audioguru2
- Apr 6, 2004
- 12,026
- Joined
- Apr 6, 2004
- Messages
- 12,026
Hi Guys, some info: ;D
The LM317 is guaranteed to have an output current of 1.5A with 15V or less across it. In order to protect itself from second breakdown, it cuts back on its output current with more than 15V across it. With a high input voltage and low output voltage, it is guaranteed to supply only 150mA.
The 3A LM350 cuts back its output current with only 10V or more across it.
These regulators shut-down when they get too hot. The TO-220 "T" package has lousy thermal conduction to its heatsink, the TO-3 "K" package is much better.
The ripple rejection is typically (no guarantee) 65dB, which is one/two-thousandth. With a 10uF cap on the adjustment terminal to ground, the ripple rejection is typically 80dB, which is one/ten-thousandth. With the cap, the ripple rejection is guaranteed to be 66dB, which is one/two-thousandth. Pretty darn good.
If you don't use a "star" separate-ground-wires wiring technique, its excellent voltage regulation and ripple rejection are ruined, due to the voltage drop across a common ground wire at high output current.
The LM317 is guaranteed to have an output current of 1.5A with 15V or less across it. In order to protect itself from second breakdown, it cuts back on its output current with more than 15V across it. With a high input voltage and low output voltage, it is guaranteed to supply only 150mA.
The 3A LM350 cuts back its output current with only 10V or more across it.
These regulators shut-down when they get too hot. The TO-220 "T" package has lousy thermal conduction to its heatsink, the TO-3 "K" package is much better.
The ripple rejection is typically (no guarantee) 65dB, which is one/two-thousandth. With a 10uF cap on the adjustment terminal to ground, the ripple rejection is typically 80dB, which is one/ten-thousandth. With the cap, the ripple rejection is guaranteed to be 66dB, which is one/two-thousandth. Pretty darn good.
If you don't use a "star" separate-ground-wires wiring technique, its excellent voltage regulation and ripple rejection are ruined, due to the voltage drop across a common ground wire at high output current.
audioguru2
- Apr 6, 2004
- 12,026
- Joined
- Apr 6, 2004
- Messages
- 12,026
I agree Ante,
But with the bypass transistor you lose short-circuit protection, unless you add an additional transistor, or let the LM317 get hot and shut-down itself and the bypass transistor.
But with the bypass transistor you lose short-circuit protection, unless you add an additional transistor, or let the LM317 get hot and shut-down itself and the bypass transistor.
audioguru2
- Apr 6, 2004
- 12,026
- Joined
- Apr 6, 2004
- Messages
- 12,026
Hi Ante,
Why do we talk about making things cheap? We sound like baby chickens! We aren't mass producing these things. Let's make it good.
There is a guy who is selling "the ultimate" audio amplifier for only $3,300US. Inside, it has only 2 National Semi's power amp ICs. You know how cheap those ICs are. Just think about how much profit that guy is making! The magazine that reviewed it said that it sounded harsh at first, then mellowed after a 6 month break-in period. The first one that the magazine tested blew-up! Then its replacement blew-up too! The designer had to take over there a 3rd amp and a replacement power supply (it was separate).
Why do we talk about making things cheap? We sound like baby chickens! We aren't mass producing these things. Let's make it good.
There is a guy who is selling "the ultimate" audio amplifier for only $3,300US. Inside, it has only 2 National Semi's power amp ICs. You know how cheap those ICs are. Just think about how much profit that guy is making! The magazine that reviewed it said that it sounded harsh at first, then mellowed after a 6 month break-in period. The first one that the magazine tested blew-up! Then its replacement blew-up too! The designer had to take over there a 3rd amp and a replacement power supply (it was separate).
To dispell any myths, I use the LM317 a lot. I am refering to the TO-220 package. I have not had problems overcoming ripple or heat. Only in extreme designs would you even need a heat sink. If you are having problems with this regulator in a design, look at the data sheet and see if you have the wrong application. Next look at how you have designed the board. Most problems I have seen from others with this regulator were actually due to crappy design or layout. This regulator has a good history and is used in many consumer products. I would never purchase a more expensive package from fear of these things listed. It is a total waste of money. Why build things cheap if you are not a manufacturer? Because what you do not spend in parts can buy a lot of beer and pizza!
