Looking for dirt cheap method DC power supply

[Sceadwian]

Space Varmint, seriously I think you may need to have your head examined closely by a specialist. Not to be offensive but 90% of tantalum failures are dead shorts they're polarized and they're intolerant to voltage spikes. Using a tantalum for an in line on mains supply is criminally negligent. The whole thing will go up in flames at the drop of a hat.

 

[Space Varmint]

Appnote 00954A.PDF is one so is TB008
**broken link removed**

Do they even make 250VAC Tantalum?

Very Nice AP Note! Looks allot like tcmtech's schematic. Hey, tcm? Did you work that out in practical application? If so, what sort of current or power did you achieve using those values? I may run with it.

Thanks

 

[Space Varmint]

Space Varmint, seriously I think you may need to have your head examined closely by a specialist. Not to be offensive but 90% of tantalum failures are dead shorts they're polarized and they're intolerant to voltage spikes. Using a tantalum for an in line on mains supply is criminally negligent. The whole thing will go up in flames at the drop of a hat.

Hey! I've seen them blow. I think you have to be careful in your ratings. I have also seen them hang in there for years. I need small.

 

[tcmtech]

I have been using that circuit for so many years I cant remember when I learned how to do it! I have built dozens of small low power line powered devices using this as the power source.
Being capacitive coupled the efficency is unbeatable! Close to 100% transfer with a near immeasurable internal power loss.

I have seen it used hundreds of times and its a very common power supply in small line control devices.
If you use the standard capacitive reluctance formulas you can figure out exactly what value you need for the C1.
IF I remember correctly at 120 VAC and 60 cycles 1 uf gives around 40 - 50 usable Ma at 5 volts.
Just make sure your zener can dissipate the wattage if the circuit has no load on it!

Granted a 5 volt 1 watt zener can hold far more than the 100ma you are planning to work with.

If C2 is about 50 times larger than C1 it can take some very nasty line spikes and noise and still have 5 volt logic IC's survive unaffected!

 

[Sceadwian]

Space, the basic fact is that 90% of them fail dead short, that simple fact ALONE should completely preclude their use when connected to mains power directly because of the (quiet serious) ball of plasma that will be produce when it fails, taking that kind of chance just because you want a small supply is completely irresponsible no matter how you look at it.

This is not the first time someone has asked about a compact inline power supply like this and the same reasons for NOT doing it stand. Can you honestly justify the risking of human life and personal property for the sake of making something smaller? What is the all encompasing concern with size anyways, I have 6 volt 100-200ma supplies that are half the size of a deck of cards, and there is no normal reason I can think of to want to try to hide something like this. If you're going to do it inline with the power cord build a real power supply and put it inside the cabinet of whatever device you're switching.

 

[blueroomelectronics]

But that simple version would not make it past UL approval. You need another two resistors and another cap as laid out in the Microchip TC008 note (yes same schematic)

 

[Space Varmint]

Space Varmint, seriously I think you may need to have your head examined closely by a specialist. Not to be offensive but 90% of tantalum failures are dead shorts they're polarized and they're intolerant to voltage spikes. Using a tantalum for an in line on mains supply is criminally negligent. The whole thing will go up in flames at the drop of a hat.

OK, got most squared away. Now, let's discuss this. First of all the voltage will be only one direction. No AC. Then we have an RMS value to consider but I think the usual safety rating is twice the peak voltage due to storage. This pertains only to full wave rectification if I remember correctly. I have had success with values slightly over twice the RMS value.

I am surprised that in recent years you guys don't feel like there have been any design improvements. I believe the success rates have increased over the years. Does anyone know right off hand if they are used in GFCI? I have seen inline ground fault circuits but not the inside of them.

 

[Space Varmint]

I have been using that circuit for so many years I cant remember when I learned how to do it! I have built dozens of small low power line powered devices using this as the power source.
Being capacitive coupled the efficency is unbeatable! Close to 100% transfer with a near immeasurable internal power loss.

I have seen it used hundreds of times and its a very common power supply in small line control devices.
If you use the standard capacitive reluctance formulas you can figure out exactly what value you need for the C1.
IF I remember correctly at 120 VAC and 60 cycles 1 uf gives around 40 - 50 usable Ma at 5 volts.
Just make sure your zener can dissipate the wattage if the circuit has no load on it!

Granted a 5 volt 1 watt zener can hold far more than the 100ma you are planning to work with.

If C2 is about 50 times larger than C1 it can take some very nasty line spikes and noise and still have 5 volt logic IC's survive unaffected!

Thank you That's very excellent information! Appreciate it much.

 

[Nigel Goodwin]

Oh my my my. What do you think the purpose of Tantalum, Coltan based devices are for? High power / compact structure.

We use 120VAC 60 cycle over here Nigel.

