Wide range input 5V power supply

[dknguyen]

These last three posts have greatly changed my decisions on this project. I have started to question myself for choosing an uoper limit of 5000V. As tsmtech said, if someone uses that much of a voltage level, they probably have their own equipment to measure it. So I decided to lower my voltage limit to 400V.

How about this, do you have any other suggestion?

You still want to be able to 400VDC as well right? Not just 400VAC? If you want to measure 400VDC, then I would not do line powered. It's difficult and expensive to turn 500VDC into 5VDC.

If it's AC measurement only and you want it line-powered (you could still use batteries of course), then I would have two selectable transformers for the regular household range (120VAC-240VAC) and another transformer for 240-400VAC. Within the input range of each transformer, it would need to step the AC voltage down enough so it could be handled by a regular AC-DC converter. So for the minimum input of each transformer, the output voltage would have to always be at least 6-7Vrms (8.4VAC-9.8VAC) so that the AC-DC converter could change it to 5VDC. And the output can't be too high at the maximum input voltage either because the AC-DC converter still needs to be able to handle it. So you have to find a AC-DC converter the minimum and maximum outputs of the transformers.

 

[tcmtech]

I would look into an LCD display system.
That alone could drop your power requirements so low that a set of good AA battery's could power it for a few years or use a simple solar panel too.
Or use a micro amp power source taken from the voltage dividers used for the voltage sensing circuits could easily power it as well.

Over all your design would be competing against a large market of self powered industrial metering devices presently available.

So far self powered LCD based voltmeters, amp meters, frequency meters, temperature gages, pressure gages, flow meters, timers, hour meters, and power meters are already available and competitively priced as well.

 

[mneary]

I would like to see the cables on that Turkish meter handle 130A for more than a few minutes.

 

[hkBattousai]

I would like to see the cables on that Turkish meter handle 130A for more than a few minutes.

I doubt it too.

You made me ask myself this question: What is the relation between the minimum required cross-section area (A) of a copper meterial which carries I amperes?

Is there a linear relation between A and I? Of so what is the I/A ratio for a copper wire?

 

[Reloadron]

I would start here and gain an understanding of resistance and resistivity with a focus on rho (ρ) and move along to this location. That should get you going.

You will see the relationships between wire gauge and ampacity.

Ron

 

[hkBattousai]

Could this circuit work?

The zener diode Zs is for generating supply voltage for the opamp. The zener diode Zr gives a reference voltage. The opamp compares this reference level with the voltage on Co capacitor, charges the capacitor until the voltage on it reaches to the reference level (5.1V). The transistor (I put the model 2N3904 symbolically, I know it is not able to stand as high as 600V) controls this charging current.

Any suggestions, or optimizations?

 

[Hero999]

That's a linear regulator, the power dissipation will be huge even at low currents.

See if you can find a way to convert 5kV-10kV to around 300V-600VDC then you can use an off the shelf mains powered SMPs to convert it to 5V.

The input power going down to DC is the real killer, for example if it were 50Hz to 1kHz then you could use a transformer.

 

[tcmtech]

If the load is equivalent to around 10K of resistance a simple voltage divider with a zener and capacitor to limit and filter the 5 volts is more than adequate.
It gets rid of 95% of that circuit while still providing the same function.

 

[hkBattousai]

That's a linear regulator, the power dissipation will be huge even at low currents.

You mean the power dissipation on Rc? Because that's the only element that would heat up. I am planning to assign a smaller value for it, less than 100Ω.

If the load is equivalent to around 10K of resistance a simple voltage divider with a zener and capacitor to limit and filter the 5 volts is more than adequate.
It gets rid of 95% of that circuit while still providing the same function.

Actually load will draw slightly more current. Load is a microcontroller which drives an LCD module, LCD has background light. Overall circuit requires 50mA current from source side. 50mA is a too much value to design this circuit with just a resistor-zener pair.

How about replacing Rc with a smaller one? That would both reduce the power dissipation and gives out more supply current?

 

[tcmtech]

At 5 volts 50 Ma is only 1/4 of a watt.

Most small zeners are .5 to 1 watt but much higher wattage ones are available as well.