MP
MP
audioguru2
- Apr 6, 2004
- 12,026
- Joined
- Apr 6, 2004
- Messages
- 12,026
No heatsink?
Oh yeah?
Let's make a simple calculation without using extremes:
1) The LM317 is rated for 1.5A, but let's choose only 1A.
2) It is spec'd with a minimum of 3V across it, but let's choose a reasonable 5V.
3) Some parts of this planet reach 40 degrees C, but let's choose only 25 degrees C.
4) The TO-220 "T" package has a Thermal Resistance for Junction-to-Ambient of 50 degrees C per Watt.
Its Max Allowable Temp is only 125 degrees C, but this non-extreme sample's junction will try to be at 275 degrees C! It will shut-down very quickly.
Try it with only 1/2A, and it will also shut-down. It needs a heatsink.
audioguru2
- Apr 6, 2004
- 12,026
- Joined
- Apr 6, 2004
- Messages
- 12,026
Hi Ante,
For years I have been using the low-dropout LM2931 for all my 9V alkaline and 7.2V Ni-Cad battery low current projects. It sqeezes all the life out of the battery. The 7805 and LM317 get into trouble when the battery is stll new and their idle current is 10 times or more what my circuits draw. The LM2931 has a very low idle current and has all the protection (and more) of the others. I use the small TO-92 size.
For years I have been using the low-dropout LM2931 for all my 9V alkaline and 7.2V Ni-Cad battery low current projects. It sqeezes all the life out of the battery. The 7805 and LM317 get into trouble when the battery is stll new and their idle current is 10 times or more what my circuits draw. The LM2931 has a very low idle current and has all the protection (and more) of the others. I use the small TO-92 size.
Nice attempt at math, audioguru.
(and before you say it, NO, I do not disagree with the manufacturer's data sheet)
This is what I refer to in my post about crappy design techniques.
I can get a half amp out of this regulator easy without a heat sink. (and using the manufacturer's design parameters).
MP
(and before you say it, NO, I do not disagree with the manufacturer's data sheet)
This is what I refer to in my post about crappy design techniques.
I can get a half amp out of this regulator easy without a heat sink. (and using the manufacturer's design parameters).
MP
audioguru2
- Apr 6, 2004
- 12,026
- Joined
- Apr 6, 2004
- Messages
- 12,026
MP,
Certainly you can get 0.5A from an LM317 in a TO-220 package with an ambient temperature of 25 degrees C, without a heatsink.
Just use a regulated input voltage that is 3.0V (no less) to 3.99V (no more) than its output voltage!
Since it will be operating at its thermal limit, it probably won't last long.
That is a crappy design technique.
Just use a reasonable input voltage and put a little heatsink on the thing.
Certainly you can get 0.5A from an LM317 in a TO-220 package with an ambient temperature of 25 degrees C, without a heatsink.
Just use a regulated input voltage that is 3.0V (no less) to 3.99V (no more) than its output voltage!
Since it will be operating at its thermal limit, it probably won't last long.
That is a crappy design technique.
Just use a reasonable input voltage and put a little heatsink on the thing.
audioguru2
- Apr 6, 2004
- 12,026
- Joined
- Apr 6, 2004
- Messages
- 12,026
Hi MP,
OOOps, sorry, I forgot.
You are probably using a fan or dry ice to cool your over-heated LM317. Then 1/2A output is easy.
OOOps, sorry, I forgot.
You are probably using a fan or dry ice to cool your over-heated LM317. Then 1/2A output is easy.
audioguru, you seem to be agitated. I design products for the real world, not as a hobbyist, so I can tell you that proven application backs up what I have told you. I see this attitude a lot in the hobbyist world. A guy wants to make sure he will get all the current he needs for a project, so he buys a BIG transformer just to be sure. But he is not getting more current by doing this. In fact, he is just dissipating the excess voltage as heat. That regulator is not turning that extra voltage into current. The extra is turned into heat. Nothing else. No other benefit.
If the regulator spec sheet says you need 3 volts more input than output, there is no reason to give the regulator more.
You save money with a smaller transformer and you save money with not having to purchase a heat sink.
I am sure you will not be convinced of anything here, but since others will also read this and can learn, let's start with the basics:
The regulator gets hot because it regulates the energy delivered to its load. A linear regulator drops an input voltage to a lower, regulated output, while conducting a current nearly equal to the load current. It must dissipate the excess energy, which is:
Pd = (Vin - Vout)*Iout (Pd is in Watts)
We see from the spec sheet the regulator needs 3 volts. We calculate Pd=(3volts) * 1/2 amp, which is 1.5 Watts.
This energy is dissipated in the form of heat - the regulator die heats up above the ambient temperature with a temperature rise that's proportional to the power dissipated, and to the thermal resistance from inside the die to the ambient environment. This formula is:
Trise = Pd*ThetaJA (ThetaJA is in degC/Watt). For this regulator, ThetaJA is 50 degrees. Thus, Trise = 1.5watts * 50, which equals 75 deg. C. This regulator has a MAX of 125 deg C. in which it shuts down to protect itself from heat over load. Our spec is a little over half of the manufacturer's Max spec. I am not worried about problems with this temperature. It will be hot to the touch and will burn your finger, but it is well within spec.
(by the way, 1 full Amp only brings you a little above the spec. Only a small heat sink needed)
I hope that others can take note of what I am saying here for their design work. You save money not having to buy a heat sink. You save money not having to buy a bigger transformer. And you are not generating a lot of heat that might affect other things in your project. You also benefit from having a smaller package in your design with no bulky heat sinks added. Good design work.
Now you have the formula. Do the math and determine if you really need the heat sink. Think twice about that big transformer.
MP
If the regulator spec sheet says you need 3 volts more input than output, there is no reason to give the regulator more.
You save money with a smaller transformer and you save money with not having to purchase a heat sink.
I am sure you will not be convinced of anything here, but since others will also read this and can learn, let's start with the basics:
The regulator gets hot because it regulates the energy delivered to its load. A linear regulator drops an input voltage to a lower, regulated output, while conducting a current nearly equal to the load current. It must dissipate the excess energy, which is:
Pd = (Vin - Vout)*Iout (Pd is in Watts)
We see from the spec sheet the regulator needs 3 volts. We calculate Pd=(3volts) * 1/2 amp, which is 1.5 Watts.
This energy is dissipated in the form of heat - the regulator die heats up above the ambient temperature with a temperature rise that's proportional to the power dissipated, and to the thermal resistance from inside the die to the ambient environment. This formula is:
Trise = Pd*ThetaJA (ThetaJA is in degC/Watt). For this regulator, ThetaJA is 50 degrees. Thus, Trise = 1.5watts * 50, which equals 75 deg. C. This regulator has a MAX of 125 deg C. in which it shuts down to protect itself from heat over load. Our spec is a little over half of the manufacturer's Max spec. I am not worried about problems with this temperature. It will be hot to the touch and will burn your finger, but it is well within spec.
(by the way, 1 full Amp only brings you a little above the spec. Only a small heat sink needed)
I hope that others can take note of what I am saying here for their design work. You save money not having to buy a heat sink. You save money not having to buy a bigger transformer. And you are not generating a lot of heat that might affect other things in your project. You also benefit from having a smaller package in your design with no bulky heat sinks added. Good design work.
Now you have the formula. Do the math and determine if you really need the heat sink. Think twice about that big transformer.
MP
audioguru2
- Apr 6, 2004
- 12,026
- Joined
- Apr 6, 2004
- Messages
- 12,026
Sure MP,
But you forgot that the IC's die heats 50 degtees C above the ambient temp of a nice and comfy 25 degrees, so the die's temp is 100 degrees (not 75) which is pretty close to its limit. Don't enclose it!
