Switch Mode PSU history

Discussion in 'General Electronics Chat' started by sirch2, Dec 26, 2013.

  1. sirch2

    Thread Starter Well-Known Member

    Jan 21, 2013
    1,008
    351
    I was chatting with someone a few days ago and he got on to wondering why SMPS had taken over the world? I gave him the usual reply about efficiency etc. But it occurred to me that I didn't really know why, about a decade ago, SMPS started to be much more common for wall-warts etc.

    My hunch is that the cost of mains voltage silicon came down to a point where it became cost effective for all PSUs to be switched mode. I am right in this assumption? Is there other reasons?
     
  2. paulktreg

    Distinguished Member

    Jun 2, 2008
    611
    120
    Size and efficiency.

    You've only to investigate the size of the transformer should gaming PC's still be powered by linear units. A 1000VA transformer weighs around 7.5kg ~ 10.0kg!
     
  3. MikeML

    AAC Fanatic!

    Oct 2, 2009
    5,450
    1,066
    And this correlates with the destruction of the HF Radio Spectrum. I am a Ham, and in the last ten or so years, the 160m to 10m amateur bands have been mostly wiped out in urban areas due to noise from all the damn switchers. I had to move to a rural environment to be able to hear signals on those bands. It used to be that hams interfered with their neigbors TV sets; now those TV sets (and their switching power supplies) interfere with users of the HF radio spectrum.
     
  4. Metalmann

    Active Member

    Dec 8, 2012
    700
    223
    "It used to be that hams interfered with their neigbors TV sets"


    I remember those days.;)
     
  5. sirch2

    Thread Starter Well-Known Member

    Jan 21, 2013
    1,008
    351
    As I said that is the reason that is always given BUT why weren't we using them in 1950 or 1990, they only became common for wall-warts over the last decade or so, what was the enabling factor???
     
  6. shteii01

    AAC Fanatic!

    Feb 19, 2010
    3,392
    497
    Maybe some others are missing your point. It seems, to me, that you are making emphasis on the wall-warts. In this case I propose this scenario. The growth in computer industry enabled the expansion and mass production of the switching psu. Since the production capacity and engineering know-how is already in place, the manufacturers of switching psu started looking for new markets, such as wall-warts.
     
  7. ian field

    Distinguished Member

    Oct 27, 2012
    4,415
    784
    TVs used to have really crappy IF strips - interference from all sources was much reduced when SAWFs came into use.
     
    Metalmann likes this.
  8. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    The biggest single reason is the development of cheap power MOSFETs. I worked as a switcher designer in the 80's and that's when we transitioned from huge (and very expensive) bipolar transistors to power FETs because the price of FETs got cheaper and they don't take as much current to drive.

    The other reason is the high level of integration of the controller/driver made switcher designs so cheap.

    In the end, the almighty dollar rules the earth and switchers are the cheapest solution.
     
    Metalmann and #12 like this.
  9. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    Check the price $$$$$$

    Pacific Rim manufacturers can crank out off-line switchers in small wall warts for a few bucks.
     
  10. sirch2

    Thread Starter Well-Known Member

    Jan 21, 2013
    1,008
    351
    Thanks bountyhunter, that's pretty much what I thought
     
  11. GopherT

    AAC Fanatic!

    Nov 23, 2012
    6,058
    3,820
    As MOSFETS got better (faster), the switching frequencies increased. Increased switching frequency means smaller inductors are possible for the same level of ripple. Also, as frequency goes up, the need for high saturation iron powder-based inductors went away. Now, at switching speeds above 1MHz, cheap and small ferrite cores are possible even at high currents.

    I was told at the Power Electronics show in 2005 that the inductors were the most expensive single part of a SMPS in an ATX supply. Followed by the heat sink. Less heat and faster switching frequency make the whole unit cheaper and smaller.
     
