Why Regulator & DC-DC Switching IC's ???

[ViReN]

Why Regulator & DC-DC Switching IC\'s ???

Hello All...

Since, it has been a while (rather few moments), .... after (working on) successful Transistor based step up circuits (with help of U great guys)....

I wonder... Why are the guys ( @ national, maxim, on-semi, torex, mircel) working so much on the Regulator & DC-DC Switching IC's ... I mean, if the IC's just do the switching (internal switching, regulation) and regulation (only) .. same can be achieved using the transistors ... (be it SMD's or normal ones) ....

as long as you get +ve di/dt across an inductor (boost circuits), what ever makes this thing happen... will do....isn't this the main principle behind these circuits. .. or take for that matter... switched capacator (charge pump circuits)

I mean... we do need an EXTERNAL inductor.... we do need an EXTERNAL (BJT / MOSFET) switches (especially in case of high current LED Drivers)... we STILL need the external components .. isn't it ? and those EXTERNAL components are the ones.. that really make the circuit work !....

I.C. = Integrated Circuit.... they should integrate every thing... except the passive components (hope my terminology is correct) (like inductors & Capacators) ? or can that also be emulated / simulated ? (inside an I.C.) ???

I may be totally wrong in the way i am thinking.... but i guess.. i have put my views to you... may be.. you guys could clear my misconceptions.... (if they really are) ....


Thanks & Regards,
ViReN

 

[jtr1962]

Re: Why Regulator & DC-DC Switching IC\'s ???

Parts count is what it's all about here. The simple two-transistor step-up driver in the other thread has only 6 parts but it's not regulating. I can add regulation with four more parts (for a total of ten) but it doesn't work quite as well as a purpose made IC. Using the CAT32 LED driver instead of discrete transistors lowers my total parts count from ten to five (and one of those is an input bypass capacitor that may not always be needed) and also lowers the total real estate needed by the circuit. In fact, the footprint of the CAT32 (and the Zetex ZXSC300) is the same as that of an SMD transistor. If you look at total cost, I may be spending $0.25 more using the driver IC but I'm saving far more than that on assembly costs if I were to mass-produce it. In addition, as you may have noticed circuits with discrete components are far more sensitive to component variations than those using purpose-made ICs with feedback loops. A circuit may work fine and then if you can no longer get a given inductor from a supplier you may need to tweak it all over again. Reliability is also higher with fewer parts.

In a nutshell the reasons to use ICs are parts count, footprint, cost, repeatability, and reliability. When LED driver ICs cost several dollars each it may have made more sense to go with discrete designs but now that many are well under a dollar I only use discrete components for low production volume designs, or in cases where I can't find an IC to do exactly what I want.

 

[NewBie]

Re: Why Regulator & DC-DC Switching IC\'s ???

Additionally, efficient designs are *much* more complex than that simplistic circuit you assembled.

On several of those switcher IC's, if you look carefully, you'll note there can be easily over 4,000 transistors (bipolar and mosfet) inside them. A good number of these are a bunch of mosfets/bipolar in parallel, to lower on resistances (lower heat generation/higher efficiencies).

Take a look at this part, you'll note at the bottom of page 16, 4344 transistors inside the part:

http://pdfserv.maxim-ic.com/en/ds/MAX5035.pdf

You can also add in, stability over temperature.

They slowly get more complex, such as this part, with 6688 transistors, and the main switching parts are not contained within the device:

http://pdfserv.maxim-ic.com/en/ds/MAX8529.pdf

 

[HarryN]

Re: Why Regulator & DC-DC Switching IC\'s ???

It is only a matter of time before one of the IC companies figures out that there is indeed a market for a totally integrated (active + passive) 1 chip solution for LED drive.

I am not sure whether or not it will achieve the efficiency levels of some of the higher end drivers that have been built by "The Dougs", but probably will be pushing 90 % pretty hard.

Every necesary circuit element already exists in existing IC's, they just use them for DC - DC voltage management, instead of single IC curent management.

 

[PeterB]

Re: Why Regulator & DC-DC Switching IC\'s ???

At the moment it is neither cost efficient nor technically usefull to integrate the inductor.
One mm2 on a silicon chip costs about 10cent and you are limited to flat inductors (increasing the area vs. a standard 3D inductor).
Inductances in the range of µH would be extremely large and therefore expensive. Typically used on chip inductances are in the range of nH (they are used for RF applications). Also the wiring is very thin (0.1µ-2µ), therefore you can't achieve low series resistance.

