SMPS needing a load to startup myth?

[Gareth Magennis]

Hmm, I keep hearing such things, but so far I haven't come across a SMPS
that will not sit there quite happily supplying voltage, without a dummy
load in sight.

Granted I only work with Audio, requiring 5v, +/- 15v for signal,and +/-
many more volts and amps for Power Amps.



Gareth.

 

One problem is that in a siongle ended flyback (rather than forward) convertor there is usually a little spike at the beginning of the pulse to be rectified. when there is a load this spike is suppressed, but if there is no load the rectigfied voltage can go too high.

If it is a leg of the supply not in the feedback loop all you gt is too much voltage there, which might be a real problem. If it is on the leg in the feedback loop then it can do strange things because it will be out of range..

But then they couldn't very well puit a bleeder on the sensed leg and let the thing blow all the other filters, and putting bleeders on all of them big enough might defeat the purpose of efficiency.

Now, as Robet said, things have changed. not only are the devices better but they know alot more about switching and all that. In fact most SMPSes now are not single ended. They are usually above resonance push pull. A totem pole type arraigement drives the transformer with pretty much brute force both way and usually full wave rectification is needed, or at least advantageous on the secondaries. I think the main thing is to avois core saturation because if the core saturates in that type of convertor it might rais the resonance so high as to take it completely out of regulation. We don't want that of course, and neither do they.

Wwhen I take pen oin hand I do things a little bit differently. I overdo everything so noting can go wrong. they can't do that on production units because they would make (considerably) less money.

 

[John-Del]

On Monday, March 30, 2015 at 5:37:33 PM UTC-4, Gareth Magennis wrote:

Hmm, I keep hearing such things, but so far I haven't come across a SMPS
that will not sit there quite happily supplying voltage, without a dummy
load in sight.

Granted I only work with Audio, requiring 5v, +/- 15v for signal,and +/-
many more volts and amps for Power Amps.



Gareth.

I've seen a couple in recent years that would float above run volts in standby but settle down to exact voltages when full load was added. Old TV techs might remember the Sonys from the late 70s that would blow up at low AC, high AC, no load, too much load, etc. Sony IMO wouldn't pay anyone royalties so used the most convoluted Rube Goldberg inspired designs of their own making. They regulated well, but were so hair trigger that only Sony semiconductors could dare be installed during repair otherwise immediate destruction would occur.

 

[RobertMacy]

On Mon, 30 Mar 2015 14:37:29 -0700, Gareth Magennis
<sound.service@btconnect.com> wrote:

Hmm, I keep hearing such things, but so far I haven't come across a
SMPS that will not sit there quite happily supplying voltage, without a
dummy load in sight.

Granted I only work with Audio, requiring 5v, +/- 15v for signal,and
+/- many more volts and amps for Power Amps.



Gareth.

back when these first came out [remember Boschert?], most OEM units did
require a load. Thus, the start of the myth.

some did NOT work during brown out and would burn up.

Their specs actually gave a 'range' of input voltage AND a 'range' of
load. Outside those ranges and catastrophe.

But that seems stupid because to get from 0 AC to ACmains values, you MUST
be outside that spec range AND to attach a load really makes sense to be
able to go through a zero to max load range, too.

Thus, as components came down in price, functionality went up, but the
memories of having a fire burn up your SMPS at 50 Vac stuck.

 

[RobertMacy]

On Mon, 30 Mar 2015 14:37:29 -0700, Gareth Magennis
<sound.service@btconnect.com> wrote:

Hmm, I keep hearing such things, but so far I haven't come across a
SMPS that will not sit there quite happily supplying voltage, without a
dummy load in sight.

Granted I only work with Audio, requiring 5v, +/- 15v for signal,and
+/- many more volts and amps for Power Amps.



Gareth.

oops, forgot. at zero load where some of the SMPS's 'skittle' around; they
will fail EMC. Good design [by definition] no problem.

 

Gareth Magennis <sound.service@btconnect.com> wrote:

Hmm, I keep hearing such things, but so far I haven't come across a
SMPS that will not sit there quite happily supplying voltage, without
a dummy load in sight.

I have one that delivers +5 V, +12 V, and a couple of other things at a
couple hundred watts total - along the lines of a PC power supply but
not the same shape. It needs a load on the +5 V to start up; I use a
12 V automotive lamp. It probably dates from the early 1990s.

