Hey Guys - I would like to use an existing dual +/-15V DC SMPS module for use in a preamp and/or headphone amp (such as The Wire or O2). There are two pre-built dual rail SMPS modules I know of:
My intention is to take the outputs of these, do some additional filtering as well as regulate them down to +/- 12.5V or so using the classic LM317 / LM 337 pair. This would make for a super-universal power supply (huge input voltage range), not to mention very efficient, tons of available current (esp with the Meanwell!) and useability basically anywhere in the world.
Can anybody suggest a specific filter I should consider given the 100khz fixed switching frequency used inside the SMPS? Also, perhaps there is a linear regulator with similar/greater specs having better rejection at this high frequency compared to the LM317/LM337? Of course, it would still need to handle the 800mA of the Meanwell and also be workable as a negative-regulator or have a counterpart like the 337.
I would like to point out that some very excellent preamps and converters use dual rails generated by SMPSs, such as the products from RME in Germany, including the justifiably popular Fireface series. Also, "funk tonstudiotechnik" in Germany makes "audio-ready" dual 15V power supply module using SMPS, but they are very expensive.
I believe it will take a power supply filtering veteran to help with this, and I realize 12.5V rails aren't always the best for preamps... But the mains and basic SMPS stuff is already taken care of with the Meanwell module from above. This could be a seriously great option for people building the The Wire or O2 headphone amps which are both very popular.
I tried searching some old threads, but none of them seem to offer suggestions for post-filtering/regulation of an existing dual rail SMPS.
I'd love to hear any suggestions!
- Traco TOM 12215 (+/- 15V @ +/- 350mA) - can't find a price
- Meanwell PD-2515 (+/- 15V @ +/- 800mA!!!) - $26.21 @ Mouser
My intention is to take the outputs of these, do some additional filtering as well as regulate them down to +/- 12.5V or so using the classic LM317 / LM 337 pair. This would make for a super-universal power supply (huge input voltage range), not to mention very efficient, tons of available current (esp with the Meanwell!) and useability basically anywhere in the world.
Can anybody suggest a specific filter I should consider given the 100khz fixed switching frequency used inside the SMPS? Also, perhaps there is a linear regulator with similar/greater specs having better rejection at this high frequency compared to the LM317/LM337? Of course, it would still need to handle the 800mA of the Meanwell and also be workable as a negative-regulator or have a counterpart like the 337.
I would like to point out that some very excellent preamps and converters use dual rails generated by SMPSs, such as the products from RME in Germany, including the justifiably popular Fireface series. Also, "funk tonstudiotechnik" in Germany makes "audio-ready" dual 15V power supply module using SMPS, but they are very expensive.
I believe it will take a power supply filtering veteran to help with this, and I realize 12.5V rails aren't always the best for preamps... But the mains and basic SMPS stuff is already taken care of with the Meanwell module from above. This could be a seriously great option for people building the The Wire or O2 headphone amps which are both very popular.
I tried searching some old threads, but none of them seem to offer suggestions for post-filtering/regulation of an existing dual rail SMPS.
I'd love to hear any suggestions!
It doesn't make much sense. You'll be paying as much or more compared to a linear, have higher inherent failure rate & shorter lifespan, more difficulty troubleshooting when it does fail, need to add additional filtering and regulation meaning another power supply board anyway, and all to do what exactly? Save a dozen mA?
The efficiency factor doesn't mean a lot when talking about mW range audio.
There are harmonics in other frequencies than 100KHz you'll need to filter. Put it on a scope and see what you are dealing with, including the non-linear amp load on it when playing.
With all due respect, it's not a great option for anyone. Being mains powered you are tethered to a wall anyway, not a portable application with space or weight constraint issues. That some expensive piece of gear does it is no proof it is better. Sometimes companies build things a particular way just to seem different, more elaborate fancy thing for visual and marketing purposes, plus they have economy of scale when manufacturing them and don't have to deal with failure years later when a warranty has ended.
Essentially you will be going to great trouble and time for an inferior result.
The efficiency factor doesn't mean a lot when talking about mW range audio.
There are harmonics in other frequencies than 100KHz you'll need to filter. Put it on a scope and see what you are dealing with, including the non-linear amp load on it when playing.
