解决方案查找器 技術支援

Loop compensation for TNY274

Posted by: treez on

hello. Figure 1, Page 1 of the DI-173
http://www.powerint.com/sites/default/files/PDFFiles/di173.pdf

......has loop compensation components R6 and C7. -However, the TNY274 datasheet states that the TNY274 uses on/off control and that no loop compensation is needed.
If no loop compensation is needed, then why has compensation components been used in the di-173 document as stated above?

回應

Submitted by Paul Lacey on 01/26/2009
The on/off control scheme implemented in the TinySwitch-III family does not require extraordinary loop compensation when implemented in a simple Flyback CV power supply, such as the one shown in Figure 14 of the TinySwitch-III datasheet. However, as active elements are added to the feedback network to adjust the overall response of the system, these components introduce characteristics to the feedback network that must be compensated for. Hence, certain compensation components may be necessary to shape the response of the feedback system when anything other than the simple CV feedback circuit of Figure 14 is used. Because DI-179 utilizes a Constant Current feedback circuit, certain components are needed to shape the feedback response of the network. For this reason, components R6 and C7 are needed to provide compensation for active element U2. Cheers! PI-Jordi
Submitted by PI-Chekov on 04/21/2009

Although the EN pin of the TNY274 is sampled digitally the control loop is still an analog system and therefore feedback loop compensation is still required. If you were to look at the current through the transistor Q1 you would see than it is a DC level just below the disable threshold for the EN pin rather than a digital square wave.

Cheers

PI-Chekov

Submitted by treez on 04/25/2009

Thankyou both very much.....in truth, i would say that using such particular values for R6 and C7 and U2 is a waste of time...its just more components for assembly staff to get wrong..................In truth, for a LED supply, who cares if there is a bit of oscillation on there?....it will not be much oscillation........it will not effect the visual appearance of the LED light......i would just suggest a simple comparator circuit with the current sense resistor.......it WILL keep the LED current easily constant ENOUGH.........And it will DEFINETELY prevent the LED current from going too high........it will do this latter better than the compensated Op-Amp solution that i referred to in my previous post..............................I admit that you may get a small bit of ripple in the LED current......but it will not have any visual effect on the LED light that you see.........................So my question is one of.......what have i missed here that makes me speak so "out-of-turn".......or are you in agreement with me?
"Perfection" is the ENEMY of "perfectly_satisfactory"

Submitted by PI-Chekov on 04/27/2009

I tend to agree - for a LED application where the load is constant if the oscillation causes and acceptable amount of current ripple through the LEDs then it's ok. For a voltage output supply - oscillation may result in poor transient response - something that doesn't occur with a fixed LED load. In general I would still recommend compensation components just to prevent something unexpected.

Cheers

PI-Chekov