offline SMPS with low "standby" power

[eem2am]

Hello,

Please can you tell me, I wish to do an isolated , offline SMPS (85 to 265VAC) , Power output = 34W.



It has one isolated output of 36V (+/-5%), 32W and one isolated output of 5V (+/- 4%) , 2W.

When the SMPS is in “standby” (i.e., its minimum load condition) the 36V rail is completely unloaded, and will be “shut down” completely, to save power.
However, the 5V rail , when the SMPS is in “standby”, will have a load on it of 200mW.

I have to design this SMPS so that it draws no more than 500mW from the mains when in standby.

If you know of a PWM controller which is suitable for this, please can you tell of it.

 

[NickS]

You are all over the map with this question
(1) So you want a SMPS that converts 85 to 265VAC at 34W
Is that 85VAC or DC? SMPS are for DC conversions you use transformers for AC conversion.

(2) You have two isolated outputs that I assume are DC(36V and 5V) what is the source voltage? Mains?

(3) you require enable circuitry for the 36V out only but do you also have to monitor the load on 36V to go into standby or will that be a manual process? I am thinking you want to detect and standby automatically but how do you plan on coming back out of standby?

(4) from your mains requirement of 500mW in standby I think I can guess that you want to power off the mains(which are AC). So you need more than a SMPS controller. You first have to deal with stepping down and rectifying your mains voltage. Then after filtering you can start your DC-DC conversions. But I still have no idea what the second line of your post has to do with any of this"(85 to 265VAC)".

 

[Mitchekj]

He means he needs a universal input (85 - 265Vac input).

There are many newer ICs available that can meet this (ridiculous, IMO) regulatory requirement of 0.5W in standby mode. Look for "Burst Mode" in the features list.

The rest of the supply will need to be engineered rather efficiently to meet this half watt requirement, btw. But there are tons of ICs that can help you get it done.

I have used the Infineon ICE3B1565 for a similar spec. Take a look at that datasheet, you'd need a different model number to accomodate your power requirement, but the family is the same.

 

[NickS]

That makes sense. thanks.

 

[eem2am]

please see a more detailed post on this called "offline SMPS with 500mW standby power"

 

[eem2am]

offline SMPS with 500mW standby power

hello,

sorry this is a long post, if you know a bit about SMPS's , it will be very straightforward though

i wish to design an isolated, offline SMPS and i am wondering if i have missed hearing about any new "low standby power" SMPS controllers?

V(in) = 85 to 265VAC
V(out)_1 = 36V +/- 10%
V(out)_2 = 5V +/- 4%

V(out)_1 and V(out)_2 must be isolated SELV outputs.

The 5V rail has maximum power of 2W
The 36V rail has a maximum "running" power of 10W.
(-but see below for its peak power of 36W)
Both 36V and 5V rails could have power levels anywhere from 200mW to their maximum, at any given moment.

In "standby", the 5V rail will have a load of 170mW.
In "standby", the 36V rail is not needed at all, and could be shut down in some way.

In "standby", the SMPS must draw no more than 500mW from the mains (yes, just 500mW).

The 36V rail is interesting.....it has a 30mF (yes, 30mF)capacitor across it.......when the SMPS is switched on at the mains, this 30mF capacitor must be charged up to 36V. -This 30mF capacitor must attain 36V within 1 second of the SMPS being switched on at the mains.

-When this 30mF capacitor is charged up to 36V, then the load on the 36V rail is 10W maximum
-This 30mF capacitor exists to provide some power to both the 36V rail and the 5V rail in the event of a mains failure.
-When the 30mF capacitor is being charged up, there is no other load on the 36V rail other than this charging capacitor.

-The big challenge with this is 1) the peak load of 38W while the 30mF capacitor is charging.....2) Even bigger challenge is the 500mW maximum power drawn from the mains when in "standby" mode............as you know, in "standby" mode, the secondary (specifically the 5V rail) has a 170mW load on it.......assuming 80% efficiency, this means a 213mW load seen at the primary........meaning that that we only have 287mW of margin before we violate the 500mW power consumption.

-Do you know of a controller thats good for this?

 

[Mitchekj]

I already gave you a good candidate. If you want me to design it for you I'm sure we can arrange payment.

 

[eem2am]

thankyou, your suggestion is good on no load..............100mW, but of course i have a 170mW load in standby when i must get below 500mW input.

also, i cannot work out how an smps auxiliary coil avoids dropping out when the regulated secondary has no load on it.................surely the auxiliary must drop out?....the laws of physics dictate?..................even if the secondary still had a small load, the coupling between isolated windings is going to be poor, and again, surely the auxiliary voltage will drop out?

 

[Mitchekj]

Here's a freebie, you take it from mains. (That IC I linked does that automatically, it's all on the die.)
What is this 170mW load, anyway? Supplying a load kind of implies you're not in standby mode, by the way.