The zener only has to dissipate the power the circuit doesn't use that is in excess of of its needs.

 

[hkBattousai]

At 5 volts 50 Ma is only 1/4 of a watt.

Most small zeners are .5 to 1 watt but much higher wattage ones are available as well.

The zener only has to dissipate the power the circuit doesn't use that is in excess of of its needs.

There is still a problem with zener. My circuit is supposed to have a variable input level, between 10V to 600V (rectified & filtered 380V AC). A zener-resistor pair designed for 600V will not work when the input is 10V, vice versa. Just because of this, I designed such a circuit.

 

[hkBattousai]

Can you please tell me onething, is there any DC-to-DC conversion technique similar to mine?
I designed this "charge capacitor to 5.1V from high voltage source" circiut from my mind, I don't even know if this design has flows or issues. What do you think about this circuit, can it work as it is supposed to?

 

[mneary]

What do you think about this circuit, can it work as it is supposed to?

When the input is 10V, the 20V supply won't happen, and the 5V zener will be starved with only 10 microamps. Even at a generous 30V, "Rs" will only provide about 200 microamps.

When the input is 600V and the load is 50mA, the "Q?" will be dissipating (595*0.05) = 29.75 watts. "Rz" will be dissipating (595^2)/470k = 3/4 watt. "Rs" will be dissipating 7.5 watts.

So the circuit won't work at 10V and at 600V it will be consuming about forty watts.

 

[hkBattousai]

Thank you for the comment.

When the input is 10V, the 20V supply won't happen, and the 5V zener will be starved with only 10 microamps. Even at a generous 30V, "Rs" will only provide about 200 microamps.

When the input is 10V, the zener will be cut-off, the opamp will be fed with 10V, I think it is enough to keep the overall system running, of course if we consider the leakage current of the zener as low as nanoamperes level.

When the input is 600V and the load is 50mA, the "Q?" will be dissipating (595*0.05) = 29.75 watts. "Rz" will be dissipating (595^2)/470k = 3/4 watt. "Rs" will be dissipating 7.5 watts.

So the circuit won't work at 10V and at 600V it will be consuming about forty watts.

I think, the high power consumption at high input voltages is the major issue of my design. I have to revise it once more.
I will try using capacitors or inductor for voltage dividing. I don't know how to do, but I had better start working on it...

 

[mneary]

When the input is 10V, assuming the op amp can function with 7.5V, it will only receive 500 microamperes. If it passes all of this current to the "Q?", the transistor gain needs to have a minimum (not typical) gain of 100.

 

[tcmtech]

If you just use a zener diode to control the gate of a higher voltage Mosfet of IGBT you can get the 5 volts output over the 10 - 600+ volt input with only around 5 - 6 components.

 

[hkBattousai]

If you just use a zener diode to control the gate of a higher voltage Mosfet of IGBT you can get the 5 volts output over the 10 - 600+ volt input with only around 5 - 6 components.

Intresting.

I remember learning about IGBTs at my power electronics course, but I've totally forgotten about them. All I remember is that they were mosfet input BJT ouput, cascaded transistor devices.

Until now, I didn't want to use mosfet transistor in my design, because they have low Rds resistance on cut-off. This problem requires selecting lower Re values which is undesirable.

 

[tcmtech]

I have used that circuit before with my wind power experiments without problems up into the 200 - 300 volt peak ranges while providing a 12 volt source for control circuits.
If the device only needs 5 volts and the Mosfet or IGBT has a gate threshold of around 5 volts then the minimum working voltage tends to be around 10 volts on up. Plus being voltage controlled the gate drive works with a 1+ meg pullup resistor so there is almost no power lost on that part of the circuit.

 

[Hero999]

You mean the power dissipation on Rc? Because that's the only element that would heat up. I am planning to assign a smaller value for it, less than 100Ω.

If Rc is only 100R how is it supposed to dissipate all the power?


The 2N3904 has a maximum voltage rating of 40V, if the input voltage is 600V, the voltage across Rc will need to be 555V so the current will need to be 5.55A so the the power dissipation will be 3.1kW.

 

[Sceadwian]

mmm portable sun. Excellent.