I think by "regular" power supply you mean without snubber, so here goes. A "regular" power supply consists of rectifier bridge(s) and smoothing/reservoir capacitors. A snubber is (usually?) a resistor in series with a small capacitor, parallel with the larger smoothing/reservoir capacitors. My research indicates that snubbers are used mainly with larger capacitors to reduce capacitor inductance and/or reduce diode ringing (high frequency noise that larger capacitors don't smooth as well).
You can use any kind of power supply you want with a chip amp, provided it's got the right voltage and current for your particular project. Voltage is determined by a combination of the chip you are using and the speaker impedance you want to drive. Current is determined by the power output at the specified voltage.
Certain types of power supply have better characteristics for audio applications. For example, most people here seem to prefer either unregulated (simple) or regulated (not as simple). Other options include switch mode power supplies (SMPS).
You can use any kind of power supply you want with a chip amp, provided it's got the right voltage and current for your particular project. Voltage is determined by a combination of the chip you are using and the speaker impedance you want to drive. Current is determined by the power output at the specified voltage.
Certain types of power supply have better characteristics for audio applications. For example, most people here seem to prefer either unregulated (simple) or regulated (not as simple). Other options include switch mode power supplies (SMPS).
Snubberized PS using 35V-CT-35V Transformer
I've heard the "power" of snubberized PSU on chip-amp, can it be implemented for MOSFET amps?
Can anyone make a calculation to make snubberized PSU using
35V-CT-35V transformer to get +50VDC/0/-50VDC ?
The big elco would be 2x (12000uF/63V) on each rail of PSU.
Thanks.
I've heard the "power" of snubberized PSU on chip-amp, can it be implemented for MOSFET amps?
Can anyone make a calculation to make snubberized PSU using
35V-CT-35V transformer to get +50VDC/0/-50VDC ?
The big elco would be 2x (12000uF/63V) on each rail of PSU.
Thanks.
While snubbers have little to do with power, they can be implemented almost anywhere across complex impedances.
The standard snubber for rectifiers and big capacitors consists of a 100 nF capacitor, where an additional resistor to set the filter quality and/or limit the current through the capacitor is omitted, or a varistor.
If you want to calculate the "optimum" snubber, you have to
The alternative is to determine the correct snubber values empirically, i. e. by Trial-&-Error like CarlosFM did with this result. Searching the forum for "CarlosFM" or "snubberized power supply" or both will turn up different versions and many threads with theories and attempts at explanations.
The standard snubber for rectifiers and big capacitors consists of a 100 nF capacitor, where an additional resistor to set the filter quality and/or limit the current through the capacitor is omitted, or a varistor.
If you want to calculate the "optimum" snubber, you have to
- - measure your PSU's resistance, capacitance and inductance.
- calculate the corresponding resonant frequency.
- calculate a snubber as if it were a notch filter for that frequency.
The alternative is to determine the correct snubber values empirically, i. e. by Trial-&-Error like CarlosFM did with this result. Searching the forum for "CarlosFM" or "snubberized power supply" or both will turn up different versions and many threads with theories and attempts at explanations.
wicho said:
If you ask the folks in the switch mode power supply arena, the answer would be that a snubber is a device which neutralizes the parasitic energy created by the capacitance of the power supply diodes and leakage inductance of the transformer. If you have a pretty good oscilloscope you can see the a high frequency ringing as the diode switches.
You can look up the diode's junction capacitance on a manufacturer's data-sheet -- it will vary with the voltage applied. The leakage inductance you have to measure with an impedance bridge or LCR meter.
The CarlosFM snubber was a combination of R and C in series/parallel with the filter capacitor. It acted to keep the impedance of the smoothing network more constant.
My interpretation of "snubber" is an impedance correction network that reduces high frequency ringing and tendency to overshoot on transients, particularly when switching is involved, eg. in a rectifier. It does this by using carefully calculated values of resistance and reactance to reduce the "Q" of another reactive circuit.
A simple film cap across a power supply is generally not a snubber, it is a "bypass" that reduces the overall impedance at high frequencies, so that an amplifier it feeds is more stable. In theory it can be a snubber, but you need to know the effective series resistance and inductance of the snubber cap being used to be able to calculate the effect of snubbing.