And good luck being 80% efficient with a 170mW load. Good thing for you, you only need to be 34% efficient on the whole to keep below .5W input power. I don't understand why you think 57% margin is low?

So, is this a school project or is someone paying you to design this?

 

[eem2am]

Offline SMPS without auxiliary coil ?

hi,

I am making an offline, isolated SMPS which must provide 38W maximum at V(out) = 30V. V(in) = 85 to 265VAC
I also must draw no more than 500mW in standby mode.

(please note, by "auxiliary coil" i refer to the coil that supplies power to the SMPS controller...also called the "bias" coil)

Sorry in advance for the length of this post!

As explained it has a "standby" mode where it is actually loaded with an average of 170mW.
-This 170mW is not a constant load, but represents a pulsed load, pulsing fully on and off with low duty , at around 4Hz
-When its in "standby", it must draw no more than 500mW from the mains.

I am sure that we all know how badly auxiliary coils are at providing the controller voltage rail when the secondary load varies over such a huge range as in this case.
....the problem of auxiliary rail voltage "wandering" is made worse by the unpredictability of the "wandering"......the voltage range of the auxiliary voltage rail depends on the amount of transformer leakage inductance.........and unless SMPS transformers are wound expensively, the leakage inductance usually has a wide tolerance of values.........so getting the right number of turns on the auxiliary coil in such cases as ours is a truly hit and miss affair.


-the only way to get sufficient voltage output from the auxiliary coil in no-load is to wind too many turns for the auxiliary, and then linearly regulate it so that it does not over-voltage the controller when the SMPS is on maximum load.

and as i said, the amount of turns is by no means a strightforward calculation, and may not work with differing amounts of transformer leakage.

-in "old" times, we simply let the auxiliary coil "drop-out" and the controller would then be supplied from the high voltage BJT regulator connected to the DC bus....but with our new 500mW maximum standby power draw regulations......this kind of thing is no longer possible, as it takes too much power.

With regard to this, i am just thinking of getting rid of the auxiliary coil altogether and using a LNK302 to efficiently provide the SMPS controller power supply.
(-LNK302 is a controller for a very low component count offline buck)
(-This is not a marketing post....since in any case powerint.com do not seem to recommend using LNK302 as i have described......on the contrary , powerint.com seem keen on the idea of using an auxiliary coil)

So in any case, my question is, since we now have these stringent demands for low standby power, why is it that every SMPS now does not just do away with the auxiliary coil and use a LNK302 to provide the SMPS power supply?

-especially when the secondary load is variable over a wide range.

After all, LNK302 is very, very cheap....and without an auxiliary coil, the transformer will be cheaper to get manufactured....and you can avoid all those extra components that auxiliary coils often need such as post regulators and filters to prevent peak charging from the leakage spike etc etc.
-Not only this, but in a typical secondary regulated SMPS, the transformer should be wound as............

[CORE] (1/2 PRIMARY) // (AUXILIARY COIL) //// (SECONDARY) //// (1/2 PRIMARY)

where:-
//// = 3500V mains isolation barrier
// = single layer of tape

....as you can tell from the above, this is expensive to manufacture, and gives more leakage inductance than when the auxilary coil is not used.

-let's look at it when there is no auxilary coil and compare:-

[CORE] (1/2 PRIMARY) //// (SECONDARY) //// (1/2 PRIMARY)

and since we now have no auxiliary, we may now get good enough primary to secondary coupling with:-

[CORE] (PRIMARY) //// (SECONDARY)

....and this is undoubtedly far cheaper to manufacture.

So in any case, my question is this:-
Since we now have these stringent demands for low standby power, why is it that every SMPS now does not just do away with the auxiliary coil and use a LNK302 to provide the SMPS controller power supply?

 

[eem2am]

The thing is, taking power from the DC bus is going to be dissipative,
Also.........due to controller bias supplies, switching losses etc, 500mW in light-load (170mW here) i feel is going to be very difficult indeed.

I understand its strange to be loaded in standby.........but thats just our system which has an RF receiver in it, ready to accept the customers incoming RF signal, saying that they want to "play"

 

[Mitchekj]

Read the datasheet. Under light load conditions you go into active burst mode. Yes it's dissipative to take Vcc from the DC bus, but in burst mode you drop Vcc's current draw to almost nothing. You only dissipate a minimum, with plenty of room for all your other losses.

Look up the eval board for those parts. They even show actual data of a 650mW input power, with a 500mW load. Can't ask for anything more. A 200mW load would get you well under the 0.5W requirement if you do it right.

 

[(*steve*)]

hi,

I am making an offline, isolated SMPS which must provide 38W maximum at V(out) = 30V. V(in) = 85 to 265VAC

Please stop creating new threads for this.