here is one way to do it
... http://focus.ti.com/lit/ds/symlink/ua7805.pdf ...
A Linear voltage regulator with wide range input voltage and very well behaved output ....
My personal application notes:
* Note that the "tab" on the TO-220 type is the same as pin#2, the ground pin. This means that the tabs can be attached to a heat sink that is at ground potential, meaning that your metal case can serve as the heat sink.
* These linear regulators can be linked in parallel and I have done this many times. Using the common "tab" / ground / heat sink of each connected together, the outputs can be in parallel ... and so can the inputs !!. Use three of these for up to 3 amps, five of these for 5 amps output current. Any differences & variations of output voltage between two to five or more regulators in parallel is nominal, as after a breakin period, an array of these regulators tend to each seek a common output voltage ... the break in time being how long it takes for the thermal charicturistics of each regulator to stabilize to the same levels. (The warmist tending to self limit, waiting for the others to "catch up" as it were.)
* Best input voltage for the -7805 is between 8 and 14 VDC = ideal for automotive use = best ripple rejection. These will stand up to +30 VDC down to as little as 1.5 VDC above the regulated output voltage, but the higher input voltages can increase heat generation.
* Note the circuits on page 6 above (see the link). The most commonly used circuit is of course the top one.
* Additional filtering can be added to the output leg (pin 3). I usually use at least 10 uF electrolitic cap on both input and output and a 0.1 uF tantalum cap on the output(s) = very good ripple rejection = decent for audio. For high resolution audio work (like the Squezzer) add more micro Farads at the output(s) and another 0.1 uF tantalum cap to the input(s) (of each regulator). All Capicitor voltages should be greater than 35 VDC rated = protecting the -78xx from surges & RF, etc.
* As a matter of prudence for the best possible performance, I over engineer the circuit, using much more capacity than may be required. === For your 2.5 Amp requirement, I would use five or six of these regulators in parallel = 5 to 6 Amps worth. This will give double what you need and much less than half the heat generated ... AND each regulator will thermally stablize better, faster ... AND your ripple & noise rejection will improve by a factor of 10x (a detailed technical explaination for this on request) ===
* I keep several dozen of these in the parts drawers in my shop. I buy them in lots of 5 to 20 at a time. The -7805 = 5.00 VDC, -7812 = 12.00 VDC, -7815 = 15 VDC and the 79xx are negative regulators (7905 = -5 VDC, etc.) ... and they can be had for between US$0.50 each to US$1.00 each ....
(For your transformer, use a 25 VA rated, primary for your voltage = 240 VAC and a secondary of 18 VAC ... fast blow fused to 1 Amp on the transformer input / line cord / plug in the wall side. This will allow you to plug in just about everywhere = from 120 VAC to 240 VAC = 9 VAC secondary to 18 VAC secondary, respectivly. Feed it through a full wave bridge rectifier ...)
... http://focus.ti.com/lit/ds/symlink/ua7805.pdf ...
A Linear voltage regulator with wide range input voltage and very well behaved output ....
My personal application notes:
* Note that the "tab" on the TO-220 type is the same as pin#2, the ground pin. This means that the tabs can be attached to a heat sink that is at ground potential, meaning that your metal case can serve as the heat sink.
* These linear regulators can be linked in parallel and I have done this many times. Using the common "tab" / ground / heat sink of each connected together, the outputs can be in parallel ... and so can the inputs !!. Use three of these for up to 3 amps, five of these for 5 amps output current. Any differences & variations of output voltage between two to five or more regulators in parallel is nominal, as after a breakin period, an array of these regulators tend to each seek a common output voltage ... the break in time being how long it takes for the thermal charicturistics of each regulator to stabilize to the same levels. (The warmist tending to self limit, waiting for the others to "catch up" as it were.)
* Best input voltage for the -7805 is between 8 and 14 VDC = ideal for automotive use = best ripple rejection. These will stand up to +30 VDC down to as little as 1.5 VDC above the regulated output voltage, but the higher input voltages can increase heat generation.
