I plan on building a couple of Power Regulators using the trusty old 78XX Voltage Regulator. The voltages for each regulator will be 5 and 12V Output. The amperage required for each project will be a fraction of 1 Amp. The power supply will be a linear power regulator of 9 and 15 Volts (for the 5 and 12 Volt Regulators respectfully).
For filtering caps, placed after the diode rectification, how much filtering capacitance will I need for a 5 Volt DC Regulator Output based on a 9 Volt Secondary AC Output, and for a 12 Volt DC Regulator Output based on a 15 Volt Secondary AC Output to eliminate ripple and supply voltage to the Regulator Input? I understand that too little capacitance will not smooth out the DC voltage during the AC Cycle. I read on this forum that too much capacitance can also be undesired as well. I am not totally rehearsed as why too much capacitance can have an ill effect - perhaps on the diodes or the transformer itself? However, from the Youtube videos I have seen regarding voltage regulators, not many worry about this.
Should I also add capacitors between the voltage regulator output to the load input to keep voltage smooth?
My only tools are a DMM and an LCR Meter, so I don't have an oscilloscope to measure any type of ripple, or noise coming from a diode.
If there is a formula I can use to pick the right amount of capacitance for my voltage needs, what would it be?
Thanks everyone for your time!!!
Thanks everyone for your time.
For filtering caps, placed after the diode rectification, how much filtering capacitance will I need for a 5 Volt DC Regulator Output based on a 9 Volt Secondary AC Output, and for a 12 Volt DC Regulator Output based on a 15 Volt Secondary AC Output to eliminate ripple and supply voltage to the Regulator Input? I understand that too little capacitance will not smooth out the DC voltage during the AC Cycle. I read on this forum that too much capacitance can also be undesired as well. I am not totally rehearsed as why too much capacitance can have an ill effect - perhaps on the diodes or the transformer itself? However, from the Youtube videos I have seen regarding voltage regulators, not many worry about this.
Should I also add capacitors between the voltage regulator output to the load input to keep voltage smooth?
My only tools are a DMM and an LCR Meter, so I don't have an oscilloscope to measure any type of ripple, or noise coming from a diode.
If there is a formula I can use to pick the right amount of capacitance for my voltage needs, what would it be?
Thanks everyone for your time!!!
Thanks everyone for your time.
The most effective way to reduce the ripple pre-reg is to use CRC or CRCRC I think you will find that with a CRC you can get much better ripple reduction with less capacitance than with a single C. eg I would say 1000uF 3r3 1000uF would be better than a single 4700uF or even a 10,000uF.
This was some experimenting I did http://www.diyaudio.com/forums/blogs/wintermute/574-yarps-finally-some-progress.html The scope shows clearly what effect the different stages have. Note this was overkill, I ended up using two 4700 caps and the final 1000uF. But you could get almost as much reduction with 1000uF.
You need to take into account the voltage drop of the R's when sizing the transformer, but if it is low current this won't be too much of a problem.
Tony.
This was some experimenting I did http://www.diyaudio.com/forums/blogs/wintermute/574-yarps-finally-some-progress.html The scope shows clearly what effect the different stages have. Note this was overkill, I ended up using two 4700 caps and the final 1000uF. But you could get almost as much reduction with 1000uF.
You need to take into account the voltage drop of the R's when sizing the transformer, but if it is low current this won't be too much of a problem.
Tony.
Last edited:
Thanks much Wintermute/Tony.
Yes, I did read your posts earlier today which triggered this question.
Thanks much for the CRC/CRCRC suggestion, as I did read in the past that multiple small caps are better at ripple and filtering than just one large cap. I was thinking about using a total of around 10,000uF in caps, even though it may be overdoing it for my project. However, would overdoing it in the caps cause more harm than good for my small needs?
When it comes to the regulator supply voltage output, I was thinking of adding smoothing caps. Should I go this route as well?
And should I connect the regulator voltage out caps in serial fashion to the positive pin, or should it be connected in parallel between output and ground (polarized) to smooth it out.
