Can I get comments on this 120v AC to 6.3v(edited) DC power supply?

MarkS

I need this for a project I'm working on. A wall wart cannot be used due to the design. Therefore, I need to convert mains power to +12v and -12v DC. The main controller board will have a 5v and 3.3v regulator and required circuitry and this power supply will supply those regulators. This is my first time dealing with AC, so I'm a little nervous and want to make sure that everything is correct.

The parts list should be pretty obvious, except for the transformer. The transformer is a Tamura 3FS-524 120v to 12v.

Any comments and criticism will be appreciated, but space is limited, so please don't bump up the parts list with anything that isn't 100% necessary.

Post Edited (MarkS) : 1/8/2008 12:21:08 AM GMT

Phil Pilgrim (PhiPi)

You're going to get about 17V (no load) from the full-wave rectifier. Is that what you want? Also, with a 1A load, you'll see about 3.6V of ripple from the 2200uF cap.

-Phil

Post Edited (Phil Pilgrim (PhiPi)) : 1/7/2008 9:46:34 PM GMT

RDL2004

If you're going to regulate this down to 5v and 3.3v then why are you using a 12v transformer? Like Phil says, you're going to get close to 17v out and your regulators are going to run hot from burning off all the excess voltage. I would look for a transformer in the 5 to 6 v range.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
- Rick

MarkS

Thanks for the info! I wanted 12 volts originally because I was expecting to need it for a relay coil. I then found a relay with a 5 volt coil that fit my needs, but never changed the transformer.

Back to work...Just out of curiosity, why would I get 17 volts from the rectifier when the transformer is outputting 12 volts and the diodes in the rectifier drops voltage?

Post Edited (MarkS) : 1/7/2008 11:02:13 PM GMT

Franklin

If you have the space most computer supplies will give you +12, -12, +5 and 3.3v and they aren't very expensive.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
- Stephen

MarkS

Franklin said...
If you have the space most computer supplies will give you +12, -12, +5 and 3.3v and they aren't very expensive.

Not really an option due to size.

While looking at Digi-Key for a new transformer, I found one from the same supplier, with the same data sheet and foot print. It outputs 6.3v from 120v, but I'd still like to know why I was told differently above about the last transformer. How did you guys come up with 17 volts?

LilDi

The way you have the transformer connected, you are going to get only +12 volts and neutral. Don't know where the system ground is going. Looks like a floating ground.
The fuse appears to be connected to nothing or is across the output of the transformer leads... Dead short!

take a look at the attachment of how it looks like this transformer need to be connected to get +-12volts.

Post Edited (LilDi) : 1/7/2008 11:37:44 PM GMT

MarkS

So, how about this?

LilDi

Yes,

Or this attachment.

Paul Baker

The reason for the 17 vs 12 Volt output is due to the way AC voltages are measured, AC is measured by the root mean square (RMS). For mains supply (ie sinusoidal) Vrms = Vpp/sqrt(2) or Vrms = Vpp/1.414 where Vpp is the peak to peak voltage. When you move from AC voltages to a DC voltage, you move from Vrms to Vdc (voltage measured wrt ground which is analogous to the peak to peak measurement). So the 12 Vrms AC signal when rectified and filtered becomes 12x1.414 = 16.968 or 17 Vpp.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Paul Baker
Propeller Applications Engineer

Parallax, Inc.

Post Edited (Paul Baker (Parallax)) : 1/8/2008 12:33:58 AM GMT

Phil Pilgrim (PhiPi)

Mark,

LilDi is right. I wasn't paying attention to your ground or to your fuse. I just saw the two windings connected in parallel. If what you really want are both plus and minus outputs, you will need separate caps. LilDi shows a half-wave configuration. For better efficiency, the configuration shown below (attached) would give you full-wave rectification for each output. In any event, the earlier comments about rectified voltage remain the same. (The reason for the higher voltage is that the transformer specs are for RMS voltage. But a capacitor-input filter under no load will provide the peak voltage as DC.)

-Phil

Post Edited (Phil Pilgrim (PhiPi)) : 1/8/2008 12:35:00 AM GMT

MarkS

Hmm...

So the corrected version of my PS would actually output roughly 9vDC. Interesting.

