Transformerless Power Suppply

[ERIK]

Hi there,

Despite the obvious safety issues etc, I am very interested in building a transformerless power supply for my PICs. I am in New Zealand where the power is 240VAC @ 50Hz so most of the App Notes/Tech briefs on the web aren't quite appropriate for these conditions. I've been using rectifier diodes and or bridge rectifiers, a Zener, and some heavy duty resistors and caps. I've asked my friend who is an electrical engineer to have a think about it and he begrudgingly agreed so I'll see what he reckons. I was just wondering if anybody else out there has built a good/simple transformerless power supply to run on 240VAC @ 50Hz?

[muddy0409]

Silicon Chip had a project about a light dimmer. That had a method of scourcing power DIRECTLY off the mains. I'll have a look for the mag and post the details if I can find it.
From memory, this small module was fitted inside a lamp with no transformer or anything else, other than direct connection to the Active mains power for its supply.

[Melanie]

Why no transformer?

A transformer is there for good reason (whether it's a conventional iron cored item or a high-frequency switcher), apart from the obvious in dropping the incoming mains supply down to something safe to play with, it also provides ISOLATION from the supply mains which is a very important feature. Without that ISOLATION your circuit is LIVE regardless what voltage you play with.

[muddy0409]

Ahh....there we are....

SILICON CHIP October 2005. You can get a reprint from their website www.siliconchip.com.au for a small reasonable price...

Hmmmmm.............I was wrong, it's a fan controller, not a dimmer.

Yes Mel, you are correct ( as always ) but in this case, built into a plastic box with no easy access, once installed in the roof, there shouldn't be a problem. I may even try this for a couple of projects at sometime in the future.
Anyway, read and see what you think.]

[Melanie]

If you must... here's part of a circuit I designed (coincidentally for a fan controller for Hospitals)... total disclaimer if you die whilst playing with it - it is LIVE - you have been warned! No big dropper Resistors (R1 is a standard 0.25W), no heat disspation problems...

The only real justification for using a circuit like this is space... The total physical real-estate a circuit like this requires is a fraction of that taken by an iron cored transformer or a switcher.

[ERIK]

Thanks for those. Yes space is exactly why I don't want a tranny. I'll try your's Mel it looks good. And I'll make sure it is well housed. I was getting some heat disspation problems as you mentioned by using all the ones with the big dropper resistors. Nice looking fan controller by the way.

[ERIK]

Hey Mel, sorry about going back to this dreaded topic again but I was wondering a couple things about your transformerless power supply circuit.

1 - What sort of 470nF 400V cap did you use? I can't seem to find a suitable one down here in my funny little country.

2 - Is the voltage dependant resistor an integral part of the circuit or a somewhat optional component?

No more questions on this subject after this I promise.

[glkosec]

Melanie i was wondering if you could suggest a manufacturer of the VDR used in your design.

[Melanie]

The VDRs (aka Varistors) we use are made by Joyin Co of Taiwan. They work out between 5-15 cents (US) each depending on the flavour. If finding Joyin is a bit elusive in your neighbourhood, then push this link...

http://www.motionnet.com/cgi-bin/sea...?a=cat&no=2149

[T.Jackson]

Early last year I designed mark II of my PC-Controlled Mains Switch. As with the original design, I opted for a transformer-less solution. 240VAC directly feed to three paralleled X2 class rated caps, then to a full wave bridge powering two shunt regs to produce 5, 12V rails respectively. The entire circuit is optically isolated from the PC at all times.

During development of the Windows-based software which drives it, one of the opto's shorted when my multimeter probe slipped. This turned out to be disastrous! It took out the entire parallel port on my notebook weighing in at the time around $3,000 !!! &*@!()K

The paramedics managed to revive me after several hours. When I came to, had a beer and faced the inevitable reality of my "accident". The notebook was sold at a bargain price to someone who wasn't terribly concerned with the absence of the parallel port.

