Hi,
I was wondering if anyone could point me in the direction of a good LPSU design for powering my Khadas Tone Board externally. I am looking at a 5V supply capable of around 500mA. I was planning on using an LM317 and am hoping to find a design that isn't too complicated as it is my first time at building a power supply. I just want soemthing to provide cleaner power than USB or my SMPS. The areas of uncertainty for me are how I should filter on the input/output.
Caps rated to 35V are relatively cheap (a 10000uf costs about $5) so i'm hoping that smart use of capacitors might provide enough filtering without going to an overly complicated design. I was thinking of using a CT transformer around 9-15VAC so there should be plenty of headroom for voltage loss.
I have seen many designs but each have their intended applications and I have not found anything specific for a DAC, which I guess means low voltage, low current and very clean.
Thanks,
Mark
I was wondering if anyone could point me in the direction of a good LPSU design for powering my Khadas Tone Board externally. I am looking at a 5V supply capable of around 500mA. I was planning on using an LM317 and am hoping to find a design that isn't too complicated as it is my first time at building a power supply. I just want soemthing to provide cleaner power than USB or my SMPS. The areas of uncertainty for me are how I should filter on the input/output.
Caps rated to 35V are relatively cheap (a 10000uf costs about $5) so i'm hoping that smart use of capacitors might provide enough filtering without going to an overly complicated design. I was thinking of using a CT transformer around 9-15VAC so there should be plenty of headroom for voltage loss.
I have seen many designs but each have their intended applications and I have not found anything specific for a DAC, which I guess means low voltage, low current and very clean.
Thanks,
Mark
Are you sure of the requirements -- 500 mA @ 5V? There are plenty of wall warts that could provide what you'd need to start; that would save you from the safety concerns of line voltages. Many have good isolation as well.
If the 1/2 Amp need is firm, an LM317 will require a decent heat sink. The rectifying and filtering isn't too complicated to work out.
Or if you don't mind some browsing and refining search specs, there are some designs around the web and on this site that have been successful.
I'm just wondering about that 1/2 Amp spec .. seems like a lot for a Khadas ..
Regards
If the 1/2 Amp need is firm, an LM317 will require a decent heat sink. The rectifying and filtering isn't too complicated to work out.
Or if you don't mind some browsing and refining search specs, there are some designs around the web and on this site that have been successful.
I'm just wondering about that 1/2 Amp spec .. seems like a lot for a Khadas ..
Regards
You are right, the DAC will draw.less than this. I have seen figures of 120mA. The 0.5amp recommendation probably comes from the max draw from a usb or something. I have a good quaity 5v smps but the manufacturer recommends a linear.power supply for better qulity sound...
I have been in your shoes not too long ago 🙂
For my raspberry streamer (also 5V) I built the Glassware power supply, a kit is available here:
LV-Regulator
It was easy to assemble and works perfectly.
Later I built another LPS, this time I chose the one from Rod Elliott, Project P05:
Power Supply for Preamps
This one is available as a PC board only, you will have to source your parts yourself. It can be configured for a number of different voltages.
I had no trouble building this one either; I put together a list of parts and asked Rod to confirm it. He responded very quickly confirming my selection of parts. This power supply may be a bit less expensive than the glassware, yet I doubt that it is in any way inferior.
Good luck, have fun. The Toneboard is on my list of projects as well 🙂
For my raspberry streamer (also 5V) I built the Glassware power supply, a kit is available here:
LV-Regulator
It was easy to assemble and works perfectly.
Later I built another LPS, this time I chose the one from Rod Elliott, Project P05:
Power Supply for Preamps
This one is available as a PC board only, you will have to source your parts yourself. It can be configured for a number of different voltages.
I had no trouble building this one either; I put together a list of parts and asked Rod to confirm it. He responded very quickly confirming my selection of parts. This power supply may be a bit less expensive than the glassware, yet I doubt that it is in any way inferior.