With all due respect, it's not a great option for anyone. Being mains powered you are tethered to a wall anyway, not a portable application with space or weight constraint issues. That some expensive piece of gear does it is no proof it is better. Sometimes companies build things a particular way just to seem different, more elaborate fancy thing for visual and marketing purposes, plus they have economy of scale when manufacturing them and don't have to deal with failure years later when a warranty has ended.
Essentially you will be going to great trouble and time for an inferior result.
Hah, ok, perhaps that's why it hasn't been done 🙂
My main reason for bringing it up though is that I'm looking for a great headphone power supply that works in Japan (100V), Colombia (110V), USA (120V) and Europe (230V to 240V), and doesn't require separate fuses and a voltage switch, and especially anything that would result in damage if the user chose the wrong settings (happened to me already to a head amp I built for a family member).
Dual primaries 115V/230V are quite common, but I don't think such a config would be good enough for Japan (100V) and perhaps getting a little too close for comfort to the linear regulators' dropping out in countries with 110V (unless you drive them a littler hotter than usual).
I think the reasons the companies I mention go with SMPS is not to really to seem different (they don't even mention it in their ads), but for more practical reasons: one design, and no confusing instructions for the end users as far as voltage settings and fuse-swapping. You're absolutely right, efficiency isn't a deal-breaker in low-power stuff, but user-friendliness is...
I guess I just want to know how hard it would be to filter these harmonics, because most of the hard work (designing the SMPS) is already done. I mean, just plopping two consecutive LM317's and LM337's in series would be great (given it's PSRR even at 100khz), the only problem is that the rails would be a little low due to the 2V dropout, and I can't find any cheap SMPS modules that are higher than 15V in dual (positive + negative) rails.
Why do I need all this international compatibility and user-friendliness? That's a whole other topic.
I could give this a rest, but I guess what bugs me is knowing that it's possible...
My main reason for bringing it up though is that I'm looking for a great headphone power supply that works in Japan (100V), Colombia (110V), USA (120V) and Europe (230V to 240V), and doesn't require separate fuses and a voltage switch, and especially anything that would result in damage if the user chose the wrong settings (happened to me already to a head amp I built for a family member).
Dual primaries 115V/230V are quite common, but I don't think such a config would be good enough for Japan (100V) and perhaps getting a little too close for comfort to the linear regulators' dropping out in countries with 110V (unless you drive them a littler hotter than usual).
I think the reasons the companies I mention go with SMPS is not to really to seem different (they don't even mention it in their ads), but for more practical reasons: one design, and no confusing instructions for the end users as far as voltage settings and fuse-swapping. You're absolutely right, efficiency isn't a deal-breaker in low-power stuff, but user-friendliness is...
I guess I just want to know how hard it would be to filter these harmonics, because most of the hard work (designing the SMPS) is already done. I mean, just plopping two consecutive LM317's and LM337's in series would be great (given it's PSRR even at 100khz), the only problem is that the rails would be a little low due to the 2V dropout, and I can't find any cheap SMPS modules that are higher than 15V in dual (positive + negative) rails.
Why do I need all this international compatibility and user-friendliness? That's a whole other topic.
I could give this a rest, but I guess what bugs me is knowing that it's possible...
The best (most optimal in time/size/expense/complexity/etc/etc power supply is one purpose specific, not universal. Define the amp you want to power then go from there.
Different voltage input is not necessarily a problem. Opamps have voltage ranges, circuits can switch input transformer windings, regulators ran at low current can still operate within thermal margins with little or no heatsink. It all starts with defining the operating requirement, then the PSU best suited. If there were only one PSU design best for everything, there would only be one PSU design that exists.
Different voltage input is not necessarily a problem. Opamps have voltage ranges, circuits can switch input transformer windings, regulators ran at low current can still operate within thermal margins with little or no heatsink. It all starts with defining the operating requirement, then the PSU best suited. If there were only one PSU design best for everything, there would only be one PSU design that exists.
At this power why not simply buy a 240V / 35V transformer.
Pre-regulate the rectified mains with a simple Zener & Transistor to bring the voltage down to 25V DC.
Then feed the 25V to your LM317.
At 240V the regulators will be dissipating 4W at 100mA.
At 100V the transformer will still be kicking out 15VAC which will still give the regulators Headroom to operate.
This method becomes very inefficient as the power goes up but is still valid.
Pre-regulate the rectified mains with a simple Zener & Transistor to bring the voltage down to 25V DC.
