Hey all,
I am just getting into designing audio amps and I've chosen to start with a class D amp. There are tons of great resources and products and I've decided to design around the TPA3251 to get started.
However, I'm not quite sure what direction to go with the power supplies so I was hoping to get some of your thoughts on the best way to go. Apologies if I missed a post that covered this.
If were to build a linear amp I'd probably go with an unregulated supply due to its simplicity, ease of design and that a well designed amp can deal with some sags or minor irregularities in an unregulated supply. I am expecting this wouldn't be true for a class D but please enlighten me if I'm wrong. Would a smps or even a regulated supply be better? Do you use off the shelf devices? Is there a consensus on the best method or psu for a class D amp? Obviously Id like to save costs but never at the expense of safety.
My design is aiming to be a stereo amp that can output 75W per channel into 8ohms.
Thanks in advance!
I am just getting into designing audio amps and I've chosen to start with a class D amp. There are tons of great resources and products and I've decided to design around the TPA3251 to get started.
However, I'm not quite sure what direction to go with the power supplies so I was hoping to get some of your thoughts on the best way to go. Apologies if I missed a post that covered this.
If were to build a linear amp I'd probably go with an unregulated supply due to its simplicity, ease of design and that a well designed amp can deal with some sags or minor irregularities in an unregulated supply. I am expecting this wouldn't be true for a class D but please enlighten me if I'm wrong. Would a smps or even a regulated supply be better? Do you use off the shelf devices? Is there a consensus on the best method or psu for a class D amp? Obviously Id like to save costs but never at the expense of safety.
My design is aiming to be a stereo amp that can output 75W per channel into 8ohms.
Thanks in advance!
I'd go for a switcher so long as there's a way to deal with the common-mode noise its going to generate. Like having trafos at the inputs (if single-ended) or you're using balanced interconnects. Then have a filter on its output to improve the ripple output (which is typically in the hundreds of mV for a SMPSU).
Otherwise go for a linear supply - use a regulator or cap-multiplier followed by some passive LC filtering.
Otherwise go for a linear supply - use a regulator or cap-multiplier followed by some passive LC filtering.
Off:
What the hell is a "switcher"? Does it mean SMPS? Or ClassD? Or what? And why creating ambiguous slang words when we already have clear terms? I can imagine only 1 reason: hiding insufficient knowledge about the subject. Correct me if I'm wrong!
On topic:
A ClassD amp can have high or low Power Supply Rejection Ratio depending on the topology and feedback parameters. No general rule. In case of a single ended, open loop design PSU must be as precise and linear as the requirement for the whole amp. High loop gain feedback can produce a very high PSRR at "low" freq, therefore less stability is required. For example UcD doesn't require stable PSU at all, even for very high performance.
But the above true only for audio frequencies. HF can call other considerations: some ClassD amps tend to be sensitive to HF noise, can interfere with it, and produce audible artifacts. In most cases this is due to less than perfect (or crappy) layout, but a newbe cannot be sure if he could avoid this or not. This means that SMPS and ClassD can be problematic in some cases.
General rule: the technology doesn't imply specific characteristics except for the obvious facts:
- switching generates some level of HF disturbance
- sampling limits bandwith
- high frequency allows smaller transformer in PSU.
In a particular case advice can be made, but globally, "What is best"?
- learning
- experimenting
I can mention a third aspect to consider: supply pumping. This is why ClassD amp always need some buffer capacitor on the rails.
What the hell is a "switcher"? Does it mean SMPS? Or ClassD? Or what? And why creating ambiguous slang words when we already have clear terms? I can imagine only 1 reason: hiding insufficient knowledge about the subject. Correct me if I'm wrong!
On topic:
A ClassD amp can have high or low Power Supply Rejection Ratio depending on the topology and feedback parameters. No general rule. In case of a single ended, open loop design PSU must be as precise and linear as the requirement for the whole amp. High loop gain feedback can produce a very high PSRR at "low" freq, therefore less stability is required. For example UcD doesn't require stable PSU at all, even for very high performance.
But the above true only for audio frequencies. HF can call other considerations: some ClassD amps tend to be sensitive to HF noise, can interfere with it, and produce audible artifacts. In most cases this is due to less than perfect (or crappy) layout, but a newbe cannot be sure if he could avoid this or not. This means that SMPS and ClassD can be problematic in some cases.
General rule: the technology doesn't imply specific characteristics except for the obvious facts:
- switching generates some level of HF disturbance
- sampling limits bandwith
- high frequency allows smaller transformer in PSU.
In a particular case advice can be made, but globally, "What is best"?
- learning
- experimenting
I can mention a third aspect to consider: supply pumping. This is why ClassD amp always need some buffer capacitor on the rails.
OK
"Switcher" is a shortened term and is used as a short form for any switch mode circuitry.
Just like scope or spec-A. (oscilloscope and spectrum analyzer)
It may apply to an amplifier or power supply.
It may take longer to type but is faster to say.
Hi, I have a very cheap 48V 10A SMPS (earthed metal case type with terminals) driving a pair of class D amps which in turn are connected to high quality PA type mids and very efficient compression drivers. 12" car type sub. Mid/high efficiency is high and ultimate volume pretty loud!! If there was any noise, it would be very noticeable.
I have no noise issues with this PSU even connected via hardwiring to a laptop on a separate mains supply.
Follow the threads for your chip regarding design errors and mods required to get the best performance. The PSU (unless really poor) is in my experience not going to give any trouble.