 

[NewBie]

Re: Why Regulator & DC-DC Switching IC\'s ???

[ QUOTE ]
HarryN said:
It is only a matter of time before one of the IC companies figures out that there is indeed a market for a totally integrated (active + passive) 1 chip solution for LED drive.

I am not sure whether or not it will achieve the efficiency levels of some of the higher end drivers that have been built by "The Dougs", but probably will be pushing 90 % pretty hard.

Every necesary circuit element already exists in existing IC's, they just use them for DC - DC voltage management, instead of single IC curent management.

[/ QUOTE ]

It is rather easy these days, if you have the skill, to blow right past 90% in reality. 93% efficiency is typical, and you can work your way up to 98%, but to get there you usually drive the cost and size up.

 

[ViReN]

Re: Why Regulator & DC-DC Switching IC\'s ???

Oh... So there is a great deal of logic....

I will still be happy.. and will eagerly wait for the IC's... that require only Passive components.. like (large...(low internal resistance)) Inductors & Capacators .. along with a few configuring resistors...
may be the total part count could be as less as 1 IC, 1 Inductor, 1 Resistor (for o/p Voltage / current configuration) and 1 Capacator (may be for o/p filter ... or in conjunction with inductor) ... totaling part count to 4 ... and YET... be rated in range of 1 - 3 Amps... having efficncy of 90 - 98 % !!!...

Thanks & Regards,
ViReN

 

[NewBie]

Re: Why Regulator & DC-DC Switching IC\'s ???

There are ICs that only take passive components, they've been around for some time now. Except add one more resistor for voltage, and one cap for the output (other is for input)

And they are rated for 4A switching, in the 95-96% efficiency range.

Those have been out for a few years.

 

[ViReN]

Re: Why Regulator & DC-DC Switching IC\'s ???

NewBie, Can you please name a few numbers ?

Thanks & Regards,
ViReN

 

[NewBie]

Re: Why Regulator & DC-DC Switching IC\'s ???

There are several, the first I recall is the TI TPS61030RSAR. You have to be careful with it, as when it is not properly compensated, it cannot protect itself, and it will go pop. Improper board layout will also cause this part to destroy itself, as noise gets into the part and it makes it difficult for the chip to decide what is going on, and will cause it to go pop also.

 

[ViReN]

Re: Why Regulator & DC-DC Switching IC\'s ???

Kool.... I will look for getting these things on the net... as these are of course not available locally.. where i live.

 

[HarryN]

Re: Why Regulator & DC-DC Switching IC\'s ???

[ QUOTE ]
PeterB said:
At the moment it is neither cost efficient nor technically usefull to integrate the inductor.
One mm2 on a silicon chip costs about 10cent and you are limited to flat inductors (increasing the area vs. a standard 3D inductor).
Inductances in the range of µH would be extremely large and therefore expensive. Typically used on chip inductances are in the range of nH (they are used for RF applications). Also the wiring is very thin (0.1µ-2µ), therefore you can't achieve low series resistance.

[/ QUOTE ]

Hi Peterb, I think the answer depends a little more on your perspective of life. Since I purchase boards and modules for my buck converters, I am paying nominally US$ 10 - 30 depending on what I buy and where I get it. An example is an LED Dynamics buckpuck 3021 - nominally $ 25 - 30 for a complete, protected module ready to go, in low volume. Georges80 makes a nice small board for $ 15-20 that I have used as well. Both appear to be well made, but there are times when they are just to large for my targeted light.

So for me personally, admittedly one with limited skills and no metrology equipment, the only thing I can do is either resistors or buy a module / finished board. So, how much would I pay for a "single integrated chip" solution? Certainly $ 10, and for one light I am playing with, maybe $ 15, although, not entirely happily.

As far as putting the inductors on chip, modern microprocesors routinely have 10 layers of metal, so at least some 3D use is possible. I think the IC foundry price for a metal layer is now between $ 500 - 1,000 / metal layer / 200mm wafers, so the cost is within a viable range.

If the package is sufficiently integrated, with large enough leads to manually solder on to, I could skip the circuit board altogether, making my life very convenient.

 

[ViReN]

Re: Why Regulator & DC-DC Switching IC\'s ???

HarryN,
There is also a posibility of using Lesser value Inductance in an On-Chip Applicaiton, where the switching frequency is about a few MHz ... but one problem will still remain... and that is of Magnetic Interference / Noice... when other components are placed so close.....

-ViReN