AC Power

BADHABIT

Could you guys look this over and tell me if I'm missing anything. Some of the components may be incorrect, like the rectifier,·and the reg shld be 3.3v. What size caps are recommended·?

Does it need any transducers or tachometers?·

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
BH

Philldapill

You might want to add at least a 100uF electrolytic on the output stide of the regulator as well. The 100nF helps to stabilize any transients, while the 100uF helps stabilizing any low freq. transients.

isleder

Looks like the transformer and the diode bridge is setup for a dual power supply because the center tap is grounded. The D1 bridge pin 4 is also grounded and will create a short and perhaps burns out D1. So either leave out the center tap or do not connect D1 pin 4 to ground.

mctrivia

the way it is displayed D1 pin 4 should go to a 7905 to give you positive and negative voltage. Remove ground on center tap or add 7905 upside down from 7805 and mirror all caps. make sure the positive side of all caps is on the ground though for negative half.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
propmod_us and propmod_1x1 are now in stock. Only $30. PCB available for $5

Need to upload large images or movies for use in the forum. you can do so at uploader.propmodule.com for free.

Dr_Acula

Well spotted isleder and mctrivia. For a single power supply (no negative volts) just leave out any connection to the centre tap. It depends on the tranformer voltage, eg if that is a 5V reg and the transformer is a 20V transformer with a centre tap, it would be better to tap off the centre and one of the arms as that would be 10V AC. What voltage is the transformer?

HollyMinkowski

Won't you be putting about 18v into the voltage regulator with that 10:1 (12vac?) transformer?
Seems high since it will be a 3.3v regulator..lots to dissipate. Maybe just use the center tap and one of the others
as the input and have like 6v into the bridge and approx 9v to the regulator?

Toby Seckshund

The transformer shows 10:1 reduction. If that is across the total secondary, leaving the centre tap off would give ((120/10) x1.414)-1 ish peak volts across the 470uF, about 15-16 volts. The output from the 7805 is, of course, 5 volts so the drop would be 10 volts at worst. At 1A the heatsink would be substantial, but...

The smoothing cap is only 470uF and so would have a large ripple voltage at 1A this would therefore reduce the overall power disipated in the reg. This is the technique used in those cheap crappy wall warts, eventually the dips would reach a point where the dropout of the reg is reached. Heat and ripple current would wear out the cap and the available "current-before-ripple on output" would get lower and lower (The great advantage of wall warts is they fit into the hand and so a good throwing range can be got) so use a good quality one with a high ripple current rating, and 105 degrees C.

If the centre tap is left as is, and the connections to (4) of the rect is dissed and transfered to that CT, so that only the diodes from 1,2,3 are used then only·about 8 Volts will be available to the reg. This is too close to the dropout of the reg, about 7V. the only thing that could be done to help this would be a much larger smoothing cap, but that would rely on low impedance from the transformer and rectifier so that the replacement energy comes a high current pulses. It would have to be at least 10x larger in value, I think it was about 2200uF per Amp per volt ripple, at 50Hz.

As mentioned already a larger cap on the output side is needed. Keep the little one as well for the high freq (ESR)

I have just notice that you want 3,3 Volts so the headroom shouldn't be so much of a problem.

Post Edited (Toby Seckshund) : 6/11/2009 10:02:18 AM GMT

Leon

I use a rule of thumb - 1000 uF per amp.

Leon

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Amateur radio callsign: G1HSM
Suzuki SV1000S motorcycle

CounterRotatingProps

> 1000 uF / amp

interesting, Leon, what's your rule of thumb for the cap voltage rating versus desired operating voltage?

These little short cuts are treasures!

thanks
- Howard

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Got Electrons?

Philldapill

I'm not Leon, but generally, you should at least double the voltage rating for the cap. In the case of a regulator, it might be a good idea to have the regulated caps be at LEAST the input voltage(i.e. voltage from the transformer). This way, if the regulator fails and sends full trans. voltage to the cap, you would have a shotgun go off in your circuit.

CounterRotatingProps

Phil, that's cool - thanks for the reply.

> you would have a shotgun go off in your circuit.

Ah, you do mean 'you *won't* have a shotgun go off ' - right?

- Howard
<<-- note the smoke [noparse]:)[/noparse])

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Got Electrons?

mctrivia

Zener diode rated at 120% of output voltage across the output is a good safety feature

Normally does nothing but in case of failure stops voltage from going to high.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
propmod_us and propmod_1x1 are now in stock. Only $30. PCB available for $5

Need to upload large images or movies for use in the forum. you can do so at uploader.propmodule.com for free.

Toby Seckshund

Power supplies are the most over looked part of any project. Too many times heartache is caused by the supply not playing nice, which is usually due to insuffient planning about what work is being done. The static current showing on your meter will just be an averaged reading, peaks may well exceed that by decades.

