Dear All:
One question i have two 100uF capacitors, one has 16V and the other has 50V wrote on the package.
The capacitor with 50V gives the circuit more stability (filter) while the capacitor of 50v gives the circuit less stability (filter)
Does the capacitor voltage on the package affect the ability of it to filter more?
My circuit has 5Volts from an ATX power supply.
Thanks.
Alfredo Mendiola Loyola
Lima, Perú
Voltage rating, itself, has little to do with stability (that is, variation with temperature). Here is the story:
There are many ceramic dielectrics. They have different combinations of dielectric constant, voltage rating, and temperature coefficient.
Generally, for a GIVEN dielectric, volume increases with capacitance and increases with voltage rating. In fact, volume is proportional (generally) to C*V (for a given ceramic) where V is the voltage RATING.
Now, generally, the ceramics with good temperature coefficient have lower dielectric constant. So, here is what happens:
For a physical size, a higher voltage rating for a capacitor of a given value requires a higher dielectric constant material. That higher dielectric constant material generally has poorer temperature coefficient.
This would seem to answer the first part of your post except for the fact that I do not know what you mean by stability.
Further, in the last part of your message. I don't know, what you mean by "filter more?" More in what way?
Jim
Voltage rating, itself, has little to do with stability (that is, variation with temperature). Here is the story:There are many ceramic dielectrics. They have different combinations of dielectric constant, voltage rating, and temperature coefficient.
Generally, for a GIVEN dielectric, volume increases with capacitance and increases with voltage rating. In fact, volume is proportional (generally) to C*V (for a given ceramic) where V is the voltage RATING.
Now, generally, the ceramics with good temperature coefficient have lower dielectric constant. So, here is what happens:
For a physical size, a higher voltage rating for a capacitor of a given value requires a higher dielectric constant material. That higher dielectric constant material generally has poorer temperature coefficient.
This would seem to answer the first part of your post except for the fact that I do not know what you mean by stability.
Further, in the last part of your message. I don't know, what you mean by "filter more?" More in what way?
Jim
Dear Jim:
I'm filtering noise from my circuit because my ADC is unstable, i'm inserting capacitors of 220uF to reduce that noise.
Thanks.
Alfredo Mendiola Loyola
Lima, Perú
When you say your ADC is unstable, do you mean a few counts? You could be fighting this forever....
What is your reference voltage and how is is configured? Looks like the problem could be in your ADC setup.
The capacitor with 50V gives the circuit more stability (filter) while the capacitor of 50v gives the circuit less stability (filter)
Does this seem to contradict itself or what?
Capacitors can vary on their effects based on their construction. A 1uf tantalum behaves far differently than a 1uf electrolytic.
In your description you have not provided ANY information on your circuit, a/d converter, nor any outboard gear so how do you expect us to give a decent answer(although The MAIN Jim did)?
As far as the voltage ratings go, there are several theories behind that and the only way you are going to get your answer is by RTFDS on the components.
This smels of a university project. I will give you one little tip. The capacitors that you are quoting of using to try and stabilize an ADC of 220uF are USELESS. Look at your design layout first.
Jim
Quote:The capacitor with 50V gives the circuit more stability (filter) while the capacitor of 50v gives the circuit less stability (filter)
Does this seem to contradict itself or what?
You still have not said what you mean when you say that one capacitor produces "more stable" results than the other. Do you mean that the ADC is more stable?
20uF is WAY out of line for filtering an ADC. What does your filter circuit look like? What is the series impedance in your lowpass filter?
Let me make some guesses. I'd bet that you have a low impedance source (maybe an op-amp) and you are NOT using a series resistor. As a result, you need a huge capacitor in combination with the really small built-in source impedance to make a filter. If this is the case, add something like a 1K resistor to your signal source and use a more modest 1000pf or so for filtering.
Now, the REAL question is the source of the variation over time. Is it part of the signal that is being digitized or is it a problem with the ADC? Try connecting a pair of 10K resistors in series from the reference source pin to ground . Connect the junction of the two resistors to your ADC input. If it is still "unstable", then it is the ADC or the supporting hardware (including traces, power supply bypassing, etc). If it is NOT "unstable, then it is probably part of your incoming signal and you would be better served to clean that up rather than brute-force filter.
Jim
Try connecting a pair of 10K resistors in series from the reference source pin to ground . Connect the junction of the two resistors to your ADC input.
Higher voltage capacitors with same capacitance of the same family generally have lower ESR. A lower ESR can help reduce ripple voltage.
oddbudman
Higher voltage capacitors with same capacitance of the same family generally have lower ESR. A lower ESR can help reduce ripple voltage.oddbudman
oddbudman wrote:Higher voltage capacitors with same capacitance of the same family generally have lower ESR. A lower ESR can help reduce ripple voltage.oddbudman
True, but the difference is not likely to be noticeable between 15V and 50V caps. Maybe these are from different families, eg. one has a high ripple current rating.