Good luck, have fun. The Toneboard is on my list of projects as well 🙂
Thanks, yes i want to have the DAC teamed up with a raspberry pi streamer and a 3e tpa3255 amp.
This is the thing. I have this SMPS
ECM60UT31 XP Power | Mouser Australia
I can easily use it for both the RPI and the DAC. The DAC claims to have very good PSRR.
I really don't know whether i'll be acheiving anything by having a separate linear supply for the DAC. It will mean 3 power supplies in total including the meanwell SMPS for the amp. Other option is to have a linear supply big enough for RPI + DAC but that might need to be about 3 amps.
This is the thing. I have this SMPS
ECM60UT31 XP Power | Mouser Australia
I can easily use it for both the RPI and the DAC. The DAC claims to have very good PSRR.
I really don't know whether i'll be acheiving anything by having a separate linear supply for the DAC. It will mean 3 power supplies in total including the meanwell SMPS for the amp. Other option is to have a linear supply big enough for RPI + DAC but that might need to be about 3 amps.
The Glassware power supply looks nice. I have looked at Rod Elliot's designs and have read alot of his articles. He puts forward a good case for not over complicating things. I was considering his P05 design with resistors tweaked for 5V but wasn't sure how suitable it would be for my application.
here is a simplistic and compact lm317 pcb design that i did long ago.
Attachments
Last edited:
Have you had a look at the D-Noizator post a few lines down? There is a single rail board with gerbers by Tombo56 here:
https://www.diyaudio.com/forums/pow...fit-upgrade-317-based-reg-16.html#post6228100
If you are looking for an onboard transformer solution, I have made one here based on Talema series transformers:
https://www.diyaudio.com/forums/pow...fit-upgrade-317-based-reg-17.html#post6239586
The price versus performance ratio is superb, and uses very easily obtained cheap parts.
12VDC wall wart @ 500mA: Mouser 562-QFWB-5-12-US01 for $6.08
L7809CV voltage regulator chip. 12VDC in, regulated 9VDC out. $0.45 at Mouser
L7805CV voltage regulator chip. regulated 9VDC in, regulated 5VDC out. $0.50 at Mouser
330 ohm "bleeder" / "guaranteed minimum load current" resistor from +5V to GND.
3 x 1.0uF electrolytic capacitors, one each at +12V, +9V, +5V.
Done.
If suffering from audiophilia nervosa, add greater numbers of capacitors at random, until you feel better.
L7809CV voltage regulator chip. 12VDC in, regulated 9VDC out. $0.45 at Mouser
L7805CV voltage regulator chip. regulated 9VDC in, regulated 5VDC out. $0.50 at Mouser
330 ohm "bleeder" / "guaranteed minimum load current" resistor from +5V to GND.
3 x 1.0uF electrolytic capacitors, one each at +12V, +9V, +5V.
Done.
If suffering from audiophilia nervosa, add greater numbers of capacitors at random, until you feel better.
Thanks everyone so far. Lots of nice designs and i'm learning a lot. I was going to attempt to do something on veroboard to start, partly for the flexibility and learning experience. Is this a lot harder to do and get right? I was looking at strip board. I have enough room in my enclosure so it is ok if it is not too compact, as long as it doesn't affect the performance.
The 5V SMPS I have also has 12V (3A). Excuse my lack of understanding, but are you saying that simply cascading 2 fixed voltage regulators after a SMPS will achieve power with ripple and noise comparable to a well designed LPSU?
If so, do you think i'd benefit from scaling up to 3A (at 5V) and supply the RPI from this as well? Or would that complicate things alot....
The 5V SMPS I have also has 12V (3A). Excuse my lack of understanding, but are you saying that simply cascading 2 fixed voltage regulators after a SMPS will achieve power with ripple and noise comparable to a well designed LPSU?
If so, do you think i'd benefit from scaling up to 3A (at 5V) and supply the RPI from this as well? Or would that complicate things alot....
You really don't even know what you want, do you?
