You guys comparing PSUs...
Why not just get a decent quality power brick that won't blow up? Perhaps then use a differential filter on the output, and it'll sound as good as any other. You can't judge PSUs by the "sound quality", that's complete and utter bollocks.
A good PSU will have a clean output, and will be able to provide the current you need. End of...
Could you clarify why would a brick "blow up"?? Based on what you made this statement? We use the PSU for audio purpose, as long as it generate "nice sound", that is all that matters, isn't it?
Regards,
As stated before, we use the PSU for sound quality. As long as the sound quality is good, that's all that matter. The comparison is based on subjective listening and measurement is not considered. Case in point, most tube amps have higher distortion than solid state amp and do not measure as good. But people still enjoy their sound. It has been debated many many time in the audiophile circle that good measurements do not necessary means "good sounding". I am not saying that measurement is not important. But bottom line, whatever the measurement results are, it has to sound "good". If a PSU measured badly and does not generate nice sound then it is not suitable for the purpose.
Regards,
I have been meaning to test my theory of whether or not these laptop bricks if identical can be paralleled for more current. I actually started a thread on this a long time ago and there was no agreement. People just said test it and see if it works. If four $6 bricks can get me 19v at 18 amps - that is really something. Maybe just adding some big fat caps at the common junction to help filter out any psu-psu interactions might work.
Where this gets interesting is taking two sets of the high current SMPS parallel supplies and tying them together with a common tap for +/- 19v dual rail low cost supply for an IRS2091 switching MOSFET class D amp. Or even a 3386 class AB chip amp and forgo the huge and expensive toroid transformer and caps which are a substantial cost of these dual rail amps. Hey, you could even use it as a dual rail for a fancy class A. Nelson Pass uses SMPS for his ACA class A amp.
In order to do a center tap, you need to get a SMPS that is not earth grounded - two prong variety so output can float. The Vaio ones I linked above are two prong and can work. I may have to try this soon.
Where this gets interesting is taking two sets of the high current SMPS parallel supplies and tying them together with a common tap for +/- 19v dual rail low cost supply for an IRS2091 switching MOSFET class D amp. Or even a 3386 class AB chip amp and forgo the huge and expensive toroid transformer and caps which are a substantial cost of these dual rail amps. Hey, you could even use it as a dual rail for a fancy class A. Nelson Pass uses SMPS for his ACA class A amp.
In order to do a center tap, you need to get a SMPS that is not earth grounded - two prong variety so output can float. The Vaio ones I linked above are two prong and can work. I may have to try this soon.
Last edited:
Paralleling a PS means that they should be able to share the current evenly. This means that they have to be the exact same output voltage for this sharing to happen efficiently with equal distribution. Easier said than done in practice. So once you have that figured out, you can proceed with caution. It is not a simple matter as just connecting them all up together.
As a safety measure you can diode isolate each PS using a low drop diode = schottky on each output that connects using a diode "OR" wiring arrangement. Of course adding the diode creates another v drop that has to compensated for.
Usually if you want to parallel supplies you want them all to share the same voltage reference, that way they are able to track since they all have the same reference. Then the feedback elements of the error amp also have to be corrected for, but that can be done using a calibration procedure.
Good luck
Rick
As a safety measure you can diode isolate each PS using a low drop diode = schottky on each output that connects using a diode "OR" wiring arrangement. Of course adding the diode creates another v drop that has to compensated for.
Usually if you want to parallel supplies you want them all to share the same voltage reference, that way they are able to track since they all have the same reference. Then the feedback elements of the error amp also have to be corrected for, but that can be done using a calibration procedure.
Good luck
Rick
Getting a good quality PSU is without question.
My point is the 3116 sounds noticeably better at 24v vs 13.8v.
Case in point. I have one hd song that distorts with a very clean upgraded 13.8v linear power supply. But when I play it with the 24v switching psu It sounds amazing, with no distortion as it should. The 3116 prefers higher voltages imo.
My point is the 3116 sounds noticeably better at 24v vs 13.8v.
Case in point. I have one hd song that distorts with a very clean upgraded 13.8v linear power supply. But when I play it with the 24v switching psu It sounds amazing, with no distortion as it should. The 3116 prefers higher voltages imo.
You guys comparing PSUs...
Why not just get a decent quality power brick that won't blow up? Perhaps then use a differential filter on the output, and it'll sound as good as any other. You can't judge PSUs by the "sound quality", that's complete and utter bollocks.