A snubber across a power supply or a rectifier diode is usually a resisitor and capacitor in series. It tends to counteract the impedance of the main power supply electrolytic capacitors or the diode switching, which have effective inductance and resistance at high frequencies.
This is not to be confused with a "Zobel" network on the output of a power amplifier. This basically provides a load to the amplifier at high frequencies so that it is more stable.
I think its important that ppl understand the difference between snubber, bypass and Zobel, and that they are used in the correct context. I've seen many examples of misuse on this forum.
A simple film cap across a power supply is generally not a snubber, it is a "bypass" that reduces the overall impedance at high frequencies, so that an amplifier it feeds is more stable. In theory it can be a snubber, but you need to know the effective series resistance and inductance of the snubber cap being used to be able to calculate the effect of snubbing.
A snubber across a power supply or a rectifier diode is usually a resisitor and capacitor in series. It tends to counteract the impedance of the main power supply electrolytic capacitors or the diode switching, which have effective inductance and resistance at high frequencies.
This is not to be confused with a "Zobel" network on the output of a power amplifier. This basically provides a load to the amplifier at high frequencies so that it is more stable.
I think its important that ppl understand the difference between snubber, bypass and Zobel, and that they are used in the correct context. I've seen many examples of misuse on this forum.
A snubber can be anything as long as it snubs disturbance signals. It can be an RC, a varistor, one or more diodes, etc.
It is a romantic notion that engineers are all day long carefully calculating optimum values. In reality they have to find a compromise between sufficiently satisfying function and the necessary investment of time and money. So instead of measuring a power supply's complex impedance to later on calculate the best capacitor and resistor value, they rather apply a low-pass filter that snubs up anything above a certain frequency. 100 nF almost always do the trick.
In some cases not. E. g. if an inductive component like a transformer is present. Then a bandpass is created between the transformer and the capacitor, and you can change the capacitor value to find the best mid frequency. Then add a resistor to adjust its bandwidth, which of course influences the Q of the total circuit. But even then, most designers would use standard values that have proven efficient in the past rather than spending too much time and effort with measuring and calculating for a secondary issue like transients above the human hearing range. High-priced high-end or DIY excluded of course, because different economic considerations apply.
It is a romantic notion that engineers are all day long carefully calculating optimum values. In reality they have to find a compromise between sufficiently satisfying function and the necessary investment of time and money. So instead of measuring a power supply's complex impedance to later on calculate the best capacitor and resistor value, they rather apply a low-pass filter that snubs up anything above a certain frequency. 100 nF almost always do the trick.
In some cases not. E. g. if an inductive component like a transformer is present. Then a bandpass is created between the transformer and the capacitor, and you can change the capacitor value to find the best mid frequency. Then add a resistor to adjust its bandwidth, which of course influences the Q of the total circuit. But even then, most designers would use standard values that have proven efficient in the past rather than spending too much time and effort with measuring and calculating for a secondary issue like transients above the human hearing range. High-priced high-end or DIY excluded of course, because different economic considerations apply.
A possible option is to first use inexpensive (non-botique) diodes and then snub them. You can do that by 1 cap per each diode from this chart: http://www.diyaudio.com/forums/showthread.php?postid=1588619#post1588619
See also Mark Houston's implementation of KBPC2504 plus 4 little caps (that's still 1 per each diode).
I think that you "may" realize a savings on heatsink expense. If so, that comes from the amplifier doing less unnecessary work.
In my opinion, an average temperature drop (at the amplifier's heatsink) without a hinderance of frequency response or output power (of the amplifier) is a certain indicator of good power supply design.
In this example, a modification to the power supply (snub diodes) has caused a problem (heat) to become reduced/absent. Its that simple (I hope).
For my own use, I've found the results to be very pleasant indeed.
See also Mark Houston's implementation of KBPC2504 plus 4 little caps (that's still 1 per each diode).
I think that you "may" realize a savings on heatsink expense. If so, that comes from the amplifier doing less unnecessary work.
In my opinion, an average temperature drop (at the amplifier's heatsink) without a hinderance of frequency response or output power (of the amplifier) is a certain indicator of good power supply design.
In this example, a modification to the power supply (snub diodes) has caused a problem (heat) to become reduced/absent. Its that simple (I hope).
For my own use, I've found the results to be very pleasant indeed.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.