SMPS not turning on at full load while it turn on at no load.

Dear All

Here i am using a TOP271 for my Project of output 12v 8amp.

it turns on no load giving me a 12.3v output also if i put the load it also draw 8 amp of current.

but if i attached load previously and trying to turn on smps it is not giving output.

Help me with remedy.

- - - Updated - - -

Condition1: If i turn on smps with no load i am getting output 12v then i will have to put the Load of 8 amp it also draws the same.

Condition 2: If i attached load to the output and try to turn on smps it does not Turn on.

WHY this is happening..??
Schematics attached in above post.

please show full circuit, including all diode part no's...also values of the opto LED resistor

- - - Updated - - -

you are missing a resistor in series with D9
(BTW, do you know that power integrations run their own forum)

- - - Updated - - -

Is your circuit the same as what PI Expert recomended?

- - - Updated - - -

what is resonant frequency of the output LC filter.?..it should really be preferably 10 times more than the crossover frequency?

- - - Updated - - -

to what voltage does the drain voltage spike up to after you suddenly load it?
To what voltage does the control pin fall to after you heavily load it.?

- - - Updated - - -

your optocoupler may have too wide ctr values...which particular one of the ps2501 family are you using?
**broken link removed**

..use a narrow CTR range one and make sure you downsize your opto LED resistor sufficiently.

What happens when you start up on full load and then suddenly totally unload the circuit?

- - - Updated - - -

What is the part number of L7?...its resistance?

- - - Updated - - -

What is you primary TVS value?....is it in conduction when you suddenly load the circuit?.....you might have to make it a higher voltage one, i cant see its part number.

- - - Updated - - -

Does the output spike up above the output cap voltages when you suddenly load it?

- - - Updated - - -

What's the highest load that you can suddenly load it with and it manages it ok?

- - - Updated - - -

Your problem may well be due to the turn-on spike exceeding the chips internal leading edge blanking time. You coudl also try reducing the snubber cap size across the output diode...(as long as you dont overvoltage the secondary schottky).
Also, the series resistor in series with the diode to the RCD may need increasing. Also, the transformer interwinding capacitance may need looking in to.

may be output capacitance is too large and soft start time of TOP not enough for complete charge of it

What's your test load? 1.5 ohm resistor?

good points, also what is your bias winding output voltage?..is it high enough?
Velkarns point may mean you need a bigger control pin capacitor.

- - - Updated - - -

OP I believe i read your point wrong (unless you updated it?), i thought you meant it WOULD start up when loaded.....SO this means that Velkarns point is probably the reason.

"what is resonant frequency of the output LC filter.?..it should really be preferably 10 times more than the crossover frequency?"

Is this a general rule for all topologies?

yes it is. There are several references on it. Some say you can get away with 3x the resonant frequency. Ill post the refs soon.

- - - Updated - - -

pg 17 of the attache AN-47 explains it

However, pg 12 of the attached document “pg 12…..” states that the TL431 connection like in that document means that the LC resonant frequency can actually be close to the crossover frequency..as long as the opto diode is connected upstream of the post filter

Velkarn said:

may be output capacitance is too large and soft start time of TOP not enough for complete charge of it

Click to expand...

Venkarn:
I am using 0.1uF/16V as suggested by PI.
for output cap: ised 1500uF//1500uF//1500uF

Does it impact and how to solve the same.

- - - Updated - - -

Also in the same design i tested it for 4 hours on 7 amp.

where i am adding another 1 amp (I.E. 8AMP) my FR107 is getting failed.
This is another Problem.

in your schematic you are using a 47uF control pin capacitor. That initially gets charged up to 5.8v by the internal startup circuit…….that 5.8v then gets used to start the converter up…however, if the converter does not get started before the control pin capacitor has discharged to 4.8V then the controller just trips out because it assumes that there is an overload or short circuit on the output.
You could try remving some output caps and try again…..does your control pin stay above 4.8V during the startup?

You might be able to use the power integrations recomended "soft-finish" circuit in order to start up with such large capacitance...search their site to see it.
...or just remove a load of that capacitance and try again.

- - - Updated - - -

FR107 is not really suitable..it has too long reverse recovery.

- - - Updated - - -

also, you are only using a 100uf dc bus cap....this isnt very big, what is the lowest excursion of the dc bus at max load?
Are you using 230vac or 120vac?

Is your load a resistive load, a constant current load, or what?

Many power supplies will not start with a constant-current (CC) load applied before startup.
But once that it has started and the feedback loop stabilized, you may apply any kind of load.

you need to design in a longer soft start in your volt feedback loop.

yes, see top right of page 17 of this AN-47...
**broken link removed**

As there no relation of control pin and fig 22..