I recall talking to someone online who had a PC power supply that had
1/4 watt resistors across the outputs, each sized to draw a few mA from
the output. This was too much power for one of them and it had burned,
but I think there were also some other failures too.

I also have a couple of 5 V, 0.5 A wall-wart cell phone chargers, and
they at least don't draw much current from the AC line/mains with no
load, according to my Kill-a-watt meter. If I unplug/replug the charger
several times, I can sometimes catch a 1 W reading on the meter for a
second, but it then goes back to 0 W.

For the cell phone chargers, I suspect they are deliberately designed to
"gulp" a little power when first plugged in, charge a cap, and then look
for a load every so many milliseconds. If the cap gets low, it takes
another "gulp", but it doesn't have to do that very often. The goal is
that you can leave them plugged in to the wall without wasting a lot of
energy.

Matt Roberds

 

[Michael Black]

On Mon, 30 Mar 2015, Gareth Magennis wrote:

Hmm, I keep hearing such things, but so far I haven't come across a SMPS
that will not sit there quite happily supplying voltage, without a dummy load
in sight.

Granted I only work with Audio, requiring 5v, +/- 15v for signal,and +/-
many more volts and amps for Power Amps.

I remember buying industrial surplus switching supplies 20 years ago at
the local surplus outlet, and some of them came with load resistors built
in. "It's cheaper to build higher current supplies than is needed, but in
case people need low current, we'll throw some of the unneeded capacity
away by adding built in load resistors".

So have computer switching supplies changed in design in recent times?
That was often where you saw the requirement for a load before they
operated, but then that was often the only switching supplies many people
came upon.

Now they are everywhere, tiny switching supplies generating low currents
(well for a switching supply 750mA or 1A is low, though for a transformer
supply, that's a fairly decent size transformer), and while I can't say
I've measured their voltage without load, they seem to be relatively well
designed.

Michael

 

[JW]

On Mon, 30 Mar 2015 22:37:29 +0100 "Gareth Magennis"
<sound.service@btconnect.com> wrote in Message id:
<umjSw.31$TC7.16@fx25.am4>:

Hmm, I keep hearing such things, but so far I haven't come across a SMPS
that will not sit there quite happily supplying voltage, without a dummy
load in sight.

Granted I only work with Audio, requiring 5v, +/- 15v for signal,and +/-
many more volts and amps for Power Amps.

I work on Tek TDS700 series of scopes quite often, they will not work if
there is no load. In fact, if you disconnect the two loads on the +25VDC
rail (The HV deflection board and the fan) it will shut down, or oscillate
on and off.

 

[Leif Neland]

Michael Black kom med denne ide:

Now they are everywhere, tiny switching supplies generating low currents
(well for a switching supply 750mA or 1A is low, though for a transformer
supply, that's a fairly decent size transformer), and while I can't say I've
measured their voltage without load, they seem to be relatively well
designed.

I've seen horror stories about the lack of quality and security if
chinese USB chargers.

You buy a $500+ phone and then a $2 charger. You get what you pay for.

--
https://www.paradiss.dk
Ting til konen eller kĂŚresten.
Eller begge.

 

[Cydrome Leader]

Gareth Magennis <sound.service@btconnect.com> wrote:

Hmm, I keep hearing such things, but so far I haven't come across a SMPS
that will not sit there quite happily supplying voltage, without a dummy
load in sight.

This was true for old switchers, but I've not seen minimum load
reqirements on modern ones.

Old switchers with no load would sometimes not work or even make horrible
sounds in the magnetics as they failed to regulate properly.

 

[DaveC]

On 31 Mar 2015, Leif Neland wrote:

I've seen horror stories about the lack of quality and security if
chinese USB chargers.

You buy a $500+ phone and then a $2 charger. You get what you pay for.

Both Dave Jones (eevblog.com) and Mike Harrison (electricstuff.co.uk) have
pulled apart these cheep chargers on their YouTube channel videos. Very scary
what’s inside.

All those CE and UL markings on the units --counterfeit and meaningless...

 

>"All those CE and UL markings on the units --counterfeit and meaningless... "

I don't know about over there, but over here the insurance companies have a very powerful lobby. Burn down a few houses here and shit will hit the fan for sure.

