I'm researching schematics and inspiration for building a stereo 2-way bi-amp component (all in one chassis and, I hope, on one power supply) in the 20w/40w range for each of the four (L+R, hi/low frequency) channels. My biggest obstacle right now is power supplies: they're expensive!
So, I'm also wanting to combine this project with designs that take advantage of the low cost of computer power supply units. 550W models meeting ATX specifications go for under $50. 300W models go for under $10, and they are available with and without fans.
Since I don't know much about what is acceptable ripple current (or voltage noise) from a regulated power supply, I wonder if anyone can comment on the suitability of the ATX specification (http://www.formfactors.org/FFDetail.asp?FFID=1&CatID=2) for audio applications (I'm also intriqued by the idea of using high quality tower cases (with their multiple, standardized bays and sleds for various ad-in devices) as chassis for multi-channel audio amplifier and control/filter equipment.
SO: I'm looking for bona fide "hi-fi" designs, suitable for a beginner, that run on +12v single-pole power. There are many high-power, great-sounding automotive amplifiers, so there must be some good designs out there, somewhere.
AND: I'm looking for suggestions on phase-linear active crossovers that would match well to drive such amps (ideally, from the same +12v power supply). I've looked at the Marchand X-1 (http://www.marchandelec.com/xm1.html) and like just about everything about it, except that it requires a +15, 0, -15v psu. Anyone know of a +12v
So, I'm also wanting to combine this project with designs that take advantage of the low cost of computer power supply units. 550W models meeting ATX specifications go for under $50. 300W models go for under $10, and they are available with and without fans.
Since I don't know much about what is acceptable ripple current (or voltage noise) from a regulated power supply, I wonder if anyone can comment on the suitability of the ATX specification (http://www.formfactors.org/FFDetail.asp?FFID=1&CatID=2) for audio applications (I'm also intriqued by the idea of using high quality tower cases (with their multiple, standardized bays and sleds for various ad-in devices) as chassis for multi-channel audio amplifier and control/filter equipment.
SO: I'm looking for bona fide "hi-fi" designs, suitable for a beginner, that run on +12v single-pole power. There are many high-power, great-sounding automotive amplifiers, so there must be some good designs out there, somewhere.
AND: I'm looking for suggestions on phase-linear active crossovers that would match well to drive such amps (ideally, from the same +12v power supply). I've looked at the Marchand X-1 (http://www.marchandelec.com/xm1.html) and like just about everything about it, except that it requires a +15, 0, -15v psu. Anyone know of a +12v
Car audio amps utilize 12VDC, but use it to power a switching power supply to crank that up to a higher voltage. Bacically you would be dropping 120VAC to 12VDC, then bringing it back up to 40-50(+)VDC (or whatever). Not very efficient. The other drawback is the the amount of current required to to this.
A few years ago, I tried using the 12V from a 300watt power supply to power a "25x2" Sony head unit (really more like 12Wx2 when measured properly). Anything over 1/4 volume, and the power supply would **** out.
Decent computer power supplies are actualy not much cheaper than a toroid/cap/bridge setup. Much like car audio, the power ratings on cheaper power supplies are typically "peak" or "surge" ratings, and not good for continuous use.
-b
A few years ago, I tried using the 12V from a 300watt power supply to power a "25x2" Sony head unit (really more like 12Wx2 when measured properly). Anything over 1/4 volume, and the power supply would **** out.
Decent computer power supplies are actualy not much cheaper than a toroid/cap/bridge setup. Much like car audio, the power ratings on cheaper power supplies are typically "peak" or "surge" ratings, and not good for continuous use.
-b
Cheap power supplies
I have seen on the web ideas for converting AT / ATX power supplies to +/- supplies for audio.
There is extensive work required however and involves dismantling and rewinding the transformer and making changes to the feed back circuitry as well. In addition extra filtering capacitors are required on the output to ensure a clean and fast supply.
You can also series connect unmodified supplies (ok if you get them for no or little cost) and achieve +/- 24v with 4 supplies using the +12v outputs. Do not use the -12 or other outputs.
But have you considered rewinding an old toroid or variac. You can find these at scrap metal recyclers inside discarded equipment (big power supplies, medical equipment, photocopiers, old mini-computers etc.) Buy a couple of these for say $10 each and you may find everything you need (transformer, capacitors, power transistors, heatsinks, bridges etc.) plus a whole lot of extra goodies.
If you look at my thread "power amp under development" you will find details of my stereo amp project with a total cash cost of around $60 US.
If you need any help rewinding or using appropriate components drop me a line.
Cheers
I have seen on the web ideas for converting AT / ATX power supplies to +/- supplies for audio.
There is extensive work required however and involves dismantling and rewinding the transformer and making changes to the feed back circuitry as well. In addition extra filtering capacitors are required on the output to ensure a clean and fast supply.