As thread starter, you are able to edit post #1 in perpetuity; go back, make some changes, and put in the specs you actually hope to obtain. 0.5A? 3A? Some other number? Make a decision and stick to it.
As thread starter, you are able to edit post #1 in perpetuity; go back, make some changes, and put in the specs you actually hope to obtain. 0.5A? 3A? Some other number? Make a decision and stick to it.
I know what I want. I want the best sound from my equipment. I just don't know how to get it.
I was hoping for some guidance to avoid going off in the wrong direction and I tried to keep it simple.
I clarified in post 5 that I will be powering a Raspberry Pi in addition to my DAC. Both happen to require 5V.
I know that DACs are sensitive to noise in general and so initially was looking for an alternative to using my SMPS for the DAC. My original post asked for LPSU suggestions and I have had some great options put forward.
My tangent was due to you providing an alternative option - i.e. to clean up the SMPS and I'd like to understand better.
It creates a new option for me.
If I am going to use a cleaned up SMPS for the DAC, I could scale up the idea and power the RPI with cleaner power too. My SMPS is rated to 12V 3A or 5V 7A.
I have not been able to get a clear answer on whether this would be hard (LM78.. do not come rate to 3A) or whether cleaner power to the RPI is likely to improve sound quality.
I did not want to use a larger LPSU to power both the DAC and the RPI due to the inefficiency compared to the SMPS. For only the DAC it would not be much.
Stick with 'just the DAC' for the intensively filtered supply. You will not be able to hear or measure (without some very expensive test equipment) the difference of running the RPI on 'special power'.
Swinging this stick at a hornets nest will surely draw some chaff, but I'm a big boy ..😉
Filter and regulate the DAC power from the +12V from the ECM60UT31 for now. You can always add another power supply later.
And remember: You will learn more by doing, than by thinking about doing.
Cheers
Swinging this stick at a hornets nest will surely draw some chaff, but I'm a big boy ..😉
Filter and regulate the DAC power from the +12V from the ECM60UT31 for now. You can always add another power supply later.
And remember: You will learn more by doing, than by thinking about doing.
Cheers
Last edited:
Sequencing -- Do you need or want to guarantee that the +5VDC supply turns on first and shuts off last, compared to all the other power supplies in the apparatus? The other way around? (5V comes up last and shuts off first)? etc.
Adjustability -- The DAC is guaranteed to operate correctly when given a power supply voltage in the range (5.0V ± T%) where "T" is the supply tolerance. But which power supply voltage is optimum?? Perhaps you need or want the ability to adjust the supply voltage between, oh let's just say, 4.4V and 5.6V. Now you can find the sweet spot which gives the very best sound from your equipment. It may very well be some voltage other than 5.00 volts. In my experience, CMOS chips exceed their datasheet specifications by the widest margin, when the supply is max ... but your DAC might not be a CMOS chip.
Reverse Protection -- If something goes wrong and an external event tries to pull the +5V supply to a negative voltage (e.g. dropped screwdriver), do you want to protect against this, and prevent the supply from falling to Negative Big Voltage? Assume the external event pulls 4 amperes so it completely swamps and totally overwhelms the "short circuit protection" in regulator ICs.
Hold Up Time -- When the electric power company has a temporary "hiccup" so the AC mains omit one or two cycles of 60 Hz power, is it okay if your +5VDC output falls below its regulated limits? Quantitatively, for how much time do you want your regulator to supply clean pure +5VDC after some dunce yanks the power cord from its mains outlet? 10 milliseconds? 100 milliseconds? 500 milliseconds?
RFI rejection -- Are you concerned about high frequency noise on the AC mains, which is not rejected by your voltage regulator IC? This noise is beyond the right hand edge of the PSRR graph in the regulator datasheet, i.e., higher frequency than the chip is designed to handle. Do you want to install countermeasures which attenuate incoming noise above 20 kHz? How extensively?
