Irs2092

[gxhcdba]

hello :
i am from china .my english is very poor
i made a class d amp with irs2092 .
now have a problem ：the ira2092 ic can't start up .
attach file is irs2092 sch, please give me some advise !

[odnaizutra]

hello,
have solved the problem?
Please,refer to oc pin adjust.First desconnect the resistor that goes to -b,leaving only the one that connect to vref and see if the amplifier works.
Maybe you will have to find mosfets with less qg or adjust the oc protection to a very value.

[gxhcdba]

Originally Posted by odnaizutra

hello,
have solved the problem?
Please,refer to oc pin adjust.First desconnect the resistor that goes to -b,leaving only the one that connect to vref and see if the amplifier works.
Maybe you will have to find mosfets with less qg or adjust the oc protection to a very value.

[gxhcdba]

Hi odnaizutra,thanks for you reply.
today,I had adjust the low side oc protection point
but the amp can't start up .
now the IC 2092's pin 5 voltage is 0v and 3v to bounce up and down.
the ic auto restart .
can you give me some advise ?
Excuse me for my poor english

[odnaizutra]

sure,
this indicate that protection is activated for some reason.
First you will have to see if all voltages are ok , because there are some under voltage protections that can activate this lock.
check ,in relation to ground , if you have unless +5 volts an -5 volts at the front end of 2092.If you have less than 4,5 volts,ic protection activate.Then ,check if you have unless 11 volts,in relation to -b at vcc.
supposing all voltages are ok,try using mosfets with less qg,since this ic activates the protection if qg of mosfets are huge.Try irfb4019 or similar.
This ic is very sensitive to layout of pcb.Try to follow ir suggestion.
Pay attention to inductor caracteristics.Dont use any unkwon ferrite at inductor.Only type 2 is ok.Other option has to be carefully studied.You could send some pictures of pcb and the hole project.

[gxhcdba]

Thank you very much!
According to your theory, I think most likely lies in pcb layout , I will layout the pcb again.(refer ir suggestion).

[odnaizutra]

Dont forget,
the output fet qg has a lot of influence at the protection activation (just the way you said , like hiccup mode).Try a low qg fet

[sphinx]

Originally Posted by odnaizutra

Dont forget,
the output fet qg has a lot of influence at the protection activation (just the way you said , like hiccup mode).Try a low qg fet

How does qg influence protection activation? Isn't the OCP sensing the voltage across Rds-on?

[Eva]

It's Vds what is sensed, and it includes not only Rds-on drop but also the drop across drain and source lead inductances (4.5nH + 12.5nH respectively for TO-220) and layout inductances. During switching events, when current is rising (negative current falling during body diode reverse recovery), voltage drop on Ls+Ld may exceed 4V (it's proportional to current rise slope) resulting in false trigger if proper blanking is not applied.


When a MOSFET is turned on, Vds does not fall to just Rds-on drop instantaneously, there is a turn-on process with several parts. In the last portion of the process, Vds is already low, below 15V usually, and gate to drain ("miller" or "feedback") capacitance increases dramatically as Vds drops further (see datasheets). As a result, the last part of the turn on process is long, particularly for parts with high gate charges. In these circumstances, Vds sensing may trigger protections erratically if blanking time is not long enough.

[sphinx]

thanks Eva..clear and concise as usual..you should write a book someday