be careful with "inrush" current when power is first applied. The current can be very high and last for several AC cycles. Slow-blow fuses are required and you might use diodes with higher current rating than you need for steady-state just so they'll survive the turn-on surge current (diode datasheets often include surge current ratings for this type of circuit). The transformer won't care - they can take large transient overcurrents as long as they don't last too long or happen too frequently.l

@bluegizmo1983 4 года назад

But how do we calculate the maths in a universe where 2+2 = 5???

@magnylizer 4 года назад

You just add 1 to every sum ;)

@davidluther3408 4 года назад

I HAVE EXTREME BACK PAIN RIGHT NOW,SPENT MOST OF WEEKEND IN BED.I DON"T SEE WHERE YOU GET VOLTAGE 510mv NUMBER?OTHER WISE MATH IS EASY TO FOLLOW

@babybear4862 4 года назад

dont see volt in or out to do any math.

@davidluther3408 4 года назад

HONEY PIE

@melplishka5978 3 года назад

Can you do more math plz bud. Ty for your time my friend.

@petekelly61 4 года назад

I use 6800 micro on my supplies. Too much?

@johnbravo7542 4 года назад

It is if your building a tube amplifier

@punjsanjeev 3 года назад

You make people confused.

@hosseinebrahimi3451 4 года назад

Thanks for video About the covid-19, in Iran about 1 month ago we would make memes and laugh about this virus. Until it riched on our doorsteps. now those memes circulate again but nobody laughs. I want to tell you what miracles I've seen in recent days. In Iran an islamic state, mosques distribute alcohol. I had a copy of 2012 movie in my library to show it to my grandchildren and brag about how i survived that day. But the videos about mosques distributing alcohol is now way more epic than 2012 movie. Today alcohol mafia can call the police and rat a bunch of people who are secretly say prayers in a mosque. This is ultimate "sci-fi came to reality" situation i can imagine.

@Jimwill01 4 года назад

Just remember, the corona virus and corona discharge are two unrelated things that you don't want in your body!

@antigen4 3 года назад

Jesus guy .ll those are. NOT your filter caps ...!! They are your reservoir caps l.. the filter caps are on the OTHER end of your circuit. Painful to watch

@peternuninba1941 2 года назад

I have capacitors connected on a board all is sixteen in numbers, how can I use it to store electricity