* Note the circuits on page 6 above (see the link). The most commonly used circuit is of course the top one.
* Additional filtering can be added to the output leg (pin 3). I usually use at least 10 uF electrolitic cap on both input and output and a 0.1 uF tantalum cap on the output(s) = very good ripple rejection = decent for audio. For high resolution audio work (like the Squezzer) add more micro Farads at the output(s) and another 0.1 uF tantalum cap to the input(s) (of each regulator). All Capicitor voltages should be greater than 35 VDC rated = protecting the -78xx from surges & RF, etc.
* As a matter of prudence for the best possible performance, I over engineer the circuit, using much more capacity than may be required. === For your 2.5 Amp requirement, I would use five or six of these regulators in parallel = 5 to 6 Amps worth. This will give double what you need and much less than half the heat generated ... AND each regulator will thermally stablize better, faster ... AND your ripple & noise rejection will improve by a factor of 10x (a detailed technical explaination for this on request) ===
* I keep several dozen of these in the parts drawers in my shop. I buy them in lots of 5 to 20 at a time. The -7805 = 5.00 VDC, -7812 = 12.00 VDC, -7815 = 15 VDC and the 79xx are negative regulators (7905 = -5 VDC, etc.) ... and they can be had for between US$0.50 each to US$1.00 each ....
(For your transformer, use a 25 VA rated, primary for your voltage = 240 VAC and a secondary of 18 VAC ... fast blow fused to 1 Amp on the transformer input / line cord / plug in the wall side. This will allow you to plug in just about everywhere = from 120 VAC to 240 VAC = 9 VAC secondary to 18 VAC secondary, respectivly. Feed it through a full wave bridge rectifier ...)
Yes = more = more
Pink: " ... Do you really need 2.5A? ..."
Gsuser set the requirement, I only offered a solution. And yes you will need more than one (1) UA7805 ... I recommended five (5) or six (6) lashed together in parallel for a steady, cool running, ripple & noise free regulator ...
You are probably right about the Squeezers not needing quite that much ... My recollection is that they need about 10 to 12 watts ... which at 5 VDC is about 2 Amps ... Yes?
Pink: " ... Do you really need 2.5A? ..."
Gsuser set the requirement, I only offered a solution. And yes you will need more than one (1) UA7805 ... I recommended five (5) or six (6) lashed together in parallel for a steady, cool running, ripple & noise free regulator ...
You are probably right about the Squeezers not needing quite that much ... My recollection is that they need about 10 to 12 watts ... which at 5 VDC is about 2 Amps ... Yes?
Well, IMHO, stringing 5 or 6 7805's in parallel is unlikely to beat the performance of whatever supply the device came with.
If you are a newbie, a super-regulator might prove too difficult. If you look at the following post and built the +ve circuit with an LM338, you might be very happy in the end:
http://www.diyaudio.com/forums/showthread.php?postid=356154#post356154
If you are a newbie, a super-regulator might prove too difficult. If you look at the following post and built the +ve circuit with an LM338, you might be very happy in the end:
http://www.diyaudio.com/forums/showthread.php?postid=356154#post356154
Oh yeah .... plenty good
http://www.diyaudio.com/forums/showthread.php?postid=356154#post356154 ... plenty good enough for this case ... and you just build the upper half of that drawing = plenty 'o current and a better parts count than my suggestion(s).
(I would bet there isn't much difference between the above and my suggestion(s) as far as ripple, surge, noise rejection goes ... but I've been wrong before.)
http://www.diyaudio.com/forums/showthread.php?postid=356154#post356154 ... plenty good enough for this case ... and you just build the upper half of that drawing = plenty 'o current and a better parts count than my suggestion(s).
(I would bet there isn't much difference between the above and my suggestion(s) as far as ripple, surge, noise rejection goes ... but I've been wrong before.)
