I concluded the brick SMPS is lacking in performace in general because both the output and input filter lack parts needed to sufficiently supress RFI. I don't want to use it in my house and in fact I'm going to check every SMPS I have in the house with a scope and if necessary install filtering as needed.
As with my DSP crossover many have changed the stock SMPS (with no output filtering at all) to a linear PS and some of them have measured the before/ after results to confirm the improved performace. There are multiple commercial power supply upgrade kits offered, so I thought all this was enough to justify my thinkering with the unit.
Good practices are exactly that for a reason and I don't see why not to follow them. The other thing is I am in this hobby because I like the experience of finding out how things work and why. Then I can recognise the shortcuts that manufacturers make in the design of equipment and correct them if I think it is appropriate.
As with my DSP crossover many have changed the stock SMPS (with no output filtering at all) to a linear PS and some of them have measured the before/ after results to confirm the improved performace. There are multiple commercial power supply upgrade kits offered, so I thought all this was enough to justify my thinkering with the unit.
Good practices are exactly that for a reason and I don't see why not to follow them. The other thing is I am in this hobby because I like the experience of finding out how things work and why. Then I can recognise the shortcuts that manufacturers make in the design of equipment and correct them if I think it is appropriate.
It's not just the noise of the original SMPS. It has bad regulation with 5V and 3V3 dropping below 4.7 and 3.1 V at times. It is followed by unadequate filterig.
A member changed the decoupling electrolythics on the board and posted harmonic distortion graphs.
http://www.diyaudio.com/forums/showthread.php?p=4398344
If the the original SMPS were adequate there'd be no change...
A member changed the decoupling electrolythics on the board and posted harmonic distortion graphs.
http://www.diyaudio.com/forums/showthread.php?p=4398344
If the the original SMPS were adequate there'd be no change...
This was the lytics on the DXO board not the power supply, your conclusion is also incorrect, the changes were very minor and were probably down to better quality caps being used with a tighter tolerance value than the original cheap caps, this will change the power delivery system and its noise spectrum slightly.
With reference to your last post, the supply regulation should not be that bad and the previous comment regarding the 40kHz switching make me suspect the power supply is damaged, is not being run in its optimal power range or is just a cheep bag of sh***, there are many good SMPS supplies out there
With reference to your last post, the supply regulation should not be that bad and the previous comment regarding the 40kHz switching make me suspect the power supply is damaged, is not being run in its optimal power range or is just a cheep bag of sh***, there are many good SMPS supplies out there
Marko, I just read through that link a bit too - the original power supply was not modified in any way - the only remark is that changing some power supply buffer caps on the DCX board was one a few subtle tweaks done in an attempt to improve performance, but the bottom line was that any performance change was pretty much negligible, and the main performance improvement were by other changes.
I'd also suggest you may need to check out how to use an oscilloscope probe to measure noise and ripple from a smps - many people get fooled in to observing much higher levels than what actually exist.
Tim
I'd also suggest you may need to check out how to use an oscilloscope probe to measure noise and ripple from a smps - many people get fooled in to observing much higher levels than what actually exist.
Tim
It'a cheep bag of dirt, lacks components, both vacant spots on the PCB and any output filtering. But I didn't give up on all SMPS supplies, just this particular one.
I allready had components to make a linear PS for the shunt regulators on hand, a little excess heat will be taken care of with better case ventillation.
And a new SMPS would cost me money which I will gladly spend on other audio related things.
Suggestion noted, will read up on the topic. It was my first outing with the scope, it had automatic Hz and Vp-p measurments but I also counted the divisions for the ripple V and f
Maybe my power supply upgrade was started by a hunch rather than measurements or an informed desidion. But the caps in the unit's SMPS were bulged and from what I read people had good experience with changing some parts of supply, like changing the 7805's or replacing the supply with a linear. I took the plunge, the DCX sounded different and I like the sound more now so...
Last edited:
Placing a cro probe anywhere near a transformer (whether mains or smps) will show a screen signal without even making contact with any circuit. As such, it is important to remove any radiated coupling into a cro measurement of ripple and noise voltages, otherwise you aren't measuring what you are aiming to measure (eg. the circuit voltage across the points of interest).
In essence, the voltage of interest is the power supply input terminals on your audio equipment. That is when the power supply itself is some distance away so as not to influence the measurement. The probe can still pick up radiated noise from many sources - even from the audio equipment itself if it has digital circuitry. So that is best managed by using a very short ground clip - an option often provided with probes that typically requires the probe's end spring clip on section to be removed, and the exposed probe tip to be used with a small springy ground point attached.
That technique can also be used for measuring a power supplies output voltage - but that is best done carefully and made across the last filter capacitor near the output terminals - or a small filter capacitor placed on the output terminals (eg. a smt cap soldered across the terminal pins on a pcb).
In essence, the voltage of interest is the power supply input terminals on your audio equipment. That is when the power supply itself is some distance away so as not to influence the measurement. The probe can still pick up radiated noise from many sources - even from the audio equipment itself if it has digital circuitry. So that is best managed by using a very short ground clip - an option often provided with probes that typically requires the probe's end spring clip on section to be removed, and the exposed probe tip to be used with a small springy ground point attached.
That technique can also be used for measuring a power supplies output voltage - but that is best done carefully and made across the last filter capacitor near the output terminals - or a small filter capacitor placed on the output terminals (eg. a smt cap soldered across the terminal pins on a pcb).
Last edited:
Yes, magnetic field in the vicinity of abrupt switching current induces spikes even in small loops. For a more accurate measuring you need a proper probe like this one
Power Tip #6: Accurately Measuring Power Supply Ripple | EE Times
Power Tip #6: Accurately Measuring Power Supply Ripple | EE Times
Cheers.- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.