A good PSU will have a clean output, and will be able to provide the current you need. End of...
Should I use a 500w halogen light bulb as a test load? I don't actually have an amp that can draw 18amps at 19v. Then what am I checking for? Voltage sag or how well each SMPS tracks? Or do I need 4 shunt resistors to track individual current from each SMPS? I have a 10ch divital strip chart recorder. Getting the feedback taps for each would be tricky. Need to dis assemble the units and know the schematic and break off the feedback trace.
Could anybody here who owns the Yang Jing board (Yuan Jing Audio - ON SALE! - TPA3116 Class-D 2.0 Stereo Amplifier Board [50W+50W] - $14.99) possibly measure the mounting hole positions on the board? I've been looking all over the internet for a layout picture of this board but to no avail. Distance between the holes is the most crucial, not the distance from the board outlines.
I'm creating my own arduino pcb in the same dimensions (ardound 100x50mm) but I would need to know the mounting hole positions too to be able to stack the two boards on top of each other.
I'm creating my own arduino pcb in the same dimensions (ardound 100x50mm) but I would need to know the mounting hole positions too to be able to stack the two boards on top of each other.
There are 3 holes, 2 at the input end, and 1 at the output.
The center to center spacing of the 2 that are close together is 25 mm, the c to c spacing from the 2 input holes to the single one at the output is 70 mm.
The hole diameters are about 3.5 mm (1/8") Hope this helps!
Last edited:
No suprise here, as your max amplitude directly corresponds to your supply-voltage. It's not like that the 3316 "prefers" higher voltages, its valid for every amp at a given volume level.
A incandescent lamp is not the best test load to use due to its characteristics, draws very high initial current, which can trip over current logic. If that is all you have, well it will have to do. Measure the cold filament, preferably using a "Kelvin" connected ohmmeter (4-wire), to see what the initial current draw would be. If not measure the lead loss and compensate for it.
"Or do I need 4 shunt resistors to track individual current from each SMPS?"
This is a good idea, so can you measure each V drop across each shunt and adjust each PS o/p V, so that the PS voltages balance out, thus it will allow for similar I draw from each. Can be a bit interactive to adjust because assumption is each PS has the same load regulation, which will not be the case 100%. What is important is that the V balancing takes place at the maximum load current, which is the goal.
Good luck
Rick
"Or do I need 4 shunt resistors to track individual current from each SMPS?"
This is a good idea, so can you measure each V drop across each shunt and adjust each PS o/p V, so that the PS voltages balance out, thus it will allow for similar I draw from each. Can be a bit interactive to adjust because assumption is each PS has the same load regulation, which will not be the case 100%. What is important is that the V balancing takes place at the maximum load current, which is the goal.
Good luck
Rick
WLX vs. SMSL amps
Thanks to the efficiency of the EPacket program from the China and U.S. Post Offices, I quickly received the Breeze Audio 3116 amp based on the WLX board. So I have been able to compare it to the SMSL SA-60 amp I already have on hand (a built amp based on the 3116).
As I've written here, I've been extremely impressed with the sound quality of the SMSL SA-60. Within its power limits (which are not always enough for hard-to-drive speakers), the SA-60 has qualities that I normally hear in high quality tube amps, including a naturalness in harmonic structure, excellent imaging, and a highly musical presentation. In addition, it has better control of the speaker than most tube amps, resulting in excellent transients and bass.
The SA-60 is the only 3116 amp I've heard up until this point, until the WLX-based amp arrived.
The WLX-based Breeze amp is not at all in the same league as the SMSL amp. It is not as good to my ears. It suffers from hardness and brashness in its overall character (noise contamination?), slowness in transient response, lack of bass, limited dynamics, and an overall less than natural musical character. Reviewing comments on this forum about the WLX board, it appears to be an incomplete implementation of the TI evaluation board for the 3116, lacking in particular parts that help limit noise contamination of the audio signal from the high frequency switching components. It probably also has a standard inexpensive pot in the audio path, that likely also limits the quality of the audio signal, which I plan to remove.
So my thought now is to mod the WLX board to see if it can be brought up to the level of the SMSL amp. I'll post again when that's done.
Sure has told me that they plan to come out with their own built amp, scheduled for January 2015. That should be an interesting comparison with the SMSL amp.