Then feed the 25V to your LM317.
At 240V the regulators will be dissipating 4W at 100mA.
At 100V the transformer will still be kicking out 15VAC which will still give the regulators Headroom to operate.
This method becomes very inefficient as the power goes up but is still valid.
Attachments
Last edited:
The easiest approach might be to insert RC or LC pi filters, using the 100kHz frequency to help determine reactance and component values.
Hi,
I can see the appeal of universal input voltage.
But there is little point regulating what is already regulated,
assuming the SMPS is half decent, the main issue is noise.
With a Class A load (constant current) the humble CRC filtering
arrangement will work extremely well, reducing residual ripple
and clobbering completely any noise issues higher up.
rgds, sreten.
I can see the appeal of universal input voltage.
But there is little point regulating what is already regulated,
assuming the SMPS is half decent, the main issue is noise.
With a Class A load (constant current) the humble CRC filtering
arrangement will work extremely well, reducing residual ripple
and clobbering completely any noise issues higher up.
rgds, sreten.
Hey Guys, thanks for all the input so far.
OK, so here is the audio circuit I want to power: The Wire by forum member OPC. There are a few versions, but schematic I'm talking about is at the bottom of the page, on PG. 1 of the last PDF entitled "THE WIRE - FINAL SCH.pdf ". We are talking 6 LME49990's as well as 2 LME49600 buffers. The voltage rails I would like should be +/-12.5V (allowing for just enough headroom).
And as for the incoming power from the outside world, I'm looking for a power supply could handle anything from 100V to 240V without requiring users to switch things that may potentially damage the unit (or themselves!). Definitely a strength of SMPS (we already know its weakness is output noise).
I haven't yet measured the exact current requirements of The Wire headphone amp, but using the suggested linear power supply, the LM317/337 regulators get nice and warm on their 0.5°C/W heat sinks when driving 300 ohm headphones. I can imagine it is even more intense for lower impedance headphones.
@KatieandDad, that is a very interesting suggestion and I truly appreciate the schematic you drew up. I am going to explore this. I'm a little worried that the power dissipated given the current draw might be a little too high for practical heat sink sizes, but I could be wrong.
@! : I appreciate your comments. Do you have any suggestions for circuits that can safely swap the input transformer windings in a fool-proof automatic fashion, and still be less complicated than designing a post SMPS filter/regulator as per my original post? I would be genuinely interested, because I have tried going down this road, but found that the circuits required ended up being quite complex and not always safe - hence my reason for posting about post SMPS filters for a generic SMPS.
@sofaspud and @sreten, thanks for these thoughts - it is sort of what I had in mind but I wasn't sure if it would be enough...
OK, so here is the audio circuit I want to power: The Wire by forum member OPC. There are a few versions, but schematic I'm talking about is at the bottom of the page, on PG. 1 of the last PDF entitled "THE WIRE - FINAL SCH.pdf ". We are talking 6 LME49990's as well as 2 LME49600 buffers. The voltage rails I would like should be +/-12.5V (allowing for just enough headroom).
And as for the incoming power from the outside world, I'm looking for a power supply could handle anything from 100V to 240V without requiring users to switch things that may potentially damage the unit (or themselves!). Definitely a strength of SMPS (we already know its weakness is output noise).
I haven't yet measured the exact current requirements of The Wire headphone amp, but using the suggested linear power supply, the LM317/337 regulators get nice and warm on their 0.5°C/W heat sinks when driving 300 ohm headphones. I can imagine it is even more intense for lower impedance headphones.
@KatieandDad, that is a very interesting suggestion and I truly appreciate the schematic you drew up. I am going to explore this. I'm a little worried that the power dissipated given the current draw might be a little too high for practical heat sink sizes, but I could be wrong.
@! : I appreciate your comments. Do you have any suggestions for circuits that can safely swap the input transformer windings in a fool-proof automatic fashion, and still be less complicated than designing a post SMPS filter/regulator as per my original post? I would be genuinely interested, because I have tried going down this road, but found that the circuits required ended up being quite complex and not always safe - hence my reason for posting about post SMPS filters for a generic SMPS.
@sofaspud and @sreten, thanks for these thoughts - it is sort of what I had in mind but I wasn't sure if it would be enough...
- Status
- Not open for further replies.