Query, Arduino, I am building a 100v-> 36-5volts DC DC buck converter

Thread Starter

Asce

Joined Feb 17, 2023
7
some years pass, I have the same problem to solve, I am building a 100v-> 36-5volts DC DC buck converter, everything is sorted out except the control circuit, given that this stuff will operate at 100 volts, an analog approach would be an insane work, a digital approach could simplify everything, except, how do I power that arduino reliably ?? well, doing some math I figured that a zener diode regulator would dissipate way too much heat to be practical.
I came up with an idea, basically use the zener regulator approach as a "kickstarter" for the system, it'll only have to work for less than a second until the arduino switches the 100 volts down into a safer range for a commercial buck converter to handle, then the arduino will switch the zener circuitry off (via mosfets), and switch the buck converter supply on (use a dead time + capacitors to keep the arduino on during that time), this will get rid of the heat dissipation issue, issue is, when I simulated a circuit for that, I couldn't get rid of a ground loop that was interfering with the mosfets, an isolated power supply is simply necessary, I am still experimenting on this project. . .
 

Thread Starter

Asce

Joined Feb 17, 2023
7
I was playing around with a mosfet gate bootstrapping circuit in falstad and I came up with this :

1676637896125.png

I used two other (N and P types) mosfets to switch the charge and discharge of the gate, a zener diode to reach the threshhold voltage for the main mosfet gate and the switch that switches the mosfets is supposed to represent an optocoupled signal.
in the simulation it seemed to work perfectly, but I haven't found anything like this on the internet so I am feeling skeptical.
do you guys think this would work in reality (especially at 110 volts ?) ?
 

Thread Starter

Asce

Joined Feb 17, 2023
7
Update, the mosfets didn't work properly in the previous circuit (I was feeding the same voltage to both mosfets) here's the new circuit:
1676639875502.png
 

LowQCab

Joined Nov 6, 2012
4,026
Are You talking about ~100-Volts DC, or 100-Volts AC ?

What is the exact description of the "100-Volt" Power-Supply for this Buck-Converter ?
How much Current is it capable of Sourcing ?

How much Current will You need from the Output at ~36-Volts ?

A Micro-Controller is not the best choice for running a Buck-Converter,
especially when there are plenty of cheap and reliable Chips that are specifically designed for the job,
and will do a much better, and more efficient job, than what is practical using a Micro-Controller.
.
.
.
 

Papabravo

Joined Feb 24, 2006
21,159
Do I understand that you want to take an input of +100 to +110 Volts DC and produce outputs of +36 Volts DC and +5 Volts DC. In addition, you want the +5 Volts DC to power the Arduino. I think this plan has at least one serious flaw. There will be a startup period where the +5 Volt DC output needs to be tightly controlled and during which time the Arduino may not be executing instructions. The conflict of these two realities may be impossible to resolve. You may need another approach.
 

Thread Starter

Asce

Joined Feb 17, 2023
7
Do I understand that you want to take an input of +100 to +110 Volts DC and produce outputs of +36 Volts DC and +5 Volts DC. In addition, you want the +5 Volts DC to power the Arduino. I think this plan has at least one serious flaw. There will be a startup period where the +5 Volt DC output needs to be tightly controlled and during which time the Arduino may not be executing instructions. The conflict of these two realities may be impossible to resolve. You may need another approach.
I am aware of that, for this, I will be using a "kickstart" circuit consisting of a simple zener regulator and a power resistor that limits the current to around 50mA enough to run the arduino and allow it to switch the MOSFET, after the output is stabilized at 36 volts, it'll switch off that "kickstarting" circuit with enormous losses, and run on a more efficient buck converter fed from the output, however, for now I am avoiding using an isolated power supply as much as possible since it'll be a while before I recieve one (they're not sold here for some reason).
I don't mind the complexity and the cost of the circuit going up for that matter.
I did experiment a bit with that, I did run into "ground loops" too so, there may be no simple way to go around this issue without using either an external power supply or an isolated one. . .
 

Thread Starter

Asce

Joined Feb 17, 2023
7
Are You talking about ~100-Volts DC, or 100-Volts AC ?

What is the exact description of the "100-Volt" Power-Supply for this Buck-Converter ?
How much Current is it capable of Sourcing ?

How much Current will You need from the Output at ~36-Volts ?