The project was finally completed on a desktop PC and published in the Silicon Chip magazine last July. Unlike the first version, it was a complete flop. No kits were produced. Double the features but twice the price to build. The system is easily expandable - (up to 8 units can be daisy chained together) - Windows-based software to boot, undoubtedly there will be no MK III, and I'm certainly in no rush again for a cost-effective transformer less solution.

*Always take extreme care when working with mains power!

http://www.siliconchip.com.au/cms/A_107104/article.html

Best regards,
Trent Jackson

[glkosec]

Melanie,
I have tried to emulate your schematic and have found that without the 5V Regulator the ciruit provides 11.7 V, but as soon as I add the regulator or put any kind of load across the circuit it drops to almost nothing. Does anyone have any idea of why this may be happeneing?

[mister_e]

and yet another SIM victim... trust me and Melanie, it's circuit works.

Just for my own curiosity, which SIM did you used?

[T.Jackson]

Interestingly, I designed a similar circuit myself early last year. Few variations with mine. Used two TO220 regs. (LM7805 & LM7812 1 amp packages) Plus I also have a bit more available current due to full wave rectification. This arrangement also makes the circuit a little safer too. The low voltage rails are tied close the Neutral line of the 240VAC mains, so provided you connect to a correctly wired GPO (240VAC socket), most of the circuit is nominally at low potential. (*Claimed by Silicon Chip publications)

Anyhow, I discovered a whole bag full of problems with this circuit on the bench. In theory, it all should have worked. In practice, I found that the 5 & 12V regs were unable to deliver the goods. The circuit was pre-calculated and should have been good for about 1.2W of power- 100mA on the 12V rail. The 12V rail failed to provide even a third of this. I have no explanation to the problem and my thoughts are entirely inconclusive. I discovered only about 20mVPP of ripple in the raw DC rail supplying the regs. This is more than an acceptable level of AC component.

In the end I settled for two zener-based shunts to produce 5 & 12V rails respectively. Circuit works well.

Trent Jackson

[Melanie]

Off-Load the VDR clips the upper voltage from climbing too high. But if your supply is crashing, then you are trying to draw too much current - that actually is an additional safety feature. If you measure the volts across the supply Capacitor, you'll find the whole lot appearing there.

If you just put a 5v Reg (like a 78L05) in circuit and your output crashes, then (a) your Regulator is defective, or (b) your main supply Capacitor is the wrong value for the current you're drawing or itself defective.

That circuit and the values given is good for 10-30mA total. You can measure how good the circuit is by putting some Resistors in series with a meter set to mA across the 11v VDR. Put another voltmeter across the VDR. Now increase the load (ie decrease the Resistance) until the voltage just starts to drop to say to 10v. That will the the maximum (total) useable current from your PSU based on the main Capacitor you're using.

[T.Jackson]

20-30mA?, lets see...

XC = 1/ 2Pie(FC)
= 1 / 6.283185307 x (50Hz x 0.00000047F)
= 1 / 6.283185301 x 0.0000235
= 6772Ω

Now, there's a 470K bleed resistor in parallel with the cap.

XC = 1 / (1/ 6772) + (1 / 470K)
= 1 / ( 0.000147666 + 0.000002127)
= 6675Ω

As you can see the 470K only knocks about 100Ω off the total reactance.
For those who don't know, reactance in an AC circuit is like resistance.

Rectified high voltage rail from RMS value (full wave rectification)

= 240VAC x 1.414
= 339.36VDC 'Peak value, has same heating effect as 240 RMS

(half wave rectification)...

= 0.637 x 339.36VDC
= 216.17VDC Average

Absolute max current (short circuit)

= 216.17VDC / 6675Ω
= 32.38mA

There's a 5V rail being produced from a LM78L05. It's input VCC must always be at least 2.5V greater than the output in order for it to maintain adequate regulation.

Absolute max current @ 5V

= (216.17VDC - 7.5VDC) / 6675Ω
= 208.67VDC / 6675Ω
= 31.26mA

Absolute max power @ 5V

= 31.26mA x 5
= 156mW

Spot on. In the perfect World, Melanie's supply should be good for about 30mA.
Component tolerances often compromise the accuracy of calculated events. Such is life.

Trent Jackson