Adjustability -- The DAC is guaranteed to operate correctly when given a power supply voltage in the range (5.0V ± T%) where "T" is the supply tolerance. But which power supply voltage is optimum?? Perhaps you need or want the ability to adjust the supply voltage between, oh let's just say, 4.4V and 5.6V. Now you can find the sweet spot which gives the very best sound from your equipment. It may very well be some voltage other than 5.00 volts. In my experience, CMOS chips exceed their datasheet specifications by the widest margin, when the supply is max ... but your DAC might not be a CMOS chip.
Reverse Protection -- If something goes wrong and an external event tries to pull the +5V supply to a negative voltage (e.g. dropped screwdriver), do you want to protect against this, and prevent the supply from falling to Negative Big Voltage? Assume the external event pulls 4 amperes so it completely swamps and totally overwhelms the "short circuit protection" in regulator ICs.
Hold Up Time -- When the electric power company has a temporary "hiccup" so the AC mains omit one or two cycles of 60 Hz power, is it okay if your +5VDC output falls below its regulated limits? Quantitatively, for how much time do you want your regulator to supply clean pure +5VDC after some dunce yanks the power cord from its mains outlet? 10 milliseconds? 100 milliseconds? 500 milliseconds?
RFI rejection -- Are you concerned about high frequency noise on the AC mains, which is not rejected by your voltage regulator IC? This noise is beyond the right hand edge of the PSRR graph in the regulator datasheet, i.e., higher frequency than the chip is designed to handle. Do you want to install countermeasures which attenuate incoming noise above 20 kHz? How extensively?
I was planning on using a relay triggered by the on-off switch to switch all power on simultaneously. I realise there will be a delay while the RPI boots up and comes online. I was waiting to see if there was any problems (clicks/pops etc) with this approach before looking at any sequencing.
It uses XMOS XU208-128-QF48 but the board has LDO regulators on board. Here is the schematic if you are interested. https://dl.khadas.com/Hardware/ToneBoard/Schematic/Tone_V13 SCH.pdf
The company claims good PSRR as seen in the graph below.
Yes I think this would be a good idea.
I don't really know my hardware will cope with this. Not sure...
I am using this power entry module which has some filtering built in. The datashett is here. https://au.mouser.com/datasheet/2/355/FN 280-707113.pdf
I have included a graph of the attenuation below.
Attachments
Thanks for posting the Khadas schem. Looks to me like there's no need to worry about the quality of power to it -- there are no fewer than 7 regulators already, each section having its own. Will have to spend a few more minutes studying it but at a glance, it may not use the incoming 5V directly at all. It will probably sound just as great with any moderately filtered 5V you would care to give it.
Thanks Rick. The SMPS has 70K switching frequency. Below are some graphs of ripple and noise (slightly different model number but same 5V primary output circuit, different voltage secondary output to mine). What approach do you think I should take to "moderately filtered" with this as a starting point.
Attachments
I would suggest trying it with no filtering -- just don't make any final decisions concerning space or other mechanical considerations.
The ADP150-3.3's have excellent PSRR -- over 55 dB, even at the higher frequencies of interest. If it turns out that I'm wrong, a ferrite bead, or small LC or RC stage should be easy to fit.
I didn't find anyplace the Khadas uses the incoming 5V directly. If you find one, by all means correct me.
One of the finer qualities of those ADP150's is the extremely low dropout voltage -- 105mV at the full 150mA load. That gives us over 1 full volt to throw away, if necessary -- lots of playroom.
Do we have a measurement of the Khadas' current demands yet?
The ADP150-3.3's have excellent PSRR -- over 55 dB, even at the higher frequencies of interest. If it turns out that I'm wrong, a ferrite bead, or small LC or RC stage should be easy to fit.
I didn't find anyplace the Khadas uses the incoming 5V directly. If you find one, by all means correct me.
One of the finer qualities of those ADP150's is the extremely low dropout voltage -- 105mV at the full 150mA load. That gives us over 1 full volt to throw away, if necessary -- lots of playroom.
Do we have a measurement of the Khadas' current demands yet?
Last edited: