Define Linear Devices & Switch Mode Power Supplies? - Electronic Circuits Projects Diagrams Free

Suraj143 · 4th July 2007, 08:28 AM

*What’s the meaning of LINEAR devices? Why they call like that?

*What is the meaning of switch mode power supply? What’s the difference between normal power supply and a switch mode power supply?

Thanks

bananasiong · 4th July 2007, 08:38 AM

http://en.wikipedia.org/wiki/Switching_power_supply
Wiki tells you the difference

Papabravo · 4th July 2007, 11:20 AM

Quote:

Originally Posted by Suraj143

*What’s the meaning of LINEAR devices? Why they call like that?

*What is the meaning of switch mode power supply? What’s the difference between normal power supply and a switch mode power supply?

Thanks

In general, the term linear, applies to any device or system where the relationship between the input and the output can be described by the equation of a straight line.

Code:

y = mx + b

where m is the slope of the line
and b is the y-intercept

That's why they call it like that.

TheVictim · 4th July 2007, 06:36 PM

Switching power supplies work in a similar manner to the Phased Locked Loop oscillator in that the actual output is compared with what the output SHOULD be, which generates a reference value known as the Error Voltage. This is feed back into part of the input to correct the problem, until there is no error voltage coming from the comparison circuit. This is why switching power supplies don't vary their voltage and PLL synthesized tuners don't frequency drift.

**Nigel Goodwin** · 4th July 2007, 07:17 PM

Quote:

Originally Posted by TheVictim

Switching power supplies work in a similar manner to the Phased Locked Loop oscillator in that the actual output is compared with what the output SHOULD be, which generates a reference value known as the Error Voltage. This is feed back into part of the input to correct the problem, until there is no error voltage coming from the comparison circuit. This is why switching power supplies don't vary their voltage and PLL synthesized tuners don't frequency drift.

Sorry, but that's completely untrue! - linear power supplies work in that exact same way, using feedback to correct for any variations in the output.

Hero999 · 4th July 2007, 09:11 PM

The difference is that switching power supplies use a switching stage on the output rather than a transistor that never saturates.

Basically linear regulators have a class A output stage and switching regulators have a class D output stage.

speakerguy79 · 5th July 2007, 05:06 PM

Actually, the true definition of 'linear' in linear systems is this:

f(x) + f(y) = f(x+y)

Which is quite interesting, because contrary to the completely natural seeming assumption that a line equation would obviously be a linear function, y = mx + b is surprisingly not a linear equation, unless b = 0. Technically, a linear system cannot have a DC offset.

This is, of course, if I remember my linear systems and signals course from college oh so many years ago. I only remember this because of a question essentially based on this on a test that killed 90+ % off the class.

Optikon · 5th July 2007, 08:52 PM

Quote:

Originally Posted by speakerguy79

Actually, the true definition of 'linear' in linear systems is this:

f(x) + f(y) = f(x+y)

Which is quite interesting, because contrary to the completely natural seeming assumption that a line equation would obviously be a linear function, y = mx + b is surprisingly not a linear equation, unless b = 0. Technically, a linear system cannot have a DC offset.

This is, of course, if I remember my linear systems and signals course from college oh so many years ago. I only remember this because of a question essentially based on this on a test that killed 90+ % off the class.

Have a look:

http://en.wikipedia.org/wiki/Linear_system

The definition does not preclude "offsets." which, makes sense. We all know y=mx + b is linear even for non-zero b.

Papabravo · 5th July 2007, 11:24 PM

Quote:

Originally Posted by speakerguy79

Actually, the true definition of 'linear' in linear systems is this:

f(x) + f(y) = f(x+y)

Which is quite interesting, because contrary to the completely natural seeming assumption that a line equation would obviously be a linear function, y = mx + b is surprisingly not a linear equation, unless b = 0. Technically, a linear system cannot have a DC offset.

This is, of course, if I remember my linear systems and signals course from college oh so many years ago. I only remember this because of a question essentially based on this on a test that killed 90+ % off the class.

Must have been a third rate school with third world TA that slipped that wowser by you. Are you unfamiliar with the concept of translation of the coordiante axes as a linear operator?

speakerguy79 · 6th July 2007, 02:37 AM

Sorry, top ten engineering school in the country. Maybe you guys should go through the math before telling someone they're wrong. Example:

f(x) = 2x + 5

f(2) + f(4) = 9 + 13 = 21

f(2+4) = 17

f(2) + f(4) != f(2+4)

Since f(x) + f(y) must equal f(x+y) for a system to be a linear system, f(x) is not linear (as far as linear system theory goes; it's still a straight line on graph paper though).

This is straight out of the B.P. Lathi textbook that most colleges use.

Papabravo · 6th July 2007, 05:54 AM

Have it your way, it's an advanced treatment.

Your basic definition apllies to linear maps from one vector space to another. In this case adding the constant makes it an affine map. This is way beyond the original poster's question and our answers were given in the context of elementary algebra. I'm pretty sure the OP did not have this higher treatment in mind when he asked his original question.

Hero999 · 6th July 2007, 01:34 PM

Smart arse.

As far as I'm concerned linear = analogue = an infinite number of quantities and switching = digital = two possible states (on and off).

Optikon · 6th July 2007, 01:41 PM

Quote:

Originally Posted by speakerguy79

Sorry, top ten engineering school in the country. Maybe you guys should go through the math before telling someone they're wrong. Example:

f(x) = 2x + 5

f(2) + f(4) = 9 + 13 = 21

f(2+4) = 17

f(2) + f(4) != f(2+4)

Since f(x) + f(y) must equal f(x+y) for a system to be a linear system, f(x) is not linear (as far as linear system theory goes; it's still a straight line on graph paper though).

This is straight out of the B.P. Lathi textbook that most colleges use.

I don't think you are applying the definition right. Look at the Wikipedia definition. That's what I see in all signals books.

You need x1(t) & x2(t) inputs and compute y1(t) & y2(t) outputs from a system under examination. The system under examination is linear if for scalars a & b,

a*y1(t) + b*y2(t) = H{a*x1(t) + b*x2(t)}

if we use x1(t) = 2*t+5 and x2(t) = 3*t+1 and choose the system to be say differentiation d/dt then try it.

I worked it out and it shows that the *SYSTEM* is linear. And we all know the operator d/dt (differentiation) is linear. If the system is constant gain of 1 just maps input to output, that too meets definition of linearity. And the inputs have offsets to them.

Your example is something else. but to test a system for linearity, you must try two functions on input.

speakerguy79 · 6th July 2007, 01:48 PM

Yeah, I can be a smart ass for sure

But to answer the OP's post:

Linear regulators take a noisy voltage in, step it down usually by 2V or more to your desired voltage, and filter out a lot of the noise. They are inefficient because they dissipate power in direct relationship to how much voltage drop you have across them, and the current through them.

Ex. A regulator with 35V in and 5V out will dissipate a whopping 45W when supplying 1.5Amps.That's 45W dissipated as heat to 7.5W delivered as power

This is of course an extreme example and no one would ever use a regulator like this.

The big plusses with linear regs are very very low ripple and noise, and good output accuract (1% easily acheivable). Also really really cheap. Good to use in analog cktry.

Switchers - don't know much about how they are designed, but they are much more efficient than linear regulators with the drawback of high frequency noise introduced onto the power supply lines. Switchers actually increase in efficiency as output power goes up, maxing out at about 80% of their full output power (at anywhere from 75 to 95 percent efficiency). Switchers can be quite a bit more expensive.

I only use switchers where efficiency is a concern or I need to generate multiple and/or positive and negative voltages from a single DC voltage source. Other than that I always use linears.

Optikon · 6th July 2007, 04:17 PM

Quote:

Originally Posted by speakerguy79

Yeah, I can be a smart ass for sure

But to answer the OP's post:

Linear regulators take a noisy voltage in, step it down usually by 2V or more to your desired voltage, and filter out a lot of the noise. They are inefficient because they dissipate power in direct relationship to how much voltage drop you have across them, and the current through them.

Ex. A regulator with 35V in and 5V out will dissipate a whopping 45W when supplying 1.5Amps.That's 45W dissipated as heat to 7.5W delivered as power

This is of course an extreme example and no one would ever use a regulator like this.

The big plusses with linear regs are very very low ripple and noise, and good output accuract (1% easily acheivable). Also really really cheap. Good to use in analog cktry.

Switchers - don't know much about how they are designed, but they are much more efficient than linear regulators with the drawback of high frequency noise introduced onto the power supply lines. Switchers actually increase in efficiency as output power goes up, maxing out at about 80% of their full output power (at anywhere from 75 to 95 percent efficiency). Switchers can be quite a bit more expensive.

I only use switchers where efficiency is a concern or I need to generate multiple and/or positive and negative voltages from a single DC voltage source. Other than that I always use linears.

This is a good general conception of linear versus switch-mode. I would only add that the more efficient switchmode power supplies can be made smaller (due to smaller magnetics due to higher frequency of operation) and that can be critical in some size demanding applications.

Any analog work I do, I avoid switchers like the plague because of the noise issues.

Also note, it takes a pretty experienced designer to squeeze every last bit of efficiency out of a switchmode design. 80% is acheivable with your eyes closed in most cases. In my experience to get into the 90%'s requires much more work. The companies who sell switchmode controllers are making it easier on designers these days to acheive the higher efficiencies.

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Latest
Need help testing AC/DC power supplies	wasman	Electronic Projects Design/Ideas/Reviews	3	25th July 2003 03:56 AM
AMP ratings on power supplies	vanhin	General Electronics Chat	2	28th June 2003 06:12 PM
RE:- help!! AC switch mode supply for low voltage lighting	mikegaylor	Electronic Projects Design/Ideas/Reviews	1	19th April 2003 02:10 PM
Help with AC switch mode supply	mikegaylor	General Electronics Chat	0	18th April 2003 06:12 PM
Switch mode power supply	Nonx	General Electronics Chat	10	7th March 2003 04:18 PM

4th July 2007, 08:28 AM	(permalink)
Suraj143	Define Linear Devices & Switch Mode Power Supplies? What’s the meaning of LINEAR devices? Why they call like that? What is the meaning of switch mode power supply? What’s the difference between normal power supply and a switch mode power supply? Thanks

4th July 2007, 08:38 AM	(permalink)
bananasiong	http://en.wikipedia.org/wiki/Switching_power_supply Wiki tells you the difference __________________ Superman returns..

4th July 2007, 06:36 PM	(permalink)
TheVictim	Switching power supplies work in a similar manner to the Phased Locked Loop oscillator in that the actual output is compared with what the output SHOULD be, which generates a reference value known as the Error Voltage. This is feed back into part of the input to correct the problem, until there is no error voltage coming from the comparison circuit. This is why switching power supplies don't vary their voltage and PLL synthesized tuners don't frequency drift.

4th July 2007, 09:11 PM	(permalink)
Hero999	The difference is that switching power supplies use a switching stage on the output rather than a transistor that never saturates. Basically linear regulators have a class A output stage and switching regulators have a class D output stage. __________________ I also post at the following sites: http://www.stop-microsoft.org http://www.heated-debates.com Screen name: Aloone_Jonez

5th July 2007, 05:06 PM	(permalink)
speakerguy79	Actually, the true definition of 'linear' in linear systems is this: f(x) + f(y) = f(x+y) Which is quite interesting, because contrary to the completely natural seeming assumption that a line equation would obviously be a linear function, y = mx + b is surprisingly not a linear equation, unless b = 0. Technically, a linear system cannot have a DC offset. This is, of course, if I remember my linear systems and signals course from college oh so many years ago. I only remember this because of a question essentially based on this on a test that killed 90+ % off the class.

6th July 2007, 02:37 AM	(permalink)
speakerguy79	Sorry, top ten engineering school in the country. Maybe you guys should go through the math before telling someone they're wrong. Example: f(x) = 2x + 5 f(2) + f(4) = 9 + 13 = 21 f(2+4) = 17 f(2) + f(4) != f(2+4) Since f(x) + f(y) must equal f(x+y) for a system to be a linear system, f(x) is not linear (as far as linear system theory goes; it's still a straight line on graph paper though). This is straight out of the B.P. Lathi textbook that most colleges use. Last edited by speakerguy79; 6th July 2007 at 02:45 AM.

6th July 2007, 05:54 AM	(permalink)
Papabravo	Have it your way, it's an advanced treatment. Your basic definition apllies to linear maps from one vector space to another. In this case adding the constant makes it an affine map. This is way beyond the original poster's question and our answers were given in the context of elementary algebra. I'm pretty sure the OP did not have this higher treatment in mind when he asked his original question. Last edited by Papabravo; 6th July 2007 at 06:35 AM.

6th July 2007, 01:34 PM	(permalink)
Hero999	Smart arse. As far as I'm concerned linear = analogue = an infinite number of quantities and switching = digital = two possible states (on and off). __________________ I also post at the following sites: http://www.stop-microsoft.org http://www.heated-debates.com Screen name: Aloone_Jonez

6th July 2007, 01:48 PM	(permalink)
speakerguy79	Yeah, I can be a smart ass for sure But to answer the OP's post: Linear regulators take a noisy voltage in, step it down usually by 2V or more to your desired voltage, and filter out a lot of the noise. They are inefficient because they dissipate power in direct relationship to how much voltage drop you have across them, and the current through them. Ex. A regulator with 35V in and 5V out will dissipate a whopping 45W when supplying 1.5Amps.That's 45W dissipated as heat to 7.5W delivered as power This is of course an extreme example and no one would ever use a regulator like this. The big plusses with linear regs are very very low ripple and noise, and good output accuract (1% easily acheivable). Also really really cheap. Good to use in analog cktry. Switchers - don't know much about how they are designed, but they are much more efficient than linear regulators with the drawback of high frequency noise introduced onto the power supply lines. Switchers actually increase in efficiency as output power goes up, maxing out at about 80% of their full output power (at anywhere from 75 to 95 percent efficiency). Switchers can be quite a bit more expensive. I only use switchers where efficiency is a concern or I need to generate multiple and/or positive and negative voltages from a single DC voltage source. Other than that I always use linears. Last edited by speakerguy79; 6th July 2007 at 01:51 PM.