The great thing about these 3116 amps is that they are reasonably priced so that it's easy to purchase and compare them without a big loss if they are below par or some unfortunate fatal accident befalls them during the modding process.
My one reservation about the SMSL amp remains - that if used as a basic amp rather than an integrated amp, the automatic volume control reset function interferes, because upon each turn off the volume is reset to a very low level and must be manually turned up again for use. As good an idea as this may be for use as an integrated amp to protect one's speakers and ears, it makes it difficult to use several of these amps in a multichannel system.
Thanks to the efficiency of the EPacket program from the China and U.S. Post Offices, I quickly received the Breeze Audio 3116 amp based on the WLX board. So I have been able to compare it to the SMSL SA-60 amp I already have on hand (a built amp based on the 3116).
As I've written here, I've been extremely impressed with the sound quality of the SMSL SA-60. Within its power limits (which are not always enough for hard-to-drive speakers), the SA-60 has qualities that I normally hear in high quality tube amps, including a naturalness in harmonic structure, excellent imaging, and a highly musical presentation. In addition, it has better control of the speaker than most tube amps, resulting in excellent transients and bass.
The SA-60 is the only 3116 amp I've heard up until this point, until the WLX-based amp arrived.
The WLX-based Breeze amp is not at all in the same league as the SMSL amp. It is not as good to my ears. It suffers from hardness and brashness in its overall character (noise contamination?), slowness in transient response, lack of bass, limited dynamics, and an overall less than natural musical character. Reviewing comments on this forum about the WLX board, it appears to be an incomplete implementation of the TI evaluation board for the 3116, lacking in particular parts that help limit noise contamination of the audio signal from the high frequency switching components. It probably also has a standard inexpensive pot in the audio path, that likely also limits the quality of the audio signal, which I plan to remove.
So my thought now is to mod the WLX board to see if it can be brought up to the level of the SMSL amp. I'll post again when that's done.
Sure has told me that they plan to come out with their own built amp, scheduled for January 2015. That should be an interesting comparison with the SMSL amp.
The great thing about these 3116 amps is that they are reasonably priced so that it's easy to purchase and compare them without a big loss if they are below par or some unfortunate fatal accident befalls them during the modding process.
My one reservation about the SMSL amp remains - that if used as a basic amp rather than an integrated amp, the automatic volume control reset function interferes, because upon each turn off the volume is reset to a very low level and must be manually turned up again for use. As good an idea as this may be for use as an integrated amp to protect one's speakers and ears, it makes it difficult to use several of these amps in a multichannel system.
Les H,
Thanks for the review and comparo. The Breeze amps are not know for their quality parts. Probably swapping the main power caps to OSCON's will help immensely. Also, the SMSL is one of the few amps with a bootstrap snubber and that makes a huge difference in the harshness. So adding the bootstrap snubber mod to your Breeze may help a lot.
Good luck.
Thanks for the review and comparo. The Breeze amps are not know for their quality parts. Probably swapping the main power caps to OSCON's will help immensely. Also, the SMSL is one of the few amps with a bootstrap snubber and that makes a huge difference in the harshness. So adding the bootstrap snubber mod to your Breeze may help a lot.
Good luck.
***
Thanks for the input. I plan to order the parts from DigiKey and implement the mods you've suggested for a new comparison. It appears that most of the inexpensive 3116 boards do not fully or correctly implement the TI design, though the SMSL and Sure seem to come closest.
I ended up using a portable cook top as the resistive load. I wired all four SMPS in parallel and turned them on one by one. The voltmeter read 19.71 volts for any single SMPS alone or in parallel. With the hot plate turned on, the voltage dropped to 19.68 volts and the plate got warm. Nothing switched off and nothing blew up. I will have to figure out a way to see how many watts are delivered. Maybe heat water and measure temperature rise. If this works, we have a low cost ($24) 18 amp 19v power supply with lots of potential - maybe bass transients on TPA3116D2 in PBTL 2ohms will be better?
Here are the supplies before the plug got snipped off:
Here are the supplies soldered to a single parallel connection (a 120 v lamp cord):
The next thing I want to try is a +/- 19v dual rail supply with 9A capability. Could be useful for a chip amp or IRS2902.
Attachments
Last edited:
OSCON's as an upgrade is one of the most discernible instant gratification upgrades there is. If you already have good Nichicon's or Panasonic FM's it may not be as noticeable as stock Chinese caps. Most folks will agree that this is a great mod.