You can put any reasonable amount of capacitance you like on there to filter it!

It will work.... sort of. Why don't you try it? Put say 10uF 40V across the +/- of the bridge and scope it, you will find a very nice clean DC voltage.

That's all well and good but if you add a load, say a 4K7 resistor, scope it and see what happens.

What you are seeing is the effect of the load, it's trying to drag the voltage down.
Then we must add sufficient capacitance to maintain the voltage at a usable level.
Which "usable level", well from the data-sheet we can see that our voltage regulator needs 2.5V more on the input than on the output. (This is called the drop out voltage)
And since your output voltage needs to be 16V, then we must keep the input voltage at least 18.5V.

Now we have what we need to calculate the capacitor:-

C=0.7*I / (ΔV*f)

C=0.7*10 / (3*100)

C=0.023333 or 23,000uF

Wow that's a lot of capacitance!

 

You can put any reasonable amount of capacitance you like on there to filter it!

It will work.... sort of. Why don't you try it? Put say 10uF 40V across the +/- of the bridge and scope it, you will find a very nice clean DC voltage.

That's all well and good but if you add a load, say a 4K7 resistor, scope it and see what happens.

What you are seeing is the effect of the load, it's trying to drag the voltage down.
Then we must add sufficient capacitance to maintain the voltage at a usable level.
Which "usable level", well from the data-sheet we can see that our voltage regulator needs 2.5V more on the input than on the output. (This is called the drop out voltage)
And since your output voltage needs to be 16V, then we must keep the input voltage at least 18.5V.

Now we have what we need to calculate the capacitor:-

C=0.7*I / (ΔV*f)

C=0.7*10 / (3*100)

C=0.023333 or 23,000uF

Wow that's a lot of capacitance!

it funny you say that.. I have 2 28kuf 50v caps in my parts box... but they are about 2.5 inches round and 3 high.. they came from an old tv.. I might put one on my circuit and scope it..

 

ive got some new 10kuf 50v caps I think im gonna try.. do I connect between the +/-v legs or the +dc to the 0v wire and the -vdc and the 0v wires?

 

The capacitors are fixed between the (+V and 0V) and (-V and 0V)...... don't forget that the (-) terminal on the cap goes to the lower voltage. So between +V and 0V the (-ve) terminal is at 0V, and between 0V and -V the (-ve) terminal is at -V.

 

thanks so far for the help so far bance.. Ive put the 10kuf caps in my circuit and it looks good on the scope, just like the pictures in the link, looks really clean..with the caps in the voltage has increased to 30.2vdc.. And it rises even higher under load??? im not sure what the hell is going on there.. I thought under load the voltage would drop. Ive built the rectifier from 1n4744 diodes for the 8.1vac tap but haven't connected it yet to scope it. since im only using that tap for my meters, will those little diodes be ok for my rectifier? I mean I guess I could just hook it up and see if they smoke out,,lol,, I have hundreds of them..

 

any ideas what I could do with the 38.5vac tap I have left from my transformer? I was thinking of maybe keeping it ac and figuring out how to make it adjustable because I do some work ac motors for my rc quadrotor..

 

Ok, well the voltage after smoothing will rise, because as measured before the filter caps, the voltage is RMS. It should be multiplied by 1.42 to give the filtered DC voltage. As to why it rises under load that is a mystery!

Regulator next....

 

yea like 10v at 19a..lol.. I think im blaming my scope,, I was measuring amperage, voltage, and using the scope at the same time.. Im going to use different meters to measure voltage and amps and just the scope watch the wave. SO I shall start the regulator circuit next. Would it be better to build it with the lm317 only to get all the adjustable problems out of the way then add the pass transistor to increase the amperage? Also I read that I cant use 2n3055's for both the + and - voltage. the 3055 is an npn, I read I would need a pnp for -v side. The datasheet says the mj2955 is the pnp equivalent to the 3055.. is this correct?

 

Ok, well the voltage after smoothing will rise, because as measured before the filter caps, the voltage is RMS. It should be multiplied by 1.42 to give the filtered DC voltage. As to why it rises under load that is a mystery!

Regulator next....

rms versus filtered voltage,, got it.. thanks

 

Don't make the filter capacitors larger than they need to be.
A filter capacitor is a short circuit to AC across the power supply. When you increase the capacitance the peak current through the rectifier diodes will also increase.
Do the calculation for the peak diode current and you may be surprised how large it can be.

 

Don't make the filter capacitors larger than they need to be.
A filter capacitor is a short circuit to AC across the power supply. When you increase the capacitance the peak current through the rectifier diodes will also increase.
Do the calculation for the peak diode current and you may be surprised how large it can be.

how would a dummy measure that sir? the calculation for max voltage was 24000uf, since im building a variable psu I split that and used 10000uf 50v.. does this seem excessive?

 

Ive lost a good bit of drawings and schematics somehow. in the first drawing I made wont work the way its drawn. however the circuit has an lm308 (I have an nte996).. What exactly is it doing? and do I still need it? I plan on using 3 lm317t's with 2n3055 pass transistors instead of 4.

 

this is the schematic im trying to work off of but im adding pass transistors, What is the lm308(I have an nte996 that's the exact same) doing here? and will adding the pass transisitors effect its purpose?

 

See how to calculate ripple and peak diode current:

http://www.zen22142.zen.co.uk/Design/dcpsu.htm

 

Hi Adam,

Both of your circuits in post #14 can be modified to use parallel pass transistors.
The first one has the best regulation but needs more voltage from input to output to regulate so your highest output voltage will be lower.
The second one doesn't have that problem but is kind of sensitive to temperature. I think I would parallel pass transistors rather than 317's.
To figure out the size of the caps you need to use the maximum current you want at the maximum voltage you want. Then you need to make sure the ripple does not go below the voltage needed from input to output for the regulator to work.
Bance can probably explain this better.

 

ok so the 10k filters are probably to small.. I like the second picture in post 14. how would I make that adjustable? I think I have an idea. I have a 5k 10 turn and a 2k 10 turn pot. I wanted a main and a fine tune, I think I have enough heatsink to keep the 3055;s happy. Bance has been invaluable in this process. Ive been staring at the 317 datasheet and I will be posting a vouple questions on that shortly. but thanks for the help guys. I actually have a half assed working circuit with a fixed voltage of around 30vdc.

 

You can make it adjustable like attached..
You need some really big heat sinks. Each transistor is about 80 watts at 10 amps at 5 volts.. Think holding on to a light bulb while it's on.

 

You can make it adjustable like attached..
You need some really big heat sinks. Each transistor is about 80 watts at 10 amps at 5 volts.. Think holding on to a light bulb while it's on.

That looks good,, I have a good heatsink,, for what I use my psu for the bulk of my current demands will occur between 12-14vdc. I doubt Ill ever need more than 2a at 5vdc. Ive noticed that when I turn the psu off my meter still reads 30vdc from the filter caps. When the rest of the circuit is built, the voltage will turn off with the psu right?