You can also series connect unmodified supplies (ok if you get them for no or little cost) and achieve +/- 24v with 4 supplies using the +12v outputs. Do not use the -12 or other outputs.
But have you considered rewinding an old toroid or variac. You can find these at scrap metal recyclers inside discarded equipment (big power supplies, medical equipment, photocopiers, old mini-computers etc.) Buy a couple of these for say $10 each and you may find everything you need (transformer, capacitors, power transistors, heatsinks, bridges etc.) plus a whole lot of extra goodies.
If you look at my thread "power amp under development" you will find details of my stereo amp project with a total cash cost of around $60 US.
If you need any help rewinding or using appropriate components drop me a line.
Cheers
If I remember correctly, ATX supplies must have a certain sequence of 'activation' for the supply to turn on. Once this occurs, I believe there is a minimum current draw on certain circuits.
Also, computer power supplies are of a switching design instead of being linear like most of the designs in 'high' end audio, so you may have issues with noise (I don't know what freq. they switch at but would suspect around 40 to 50kHz). There is no reserve capacity to the power supply either, as the type of product it is designed to drive doesn't have transient power demands.
Or, I might be wrong.
Also, computer power supplies are of a switching design instead of being linear like most of the designs in 'high' end audio, so you may have issues with noise (I don't know what freq. they switch at but would suspect around 40 to 50kHz). There is no reserve capacity to the power supply either, as the type of product it is designed to drive doesn't have transient power demands.
Or, I might be wrong.
ATX power isn't anyways good power source for poweramps because of poor pulse current ability, unless added a few capasitors. Most modern ATX high-power powers can do happily about 140w @ +12v. But again, pulse-like current output is still an issue conserning poweramps without extra capasitors.
Radio amateurs, in my country, use those PC supplies to power transceivers.
Transceivers need 12 to 15 volts with 20 amperes.
I am not sure the method, but i heard that they found one 5 volts zener somewhere, as 5 volts output has enormous current capacity, this one is used....they put a 5 watt, with 22 ohms in series resistor in series wit the main 5 volts reference zenner and Ready!!!...there's a 13.8, 20 amperes perfect supply...using 250 Watts PC supplies.
To radio, there's a problem, there are some oscilators that create some "beating" of frequencies with the transceiver oscilators...this way...people use some LC small filtering to the output to avoid that...and put the supply 10 feet away to avoid radio frequency interference.....this rises the "idea" price, as the wire must be special and really heavy gauge, as with that current.... every ohm is a tragedy!
To make plus 14 and minus 14....or even bigger voltage...lets see....hummmm 22 plus and 22 minus, you will have to put bigger zenner related their voltages and discover the protection...the over voltage protection must be switched off...if not...supply will not go on with bigger voltage than maximum alowable by the over voltage internal protection.
I do not know how to instruct you, as i never done....but hundreds of guys are doing that here, and thousands are working in many radio amateur stations in this country.
Those switching supplies, PWM supplies created the class d Amplifier, that is some modified supply only, with double greek "P" letter filter in the output...the filter is passive.
They are wonderfull... small, lightweigth, reliable...but have one terrible defect:
- "I do not dominate them, i do not know exacty how they work", this way i stay alike an indian watching one airplane..... Yeahh..big silver bird.... God of the winds may be...Yuuuugahhh!
Maybe can put two in series to have positive , ground and negative...but really...only Eva can answer that question with 100 percent guaranteed answer quality.
If you had received some money playing cards....if some airplane fall down around you and not hitted you..... if you crash your car against some wall at 150 miles per hour, and no damage in your skin...you will be lucky enough to find this guy...sorry...this girl...EVA.
Carlos
Transceivers need 12 to 15 volts with 20 amperes.
I am not sure the method, but i heard that they found one 5 volts zener somewhere, as 5 volts output has enormous current capacity, this one is used....they put a 5 watt, with 22 ohms in series resistor in series wit the main 5 volts reference zenner and Ready!!!...there's a 13.8, 20 amperes perfect supply...using 250 Watts PC supplies.
To radio, there's a problem, there are some oscilators that create some "beating" of frequencies with the transceiver oscilators...this way...people use some LC small filtering to the output to avoid that...and put the supply 10 feet away to avoid radio frequency interference.....this rises the "idea" price, as the wire must be special and really heavy gauge, as with that current.... every ohm is a tragedy!
To make plus 14 and minus 14....or even bigger voltage...lets see....hummmm 22 plus and 22 minus, you will have to put bigger zenner related their voltages and discover the protection...the over voltage protection must be switched off...if not...supply will not go on with bigger voltage than maximum alowable by the over voltage internal protection.
I do not know how to instruct you, as i never done....but hundreds of guys are doing that here, and thousands are working in many radio amateur stations in this country.