Are you going to make a custom transformer to give exactly 3.0VDC above the output voltage? Hobbyists just pick one off the shelf.
A stock transformer doesn't cost more to have a few volts more output voltage, and certainly isn't any bigger.
Transformers have a tolerance and so does the mains voltage. Will your LM317 shut-down when they are both on the high side?
Are you going to use a huge input capacitor so that additional input voltage is not needed to allow for ripple? Hobbyists just use a reasonable cap, and let the regulator reject the ripple, the way it is designed to do.
Hobbyists just use a heatsink when required, to keep down the costs of those other things and improve reliability. Maybe you should do the same. ;D
But you forgot that the IC's die heats 50 degtees C above the ambient temp of a nice and comfy 25 degrees, so the die's temp is 100 degrees (not 75) which is pretty close to its limit. Don't enclose it!
Are you going to make a custom transformer to give exactly 3.0VDC above the output voltage? Hobbyists just pick one off the shelf.
A stock transformer doesn't cost more to have a few volts more output voltage, and certainly isn't any bigger.
Transformers have a tolerance and so does the mains voltage. Will your LM317 shut-down when they are both on the high side?
Are you going to use a huge input capacitor so that additional input voltage is not needed to allow for ripple? Hobbyists just use a reasonable cap, and let the regulator reject the ripple, the way it is designed to do.
Hobbyists just use a heatsink when required, to keep down the costs of those other things and improve reliability. Maybe you should do the same. ;D
Well, like I said....I did not think that you would get much out of it..... and I think your original comment was that it could not be done at all? You seem to be changing as we go on.
More voltage is not needed to allow for ripple rejection. As I said earlier, the added voltage only goes to heat transfer. You get no benefit from it at all.
I really spent the additional time to explain this for the others who were interested.
BTW, not special transformers to tweak the input voltage to the reg. There are other cheap ways to do this also if it is needed. It is all part of the good design process.
MP
More voltage is not needed to allow for ripple rejection. As I said earlier, the added voltage only goes to heat transfer. You get no benefit from it at all.
I really spent the additional time to explain this for the others who were interested.
BTW, not special transformers to tweak the input voltage to the reg. There are other cheap ways to do this also if it is needed. It is all part of the good design process.
MP
Emiliano86
- Aug 19, 2004
- 12
- Joined
- Aug 19, 2004
- Messages
- 12
hello...i have read all the comments adn i have some questions:
1- i want to design a pSU with 1.2-30V output at the maximun load possible, so what can i do to improve the ripple rejection?, more filter caps?
2- if the soulutio is to use more filer caps, i must use a bigger bridge, do not I?
3- if i use a 15 + 15 seconadary transformer, i can use only one side of it if i need a voltage in the ouput between 1.2 - 15V, and the two sides of the transformer if i need more than 15V, i think that it can improove the current output, but how can i do a voltage detector or selector or something like that?....
OK Thanks!!!!
1- i want to design a pSU with 1.2-30V output at the maximun load possible, so what can i do to improve the ripple rejection?, more filter caps?
2- if the soulutio is to use more filer caps, i must use a bigger bridge, do not I?
3- if i use a 15 + 15 seconadary transformer, i can use only one side of it if i need a voltage in the ouput between 1.2 - 15V, and the two sides of the transformer if i need more than 15V, i think that it can improove the current output, but how can i do a voltage detector or selector or something like that?....
OK Thanks!!!!
Emiliano,
1.Yes bigger filter capacitors will improve the ripple rejection but so will a series inductor (choke).
2.The inrush current will increase with bigger caps so a bigger bridge is necessary.
3 It is possible to use one half of the transformer for a
1.Yes bigger filter capacitors will improve the ripple rejection but so will a series inductor (choke).
2.The inrush current will increase with bigger caps so a bigger bridge is necessary.
3 It is possible to use one half of the transformer for a