- - - Updated - - -

treez said:

in your schematic you are using a 47uF control pin capacitor. That initially gets charged up to 5.8v by the internal startup circuit…….that 5.8v then gets used to start the converter up…however, if the converter does not get started before the control pin capacitor has discharged to 4.8V then the controller just trips out because it assumes that there is an overload or short circuit on the output.
You could try remving some output caps and try again…..does your control pin stay above 4.8V during the startup?

You might be able to use the power integrations recomended "soft-finish" circuit in order to start up with such large capacitance...search their site to see it.
...or just remove a load of that capacitance and try again.

- - - Updated - - -

Treez:
As per the PI Application note:
In addition to the 100 nF capacitor connected to the CONTROL
pin, a series combination of a 6.8 W resistor and a 47 mF
electrolytic capacitor is required to be connected between the
CONTROL pin and the SOURCE terminal of the TOPSwitch-JX.
The 47 mF capacitor acts as an energy reservoir and provides
power to the TOPSwitch-JX internal circuitry during start-up

Now who will make it go 4.8v..??

Click to expand...

You mean 47uF don’t you?

The topswitch bias supply current comes from the feedback optocoupler.
If your output doesn’t get into regulation quick enough , due to having too big output caps, then the series resistor in series with the opto-LED may well prevent sufficient current to get through the opto to power the topswitch…then the topswitch will be forced to power itself from whatS left in the 47uf cap. If this cap falls to 4.8v before the topswitch starts self supplying via the opto then it will just trip out and never get into regulation.

It is called “overload protection”, and this tripping out if the smps stays out of regulation for long enough is a standard feature in virtually every offline smps controller on the market.

The startup of such offline smps is…

1….Linear regulator powers the controller by charging up a reservoir cap (in your case the 47uf)
2….the linear regulator gets turned off when the reservoir cap gets to x volts.
3….the smps starts working , (using the energy in the reservoir cap) and when its bias winding has risen high enough, it then self supplies, not anymore needing the linear regulator.
4…BUT if there are big output caps then the smps may never start self supplying before the reservoir cap is too discharged….in this case the controller assumes output overload and just latchs off or goes into ‘hiccup short circuit protection’ mode

So what is remedy.

I am using solenoid at 12 volt outputs and some fans.

What will lead to solve my problem..?

remove output caps till it works for now.
Also What is the resistance of the solenoid coil?
And What is the current draw of the fans at 12v?

Bjtpower said:

Dear All

Here i am using a TOP271 for my Project of output 12v 8amp.

it turns on no load giving me a 12.3v output also if i put the load it also draw 8 amp of current.

but if i attached load previously and trying to turn on smps it is not giving output.

Help me with remedy.View attachment 127765View attachment 127766

- - - Updated - - -

Condition1: If i turn on smps with no load i am getting output 12v then i will have to put the Load of 8 amp it also draws the same.

Condition 2: If i attached load to the output and try to turn on smps it does not Turn on.

WHY this is happening..??
Schematics attached in above post.

Click to expand...

The attached circuit work for the 8 days with full Load and functionality.
now it is Failed. and it it turning on/off.

I checked the Top switcher,Output Diode as well as flyback Components.

As all are good.

Dont know where is problem..??
Kindly Help Me.

FlapJack said:

"what is resonant frequency of the output LC filter.?..."
Is this a general rule for all topologies?

Click to expand...

These are poles and it is good to be away from the poles. Under steady state conditions, you can safely operate close to a pole but you can never be sure during the brief period of start-up or shut-down.

They will drink energy like black holes and will cause strange failures very tough to debug.

Unless you are going to use the resonance effect, you should be better safe and stay away from all poles.

C_Mitra:

i have not considered any poles and zeros for my design.

the PI is having there Ready made IC'S,That we have used for our design.

It was worked well under designed load now it is giving me a problem for the Debugging.

I am not getting from where i shud start for the designing.

Help pls.

- - - Updated - - -

C_Mitra:

i have not considered any poles and zeros for my design.

the PI is having there Ready made IC'S,That we have used for our design.

It was worked well under designed load now it is giving me a problem for the Debugging.

I am not getting from where i shud start for the designing.

Help pls.

- - - Updated - - -

C_Mitra:

i have not considered any poles and zeros for my design.

the PI is having there Ready made IC'S,That we have used for our design.

It was worked well under designed load now it is giving me a problem for the Debugging.

I am not getting from where i shud start for the designing.

Help pls.

- - - Updated - - -

C_Mitra:

i have not considered any poles and zeros for my design.

the PI is having there Ready made IC'S,That we have used for our design.

It was worked well under designed load now it is giving me a problem for the Debugging.

I am not getting from where i shud start for the designing.

Help pls.