Double insulated laptop supplies can give noise trouble though when hardwired to other double insulated devices such as a laptop. Circulating currents picking up in the inputs. The solidly earthed 48V PSU does not exhibit this.
A vinyl source or AV equipment - I've found they definitely get noisy with a double insulated laptop supply for the amp. Worse than a laptop. I could have experimented with additional earth bonds or loop isolators, but Bluetooth did the same thing with a 10ft air gap
Earthed, terminal type PSU (despite being stupidly cheap) = no issues. 7p per Watt LOL.
I have no noise issues with this PSU even connected via hardwiring to a laptop on a separate mains supply.
Follow the threads for your chip regarding design errors and mods required to get the best performance. The PSU (unless really poor) is in my experience not going to give any trouble.
Double insulated laptop supplies can give noise trouble though when hardwired to other double insulated devices such as a laptop. Circulating currents picking up in the inputs. The solidly earthed 48V PSU does not exhibit this.
A vinyl source or AV equipment - I've found they definitely get noisy with a double insulated laptop supply for the amp. Worse than a laptop. I could have experimented with additional earth bonds or loop isolators, but Bluetooth did the same thing with a 10ft air gap
Earthed, terminal type PSU (despite being stupidly cheap) = no issues. 7p per Watt LOL.
Last edited:
Yes just my experience. The little old Packard Bell netbook I am using here has no connection to earth with it's standard supply. Same goes for my HP laptop.
The plug has 3 pins and so does the PSU, but there is no earth connection on the secondary side. Several hundreds of KOhms.
The 24V versions i bought from EBAY for the TPA3116 also have no secondary earth and caused issues. 48V aluminium cased 'industrial' style PSU is solidly earthed and has no problems with TAS5630 amps.
I imagine TWO earthed supplies (one in the source and one in the amp) would cause some mains hum and buzzing. Nothing earthed at all seems to cause high frequency noise as well as buzzing.
There are that many combinations of supplies and sources and I totally agree that there is no single rule and you need to try it and see.
Last edited:
Don't get that. Any laptop AC-Adapter with 2 prong plug must be double insulated, and the mains outlet delivers per definitionem a "low voltage" (have a look at the "low voltage directive"
Last edited:
The adapters I have generally have (UK) have 3 pins on the primary side and + - on the secondary, neither of which are connected to earth. Hence my description of double insulated.
Any metal part on the secondary is not connected to earth.
So the description of double insulated is correct.
The low voltage directive is something different and is related to potential to earth and protection of persons. I wasn't going to argue the point before your reply though.
Noisy audio...... Isn't related to the LV directive. EMC and equipotential bonding, yes. Ground or circulating currents yes.
Any metal part on the secondary is not connected to earth.
So the description of double insulated is correct.
The low voltage directive is something different and is related to potential to earth and protection of persons. I wasn't going to argue the point before your reply though.
Noisy audio...... Isn't related to the LV directive. EMC and equipotential bonding, yes. Ground or circulating currents yes.
Last edited:
Because a cheap, available SMPS does an excellent job for little outlay and no time or design input.
My thoughts are that if you sort the amplifier design (which invariably has faults on Chinese implementations) then the PSU is of secondary concern.
I'm amazed by what can give superb audio at low cost. I'm also upset that stupid errors in implementation can cause bad results. A few intelligent alterations and the results are great.
Last edited:
Sorry, I ment Safety Extra Low Voltage or Extra Low Voltage, but I don't have time and fast internet to investigate the exact limit. Laptops itselfs are not connected to mains, and not doubly isolated as you know. This is the point.
Dogshome wrote:
But laptops are not double isolated devices.
Being not connected to earth is absolutely not a satisfactory condition for being doubly isolated. This is a very dangerous misbelief, please don't spread it!
"Double isolation" is part of the low voltage directive. Not "something different".
OK
"Switcher" is a shortened term and is used as a short form for any switch mode circuitry.
Just like scope or spec-A. (oscilloscope and spectrum analyzer)
It may apply to an amplifier or power supply.
It may take longer to type but is faster to say.
Here "switcher" also is a somewhat derogative term for SMPS. That is why I like it
Like in : "Are you going to build a proper power supply or will it be a cheap switcher ?"
Last edited:
Here "switcher" also is a somewhat derogative term for SMPS. That is why I like it
Like in : "Are you going to build a proper power supply or will it be a cheap switcher ?"
Those that have a PE connection generally generate less EMI is what I observed. Possibly because the filter behaving better guiding pollution to PE ?! Results with open frame earthed SMPS were always better than external power adapters till now. Probably for ease of use there seems to have grown a silent standard with external power bricks and those silly round 5.5 mm connectors always getting plugged in at the wrong moment causing sparks. I do keep PE/earth separated from audio GND/amplifier GND so I mount the open frame SMPS on isolating standoffs (PE connected of course!) or mount the boards with no connection to PE (harder). A real 230V mains switch and I am sure that the device really is off when I don't use it. Having experienced nice effects of caps going out with a bang I prefer to have no adapters plugged in while I am away.
(disclaimer: as a total laymen not understanding SMPS, not seeing the apparent beauty of SMPS, not seeing that transformers are old news, turning a blind eye to efficiency etc. etc.).
(disclaimer: as a total laymen not understanding SMPS, not seeing the apparent beauty of SMPS, not seeing that transformers are old news, turning a blind eye to efficiency etc. etc.).
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.