Everything nowerdays is expected to be so small that it will fit into a matchbox, in the olden days the power supply was assesed and the doubled/tripled for long term reliability, even if it did take up half the room. And it would never be allowed to become any more than pleasently warm.

But then ... A laptop built to these maxims would be a tad heavy

mctrivia

We definetly build things closer to there limits today. The power supply on my propmod-us_ps is rated for 800mA and fused. In the past it would be double or more but when the supply is 90% efficient and you only have 400x600 mil of space fusing just under the limits is acceptable.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
propmod_us and propmod_1x1 are now in stock. Only $30. PCB available for $5

Need to upload large images or movies for use in the forum. you can do so at uploader.propmodule.com for free.

Toby Seckshund

Mind you, there if a satifactionto be had on hearing a muffled thump from the local substation, all because your PSU's crowbar has fired

Many years ago, I worked on 300KW HF TX, the 11KV 3 phase mains was full wave rectified, smoothed a bit (doesn't need much) and stuffed onto the anodes. Faults were brutal!! The fuses were nicked to make rolling pins. The power rooms, for these sets, were rooms.

No surface mount nonsence there, you needed spanners.

BADHABIT

· Thanks for the replies.

After reading through everything I have made a few changes. Some of the component #'s might be wrong bc I have no idea what some of the them mean. When building it I would get the correctly rated items (hopefully).

I am thinking about using p/n 030-7329-0·from Pulse Eng. as the transformer.·Is that what I shld use, or would something else be better?

I will also be adding a 3v reg after the 7805. What should·the cap config before and after that look like?

This will be powering a mcu, 3-5" LCD, keypad, and some small relays.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
BH

Philldapill

I don't see any problems. However, you might want to make sure that secondary winding on the transformer IS in fact 9V. The one you had before was a center tap transformer, and if it was 9V from center tap to each leg, then it's 18V from leg to leg. You might want to make sure. Your design should still work, but you might be dropping excessive voltage if it's 18V, which would result in a hot regulator.

Toby Seckshund

If you can account for the heat, and the reg can stand it, then excess headroom volts does give you a sort of power backup in that, if a small hole in the mains comes along there should be some energy stored in the smoothing cap to breeze past it.

Watch out for the backlight of the LCD it will probable be the biggest current consumer, next will be the relays, low volts=high current , budget for worst case ie everything on together

BADHABIT

The reason for the center tap was because it was the only one in there that said what it was, but no matter what I end up w/I will check the output

"next will be the relays, low volts=high current"

If I use a charge pump and higher voltage relays would that reduce the current draw?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
BH

Toby Seckshund

Please don't let me frighten you I was just pointing out that if the relays are to be 5V powered ( I have never seen a 3.3V relay) then they can take 30mA for "high sensitivity" ones to 100mA for normal.
This could be a problem if 32 of them ended up closed at the same time.

There is nothing to say that the relays do not get powered pre regs, whatever the tfmr you choose will dictate the voltage of the relays. Relays thake a big pull to get them closed but less to keep them closed, so a 12 relay will take 3/4+ of that to close smartly. They certainly do not req reg volts. obviously use a transistor to drive them anyway and place a reversed diode across the winding to catch the back EMF ( huge reverse voltage spike when the magnetic field collapses ) a common 1A diode should be ok. ULN2803s are built for this

Charge pumps cannot create energy the extra volts come a the cost of higher current so you just loose by the inefficiency.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Style and grace : Nil point

Post Edited (Toby Seckshund) : 6/14/2009 7:46:23 PM GMT

mctrivia

Get 9v relays and run them off unregulated power. Then you almost don't need to worry about the current. I say almost because rectifier and traces need to take current

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
propmod_us and propmod_1x1 are now in stock. Only $30. PCB available for $5

Need to upload large images or movies for use in the forum. you can do so at uploader.propmodule.com for free.

BADHABIT

mctrivia said...
Get 9v relays and run them off unregulated power. Then you almost don't need to worry about the current. I say almost because rectifier and traces need to take current

That's a great idea.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
BH

Sapieha

Hi BADHABIT.

If You will not have problem use 2 rectifiers parallel.
One to regulated and one to unregulated power.
Eles if You use power to drive relays from traces before regulator all EMF from them come directly in reg.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Nothing is impossible, there are only different degrees of difficulty.
For every stupid question there is at least one intelligent answer.
Don't guess - ask instead.
If you don't ask you won't know.
If your gonna construct something, make it·as simple as·possible yet as versatile as posible.


Sapieha

BADHABIT

Fer Shizzle.

I have been thinking from the beginning that it looked very sparse.

Any ideas for Bridge Sizes - or are those 9mm singles in 4's more economical?

I just remembered a board I found. I always liked it for it's Rigid Simplicity. But this is pretty much the power supply I would want for the relay stage?


AC comes in the 2 tabs @ the front

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
BH

Post Edited (BADHABIT) : 6/15/2009 3:21:31 AM GMT