 

[whit3rd]

On Monday, March 30, 2015 at 2:37:33 PM UTC-7, Gareth Magennis wrote:

Hmm, I keep hearing such things, but so far I haven't come across a SMPS
that will not sit there quite happily supplying voltage, without a dummy
load in sight.

The early SMPS units included frequency-modulating schemes (high output
current required higher oscillator frequency). Modern ones mainly use
pulse-width modulation (duty cycle modulation) at a fixed frequency.

An FM-SMPS power supply has greater ripple at low frequencies (unless one
overspecifies the filter capacitors), and is slow to correct output
deviations, so is likely to have unsuitable output
and/or to trigger fault detection circuitry and shut down.

PWM-SMPS circuits, on the other hand, have no particular difficulty
when the duty cycle gets very low, and the low-output-current load
that causes low duty cycle, keeps the output capacitor nearly ripple-free.

 

[Leif Neland]

jurb6006@gmail.com kom med følgende:

"All those CE and UL markings on the units --counterfeit and meaningless...
"

I don't know about over there, but over here the insurance companies have a
very powerful lobby. Burn down a few houses here and shit will hit the fan
for sure.

I don't think your insurance company can do much about what you import
yourself from China.

Leif

--
https://www.paradiss.dk
Ting til konen eller kĂŚresten.
Eller begge.

 

On Friday, April 3, 2015 at 3:58:56 AM UTC-7, Pimpom wrote:

"DaveC" <not@home.cow> wrote in message
news:0001HW.1ACB927A00A0706310B1C03CF@news.eternal-september.org...
On 31 Mar 2015, Leif Neland wrote:

I've seen horror stories about the lack of quality and
security if
chinese USB chargers.

You buy a $500+ phone and then a $2 charger. You get what you
pay for.

Both Dave Jones (eevblog.com) and Mike Harrison
(electricstuff.co.uk) have
pulled apart these cheep chargers on their YouTube channel
videos. Very scary
what's inside.

All those CE and UL markings on the units --counterfeit and
meaningless...

I once bought a 5-metre strip of 300 3528 LEDs to play with. It
came bundled with a 12V 1A power supply although in theory
driving the whole 5-metre length should need 2A. Anyway, I took a
look inside the PSU and guess what? The final rectifier which was
supposed to be able to drive loads up to 1A was a single 1N4148
diode!!

To return to the OP's question, I've occasionally come across
computer ATX PSUs that need a load to start up at all. I haven't
made any attempt to determine the minimum load needed but
something like an old CD-ROM drive works fine.

How do these 'no-load needed' ATX-PSU stay running without the Power_Good signal coming from the motherboard to the PSU, so they can work without being hooked up to the motherboard?

I've been curious how to trick that for awhile.

 

[Pimpom]

"DaveC" <not@home.cow> wrote in message
news:0001HW.1ACB927A00A0706310B1C03CF@news.eternal-september.org...

On 31 Mar 2015, Leif Neland wrote:

I've seen horror stories about the lack of quality and
security if
chinese USB chargers.

You buy a $500+ phone and then a $2 charger. You get what you
pay for.

Both Dave Jones (eevblog.com) and Mike Harrison
(electricstuff.co.uk) have
pulled apart these cheep chargers on their YouTube channel
videos. Very scary
what's inside.

All those CE and UL markings on the units --counterfeit and
meaningless...

I once bought a 5-metre strip of 300 3528 LEDs to play with. It
came bundled with a 12V 1A power supply although in theory
driving the whole 5-metre length should need 2A. Anyway, I took a
look inside the PSU and guess what? The final rectifier which was
supposed to be able to drive loads up to 1A was a single 1N4148
diode!!

To return to the OP's question, I've occasionally come across
computer ATX PSUs that need a load to start up at all. I haven't
made any attempt to determine the minimum load needed but
something like an old CD-ROM drive works fine.

 

[JW]

On Fri, 3 Apr 2015 07:48:36 -0700 (PDT) rev.11d.meow@gmail.com wrote in
Message id: <89709d02-0c9d-410a-8e09-b30fa4b24020@googlegroups.com>:

How do these 'no-load needed' ATX-PSU stay running without the Power_Good signal coming from the motherboard to the PSU, so they can work without being hooked up to the motherboard?

I've been curious how to trick that for awhile.

The power good signal (PWR_GD) is an input to the motherboard. The power
supply on signal (PS_ON) is an input to the power supply. If you short
PS_ON to GND, the supply should turn on.