LM5177: Unstable CC CV Mode Compensation and Slope Settings

Hello Luke,
Thanks for splitting this topic from the previous thread.

Stefan, who knows this device best, is currently out of the office and will be back on Friday.
He may have many open tasks from the whole week on his table, so I cannot guarantee that he will be able to get back to you before the end of this week.
Please allow for Monday next week for a reply from Stefan.
I am hoping for your understanding.

Thanks and regards,
Harry

Thank you for notifying me of the case Harry...

Meanwhile, I will add more details....

After seeing the bode plot change with the input voltages, I changed Cc1 from 22nF to 47nF to get a better phase margin throughout the whole range of input voltages....

Also, I have 3 - 22mOhm ESR polymer capacitors and 9 - 10uF Ceramics on the Output, so I changed the ESR Cout to 1mOhm from 10mOhm, and now there is a huge spike in the phase margin around the switching frequency... Is this normal?

I changed the slope resistor back to the recommended 280k Ohm and now the battery will charge at the full CC until it reaches around 18.9V - 20V... The point where the charging current slows down varies with input voltage... After it slows down, depending on the input voltage it stops charging completely, or slows way down from 0.3A - 2.5A....

The audible noise is still an issue, and the buck mode still works same,  I cannot understand why the boost mode is so noisy and why the charging current and voltage drops so badly when it reaches nearly the setpoint....

Hello Luke,
It is expected that the behavior of the power stage will change quite a lot when you change the capacitors.
So, you will need to change the compensation network accordingly, every time you modify the components of the power stage.

Best regards,
Harry

Harry,
Yes I understand that, but when I look at the Bode Plot across my input voltages, I have the best results with the Compensation network in the attached calculator... So I cannot understand why it looks good in the calculator and yet performs so badly in my circuit... I have attached the most recent calculator...8446.LM5177 Buck-Boost - In Use.xlsm

Hello Luke,

The calculator is assuming ideal componetns whereas the real board has a lot of parasitics (in the components and in particular in the PCB itself).
The capacitors of the compensation network are pretty small (dual digit Nano Farad or even Pico Farad) and the parasitics are in about the same ballpark.
Therefore, a calculation will never reflect the reality and only the measurements (or tests/experiments) on the real board can tell you the best parameters.

All the calculation can do, is helping you to select initial components, so that the board will not get destroyed right away when you turn it on for the first time.
This will also give you a good starting point from where you can start tweaking/improving the compensation network.

Best regards
Harry

Harry,

I realize that it will not be perfect and will need tweaking, but I find it hard to believe that the compensation network is the only thing wrong with this design, as it is not even close to working nicely anywhere in the boost mode.... 

So it works fine when it is in CC mode far below the battery voltage(On charging a 20V Battery, the battery is up to 19V and still performing perfectly)

After the battery reaches 19V, depending on the input voltage, the output current will drop down to 2.5A-1A charging current, and then the battery takes a very long time to reach full charged.... 

I don't really know that this can be a COMP issue, the COMP is probably the reason for all the inductor noise though...

Hi Luke,

if i understood right you are using the LM5177 and all the test where done in this mode only.

As there are two compensation loops working together, it would be good to just focus on one first.

So can you check out the voltage loop first:

- do not use the Battery as load, e.g. use an electronic load

- is the system working within the full input voltage range, when not reaching the current limit ?

- if you see issues there, can you disable the current limit function (change CFG resistor to disable the limiter -> enable the monitor or put IMONOUT to VCC)

Best regards,

 Stefan

Hello Stefan,

Stefan Schauer said:

you are using the LM5177 and all the test where done in this mode only.

I think you mean in this CFG setting, and you are correct, I am using a 20.5k CFG resistor setting in all of this testing...

Stefan Schauer said:

do not use the Battery as load, e.g. use an electronic load

Okay, I only resorted to the battery after it seemed the electronic load was causing issues...

Stefan Schauer said:

is the system working within the full input voltage range, when not reaching the current limit ?

No, only in the buck mode... In boost or buck-boost mode, a load of 5A will drop the output voltage down to 18Vish depending on the exact input voltage, and a 3A load will drop the output down to 19.6Vish depending on the exact input voltage....

Stefan Schauer said:

if you see issues there, can you disable the current limit function (change CFG resistor to disable the limiter -> enable the monitor or put IMONOUT to VCC)

Okay, so I put IMONOUT to VCC, and still the same story, not quite as much output drop, but still down to 19V with a 5A load, and 20V with a 3A load in buck-boost and boost modes....

Hi Luke,

thanks for the good summary.

Can you please provide some scope plots with that latest setting (current limit disabled) / 19V / 5A

- what was the input voltage for this check ( keep it and let me know)

- Vin / VOUT    (DC and AC coupled over a few switching cycles and ~ 10ms)

- COMP / SW1 / SW2  (DC coupled over a few switching cycles and ~ 10ms)

Thanks,

 Stefan