Dear All:
My problem is related with a noise in the whole circuit, i have a circuit that has an atmega8 for
multiplex seven digits display (5 digits) and has an atme16 to manage ADC and digital inputs.
I'm using transistor NPN BC548 with common anode displays for mulplexing, the ADC works well until
i turn on the multiplexed circuit, when i shut down the multiplexed circuit the ADC becomes stable,
even just removing the npn transistor my adc becomes stable.
When i placed two capacitors of 220uF in parallel between 5V and GND close to the atmega16 the ADC
noise disappears
I think the multiplexing hardware is causing an electric noise, ripple i 'm not sure because i don't have
an oscilloscope, but if i add capacitors on the circuit those solve the ADC stability i guess it is related
with electric noise, ripple...
The ADc is set up with internal voltage 2.56v because with 5V the adc stability was low.
I'm using aref to connect my pots, avcc has a capacitor of 100nF.
the ADC input pin also has a 100nF capacitor.
My ATX power supply (360W) gives me 5.25volts, the circuit uses a 220uF capacitar and an
100nF capacitor between 5V and GND close to the power plug.
Everything was working well until i added the multiplexing circuit which consumes 300mA.
Please help me.
Alfredo Mendiola Loyola
Lima, Perú
OK, the situation you seem to describe is usually called "bypass" capacitors rather than "filter" (though it does filter). I think most of the responses, including mine, assumed that these caps were part of a filter at the ADC input.
Now, 220uf (actually 440uf when two are in parallel) for power supply bypass with the kind of currents you describe is quite reasonable. In fact, I would be very surprised if it works without the caps in this situation.
There are several ways that this issue can be dealt with. One way is to use on-board voltage regulators and use a separate regulator for the LEDs.
You also need to be paranoid about your ground circuit. Use separate grounds for the LED system and the processor + ADC + analog input. Connect these grounds only at the power supply. Ditto Vcc.
Jim
OK, the situation you seem to describe is usually called "bypass" capacitors rather than "filter" (though it does filter). I think most of the responses, including mine, assumed that these caps were part of a filter at the ADC input.Now, 220uf (actually 440uf when two are in parallel) for power supply bypass with the kind of currents you describe is quite reasonable. In fact, I would be very surprised if it works without the caps in this situation.
There are several ways that this issue can be dealt with. One way is to use on-board voltage regulators and use a separate regulator for the LEDs.
You also need to be paranoid about your ground circuit. Use separate grounds for the LED system and the processor + ADC + analog input. Connect these grounds only at the power supply. Ditto Vcc.
Jim
Dear Jim:
Can i use capacitor of 16V with 470uF without problems in my circuit? What is the upper limit in capacitance?
If i use inductors will it improve the circuit stability?
Is there a way to isolate the display part from the analog part of the circuit using electronic components. I don't understand well how to isolate the grounds.
I'm using a common cable that supply 5 volts to each circuit and a common ground to each circuit, like a bus.
Thanks,
Alfredo Mendiola Loyola
Lima, Perú
There is no real upper limit on bypass capacitance other than size, cost, and how long it take to charge up when you turn on power. In this application, inductors probably won't help much.
To separate grounds, run TWO ground wires from your supply to the device. All of the ground connections for the LED multiplexing circuit connect to each other and ONLY to the power supply ground. Ditto processor and and analog circuits. It MIGHT be necessary to connect these two ground systems together at one point in your device.
An second option might be a single ground wire from your supply. Two local grounds as described in the previous paragraph. Connect them together ONLY at the point where the power supply ground wire comes into the device.
Jim
There is no real upper limit on bypass capacitance other than size, cost, and how long it take to charge up when you turn on power. In this application, inductors probably won't help much.To separate grounds, run TWO ground wires from your supply to the device. All of the ground connections for the LED multiplexing circuit connect to each other and ONLY to the power supply ground. Ditto processor and and analog circuits. It MIGHT be necessary to connect these two ground systems together at one point in your device.
An second option might be a single ground wire from your supply. Two local grounds as described in the previous paragraph. Connect them together ONLY at the point where the power supply ground wire comes into the device.
Jim
Dear Jim:
I solved my problem using a low pass filter circuit
using a 100uF inductor, 100nF and a 10uF capacitor.
The low pass filter is between my Power supply and my ADC circuit.
It works well even without capacitors, but just in case i will use the capacitors...
Thanks.
Alfredo Mendiola Loyola
Lima, Perù