Those switching supplies, PWM supplies created the class d Amplifier, that is some modified supply only, with double greek "P" letter filter in the output...the filter is passive.
They are wonderfull... small, lightweigth, reliable...but have one terrible defect:
- "I do not dominate them, i do not know exacty how they work", this way i stay alike an indian watching one airplane..... Yeahh..big silver bird.... God of the winds may be...Yuuuugahhh!
Maybe can put two in series to have positive , ground and negative...but really...only Eva can answer that question with 100 percent guaranteed answer quality.
If you had received some money playing cards....if some airplane fall down around you and not hitted you..... if you crash your car against some wall at 150 miles per hour, and no damage in your skin...you will be lucky enough to find this guy...sorry...this girl...EVA.
Carlos
+13.5V 14Amp cont. 20 A for 20 sec.
See the link,Tried it 2 years back it works fine...Still used as a battery charger.
http://www.webx.dk/oz2cpu/radios/psu-pc1.htm
See the link,Tried it 2 years back it works fine...Still used as a battery charger.
http://www.webx.dk/oz2cpu/radios/psu-pc1.htm
Hi!
As a amateur radio operator I have also looked into this conversion. Most PC SMPS are noisy in one of them I measured about 50mV - 100mV peak high frequency ripple.
While using with HF ie shortwave transceivers the radio goes CHuuurrrr every 40 Khz. For VHF radios operating on 144-146Mhz the noise is not noticeable. One such converted supply gave about 11amps at 13.6V . This was good enough for normal 25-40W VHF transceivers.
Here is another link describing such Mod:
http://www.qsl.net/aa3sj/Pages/PC-Supply.html
BTW Icom kenwood and many other companies manufacture such supplies a search over the net can provide their schematics. PS 85 is one such PSU.
Regards
Rahul ( Wireless call VU3WJM)
As a amateur radio operator I have also looked into this conversion. Most PC SMPS are noisy in one of them I measured about 50mV - 100mV peak high frequency ripple.
While using with HF ie shortwave transceivers the radio goes CHuuurrrr every 40 Khz. For VHF radios operating on 144-146Mhz the noise is not noticeable. One such converted supply gave about 11amps at 13.6V . This was good enough for normal 25-40W VHF transceivers.
Here is another link describing such Mod:
http://www.qsl.net/aa3sj/Pages/PC-Supply.html
BTW Icom kenwood and many other companies manufacture such supplies a search over the net can provide their schematics. PS 85 is one such PSU.
Regards
Rahul ( Wireless call VU3WJM)
Rahul !!!
Thanks for ur input. Switching supplies produce noise and the harmonics may even extend to HF Range, and even upto 250Mhz.U may notice this only in AM ie in HF,and not
in 2 meter amateur VHF band the sound is Freq Modulated.
I noticed noise in picture of TV,Which was solved by using a high PYE filter at the output of SMPS and a double PYE @ the mains input end.(filter from an old TV was used successfully.
Thanks for ur input. Switching supplies produce noise and the harmonics may even extend to HF Range, and even upto 250Mhz.U may notice this only in AM ie in HF,and not
in 2 meter amateur VHF band the sound is Freq Modulated.
I noticed noise in picture of TV,Which was solved by using a high PYE filter at the output of SMPS and a double PYE @ the mains input end.(filter from an old TV was used successfully.
Dear Sivan,
Hi!
Sorry for late reply was out of town. I have come to conclusion after going thru many ckts that common mode at AC as well as DC out put is solution to the noise problem along with extensive lead screening using high mu ferrite cores.
Recently while at N Delhi I have procured ferrite cores as well as powdered iron torroids to make a 25amp 13.6V supply. Small double section bobbin and cores have been also procured for input filtering as you suggest.
I would be obliged if you can give some tips for the same specially on choise of transistor. Topology used is half bridge as in common PC supplies.
Regards
Rahul
Hi!
Sorry for late reply was out of town. I have come to conclusion after going thru many ckts that common mode at AC as well as DC out put is solution to the noise problem along with extensive lead screening using high mu ferrite cores.
Recently while at N Delhi I have procured ferrite cores as well as powdered iron torroids to make a 25amp 13.6V supply. Small double section bobbin and cores have been also procured for input filtering as you suggest.
I would be obliged if you can give some tips for the same specially on choise of transistor. Topology used is half bridge as in common PC supplies.
Regards
Rahul
Rahul !!!
BUH1215 Transistor = 600watts SMPS A/C i/p =250volts
IFR840= Hexfet= 500 watts SMPS A/C i/p =250volts.
both in half bridge config.
My choice IRF840, Can use the same EE16 core (as in PC SMPS) for gate driving but with modiified turns ratio.
PS: I was really amazed to see ur web site..congrats and 73's.
