Help with obtaining constant DC current from buck converter, please!

Thread Starter

selvac19

Joined Dec 23, 2017
26
Hello peeps!

Does any of you know how I can obtain a constant DC current using a buck converter configuration circuit?
If additional info is required, I can provide further. The idea is to see if knowing the basic concepts of a buck converter, it would be possible to visualize the buck principle as a constant DC current source supply?

Thanks in adavance for any insight.
 

GetDeviceInfo

Joined Jun 7, 2009
1,741
If your converter had differential input into the feedback, then easy. Otherwise, you need to converate the differential ( voltage over a resistor), and convert that to a single ended input for feedback
 

crutschow

Joined Mar 14, 2008
25,690
Basically instead of using feedback from the output voltage to control the converter PWM duty-cycle, you use the voltage across a shunt resistor in series with the load output, which gives a voltage proportional to the output current, to control the converter PWM duty-cycle.
Thus the output is regulated to a constant current instead of a constant voltage.
 

MrAl

Joined Jun 17, 2014
7,849
Hello peeps!

Does any of you know how I can obtain a constant DC current using a buck converter configuration circuit?
If additional info is required, I can provide further. The idea is to see if knowing the basic concepts of a buck converter, it would be possible to visualize the buck principle as a constant DC current source supply?

Thanks in adavance for any insight.
Hello there,

It depends highly on what parts you are willing to add to the project, and also how much accuracy you need, and if you can afford a slight increase in overhead voltage.

If you dont need super accuracy a transistor and resistor can be used to sense current and bypass the voltage regulation (even though you can keep that too) and turn it into current regulation. This is a super simple way to do it partly because the transistor is so fast that it does not interfere with the normal stability, which is an important consideration when adding current feedback.
The slight downside is the accuracy is maybe around 5 percent over some temperature variation and there is a slight increase in overhead voltage (0.6v roughly) so you need to have a decent input supply. There is also a decrease in efficiency which may or may not be significant depending on output voltage.
I've used this trick for current limiting in Simple Switcher applications that dont come with current feedback.
 

Thread Starter

selvac19

Joined Dec 23, 2017
26
Alright, I understand your insights and highly appreciate them all. Now, I have a current regulator design someone worked on, however, I still do not completely understand the mechanism of this buck converter adpation in light of what each component is doing in each of their corresponding positioning. The idea is: 1. To understand the operation of this particular circuit which is comprised of 9 converters (so taking 1 of the 9 is enough to understand as each provides a constant DC current of 85 A) and 2. To verify at a glance if this design is stable or reliable enough to accomplish with the oriented purposes.

Thanks again to all and any other insight would be highly appreciated.

upload_2018-6-27_18-2-27.png
 
Last edited:

MrAl

Joined Jun 17, 2014
7,849
Hello again,

With a quick look it looks like it could just be an interleaved design in order to be able ot get much higher current operation from lower rated parts. For this view i assume that the different parallel sections have all the same values or nearly so.
 

Thread Starter

selvac19

Joined Dec 23, 2017
26
Hello again,

With a quick look it looks like it could just be an interleaved design in order to be able ot get much higher current operation from lower rated parts. For this view i assume that the different parallel sections have all the same values or nearly so.
Hello!

Thanks for the insight, and yes, you are right. All the different parallel sections have the same value of 85. 5 ish A. I am nearly confused about the actual operation of the capacitor as of the inductor is directly energized and de-energized from the power supply itself.
All component values were calculated not using the Duty Ratio variable. How does that affect or makes an impact in the operation of the circuit in light of pretending to behave as a constant DC current regulator?

Thanks again for any insight.
 

Thread Starter

selvac19

Joined Dec 23, 2017
26
Hello again,

With a quick look it looks like it could just be an interleaved design in order to be able ot get much higher current operation from lower rated parts. For this view i assume that the different parallel sections have all the same values or nearly so.
Also, would this design be considered stable as of its orientation of being a DC constant current regulator? If more info is required, I would be more than glad to provide so. Thanks again.
 

MrAl

Joined Jun 17, 2014
7,849
Also, would this design be considered stable as of its orientation of being a DC constant current regulator? If more info is required, I would be more than glad to provide so. Thanks again.
Hi again,

The only way i see that drawing being constant current is if the load is also constant. That is because there is no feedback. Could it be possible that it is not drawn right? If the line on the right going to the control circuit was drawn connected to the output, i think it would make more sense, bu then it would be voltage regulated unless the load was constant.
For it to be true constant current there has to be some feedback that measures current.

