Battery charger current sense 4

[treez]

hello,

I am doing a switch mode, lead acid battery charger with UC3909 Charger IC.....

Datasheet.....

http://focus.ti.com/lit/ds/symlink/uc3909.pdf

..on pg 7 of the datasheet for UC3909, it shows the charge current sense resistor being placed at the non load side of the buck converter.....can any reader understand why it is placed here?

I thought that in this position it is having to take the ripple current through the output capacitor aswell as the charging current,....and that surely is bad ?

..surely this current sense resistor should be on the battery side of the output caspacitor, then it will only experience the DC charging current, which i thought would be right?
------------------------
Relevant to this, the following app note on this UC3909...

http://focus.ti.com/lit/an/slua058/slua058.pdf

...on page 4 it shows a flyback charger circuit, and again shows the sense resistor the "seemingly wrong" side of the output capacitor...(they then use Rsf and Csf to filter out the flyback ripple current).

I cannot see the point of this...why bother having to filter it?...why not i think just put the sense resistor the other side of the output capacitor ?

In other words, surely its best to keep the sense resistor outside the rectifier loop and in the output loop , where its nice smooth DC?

grateful for any thoughts.

 

[itsmoked]

I haven't poured over the data sheet so take these as general answers.

One reason may be for biasing. Not a lot of head room.

The other is; Many of these controllers use the current thru the inductor to control the turn off of the power elements each cycle. Heavily filtering it may cause complete failure of the control, whereas some filtering may be needed for winnowing out the noise.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[GonzaloEE]

From Fig 2/page 7 of the uc3909 datasheet, the current sense pins go to an internal diff. amplifier opamp, with one of their nodes connected to a fixed 2.3V reference. So leaving the sense resistor near ground makes sense as it keeps the input signal below that voltage.

 

[DHambley]

By sensing the inductor current, high speed loop response is possible which you can't get after the capacitor. In particular, look at the resistor before the cap at pin 14 on page 7. This value is critical to effect slope compensation for current mode control. CMC makes the inductor look like a current source, removing one pole from the loop.

 

[benta]

There is one important reason for placing it there (don't ask me how I know).
There is an overcurrent failure mode that is not obvious, in fact you need a storage scope to identify it.
This happens when the regulator is recovering from a short-circuit on the output. The output capacitor voltage is 0 (due to the short) and the short is then removed. If you are sensing the output current after the cap, the resistor will not see the charging current going into the cap to bring it up to the required output voltage.
Worst case, your inductor saturates and your switch will have to try to survive a large current spike going into the cap.
By placing the resistor before that cap, the IC will see the spike and turn off the switch accordingly.

Benta.

 

[treez]

hello,
Thanks for looking into this.

-Actually , ive just noticed that on pages 8 and 9 (figures 3 and 4) of the datasheet for UC3909.....

DATASHEET:

http://focus.ti.com/lit/ds/symlink/uc3909.pdf

...they actually do have the current sense resistor outside of the rectifier loop. Do you belive these are mistaken figures?

 

[GonzaloEE]

IMO, Rs does the same than before: current sensing at the return path of the load (battery). It is in series with the sec. diode, so it's part of the rectif. loop.

The IC is powered by default from the bias winding, with battery backup through that diode close to "Vcc".

 

[benta]

My comment on the switch overcurrent failure is only relevant to a buck converter. A flyback is something completely different.

Benta.

 

[gabisuciu]

they use the average current mode regulation method, in which by placing the shunt inside the H.F AC power loop the loop compensation becomes much easier (some close to peak current mode)
actually by measuring the AC H.F current attached to DC output current a better pk-to-pk fast regulation can be achieved.