Hi,
I'm going to build a power amplifier based on 3eAudio's TPA2355 board (2 channel 260W per channel class D amplifier).
I'd be grateful to hear people's views on what power supply would be best to use with it. 3eAudio offer a Quasi-Resonant Power Supply or alternatively I have built non-regulated power supplies before or I could look at a regulated power supply.
Any suggestions most appreciated.
I'm going to build a power amplifier based on 3eAudio's TPA2355 board (2 channel 260W per channel class D amplifier).
I'd be grateful to hear people's views on what power supply would be best to use with it. 3eAudio offer a Quasi-Resonant Power Supply or alternatively I have built non-regulated power supplies before or I could look at a regulated power supply.
Any suggestions most appreciated.
It also will depend in the audio sources. Switching are pretty good and small size, but for listening radio (AM or FM) will create lots of problems. Linear regulated are costly, bulky and have low efficiency. Non regulated, are a poor choice because of hum and regulation.
Thanks for the reply Osvaldo, I would be using a Thorens turntable and an Apple Airport Express as a sources with an old Realistic integrated amplifier as a pre amp (it has a preamp out).
Then, IMHO, a standard SPMS of slightly greater power handling than those of the amp, would be my choice.
I've been wondering about that too. Although SMPS seem adequate for this job, they are known to feel uncomfortable when the current consumption is not close to their max output. How a SMPS would behave when a class D amp is at idle i.e. much lower current demand?
It will depend on the SMPS topology and its design.
Most forward (1 and 2 transistors), half and full bridge, need to rectify on a choke input filter, to run properly. If the current at the output is too low, the inductor runs dry and then it "disappears" from the circuit, and changes the mode of working of the SMPS. This minimum usually is between 10 and 20% of the full rated output current. So, unless the SMPS has an internal load, the light load or no load of a SMPS is disliking for it. It is said that the SMPS cycles. The logic starts the power stage, then the outputs cap`s are charged to a level higher than the normal output voltage, as a consequence that the indcutor doesn't work. The overvoltage protection stops the logic, leaving the caps to discharge tough the leak circuitry (Voltage sense, power led/relay, etc). When the voltage goes under the trip voltage, the logic starts it again and so on. So the SMPS starts and stops running entire cycles. If the SMPS is not prepare to this mode of operation, the power stage may blow due to continuously starting/stopping at audiofrequency ranges. Some IC's are designed for this job, and is usually called Burst mode. The difference between the two cases is that the burst mode IC's start with the soft start function, while standard IC's don't.
In the other hand, a slight overload is tolerable, and a 5% of overload is safe for them.
Some Class D amps may cause "Bus bumps" at lower frequencies, which is a devolution of the RF energy at the filter of the class D amp to the supply bus(es) via the output's MOSFETs causing the voltage to rise with no control. This also can affect the SMPS job.
Most forward (1 and 2 transistors), half and full bridge, need to rectify on a choke input filter, to run properly. If the current at the output is too low, the inductor runs dry and then it "disappears" from the circuit, and changes the mode of working of the SMPS. This minimum usually is between 10 and 20% of the full rated output current. So, unless the SMPS has an internal load, the light load or no load of a SMPS is disliking for it. It is said that the SMPS cycles. The logic starts the power stage, then the outputs cap`s are charged to a level higher than the normal output voltage, as a consequence that the indcutor doesn't work. The overvoltage protection stops the logic, leaving the caps to discharge tough the leak circuitry (Voltage sense, power led/relay, etc). When the voltage goes under the trip voltage, the logic starts it again and so on. So the SMPS starts and stops running entire cycles. If the SMPS is not prepare to this mode of operation, the power stage may blow due to continuously starting/stopping at audiofrequency ranges. Some IC's are designed for this job, and is usually called Burst mode. The difference between the two cases is that the burst mode IC's start with the soft start function, while standard IC's don't.
In the other hand, a slight overload is tolerable, and a 5% of overload is safe for them.
Some Class D amps may cause "Bus bumps" at lower frequencies, which is a devolution of the RF energy at the filter of the class D amp to the supply bus(es) via the output's MOSFETs causing the voltage to rise with no control. This also can affect the SMPS job.
Last edited:
So, the phenomenon is not proportional to the current demand? Is there a threshold say at 20% where it completely stops happening?
