Linkswitch-PL 458 Transient problem

14 posts / 0 new

Dear Support,

I'm designing drivers using PI ICs for market use a year ago. Most of our design are buck topology which don't have a simulation tool or Pspice model. Now I'm just using PIXls Designer 9 for Linkswitch-PL and have some inquires:
1- What's best way to select the feedback filter "RC" parameters and what's it's effect on the transient time when moved suddenly from 265Vin t 180Vin.
2- When I put C5=1UF this problem occurred and when I reduced it's value to 0.5uf it's slightly moved.
3- How to reduce the transient time and deviation?

I faced another issue , while transient fron 265Vin to 180Vin to power consumed increased by 1W , I m3an the output current increases from 120 to 145mA and still like that till restart the input

I apologize for the late reply. We were on holiday break and I was not able to reply earlier. For the transient response,
LNK-PL has an inherent slow loop response and the undershoot is a normal behavior. The R-C values you originally have are already optimized. However, reducing C5 will only worsen the regulation and I think that's the reason why the output current increases that high from 265V to 180V. Can you try to increase the capacitor to 2.2uF and check if the regulation improves? You also can't put too high value as this may cause oscillation.

Hi Pato,
Please note that for a single stage with good PF and THD solution converter the transient response is a little bit slower than the non-PF solution.

Regards,
PI-Jono

Dear,
You say that R-C values ara alrady optimized!, what do you mean by that? is these values are the same for any design. I believe this is a low pass filter which it's design affect the cutoff frequency, the regulation. So How you design it mathematically?

I made more than 12 tries and error using three cap vlues 0.5uf, 1uf, 2.2uf and 4.4uf with different values for resistors 1.5kohm, 2.2K, 3.3k, 4.7k and all the cases gives very bad regulation more than 17%. Besides the high transient peak current and low response which causes flicker.

Please I need a detailed scientific and logic reply.

Thanks.

Please help and reply
9.5 w input Buck topology using lnk458 gives very bad regulation

hi,

In a buck topology using LNK-PL, the output current is inversely proportional to the input line voltage because of the way it senses the current which is through freewheeling diode only. Thus the output current is not sampled during the switch ON-time. To get a better regulation, I recommend using buck-boost topology.

Dear,
you recommended to use buck-boost to get a better regulation than buck topology although the way of sensing the output current in both topologies are the same " through the freewheeling diode", So there is no sampling of the output current through the ON-time of the high power switch.
I need to know what's the reason which lead to better regulation? , Is the duty cycle affects that ?
Please clarify the reason ??!!!!

Thanks.

I tried to design a buck-boost topology by LNK458 using PIxls designer but it gave me a warning that the maximum current limit exceeds device Ilim_min. although the power is lees than the IC rated.
These are my parameters:
VACMIN: 180V
VACNOM: 220V
VACMAX: 265V
FL: 50Hz
Vout Nom: 65V
Iout: 0.14 A
PO: 8.96 W

Whats your recommendations?

Thanks.

There is currently no option to change Kp or Lp in the spreadsheet. However, you can increase Lp to 680uH so that the peak current is reduced without going into continuous mode territory. I don't know which transformer you are planning to use but you can just calculate the required turns for <3000Gauss flux density. Use Ilim_min in the calculation.

Thank you for your clarification.

But What if I use buck topology and change the sense resistor place to be between the high power switch and the diode so that the current can be sampled with the ON state time that will lead to a better regulation.

but is this affect any other parameter ???!!!!

Thanks

LNK-PL only works with positive FB signal which is why the sense resistor is placed as such.

Dear,
1- you recommended to use buck-boost to get a better regulation than buck topology although the way of sensing the output current in both topologies are the same " through the freewheeling diode", So there is no sampling of the output current through the ON-time of the high power switch.
I need to know what's the reason which lead to better regulation? , Is the duty cycle affects that ?
Please clarify the reason ??!!!!
2- when I put the sense resistor series with the inductor " after doubling it's value "nnot the freeweeling diode in a buck desig it gave me a very good regulation. But the feedback max. voltage "sined shape" across the the sense resistor were around 2V or higher .. is this technique acceptable??

Thanks.

hi,
You are right that in both topology, the sensing is thru the freewheeling diode. The FB senses the average diode current. However for buck, IOUT is the average inductor current which takes supply in both ON-time and OFF-time whereas the current only flows thru the diode during off time and is a function of duty cycle. For buck-boost, on the other hand, you are sensing IOUT thru the diode during OFF time. This is why I suggested buck-boost for better regulation.

If you check the datasheet, there are several operating modes of FB depending on the input it is sensing. Above 520mV, the unit will enter skip-cycling mode which affects PF and THD. Above 2V, the device will auto-restart. These are the consequences when the FB is not operated properly.