This would be my first voltage regulator project. I am confident on gathering the parts and building it. I just want to make certain that how I do it is as effective as possible.
Thanks again!!!
Yes, I did read your posts earlier today which triggered this question.
Thanks much for the CRC/CRCRC suggestion, as I did read in the past that multiple small caps are better at ripple and filtering than just one large cap. I was thinking about using a total of around 10,000uF in caps, even though it may be overdoing it for my project. However, would overdoing it in the caps cause more harm than good for my small needs?
When it comes to the regulator supply voltage output, I was thinking of adding smoothing caps. Should I go this route as well?
And should I connect the regulator voltage out caps in serial fashion to the positive pin, or should it be connected in parallel between output and ground (polarized) to smooth it out.
This would be my first voltage regulator project. I am confident on gathering the parts and building it. I just want to make certain that how I do it is as effective as possible.
Thanks again!!!
Regarding the CRC on the input: the higher the R, the lower your ripple, at least at constant load. You want the R to be low enough to still deliver enough input voltage to the 78xx at maximum load current.
Example: 9V in, 5V out, let's assume a maximum load current of half an amp. The 78xx needs about 2V dropout, so minimum voltage at its input is 7V. 9V-7V = 2V, which is the maximum allowed voltage drop at the R. At 500mA this means R must be less than 2V/500mA=4R.
Ben
Example: 9V in, 5V out, let's assume a maximum load current of half an amp. The 78xx needs about 2V dropout, so minimum voltage at its input is 7V. 9V-7V = 2V, which is the maximum allowed voltage drop at the R. At 500mA this means R must be less than 2V/500mA=4R.
Ben
I've been told that using big caps can be a negative because they themselves can pick up Hum. I'm not sure on this though. I used 4700 3r3 4700 3r3 1000 on the input to my regulated supply. There is quite a voltage drop though due to the resistors (which gets bigger the more current you draw) so you need to be careful and calculate based on the max current draw you expect, to ensure your input voltage is high enough as FlexBen has said.
The other things that can be a problem with too much capacitance is inrush current (the transformer and diode bridge can be overloaded in the initial charging, or you can blow fuse) but 10,000 uF total shouldn't be an issue.
With regards to capacitance on the output, you need to be careful. If you don't have any the circuit can have problems. If you have too much the circuit can have problems. If it is too low an ESR then you will have problems.
Generally the datasheets recommend a cap between 10uF and 100uF as long as you don't use a super low ESR type you should be ok, The cap should be from positive to zero volts (or ground if a single rail supply). A cap will not pass DC so if you put it in series you will not get any output voltage 🙂
Have a read of the datasheets for the reg's. They tend to have layout notes and recommendations as to what to use where.
Tony.
The other things that can be a problem with too much capacitance is inrush current (the transformer and diode bridge can be overloaded in the initial charging, or you can blow fuse) but 10,000 uF total shouldn't be an issue.
With regards to capacitance on the output, you need to be careful. If you don't have any the circuit can have problems. If you have too much the circuit can have problems. If it is too low an ESR then you will have problems.
Generally the datasheets recommend a cap between 10uF and 100uF as long as you don't use a super low ESR type you should be ok, The cap should be from positive to zero volts (or ground if a single rail supply). A cap will not pass DC so if you put it in series you will not get any output voltage 🙂
Have a read of the datasheets for the reg's. They tend to have layout notes and recommendations as to what to use where.
Tony.
Last edited:
Large first capacitors decrease the conduction angle forcing high current flow for short periods within the cycle. The current decay waveform may promote transformer ringing and put stress on other components.
1mF per ampere of load is about the minimum.
Many will use 2mF/A
really low ripple is seen (on a scope) when 5mF/A is used.
Pre-amps may use as high as 10mF/A
BTW,
a single C is not a filter.
you need a preceding impedance to form a filter. The simplest is RC where only resistance is provided.