The placement of the capacitors is strange too. Normally you dont want a cap there because it caues what looks like a longer recovery time for the diode. That is unless of course that cap represents the internal capacitance of the diode itself.

You can analyze one section at a time if you like. That will give you more insight into how each section works. Then just reconfigure as an interleaved design if you like.
 

Thread Starter

selvac19

Joined Dec 23, 2017
26
Hi again,
Hi again,

The only way i see that drawing being constant current is if the load is also constant. That is because there is no feedback. Could it be possible that it is not drawn right? If the line on the right going to the control circuit was drawn connected to the output, i think it would make more sense, bu then it would be voltage regulated unless the load was constant.
For it to be true constant current there has to be some feedback that measures current.

The placement of the capacitors is strange too. Normally you dont want a cap there because it caues what looks like a longer recovery time for the diode. That is unless of course that cap represents the internal capacitance of the diode itself.

You can analyze one section at a time if you like. That will give you more insight into how each section works. Then just reconfigure as an interleaved design if you like.

The only way i see that drawing being constant current is if the load is also constant. That is because there is no feedback. Could it be possible that it is not drawn right? If the line on the right going to the control circuit was drawn connected to the output, i think it would make more sense, bu then it would be voltage regulated unless the load was constant.
For it to be true constant current there has to be some feedback that measures current.

The placement of the capacitors is strange too. Normally you dont want a cap there because it caues what looks like a longer recovery time for the diode. That is unless of course that cap represents the internal capacitance of the diode itself.

You can analyze one section at a time if you like. That will give you more insight into how each section works. Then just reconfigure as an interleaved design if you like.

upload_2018-7-2_11-55-6.png

This is the feedback version of the same circuit. I would appreciate any insight with this new model.
I highly appreciate all your insights provided by far.

Thanks indeed.
 

MrAl

Joined Jun 17, 2014
7,849
Hi,

That looks a little better, but the non standard symbols make it harder to figure out what is happening.
"DLL" what is that for example.
 

Thread Starter

selvac19

Joined Dec 23, 2017
26
Hi,

That looks a little better, but the non standard symbols make it harder to figure out what is happening.
"DLL" what is that for example.
"The general DLL block allows users to write code in C or C++, compile it as a Windows DLL, and link to PSIM. Unlike the simple DLL blocks with fixed number of inputs and outputs, the general DLL block provides more flexibility and capability in interfacing PSIM with custom DLL files." Directly extracted from the PSim website.

Basically, what I am more into understading is the circuitry as itself since I am deeply confused about that specific configuration as if that proposed configuration is correct or at least stable enough.

Thanks again for any insight.
 

MrAl

Joined Jun 17, 2014
7,849
"The general DLL block allows users to write code in C or C++, compile it as a Windows DLL, and link to PSIM. Unlike the simple DLL blocks with fixed number of inputs and outputs, the general DLL block provides more flexibility and capability in interfacing PSIM with custom DLL files." Directly extracted from the PSim website.

Basically, what I am more into understading is the circuitry as itself since I am deeply confused about that specific configuration as if that proposed configuration is correct or at least stable enough.

Thanks again for any insight.
Hi again,

To be able to tell if a design is stable we have to have more complete information about the system. If there are any DLL's to be written they have to ba already written for example.

What i can tell you is that it looks like a preliminary design that does measure output current and use that as feedback, so it looks like it could work. To know for sure, we have to know all the details and even then it may not be easy.

I am assuming that the symbol at the top where the line connects to the Vcc line is a current sensor of some kind. That would be what measures the current and provides feedback to the system. That would be typical for a current controlled regulator.
Now that i look at your original drawing i see that was also had a current sensor but at the time it looked just like a regular dot so it did not look like current was being measured. Now that we can assume it is a current sensor, at least we can say that there is some attempt to measure load current and use it as feedback.
When non standard symbols are being used it is best to label them clearly.
Also a little questionable is why some of the transistors look like bipolar.

The only guess i have about the caps across the diodes is maybe help with radiated EMI. There are no values given however so it's hard to say for sure. Normally we dont want caps there because it makes the recovery time for each diode look longer and that's bad. In this case i would ask the user to go back to the original designer and ask them what they had in mind. Someone had to have drawn that schematic, and they put the caps there for a reason. It could be a mistake or they had some specific reason for doing so that could be either valid or not valid theoretically.
 
Last edited:
Top