  12. #12

    Expert

    Nov 30, 2010
    16,298
    6,810
    I'm with bountyhunter. Fast, cheap mosfets and IC's to drive them became available. Try making a switching supply with only bipolar transistors, no IC's, and you will see the problem.
     
  13. vk6zgo

    Active Member

    Jul 21, 2012
    677
    85
    Bipolar transistor switch mode power supplies were pretty much universally used in TV sets from the early '70s.onward.

    Early example:philips K9 series.

    It is only much more recently,with the advent of high power switching FETs that switch modes have become viable in "wallwarts" & the like.

    Huge?,Expensive?

    BU208D transistors,(same package size as a 2N3055),or similar,were used in many 1980s TV switch modes,& were about $3.00 in "one off" quantities at Dick Smith's.
     
    Last edited: Dec 26, 2013
  14. Brownout

    Well-Known Member

    Jan 10, 2012
    2,375
    998
    THE_RB likes this.
  15. vk6zgo

    Active Member

    Jul 21, 2012
    677
    85
  16. paulktreg

    Distinguished Member

    Jun 2, 2008
    611
    120
    No IC's!

    How far back are we going here?
     
  17. THE_RB

    AAC Fanatic!

    Feb 11, 2008
    5,435
    1,305
    If you want to go FAR back in time, I came across some mechanical SMPS in a junkyard. I think they were Korean war era.

    They used an oscillating solenoid core within some cylindrical coils, and a high quality adjustable contact set at the end to make it oscillate and probably to do rectification too, I can't remember the specifics. They were military spec equipment labelled as DC-DC converters.

    So you don't even need semiconductors or vacuum tubes to make a SMPS! ;)
     
  18. vk6zgo

    Active Member

    Jul 21, 2012
    677
    85
    In 1950,Semiconductors were something with a great future--somewhere a decade or so down the track.
    Effectively,the only game in town was to use Thermionic devices (vacuum & gas tubes).
    I won't say there never was a vacuum tube switchmode supply,as someone will probably find one on the 'Net!:D
    It would be very large,& probably considerably less efficient than a transformer supply.

    Later in the 1950s,as Semiconductors started to invade the market,interestingly enough,one of the first practical applications of power semiconductors was the use of germanium power transistors in Inverter circuits replacing electromechanical "vibrator" supplies to run tube equipment from 6 & 12v batteries.

    These were switchmode supplies in all but name!
    Early transistor voltage ratings were not sufficient to allow the design of switch mode Mains supplies.

    Up into the late 1970s "Plugpacks/Wallwarts" were very rare,as most Mains equipment included its power supply,(analog or switchmode) inside the case,with different models made for different markets.

    It became cheaper to produce a single device for many markets & use a "wallwart" as only the latter needed to meet the different specs & approvals for each country.

    Initially,the power requirements were modest,so it was easier to use transformer supplies,but as the power demand became greater,& power FETs became common,wallwarts moved to switchmode technology.
     
  19. vk6zgo

    Active Member

    Jul 21, 2012
    677
    85
    Sounds like a "vibrator" type supply.
    These were common back then,--smaller ones were used in car radios.

    In the early '60s,"Van Ruyten" 32V DC to 240V AC Inverters were widely used in rural Australia to run Mains type TV sets from farm home lighting supplies.
    These things ran at 50Hz, & the "vibrators" were huge (don't go there!!:D).
    I think the car ones were 400Hz.
     
  20. #12

    Expert

    Nov 30, 2010
    16,298
    6,810
    If only I had known this in 1975 I could have smoked those hot shot engineers that kept sending in toroid designs that didn't have enough hole to put all the turns through.:D

    But seriously, I was thinking of a power supply company I worked for in 1975, and the idea of producing a switching design for all of their products still seems daunting to me. How about +/- 35V @6A +/- 1/2% followed by a 28V @4A +/- .01% regulation for line and load combined, with .06 mv RMS ripple at full load. I could do that in analog, but switchers are still a study in progress.
     
    Last edited: Dec 27, 2013
Loading...