Is this a safe voltage divider schematic, and where can I learn about how/why?

[eihli]

I'm at a stage of learning about electronics where I don't know what I don't know and I'm a bit worried about doing something that could cause a fire. Can anyone point to a good resource for gaining an overview of electronics safety? A book or an online course?

The thing that prompts that question is I'm trying to re-use an old weed-eater battery to power a little thermal printer. The weed-eater battery is 40v and the thermal printer needs 12v and about 2 amps.

I looked up schematics for a voltage divider and have a pack of 1/4w resistors. Would the following circuit work?

[Fronberry]

It's not recommended to use a voltage divider to "step down" voltage for a load.

 The resistors have only 1/4W rating, and will have 2 amps flowing through them.  P = I² R

So for those two resistors you'd have upwards of 1600 watts.  They will not be resistors for long!

The most common circuit for this type of application would be a "step down" or "buck" converter:  https://en.wikipedia.org/wiki/Buck_converter

As for safety, ohm's law and datasheets are your best bet.  Measure thrice, cut once!

[ataradov]

No, this will not work with a real variable load. This only works because voltmeter presents very little load. But once you connect your thermal printer, its input resistance will be connected in parallel with R1, and overall resistance will change.

You need a proper voltage regulator. The simplest ones are sold all over the place as DC-DC converter modules. Although 40 V is too high for most common units (based on LM2596), they typically rated for 35 V max input voltage.

[John B]

It's not recommended to use a voltage divider to "step down" voltage for a load.

 The resistors have only 1/4W rating, and will have 2 amps flowing through them.  P = I² R

So for those two resistors you'd have upwards of 1600 watts.  They will not be resistors for long!

The most common circuit for this type of application would be a "step down" or "buck" converter:  https://en.wikipedia.org/wiki/Buck_converter

As for safety, ohm's law and datasheets are your best bet.  Measure thrice, cut once!

R2 will limit the power to around 1.6W, not 1600W as it's impossible for a load to draw 2A in this case.

But it's power-excessive for a voltage bias reference and useless as a power supply.

Many common DC-DC converters will also have a higher voltage version around 60V input rated. But if the OP is just starting to design stuff, rolling your own DC converter is pretty optimistic. You could try to find a step down module, but I'd be very wary of anything on ebay/ali express or whatever. Those ICs are fake at the best of time and even if they claimed to be a HV version, there's no way I'd power them with 40V.

Maybe something marketed as a step down converter for 48V supplies would be a better choice.

[james_s]

No, this will not work with a real variable load. This only works because voltmeter presents very little load. But once you connect your thermal printer, its input resistance will be connected in parallel with R1, and overall resistance will change.

You need a proper voltage regulator. The simplest ones are sold all over the place as DC-DC converter modules. Although 40 V is too high for most common units (based on LM2596), they typically rated for 35 V max input voltage.

The LM2596HV will do it, and it's a part that I've used before and found it to work very well, max of 60V input. It should be noted that the VAST majority of cheap "LM2596" modules you can buy cheaply are fake, they are real voltage regulators and they do work ok but they are not real LM2596 parts, they don't even operate at the same frequency.

[ataradov]

Yes, "LM2596" is more of a search term than a part number. Jumping from a voltage divider as a power supply to making a custom board based on an IC sounds like a stretch, so whatever solution is proposed should be as easy as those modules.

[John B]

This solution might be pricier than the OP was hoping, but I'd be looking at modules on a parts supplier like digikey or element 14.

Those cheap modules lack output protection and you risk the full ~40V being supplied to the load in the event of a failure. (Pretty likely let's be honest)

[Ian.M]

There's also an older version, the LM2576HV [datasheet], which operates at 52KHz rather than the 150KHz of the genuine LM2596 series regulators.

Usually the LM2596 series parts are preferable, as the LM2576 needs larger capacitors and inductor, so is more expensive to implement, but for a near novice trying to build a switching regulator on stripboard or other protoboard, the lower switching frequency make it more forgiving of a less than ideal circuit layout.
I have built LM2576 buck regulator circuits using the TO220-5 through hole package on stripboard, and as long as you do the math properly for the design equations in the datasheet 'Application and Implementation' section, choosing the next higher preferred value for the inductor and capacitors, get an inductor that can handle the calculated peak current without saturation, and follow the layout hints in Fig. 8.9 (star grounding and direct connection of the input capacitor and catch diode to the device pins), you can expect it to 'just work'.

Also "LM2576HV" as a search term for prebuilt modules avoids all the 'results noise' due to the fake LM2596 chips!

If you do choose to use a cheap off-brand module, I strongly recommend a crowbar circuit set for 10% overvoltage at the output.  Use a heavy duty TRIAC with a surge rating several times that of the fuse (hint: estimate the loop impedance and thus the fusing current and compare I²t curves) See the TL431 precision reference datasheet, fig. 30.

[BeBuLamar]

With a voltage divider you waste a lot of power in the resistors and they have to be big because they do get very hot. I think the best bet for your problem is to use something like this
https://www.amazon.com/Adjustable-Programmable-Step-Down-Regulated-Converter/dp/B01N0VKTC2/ref=asc_df_B01N0VKTC2/?tag=hyprod-20&linkCode=df0&hvadid=241973970700&hvpos=&hvnetw=g&hvrand=2284932261684792355&hvpone=&hvptwo=&hvqmt=&hvdev=c&hvdvcmdl=&hvlocint=&hvlocphy=9026816&hvtargid=pla-399690329417&th=1
 
Of course Ruideng  sells cheaper else where.

[james_s]

Yes, "LM2596" is more of a search term than a part number. Jumping from a voltage divider as a power supply to making a custom board based on an IC sounds like a stretch, so whatever solution is proposed should be as easy as those modules.

A bit of a stretch yes, but not impossible. I have built one on perfboard and they are simple enough that we can essentially give him a recipe to follow if he's interested in that route.

[Zero999]

Yes, "LM2596" is more of a search term than a part number. Jumping from a voltage divider as a power supply to making a custom board based on an IC sounds like a stretch, so whatever solution is proposed should be as easy as those modules.

A bit of a stretch yes, but not impossible. I have built one on perfboard and they are simple enough that we can essentially give him a recipe to follow if he's interested in that route.

An option is to buy a cheap module, just for the PCB. Remove the fake LM2596 and replace it with a genuine LM2596HV. Needless to say, the electrolytic/tantalum capacitors should be changed to decent ones with sufficient voltage ratings.