To fuse or not to fuse

This is a rough draft of my power supply circuit. I was wondering if I should place a fuse (0.75A) before my voltage regulators. I'm using 3 voltage regulators because the 74vdc system that I'm using is quite noisy so I figured the more the better.

Any suggestions would be appreciated, thank you

... apparently I'm having issues loading the attachment for the schematic. Sorry for the inconvenience. I'm working from an iPhone.

without seeing your circuit.

Hypothetical scenario,
unattended device, circuit fault, dead short
with fuse
fuse burns
without fuse
house burns

Err on the safe side would be my choice.

74volt positive to a voltage divider.
1k ohm to 100 ohm. Roughly 7.4vdc between the resistors.

Resistor dividers are for measuring voltages and conditioning signals, not for power conversion.

Perhaps a DC-DC converter down to 8V or so before the linear regulators would be better?

Ok 3 questions.
A) where could I get one or how would you recommend making one that converts 70-76vdc to 8vdc
B) so would the schematic I drew work for getting signal. There is another part of this project that needs to read when there is 74vdc at a particular spot in the 74v system. The only difference between the schematic I drew for power supply and the schematic for "signal" is two less voltage regulators.
C) would I still need to fuse the signal circuit

It depends on what kind of failure or damage you're trying to protect against...

A fuse is generally there to prevent a fire or to prevent wires from burning-up, etc. Solid state electronics can blow faster than a fuse, so sometimes the transistor/MOSFET/etc. shorts and then the fuse blows preventing any further damage. Or, solid state electronics can be killed by heat without the fuse ever blowing.

I always fuse the AC side when I build a power supply. If it's a "bench" supply or something that might be connected/disconnected to different loads it might make sense to fuse the DC output or to use some other kind of short-circuit protection. (Many voltage regulators chips are short-circuit-and over-temperature protected, so you might not have to add any additional circuitry.)

Adding a fuse can (usually) never hurt, especially with a source that may not be protected from short / overloading. The only negative I can see is the existence of a fuse might give your a false sense of security / safety.

Regarding your circuit. Without knowing the output current requirement and variations in output current one cannot know if your resistor divider scheme can work. Could be iffy

I don't think paralleling 3 regulators is a good idea. Likely one regulator will be slightly higher that the other, the regulator with the lower output will see its output is higher that it wants and start to shutdown.
Final result the regulators will not share the load evenly.

You might look at ...TL783 High-voltage Adjustable Regulator. I don't know anything about it except I read the 1st page, but I expect there are more options out there.

Curious, what were your thoughts then adding the two diodes?

The diodes were placed incase somebody hooks it up backwards.
If hooked up backwards the voltage divider would be incorrect

Just try a 5volt cellphone charger.
The mains powered 5volt supplies I tried worked from ~50volt DC.
Current capability could be an issue.

ok, so this was my figuring...
the Arduino is going to activate (digitalWrite) two solid state relays with one LED parallel to each SSR, they will be activated via NPN transistors. one for each ssr.

MarkT said voltage dividers shouldn't be used as power supplies.

Ohm's law E(74) / R(1k) = I(0.074) so probably not enough

but what if I changed the voltage divider to 100ohms into 10ohms. that would make I(0.74)

and that should be enough to power my circuit, yes?

How much current can the 74V supply source? Also don’t think it wise to place a diode on the regulator ground. That puts the regulator ground reference .7V above supply ground reference.

What is this 74V anyhow? Is it AC, DC, something in between?

Below the ---- line is what I was going to say but looking at your schematic again there are a few things bothering me.......

  1. your divider results in 6.7 volts into the regulators. (I'm ignoring the diodes for now)
  2. you have 3 regulators so I assume you need more current that one regulator will supply.

What will the 100 ohm resistor voltage be when the regulators start supplying current? This current will have to go through the 1k resistor causing more voltage drop and likely resulting in too low a voltage into the regulators.

You should probably post some current requirements.

If you must use a resistor divider I would suggest a hefty zener in place of the 100 Ohm resistor.

I would start with a Zener value of about 80% of the max voltage of the regulators. Unless the regulators cannot handle that much power. Then you will have to post some more information else we will be chasing our tail trying to balance zener current with regulator capability.

Ideally you would calculate your series resistor so

at the 74V minimum (assuming there is some variability)
AND the maximum load
The zener current is minimal, maybe even zero.

Then at no load and circuit current minimum you check to see you aren't burning up the zener. If you are, You will need to get a bigger Zener.

Now if you want to get a little fancier but not go to the high voltage regulator, with a casual google for high voltage regulator I found several app notes that would use a LM317 with external transistors to create a high voltage regulator.

The LM317HV is 60volt differential max.
An TL783 (125volt) seems to be commonly used for high medium/voltage linear supplies.

But why.
As said in post#9, a mains powered 5volt cellphone charger will porbably do the job.
Just tested one from the junk box, and that just was able power an Uno from a 30volt lab supply.
~74volt (on the mains power terminals) shouldn't be a problem.

What application needs you to do this voltreg circuit.
Can't you find a lower DC supply to supply your regs?

To convert 74Vdc to 5Vdc you would be better to use opto-isolation.
74Vdc is considered high voltage and dangerous.

Tom... :slight_smile:

thanks for the comments.
@TomGeorge...74vdc is the "lowvoltage control" side of locomotives, especially anything before 1985.

I was attempting to get power from that source since it is a common voltage on locomotives.
charging system 74vdc
Battery 64vdc either by two 32v batts or eight 8v batts.

I have a partner for this project whom I spoke with earlier today, he informed me that this particular locomotive has as 74v to 12v converter for the communications radio (that info was not given to me at the beginning of this post)

I guess i'll pull my power from the 12vdc. I can attach directly to the 5v voltage regulators.

with that being said, I still need a circuit that senses when a part of the 74v locomotive system has 74v applied to it.... would my schematic above work for that?? (only using on VR instead of 3) or should perhaps coming from the voltage divider, use that signal and attach it to an NPN to control a 5v signal to the arduino's pin??

I see a number of possibilities but before making suggestion I have some questions.

  • Is the 12V power you can access have the same ground as the 74V?
  • If yes are the two grounds physically close? i.e. in an automobile, the grounds for different modules can be up to +/- 2 volts different
  • I'm going to assume the 74V is not "signal" quality but likely has spikes in excess of 150V
  • You don' need any power from the 74V line but only need to know when it is on or off. (Is this correct)
  • Is it acceptable to know if the 74V line is at 0 or say above 40V?
  • Your sensing input is the input of an arduino type board running at 5Vdc

To convert 74Vdc to 5Vdc you would be better to use opto-isolation.
74Vdc is considered high voltage and dangerous.

A mains power cellphone supply is transformer isolated.