Phil Pilgrim (PhiPi) said...
Mark,

LilDi is right. I wasn't paying attention to your ground or to your fuse. I just saw the two windings connected in parallel. If what you really want are both plus and minus outputs, you will need [noparse][[/noparse]two] rectifiers and two caps. LilDi shows a half-wave configuration. For better efficiency, the configuration shown below (attached) would give you full-wave rectification for each output. In any event, the earlier comments about rectified voltage remain the same. (The reason for the higher voltage is that the transformer specs are for RMS voltage. But a capacitor-input filter under no load will provide the peak voltage as DC.)

-Phil

Thanks again. More work to do.

MarkS

Last revision?

LilDi

You can use one full wave bridge rectifier and one 2200uf capacitor if you look at my last posted attachment

Phil Pilgrim (PhiPi)

I think I'd stick with two caps. The reason is that with only one, neither filter is tied to ground, and load transients on one output could cause voltage fluctuations on the other.

-Phil

MarkS

Phil Pilgrim (PhiPi) said...
I think I'd stick with two caps. The reason is that with only one, neither filter is tied to ground, and load transients on one output could cause voltage fluctuations on the other.

-Phil

So, to recap...

The last design I posted would work? Or will I blow something up? Also, using the formula Paul posted, I'll be getting 8.9v and -8.9v DC with a transformer rated for 6.3v, correct?

Phil Pilgrim (PhiPi)

It should work. You'll get about 8.3V (8.9V - 0.6V rectifier forward voltage) at no load.

What will be the maximum load each output has to drive? This is important in picking the final capacitor value, so the minimum output voltage due to ripple doesn't drop below the input voltage required by your regulators.

BTW, what is the negative output being used for?

'Just noticed: The transformer end of your fuse is way too close to the chassis ground!

-Phil

MarkS

I haven't calculated the load yet. I guess, to be safe, figure full load of 1 amp. The system is not much more than a Prop with a few relays and transistors + a couple of ADCs, OpAmps, temperature probes (LM34), the 1302 time keeping chip and a LCD (as yet to be determined).

The negative voltage is for the OpAmps, but after looking at some diagrams with a LM34 and ADC, I probably wont be needing them.

LilDi

I'm thinking the fuse should be about 1/2 amp to·3/4 amp. 250 volt
Phil can probably give you a more accurate fuse type.

Post Edited (LilDi) : 1/8/2008 2:49:16 AM GMT

MarkS

I'm thinking that after all this trouble and what will be more than a little expense, it would be better for me just to buy and use a open frame Mean Well power supply.

Phil Pilgrim (PhiPi)

I couldn't agree more! I've used their supplies, they work great, and they sure beat the hassle — and danger — of rolling your own! Good choice! I believe Jameco still carries them.

Phew!

-Phil

MarkS

Jameco still carries them and it will be far cheaper than buying the parts and having the PCB fabbed. I still learned some stuff today, so this thread is not a total loss.

MarkS

I hate to bring up an old thread, but I went to Radio Shack and bought the components to make this. At no load, save for the voltmeter, I get 5.0V coming out of a 7805 regulator. This is, of course, to be expected. Now, what is the best way to add a dummy load? The regulator has a maximum rating of 1A, but I seriously doubt I'll ever need to draw more than 800mA out of this.

Phil Pilgrim (PhiPi)

A couple of these in parallel will do nicely for a dummy load.

-Phil

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
'Still some PropSTICK Kit bare PCBs left!

LoopyByteloose

Make sure the capacitors are rated above the Peak Voltage and not the RMS or they may quickly fail. This is a common beginner's error. When in doubt, it doesn't hurt to provide a rating that is double the expected output voltage. In fact, I usually buy 50 volt electrolitics to use in everything because I don't have to worry about cutting voltage ratings too close.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
PLEASE CONSIDER the following:

Do you want a quickly operational black box solution or the knowledge included therein?······

···················· Tropically,····· G. Herzog [noparse][[/noparse]·黃鶴 ]·in Taiwan

MarkS

They're 35 v caps axial lead caps. Right now everything is on a bread board for testing. I'll use radial lead caps when I get the PCB made, so I'll get 50v caps then. I've heard that I should always de-rate a cap by 50% to be safe. I have roughly 9.5v coming out of each DC leg of the rectifier, which is almost half of the 50% de-rating of the 35v caps.