I'm quite aware of what your mains is - but I'm also aware that you can't use a polarised capacitor for a mains dropper - as is blindingly ovious

It's essential that you use a correctly rated capacitor that's specifically designed for this exact use.

I'm presuming you are aware that such circuits leave your project live to the mains?.

 

[tcmtech]

Is this a personal project I assume?
If so you dont actually need to worry about the UL and other stuff unless it going into a production application.
As far as the size of C1 I have some 2.2 uf 200VAC rated snubber caps I have used for circuits like this that are about .250" x .500" x 1" . The three diodes and the 6.3 VDC cap take up another .250" past the .5" side.
So the total circuit is still under .250" x .750" x 1" inches.
Less total volume than a match book!

But if absolutely needed a 2.2 meg resistor in parallel to C1 and one 100K in parallel to C2 work just fine too!
I just went by your asking for the dirt cheap and simplest design request for the posted schematic!

 

[Sceadwian]

Nigel I think he intends to half wave rectify the incoming AC.

Space you can justify it any way you want, no improvement in tantalum construction can override the simple fact that if it DOES fail there's a 90% chance of a fire, it will draw more than enough current to create a plasma ball but not enough to cause the breaker to trip before it goes up. Also you're forgetting mains spikes, which will destroy tantalums. Single or sub cycle transients on 120 volt lines can be in the thousands of volt range, say goodbye to your tantalum and yer house if no one's home.

 

[Nigel Goodwin]

Nigel I think he intends to half wave rectify the incoming AC just failed to mention it.

Space you can justify it any way you want, no improvement in tantalum construction can override the simple fact that if it DOES fail there's a 90% chance of a fire, it will draw more than enough current to create a plasma ball but not enough to cause the breaker to trip before it goes up. Also you're forgetting mains spikes, which will destroy tantalums. Single or sub cycle transients on 120 volt lines can be in the thousands of volt range, say goodbye to your tantalum and yer house.

LIke I said - silly idea!.

 

[Speakerguy]

There is a very critical piece on the Microchip app note written here by a very knowledgeable fellow:

**broken link removed**

Just thought I'd throw it out there and see what the group thinks.

 

[tcmtech]

Thanks for the post speakerguy79! I actually put it in my tech info archive for future reference.
I personally have liked using the snubber caps because they had the inline resistor built right in! I completely forgot that part on the schematic! OOPS!

Their C1 and C2 values are a bit different than what I use but I can see their reasoning for it. Most of what I have used these circuits for are not that picky about ripple on the load side but I can see that if micro controller was running at a marginal performance level it could need the extra C2 value.

But in practical application I have found that a little extra C1 value helps if there are any momentary loads that go above the calculated amps value. And any larger size of C2 will work too if the space allows it!

 

[blueroomelectronics]

There is a very critical piece on the Microchip app note written here by a very knowledgeable fellow:

**broken link removed**

Just thought I'd throw it out there and see what the group thinks.

I like the line

the circuit described violates the wiring rules of every developed country on the planet!

So it's fine for the undeveloped countries.

 

[Speakerguy]

Could a TL431 or similar shunt regulator be used in place of the zener? Could give much better ripple rejection. TL431 is good for 100ma.

 

[Hero999]

How much power do you require?

Does it need to be isolated?

What are the regulation requirements?

What about the classic Royer converter?

A couple of high voltage transistors, a ferrite core with some aditional resistors and capacitors.

It's not regulated but you can get quite a lot of power out.

 

[Willbe]

Thats an easy one!
This is as simple of regulated power source you can get thats directly line powered.
The AC cpacitor value should be matched to give you about 20 Ma or so higher than what the circuit uses at peak amps draw.

100 mA at 120v is 12w. You could also use an incand. bulb as a dropping resistor.

 

[tcmtech]

100 mA at 120v is 12w. You could also use an incand. bulb as a dropping resistor.

Capacitive coupled reluctance does not drop the voltage while maintaining the current. All the actual power loss is at the 5 volt DC part conversion part. In this case 100 ma at 5 volts = .5 watts.
The "dropping" device is the capacitor and the actual wattage losses in it at 60 Hz at 120 VAC is microscopic.

Think of it as a very small power factor correction capacitor with a 5 volt drop across a resistor at the one end. That capacitor itself does not dissipate any energy but the 5 volt drop across the resistor does!

There is no I/R losses in the capacitor. Or at least they are very very small! Thats how it gets an apparent near 100% efficiency.

Its a weird one but a scientific fact non the less! An apparent voltage drop with a measurable amp flow but yet no wattage is lost!

Try it, It works!

 

[Willbe]

Capacitive coupled reluctance does not drop the voltage while maintaining the current.
It has to, through the Z of the capacitor.

An apparent voltage drop with a measurable amp flow but yet no wattage is lost!

Try it, It works!

It's called reactance.