All single capacitor input PSU actually include a preceding R value, but this R value is quite low. The R comes from the Primary transformer resistance + Secondary transformer resistance + wiring resistance to the rectifiers + the rectifier dynamic resistance + wiring resistance to the smoothing capacitors. It all adds up to a few tens of milli-ohms, to maybe as much as a couple of ohms.
I show it by referring to rC instead of C and using rCRC instead of CRC.
This shows it all graphically. The circuitry to the left is the transformer, the bridge and the reservoir cap. To the right (the yellow box) is a 12 volt regulator. Just imagine that is the 7812. The load resistor set at 12 ohms is pulling one amp. The transformer is set to be a 15 v rms type.
This shows the ripple for a cap value of 1000uf (note the regulator output is being affected by the ripple falling below the regulators minimum input voltage).
Next we have a 2200uf cap. Now the regulator is fine because the input is always higher than the regulators minimum voltage requirement.
I've attached the sim file so you can play around and see what happens (click my signature line to see how to use it).
This shows the ripple for a cap value of 1000uf (note the regulator output is being affected by the ripple falling below the regulators minimum input voltage).
Next we have a 2200uf cap. Now the regulator is fine because the input is always higher than the regulators minimum voltage requirement.
I've attached the sim file so you can play around and see what happens (click my signature line to see how to use it).
Attachments
Thanks all for your comments.
If a bleeder resistor is required for each capacitor, is there a formula I should use when selecting a resistor for each capacitor? I plan to use an CRC, or a CRCC (3 caps)filter capacitor layout.
Do I place the bleeder resistor on the capacitor leads for each cap or do I place it parallel to the cap?
Thanks again for everyones responses.
If a bleeder resistor is required for each capacitor, is there a formula I should use when selecting a resistor for each capacitor? I plan to use an CRC, or a CRCC (3 caps)filter capacitor layout.
Do I place the bleeder resistor on the capacitor leads for each cap or do I place it parallel to the cap?
Thanks again for everyones responses.
The amplifier or other circuit does not need a bleeder resistor.
If there is no need to be satisfied, it follows there can be no model that needs simulating.
If there is no need to be satisfied, it follows there can be no model that needs simulating.
Low voltage circuits don't normally need bleed resistors. They are more for safety to discharge high voltage caps when power is removed.
Is that what you meant ? or did you mean the resistor values making up the CRCRC filter.
Is that what you meant ? or did you mean the resistor values making up the CRCRC filter.
Hi Mooly and Andrew.
Thanks for your responses. Mooly, yes, that is what I meant. But now that you bring it up, I plan to perform a CRCC filter cap of various sizes thinking it will perform better than a CRC (either same or different size caps). Would my thinking be correct that 3 smoothing caps are better than two?
If I plan to go CRCC, would I need the other resistor for better performance - CRCRC configuration? If so, what would be the formula for it?
Thanks everyone for your help.
Thanks for your responses. Mooly, yes, that is what I meant. But now that you bring it up, I plan to perform a CRCC filter cap of various sizes thinking it will perform better than a CRC (either same or different size caps). Would my thinking be correct that 3 smoothing caps are better than two?
If I plan to go CRCC, would I need the other resistor for better performance - CRCRC configuration? If so, what would be the formula for it?
Thanks everyone for your help.
I'm not so sure there is any specific formula for all this. You have to decide what the maximum allowable volt drop would be across the combined resistor network at maximum regulator output (so the reg doesn't drop out and so the resistors aren't dissipating to much unwanted heat) and also what your maximum wanted or allowable ripple voltage would be.
For 78xx regs with a max current of 1amp you could generalise and say a 1 ohm series R would only 'lose' 1 volt.
Here is our simulation again, this time with 2 ohms series resistance. Look at the ripple at each point. You have to decide the ripple values that are acceptable to your design brief and then design the filter to give those values.
For 78xx regs with a max current of 1amp you could generalise and say a 1 ohm series R would only 'lose' 1 volt.