Re: Oh yeah .... plenty good
There is. Those older fixed regulators are noisier because they use a cheaper voltage reference. Why do you think you see so much use of the LM317 on this forum? You think all those countless DAC designs use the LM317 for 5VDC because the designer is too stupid to figure out that a 7805 exists?
FastEddy said:[B(I would bet there isn't much difference between the above and my suggestion(s) as far as ripple, surge, noise rejection goes ... but I've been wrong before.)
There is. Those older fixed regulators are noisier because they use a cheaper voltage reference. Why do you think you see so much use of the LM317 on this forum? You think all those countless DAC designs use the LM317 for 5VDC because the designer is too stupid to figure out that a 7805 exists?
Old uA7805 v. "newer" LM317
leadbelly: " ... Those older fixed regulators are noisier because they use a cheaper voltage reference. Why do you think you see so much use of the LM317 on this forum? ..."
I am sure you are right on here ... the uA78xx and uA79xx series of fixed voltage 3-pin chip regulators have been around for decades. In fact they were originally a "flagship" product of the original Fairchild company in the late 1970's, possibly earlier.
Modern iterations of the -78xx/-79xx series from TI, National and others do have a "noisey" voltage reference, only good out to a couple of decimal places. But that is the reason I usually use a tantlum capacitor on the output leg = to "tame" this schrew and make it work very well for audio work.
We are talking here of a simple substitute for an existing power supply wall wart and not the ultimate heavily filtered supply of a higher performance output amplifier. The Squeeze Box 2 & 3 does have very well regulating internal filters = good enough to "tame" the wall wart for most consumers. When the Squeeze's internal filters are coupled to a 78xx based linear supply (with a tantlum snubbing cap in place), the Squeeze Box power bus will be very well mannered indeed = with a PS noise floor well below 100 db in virtually every situation. (The Squeezer is a "pure" digital device after all = no need to get too picky here = we only need to be helpful to the DAC's downstream.)
Granted the LM317 from National (and TI and others) is an almost perfect solution for this type of job and may be nominally as good as or possibly a little better than the -78xx, the truth is that these two options are so close in specs and results ... as to be splitting hairs very, very finely in this case.
I have used both and agree with your recommendations but I know that either would do the job very nicely. Truth is one, paths are many ... I say, pick 'em ... the end results will be so similar as to indistingishable, the -78xx circuit being the less complicated choice.
leadbelly: " ... Those older fixed regulators are noisier because they use a cheaper voltage reference. Why do you think you see so much use of the LM317 on this forum? ..."
I am sure you are right on here ... the uA78xx and uA79xx series of fixed voltage 3-pin chip regulators have been around for decades. In fact they were originally a "flagship" product of the original Fairchild company in the late 1970's, possibly earlier.
Modern iterations of the -78xx/-79xx series from TI, National and others do have a "noisey" voltage reference, only good out to a couple of decimal places. But that is the reason I usually use a tantlum capacitor on the output leg = to "tame" this schrew and make it work very well for audio work.
We are talking here of a simple substitute for an existing power supply wall wart and not the ultimate heavily filtered supply of a higher performance output amplifier. The Squeeze Box 2 & 3 does have very well regulating internal filters = good enough to "tame" the wall wart for most consumers. When the Squeeze's internal filters are coupled to a 78xx based linear supply (with a tantlum snubbing cap in place), the Squeeze Box power bus will be very well mannered indeed = with a PS noise floor well below 100 db in virtually every situation. (The Squeezer is a "pure" digital device after all = no need to get too picky here = we only need to be helpful to the DAC's downstream.)
Granted the LM317 from National (and TI and others) is an almost perfect solution for this type of job and may be nominally as good as or possibly a little better than the -78xx, the truth is that these two options are so close in specs and results ... as to be splitting hairs very, very finely in this case.