This fenomena is proportional to load current, and stops between 10 and 20% unless the PSU has minimum internal loading, as most of them. PC ATX's for example, has problems in the unsensed voltage outputs when no load, the 12V may be, perhaps, as low as 8-9V without loading in the 5V out.
Isn't it true that this is not an issue if the amps are running BTL? This appears to be the case if the amp has 2 channels at 260w per channel.
It's true in this case and when a stereo amplifier has audio signals inverted. But in case of mono or normal stereo, bus overvoltages still occur.
Thanks for all your comments. So if I’ve understood correctly, the Quasi-Resonant Power Supply should be fine for the amp board I’m looking at as the Amp is BTL configured.
So my next question is would I hear and significant difference in sound quality between this and any other type of power supply?
So my next question is would I hear and significant difference in sound quality between this and any other type of power supply?
I can't be sure of this because I haven't musical ears nor musical knowledge, I'm only an Electronic Engineer. But surely, having the buses well regulated, the optimal performance of the amp is warranted from the electrical point of view, but some musicians find interesting the sag in the power lines, mainly in tube amplification. I am not in agree with this concept, technically is BAD.
It is true that bridge class D amps haven't bus bumps, as when one is draining current, the other is sinking it, thanks to the 180 electrical degrees phase inversion between them.
It is true that bridge class D amps haven't bus bumps, as when one is draining current, the other is sinking it, thanks to the 180 electrical degrees phase inversion between them.
Thanks for all your comments. So if I’ve understood correctly, the Quasi-Resonant Power Supply should be fine for the amp board I’m looking at as the Amp is BTL configured.
So my next question is would I hear and significant difference in sound quality between this and any other type of power supply?
So my next question is would I hear and significant difference in sound quality between this and any other type of power supply?
Almost any SMPS will be fine for BTL. But, paricularly the fly-back topology has the peculiarity of having the widest dynamic range, because it hasn't output inductor, classic PWM or resonant. This kind of SMPS would be my prefered, current mode still better, and with PFC still better.
Class-D amplifiers have proven, to me, a legitimate choice to try, after my experience with Breezy Audio TPA3116 audio small power amplifier.
The low price and the numerous praises got me curious, and I bought one, along with a Nobsound, apparently marketed (made?) by Douk Audio.
The DC-DC supply seemed ready made for the TPA3116, with its 19v/3.42Ah capability.
The only problem I had was with poor heatsink contact on the TPA chip, easily solved with grease.
Now I wonder if a sturdier, larger, better somehow DC-DC power supply might be in order, particularly one that would allow me higher voltages and other Class-D power amps.
Today I got two offers from Banggood:
gophert nps-1601 0-32v 0-5a 110v/220v 160w switching digital adjustable dc power supply Sale - Banggood.com
rd6006/rd6006-w led switching power supply s-400w-48v/dc12v/15v/24v/36v/60v 8.3a-33.3a support monitoring transformer lighting Sale - Banggood.com
Capacities are quite different, and the 400W option seems like an overkill. But overkill in power supplies was usually a good thing.
I wish there was some place that did test DC-DC supplies for noise and other issues, particularly when powering audio stuff.
Has anyone tried any of these options or was considering to?
Any other recommendation to look for?
Thanks!
The low price and the numerous praises got me curious, and I bought one, along with a Nobsound, apparently marketed (made?) by Douk Audio.
The DC-DC supply seemed ready made for the TPA3116, with its 19v/3.42Ah capability.
The only problem I had was with poor heatsink contact on the TPA chip, easily solved with grease.
Now I wonder if a sturdier, larger, better somehow DC-DC power supply might be in order, particularly one that would allow me higher voltages and other Class-D power amps.
Today I got two offers from Banggood:
gophert nps-1601 0-32v 0-5a 110v/220v 160w switching digital adjustable dc power supply Sale - Banggood.com
rd6006/rd6006-w led switching power supply s-400w-48v/dc12v/15v/24v/36v/60v 8.3a-33.3a support monitoring transformer lighting Sale - Banggood.com
Capacities are quite different, and the 400W option seems like an overkill. But overkill in power supplies was usually a good thing.
I wish there was some place that did test DC-DC supplies for noise and other issues, particularly when powering audio stuff.
Has anyone tried any of these options or was considering to?
Any other recommendation to look for?
Thanks!