Cable-drop compensation with INN2125K

Hi Fla,

Thank you for using our InnoSwitch family of devices.

May I request that you send the schematic and layout? I would like to know how the current sense resistor was connected to the IS-pin. The resistor should be placed close to the pins. How did you measure the IS-pin, through an oscilloscope or a DMM?

As to the IS-pin filter, additioanal filtering might add some delay and will cause error in sensing the current during transient conditions.

Regards,

PI-Picard

Hello. Thank you for your reply.

In the schematic, please check the sense resistor and how it is connected to the IS pin, through the RC-filter. Previously, the sense line was connected directly to the IS pin (no filter).

Unfortunately, I can not show the PCB-board. I see the point that the sense resistor should be as close as possible to the regulator, to measure millivolts effectively; but I can ensure that we implemented proper layout-techniques for current-sensing.

The threshold was measured with a DMM (DC-reading).

Back to our original question, we use the sense-resistor only to implement cable drop compensation, i.e. we don't need precise constant-current operation. So, the delay caused by the RC-filter should not be a problem. In this particular case, is it allowed to use the filter?

Thanks again.

Hi Fla,

As mentioned, we don't recommend filtering the Isense signal. But since, you don't need accurate CC you may do so provided you test transient\dynamic loading, start-up (full load, output short circuit) and running output short-circuit.

Are those RC values already optimized?

Regards,

PI-Picard

Hello. Thank you for your reply.

The values of the RC-filter (1kΩ and 100nF) gave good results on the prototype (i.e. accurate cable drop compensation).

We will optimize the values on the final version of our PCB, which we will receive during this week.

Regards.

Hello.

I re-open this topic because I have more details to add.

With the final PCB, we experienced the same problem with cable-drop compensation.

For this first lot, we did not require cable-drop compensation (even if it would be good to harmonize our BOM's). So, we disabled it by using 0-ohms current sense resistor and connecting the IS-pin to ground.

It turned out that the output voltage varies in similar way. The output voltage (measured at output pads, no drop-out due to connector) is as  follows at different loads (increasing from 0 A upwards):

Is that a problem with regulation, that was not evident when using cable-drop compensation?

If you request to see the layout, we can proceed only with private contacts.

Thank you.

Hi fla,

What is the difference between the final PCB with the one that has the CVCC working? What is your full load output current?

What do you mean by private contacts?

Regards,

PI-Picard

Hello.

CVCC operation was not working as expected even during previous tests with the same PCB, but cable-drop compensation was implemented according to the schematic I posted (IS-filtering was required to avoid issues with early CC-operation).

We decided to remove cable-drop compensation (not needed for our first lot, but will be required later) and we still saw the same behavior, as described by the table above. Of course, we do not have cable-drop compensation superimposed, so it is not responsible of the behaviour we see, as we originally thought.

Full-load current is 4.2 A.

If you need to see the layout to provide suppot to us, I am not allowed to show it on the public forum and I would need direct contact with the responsible persons at PI. 

If you need particular measurements, please advise.

Thank you.

Regards.