A Micro-Controller is not the best choice for running a Buck-Converter,
especially when there are plenty of cheap and reliable Chips that are specifically designed for the job,
and will do a much better, and more efficient job, than what is practical using a Micro-Controller.
.
.
.
I am talking about 110VDC and the output @ 10Amps to be exact,I couldn't find chips that are rated for such current locally, so, I decided to design my own, to keep things simple enough (with those voltages and currents) I figured a digital system is the best suited for this, if you have references for chips thatcan handle these voltages and currents, may you please suggest me some I can order ?
 

ronsimpson

Joined Oct 7, 2019
2,989
Almost every company makes parts for an AC to DC power supply like this.
This one regulates without looking directly at the output voltage, so its regulation is a little loose.
The point is there is application notes for AC to DC isolated supplies. Remove the bridge diodes and feed the 100V in at that point. Re-work the math for 100V not 120AC. (lower the primary turn count a little)
1676655384288.png
Some commercial supplies like this will work at 100VDC just fine. You might consider using an off the shelf approach.
 

Papabravo

Joined Feb 24, 2006
21,159
I am aware of that, for this, I will be using a "kickstart" circuit consisting of a simple zener regulator and a power resistor that limits the current to around 50mA enough to run the arduino and allow it to switch the MOSFET, after the output is stabilized at 36 volts, it'll switch off that "kickstarting" circuit with enormous losses, and run on a more efficient buck converter fed from the output, however, for now I am avoiding using an isolated power supply as much as possible since it'll be a while before I recieve one (they're not sold here for some reason).
I don't mind the complexity and the cost of the circuit going up for that matter.
I did experiment a bit with that, I did run into "ground loops" too so, there may be no simple way to go around this issue without using either an external power supply or an isolated one. . .
I think this will be a mistake and cause no end of problems which you may or may not be able to debug. A better approach might be to run the Arduino from a battery/charging circuit. In that fashion your circuit is always under control. But hey, you do you if that is what you want. Who am I to argue with you?
 

LowQCab

Joined Nov 6, 2012
4,026
It doesn't have to be complex, and "unavailable",
that is, if You don't mind the package being physically larger that a polished commercial product.

You still have only given little snippets of information
regarding the whole setup and it's ultimate purpose in life.

This forces every response to be a vague generality.

An additional question .......
Why are You set on using a Micro-Controller as a Buck-Regulator ?
This makes no sense, unless You are just trying to see if You can pull it off.

This project can be done using standard, in-stock, Jelly-Bean-Parts.
.
.
.
 

Thread Starter

Asce

Joined Feb 17, 2023
7
Almost every company makes parts for an AC to DC power supply like this.
This one regulates without looking directly at the output voltage, so its regulation is a little loose.
The point is there is application notes for AC to DC isolated supplies. Remove the bridge diodes and feed the 100V in at that point. Re-work the math for 100V not 120AC. (lower the primary turn count a little)
View attachment 287767
Some commercial supplies like this will work at 100VDC just fine. You might consider using an off the shelf approach.
that's the catch, the circuit I am going for is not AC-DC, it's DC-DC (The input has no alternance for a transformer to work)
however, what you just said gave me an idea, I'll try to figure it out, if it works I'll share it in here. . . thank you !
 

Thread Starter

Asce

Joined Feb 17, 2023
7
I think this will be a mistake and cause no end of problems which you may or may not be able to debug. A better approach might be to run the Arduino from a battery/charging circuit. In that fashion your circuit is always under control. But hey, you do you if that is what you want. Who am I to argue with you?
yeah this makes sense, this was my final approach tbh, I was gonna make the arduino run on a seperate li-ion battery, but it just seemed like an inconvenience to have a bettery in your system while it could somehow power itself by itself.
thanks anyways !
 

LowQCab

Joined Nov 6, 2012
4,026
The solution depends entirely on exactly what is desired as a "valuable-final-product",
which we, unfortunately, are still in the dark about.
.
.
.
 

dendad

Joined Feb 20, 2016
4,452
For the 5V supply, I would try an old universal mains phone 5V plug pac. It will probably work well enough for the Arduino current needed.
They work on AC and DC input as the first part in them is usually a rectifier bridge. 100VDC is below the min input of generally 90AC, but give it a go. You will pick one up for nothing usually as there are tons around. If you are like most electronics enthusiasts, you probably have a box of them already.
 
Top