BUH1215 Transistor = 600watts SMPS A/C i/p =250volts
IFR840= Hexfet= 500 watts SMPS A/C i/p =250volts.
both in half bridge config.
My choice IRF840, Can use the same EE16 core (as in PC SMPS) for gate driving but with modiified turns ratio.
PS: I was really amazed to see ur web site..congrats and 73's.
I've tried BUH1015 and BUH1215 and they are very slow so switching losses become quickly unacceptable. Don't waste money with them
I've also tried MJE18008, that worked much better, and MJE13009 that is the best I've ever tried. Note that these devices were originally made by On Semiconductor but have been cloned later by a lot of companies. I've not tried the clones but I have reports stating that they show inferior performance in comparison with the originals, making them unusable in high power SMPS
I've succesfully used four MJE13009 in a full-bridge configuration to produce 1000W of continuous output at an Ic of 6A. Now I'm gathering more parts and redesigning the prototype to use two power transformers instead of one and three pairs of MOSFETs instead of one for synchronous rectification of the 14.4V output. My target is 1500W [Ic=9A] or more if the MJE13009 allow them reliably. This is 14.4V 108A while the current prorotype produces only 72A
Base-drive transformer design is very tricky. I'm using its own magnetizing inductance to borrow some Ib at the end of the cycle for faster turn-off and get that Ib returned back in the beggining of the next cycle to speed up turn-on saturation process
This works very well and makes redundant the saturation-helping circuit I was previously using,
that provided 1A of additional Ib during the first 1uS of each cycle. The difference is easily felt on the heatsink
When finished, this project will also make a nice and simple high efficiency SMPS for audio amplifiers [with suitable transformer secondaries] and it does not require expensive nor exotic parts. Only SMPS transformer winding skills will be required to build it
I'm also planning to use a PFC front-end to power the full bridge with 340V or 420V pseudo-regulated and get up to 2.8KW output with just four MJE13009 or MJE18008 bipolar switches
I've also tried MJE18008, that worked much better, and MJE13009 that is the best I've ever tried. Note that these devices were originally made by On Semiconductor but have been cloned later by a lot of companies. I've not tried the clones but I have reports stating that they show inferior performance in comparison with the originals, making them unusable in high power SMPS
I've succesfully used four MJE13009 in a full-bridge configuration to produce 1000W of continuous output at an Ic of 6A. Now I'm gathering more parts and redesigning the prototype to use two power transformers instead of one and three pairs of MOSFETs instead of one for synchronous rectification of the 14.4V output. My target is 1500W [Ic=9A] or more if the MJE13009 allow them reliably. This is 14.4V 108A while the current prorotype produces only 72A
Base-drive transformer design is very tricky. I'm using its own magnetizing inductance to borrow some Ib at the end of the cycle for faster turn-off and get that Ib returned back in the beggining of the next cycle to speed up turn-on saturation process
This works very well and makes redundant the saturation-helping circuit I was previously using,
that provided 1A of additional Ib during the first 1uS of each cycle. The difference is easily felt on the heatsink
When finished, this project will also make a nice and simple high efficiency SMPS for audio amplifiers [with suitable transformer secondaries] and it does not require expensive nor exotic parts. Only SMPS transformer winding skills will be required to build it
I'm also planning to use a PFC front-end to power the full bridge with 340V or 420V pseudo-regulated and get up to 2.8KW output with just four MJE13009 or MJE18008 bipolar switches
Thanks Sivan & Eva,
Most of the commercial supplies that I have gone thru in this class I found using transistors in TO247 type of casing. In one of them the transistor was 2SC2625. I will have to check its specs and availibility in local mkt.
Eva as you have rightly noticed the driver transformer design is very critical specially where saturation at start up is concerned. A recent article in Dec 2004 QEX deals with this issue and component wise detailed analysis and modification of an ATX power supply. It has been found that ones used have more than the required inductance so a replacement has been designed using ferrite torroid.
If needed drop a line for that article, its about 1.4meg.
I do remember repairing a small 40Kv unit for xray using IRF840 in half bridge configration, remember tracing it out will look around. This appears to be the cheapest way to go.
Regards to all.
Rahul
Most of the commercial supplies that I have gone thru in this class I found using transistors in TO247 type of casing. In one of them the transistor was 2SC2625. I will have to check its specs and availibility in local mkt.
Eva as you have rightly noticed the driver transformer design is very critical specially where saturation at start up is concerned. A recent article in Dec 2004 QEX deals with this issue and component wise detailed analysis and modification of an ATX power supply. It has been found that ones used have more than the required inductance so a replacement has been designed using ferrite torroid.
If needed drop a line for that article, its about 1.4meg.
I do remember repairing a small 40Kv unit for xray using IRF840 in half bridge configration, remember tracing it out will look around. This appears to be the cheapest way to go.
Regards to all.
Rahul
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.