Current cource with LM5145 ?

I need a regulated current source 0.1-20A (output power 5W-500W) which will be able to supply current permanently (couple of minutes) , but also pulse with adjustable width and pulse frequency from 100ms.
Pulse operation will be resolved by a transistor output at the control processor.
My idea is use 75V Synchronous Buck Controller With Wide Input Voltage and Duty Cycle Ranges with LM5145, any as this
Input voltage 48V, on output high side current monitor for current feedback and high side ultralow RdsOn MOSFET for pulse operation.
The question is, do you consider it a good idea or am I just getting into trouble?

The manufacturer designed the IC to be versatile but it can't do everything. If you draw abrupt On-Off pulses from the output, it's like asking the device to turn its operation on and off very quickly. The instructions say:

The load should be an electronic constant-resistance (CR) or constant-current (CC) mode

Click to expand...

Consider installing filtering circuitry between the IC and your pulse generator. Possibly a T-shape LCL. The aim is to draw smooth current from the IC.

Yes at the output I was considering induction. something like see. picture.
It's 17 threads of 10x3mm copper
It could be around 1mH
And 20 x 10uF 100V X7R 1210 capacitors
and for switching betwenn C a L MOSFET IAUT300N08S5N012 80V 1,2mOhm 300A

It could be around 1mH

Click to expand...

Surely not milliH. The calculation tool screenshot says µH.

Of course uH.
In reality I hope it will be tens of uH (about 20uH), because the calculation is for the air coil and this one has a ferrite rod in the core.
If anyone can estimate the induction of such a coil I will be happy.

https://coil32.net/online-calculators/ferrite-rod-calculator.html

I tried to design a source.
Please take it as the first shot.
I tried universal, the source should know both CV and CC
Sources can be ordered in parallel
The LM5145 connection itself is more or less a catalog issue
Let's pay attention to the feedback.
( For control of PS will be used MCU any as STM32F3

CV mode
Output voltage is sense over resistor divider R12/R15 and to this voltage is over R23 add voltage from DAC2. DAC2 drive output voltage in CV mode.

CC mode

Output current is sense on high side on resistor R13 || R17 2.5m Ohm,
There's a drop in the resitors is amplified on current sense amplifier INA168, in second stage on amplifier U4 MCP6V71.
This voltage is compere with voltage DAC1 on error amplifier U1A
It also depends on whether the resistor R5 or R28 will be installed

Over R5 is output of error amplifier connected to FB input LM5145 is over FB input.
Over R28 is output of error amplifier connected to Soft-start input of LM5145. Only one mode will be choice.


Do you find this a big nonsense?

the o/p filter is 4kHz, for general stability, the control loop should have gain < 1 by 400Hz for volt mode, and <1kHz for well designed current mode

the discharge time for the load must be considered if you want sharp edged pulses - ( and the charge time )

e.g. 200uF and 0.1 ohm load = 60uS for current ( and Vout ) to fall to ~ 10%

For safety have a separate over volt detect ckt on the output that shuts the control IC down ...

There is likely too much gain in your FB components ...

- - - Updated - - -

at 100kHz and 48 to 25V, the delta-I in the 8.2uH choke is 30A, i.e. 20A out +/- 15A - can the choke and fets handle this ...? ( 35A pk)

As transistors I have in the design BSC040N10NS5 and BSC035N10NS5, they are 100A monsters that might be able to handle 35A and should do with a relatively small cooler.
Induictor is HCI2212-8R2 yes her is saturration current only 30A .
Capacitor is ceramics X7R size 1210 and ceramic capacitor the capacity decreases with DC voltage.
Ti have por LM5145 design sw in **broken link removed**

For something like this, you want a controller with a tracking function, meaning you can change the output by adjusting the reference voltage to your feedback controller. The LM5145 offers such a functionality using the SS/TRK pin. That becomes your current control input.

For stuff like this where you need a very versatile PWM controller, I recommend the UC2823. It includes all the basic building blocks for various types of PWM controller, but they aren't internally connected so the user can implement things exactly how they want. For example, all three pins of the error amplifier are exposed, allowing it to implement tracking easily.

UC2823 it is perhaps originally a Unitrode of PWM controller, is not bad, but driver for Synchronous driven is not.

The old unitrode controllers are best used with external gate drivers. Generating complementary outputs for a synchronous rectifier is fairly trivial.