I have used both and agree with your recommendations but I know that either would do the job very nicely. Truth is one, paths are many ... I say, pick 'em ... the end results will be so similar as to indistingishable, the -78xx circuit being the less complicated choice.
noisey?
mrjam: " ... The reason isn't noise ..."
wine&dine: " ... 2) Is he going to buy an ST Micro brand 78xx or some cheap noisy clone? (What are the odds?) ..."
Odds are good = cheap knockoffs are too common.
BUT with the inclusion of a decent snubbing cap of the plastic or tantlum variety on the output and input legs, the differences between the LM317 and the -78xx disappear.
... and that's why God gave us osciloscopes.
mrjam: " ... The reason isn't noise ..."
wine&dine: " ... 2) Is he going to buy an ST Micro brand 78xx or some cheap noisy clone? (What are the odds?) ..."
Odds are good = cheap knockoffs are too common.
BUT with the inclusion of a decent snubbing cap of the plastic or tantlum variety on the output and input legs, the differences between the LM317 and the -78xx disappear.
... and that's why God gave us osciloscopes.
Many thanks for all these replies, not sure I understand everything... 
In fact, I checked the Squeezebox, and it constantly needs 1100 mA (with wifi and display on). That's why I thought about 2.5A for this power supply.
The "positive-voltage regulators" shown by FastEddy will be too difficult for me to built (first DIY work
)
I think I can handle with leadbelly's link to "enhanced regulator Altus Audio" but what regulator should I use ? And what values for the other components ? I don't know what kind of transformer will be the best one (toroidal or other) Are 9v 25VA right ?
Thanks again for your help !
In fact, I checked the Squeezebox, and it constantly needs 1100 mA (with wifi and display on). That's why I thought about 2.5A for this power supply.The "positive-voltage regulators" shown by FastEddy will be too difficult for me to built (first DIY work
)I think I can handle with leadbelly's link to "enhanced regulator Altus Audio" but what regulator should I use ? And what values for the other components ? I don't know what kind of transformer will be the best one (toroidal or other) Are 9v 25VA right ?
Thanks again for your help !
what's the brick say?
So, what exactly is written on the existing wall wart (brick) ... 9 VDC ?? 5 DVC ???
What exactly is the output of the existing wall wart ?? ... 25 watts?? 15 watts?? 10 watts??
Knowing this, it is quite easy to add extra filtration to your existing wall wart without resorting to building a complete power supply from scratch ....
So, what exactly is written on the existing wall wart (brick) ... 9 VDC ?? 5 DVC ???
What exactly is the output of the existing wall wart ?? ... 25 watts?? 15 watts?? 10 watts??
Knowing this, it is quite easy to add extra filtration to your existing wall wart without resorting to building a complete power supply from scratch ....
Obviously a switching supply.
Obviously a switching supply.
My favorite inline filter for this type is two inductors, one in each line (+ & -) with a 1.0 uF. Poly or MTF Capacitor (to 35 Volt rating) in between the two cheap inductors ...
-/\/\/\------------------ (+ 5)
...........=|=
---------------/\/\/\ --- (- 0)
Add line connectors to suit ...
I'll do a decent graphic if anyone is interested.
Obviously a switching supply.
My favorite inline filter for this type is two inductors, one in each line (+ & -) with a 1.0 uF. Poly or MTF Capacitor (to 35 Volt rating) in between the two cheap inductors ...
-/\/\/\------------------ (+ 5)
...........=|=
---------------/\/\/\ --- (- 0)
Add line connectors to suit ...
I'll do a decent graphic if anyone is interested.
increases impedence?
Yes, you are right, it does ... and that is exactly what you want from a power supply's output filters = increased impedence to any frequency above 1 Hertz ! === Increasing impedence to noise and hum.
Note that the use of inductors and/or capacitors does not "raise the output impedence" of the output DC, direct current/voltage.
Yes, you are right, it does ... and that is exactly what you want from a power supply's output filters = increased impedence to any frequency above 1 Hertz ! === Increasing impedence to noise and hum.
Note that the use of inductors and/or capacitors does not "raise the output impedence" of the output DC, direct current/voltage.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.