simple low pass RC filter noob question

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Standard low pass filter:
lpf.gif



Can you use more than one capacitor in parallel in the above circuit? (to give you more capacitance)

I need 10,000uF for the filter. Instead of using one big cap, can I use 3*3300uF caps?

Will the filter work the same? Will it be better/worse? etc


Thanks.
 
There is no harm in paralleling caps... it will often result in a lower ESR & ESL than would be achieved in an equivalent single cap... it's actually a good thing.

Are you laying these on a board or using point to point wiring?

I ask because there is an important detail often overlooked when big caps are wired in parallel.

:)
 
I have attached a very poor drawing. This is how you wire the caps... PCB or P to P.

The drawings on the right are the WRONG, but typical, way to do it. The problem is the cap marked "A" will see more of the ripple current and do more than its share of the work.

The idea behind the methods on the left is that each cap sees the same length of connecting wire. There is inductance and resistance in the wiring which is on a par with the ESL & ESR of the caps.

Does this makes sense? Explained well enough?

:)
 

Attachments

  • zee wiring.gif
    zee wiring.gif
    11 KB · Views: 279
Hey Andrew,

It is a fine point at 120 Hz, but it is good practice. Just sketch out the current waveform in a PSU cap... there are lots of higher harmonics present. The current is what causes all the heating and wear, so it makes sense to spread it out evenly.

I build high power inverters (100 kW) with huge banks of caps, at these levels, "zee wiring" is simply a must. I have to use laminated copper sheet for low L busbars... they are a real hassle, but they look cool!

You do this with batteries as well. the small voltage differences have a huge impact on charging efficiency and quality.

:)
 
I have one more question.. if the low pass filter was used in a power supply, would the capacitor only be involved in filtering or would it also be able to supply power when it is needed?

For example, if the resistor was not there and you just had the capacitor smoothing out the DC, the capacitor would also be able to provide a large amount of current on demand. But when you put the resistor in to make the filter, does the capacitor loose this dual funtion and only be involved with filtering?


Does that make sence??
 
Hi,
your RC filter does both jobs.

With or without the R the cap will try to keep the voltage at the same level until it's charge changes.

i.e. when the current demand increases the input voltage falls, but as soon as the voltage has fallen below the voltage level of the cap, the cap starts to discharge into the increased current demand.
Similarly if the current demand reduces then the input voltage will increase and the capacitor will absorb current (charge) until it's stored voltage matches the new voltage supply.
 
Simple RC networks like this a extremely easy to simulate using spice. You might consider downloading SwitcherCAD from Linear Technology (FREE!) and learn how to use it. You will be able to answer all your own questions in a matter of seconds, in far greater detail than you can imagine.

Example:
 

Attachments

  • rc1.jpg
    rc1.jpg
    60.7 KB · Views: 118
mr.duck said:
I have one more question.. if the low pass filter was used in a power supply, would the capacitor only be involved in filtering or would it also be able to supply power when it is needed?

For example, if the resistor was not there and you just had the capacitor smoothing out the DC, the capacitor would also be able to provide a large amount of current on demand. But when you put the resistor in to make the filter, does the capacitor loose this dual funtion and only be involved with filtering?

The cap can supply current far beyond the regulated supply's nominal capability. I recently made a stroboscopic illuminator for a microscope that uses a 5W LED. I am driving the LED with 12A pulses even though the power supply in the strobe is a 15V, 5A regulated supply. I simply added a 10,000 uF cap on the supply rail at the FET that switches the current through the LED. This works fine for this application because the duty cycle is kept below 1/15. The high current discharge is brief (variable from 5-250 us) and the recharge time relatively long, so I can flash the LED up to about 250 Hz and still maintain the pulses at 12A.

I_F
 
thread dig rather than starting a new thread as OP has posted up the LP filter type i'm planning on using on a rough project.

I've been calculating values required for the ballpark cutoff frequency I'm after and experimenting with the handful of capacitors I have on hand (don't want to use inductors as they are "expensive" relative to what i'm building) which has taught me enough to know what value caps to order.

here are my thoughts and some questions:

Using a Higher value resistor means you can use a lower value capacitor, but this is at the cost of attenuation

Using a low value resistor means you need a higher value capacitor (not really a problem in my case) without much attenuation of the desired low frequency noises.

Will using too low a resistance (I tried 0.5 ohms, then omitting the resistor altogether) present an unsafe (low impedance) load to the amplifier?

It seems to me that the obvious solution to this issue would be to put the resistor between the capacitor and "ground", but there must be a reason not to do this.
(EDIT: I've just drawn this out and the answer is pretty simple, with a resistor in series with the capacitor the filter won't "block" high frequency signals, just attenuate them as there will be a voltage divider between the additional resistor and the loudspeaker)

If the low load isn't a problem (I think it probably would be) then I'm surprised this method isn't more common, one source suggested you can get nth order filters by cascading filters, but almost all passive low pass filters I've seen use inductors or inductors and capacitors.

I think I'll just learn by doing, the amplifier is inside an ebay $30 headunit, and I don't like it much anyway.

sorry if this is poorly worded, I did study this stuff 15 years ago, will try to add some diagrams to better explain what i'm getting at.
 
Last edited:
I'm not that knowledgeable but I'll have a go at answering.

Capacitors would be like a short circuit to higher frequency signals, so a really big cap and low R would be a problem for an amplifier I'm guessing.

Cascading filters doesn't change the cut-off freq, only the steepness of the filter. So unless you want to channel out a specific frequency portion, you don't need this. I don't know if there is some other undesirable problem around the cut-off freq like a phase shift too which is made worse by cascading filters. Or simply you don't get the expected results with simple passive filters, where the output impedance of one affects the performance of the next.
 
I have attached a very poor drawing. This is how you wire the caps... PCB or P to P.

The drawings on the right are the WRONG, but typical, way to do it. The problem is the cap marked "A" will see more of the ripple current and do more than its share of the work.

The idea behind the methods on the left is that each cap sees the same length of connecting wire. There is inductance and resistance in the wiring which is on a par with the ESL & ESR of the caps.

Does this makes sense? Explained well enough?

:)

Explained well, but incorrectly drawn. If lead length is critical, then both top and bottom leads should be equal length before terminating at the "first" cap. As drawn, the leads are equal between caps, and it's good that two caps share being "first" in line, but the upper lead is drawn much longer than the bottom lead.

Or am I not seeing this correctly?

Peace,
Tom E
 
I'm not that knowledgeable but I'll have a go at answering.

Capacitors would be like a short circuit to higher frequency signals, so a really big cap and low R would be a problem for an amplifier I'm guessing.

yeah that's what I figured. I ended up just series wiring the two inductors in some OTS crossovers I had. I figured that the inductors will knock out a fair bit of the signal power so the capacitors won't have to sink as much. my stupid woofer boxes have a really strong resonant peak at "annoying" Hz, so I need to fix that before worrying too much about the filter. might throw the caps on just to see what happens anyway.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.