Here is our simulation again, this time with 2 ohms series resistance. Look at the ripple at each point. You have to decide the ripple values that are acceptable to your design brief and then design the filter to give those values.
Attachments
Notice TWO effects.
1. the peak to peak value of the ripple decreases as one progresses through the RCRC
2. the curves become more rounded due to attenuation of the HF content of the ripple. The sawtooth with sharp corners has a very high proportion of HF in the basic 100Hz ripple.
The sawtooth with rounded corners has less HF content in proportion to the now smaller 100Hz ripple.
Now to the capacitors.
I would use 2 off 470uF for C1 and 2 off 1mF for C2.
The extra ripple current capability of the 2*470uF will help with keeping them cooler. It's the first C of a rCRC or rCLC PSU that suffers the highest ripple current.
The last C is split into 2 to give a lower output impedance than a single cheap/normal 2mF electrolytic. It's this last C that meets the majority of the transient demand of the load. This gets towards your
It's the same reasoning that led to me buying a box of 1000 off 1r0 600mW 1% metal film resistors. My latest output stage using 2drivers and 4 TT output devices has 29 off 1r0 resistors that cost a total of <30p (UK pence) for the 4 emitter resistors and 4 base resistors and the drivers' emitter damping resistor.
1. the peak to peak value of the ripple decreases as one progresses through the RCRC
2. the curves become more rounded due to attenuation of the HF content of the ripple. The sawtooth with sharp corners has a very high proportion of HF in the basic 100Hz ripple.
The sawtooth with rounded corners has less HF content in proportion to the now smaller 100Hz ripple.
Now to the capacitors.
I would use 2 off 470uF for C1 and 2 off 1mF for C2.
The extra ripple current capability of the 2*470uF will help with keeping them cooler. It's the first C of a rCRC or rCLC PSU that suffers the highest ripple current.
The last C is split into 2 to give a lower output impedance than a single cheap/normal 2mF electrolytic. It's this last C that meets the majority of the transient demand of the load. This gets towards your
or you could buy lots of small same value caps at a quantity discount. eg 100 off 470uF, will be much cheaper per unit, than buying 2 off 470uF and 4 off 1mF
It's the same reasoning that led to me buying a box of 1000 off 1r0 600mW 1% metal film resistors. My latest output stage using 2drivers and 4 TT output devices has 29 off 1r0 resistors that cost a total of <30p (UK pence) for the 4 emitter resistors and 4 base resistors and the drivers' emitter damping resistor.
Last edited:
Let's start a 'save the mains sine wave from decapitation' movement.
Replace the crcrc.... with small c, and just use the regulator to suppress ripple. If still concerned, and have the measurements skills to know what residuel ripple exists, then add a finishing regulator.
Ciao, Tim
Replace the crcrc.... with small c, and just use the regulator to suppress ripple. If still concerned, and have the measurements skills to know what residuel ripple exists, then add a finishing regulator.
Ciao, Tim
Hi...& to calculate the power value of that resistor?
2V(v.drop by R)×0.500A=1w ??
I know Ohm's law but i am kind of dumb 😛 🙂
Regards
Last edited:
> 2V(v.drop by R)×0.500A=1w ??
Yes.
When you _buy_ the resistor, double the calculated power dissipation and round up. A resistor running AT its rating runs HOT! And won't live forever.
For a "critical" furnace-fan controller I calculated 5 Watts and bought 25W resistors. I would not go that far for a hi-fi: if the hi-fi resistor fails my nose and toes won't freeze. But double at least, and check/see if even higher wattage is selling for about the same money.
Yes.
When you _buy_ the resistor, double the calculated power dissipation and round up. A resistor running AT its rating runs HOT! And won't live forever.
For a "critical" furnace-fan controller I calculated 5 Watts and bought 25W resistors. I would not go that far for a hi-fi: if the hi-fi resistor fails my nose and toes won't freeze. But double at least, and check/see if even higher wattage is selling for about the same money.
- Status
- Not open for further replies.