Building a Power Supply

[vne147]

Hey everyone. I’m helping a friend of mine build a bench power supply and I’d like some verification that what I’m thinking is correct. We are going to use the NTE956 (LM317 equivalent).

First I want to determine the input voltage to the NTE956. The transformer we are using says 110 VAC 60/50 Hz on the primary side and 26.2 VAC on the secondary side. I’m assuming (maybe incorrectly) that 26.2 VAC is RMS voltage. If that is a correct assumption then when rectified, 26.2 VAC will become roughly 35 VDC.

(26.2V)*(√2) – 2V ≈ 35V

The 2V I subracted accounts for the forward voltage drop across the diodes in the bridge rectifier.

So with an input voltage of 35 VDC, we should be able to get up to 32 VDC (+/- a little) at the output of the NTE956.

Is this all correct so far?

Next question is that my friend wants to use an analog panel meter like **broken link removed** one. It has a range of 0 to 30 VDC. Could we damage the meter by applying 32V across it?

Thanks in advance for your input.

 

[ronv]

supply

You almost have it. The only thing you missed is the ripple voltage at the input to the regulator. You will need a large cap. (Size depends on the load on the supply) to smooth out the ripple from the rectifier. Chances are you will loose a few volts here and end up with a 1.2 to 25 volt supply so the meter will be perfect.

 

[vne147]

You almost have it. The only thing you missed is the ripple voltage at the input to the regulator. You will need a large cap. (Size depends on the load on the supply) to smooth out the ripple from the rectifier. Chances are you will loose a few volts here and end up with a 1.2 to 25 volt supply so the meter will be perfect.

We will be using a 3300µF smooting capacitor. I know the capacitor will never charge all the way up to the peak voltage ouptut from the bridge rectifier but it should be close, shouldn't it? You said I would end up with a 1.2 to 25V output from the NTE 956. Where am I losing 7V from?

 

[ronv]

Power supply

Well, I'm not sure cause I don't know what the load is. (How many amps you want at 30 volts) But lets say you want the 1.5 amps that the chip puts out. Your numbers are correct, but the ripple will be as low as 30.75 volts at 1.5 amps. If you allow 3 volts for the regulator to regulate (the only spec. I see for this one) then you are down to 27.75 volts. If you want to take in low line voltage you might be as low as 25 volts that is actually regulated.
Several things you can do to get a little more:
Schottky diodes in the bridge. .8 v
2X3300 Ufd. 1.25 v.
Lower drop out regulator 1 v. (Careful most LDOs are only good for 30 volts)
Ignore low line 2.75

 

[bountyhunter]

With an unregulated DC input of 30V, internal IC circuitry will prevent you from getting an Amp of output current from an LM317 type device due to internal SOA (safe operating area) limiting. There are also major power dissipation problem operating at lower output voltages. The way such power supplies are designed is we use multiple power transistors in parallel to share current (and spread power around).

FYI, LM317 type devices are useful for generating regulated voltages but pretty useless for designing a bench supply where you need a wide output voltage range at usable current values. I know the semi makers put such circuits out but they are pretty bogus. Look CLOSELY at the published data curve showing available current vs input-output voltage differential and you will see why.

 

[vne147]

Well, I'm not sure cause I don't know what the load is. (How many amps you want at 30 volts) But lets say you want the 1.5 amps that the chip puts out. Your numbers are correct, but the ripple will be as low as 30.75 volts at 1.5 amps. If you allow 3 volts for the regulator to regulate (the only spec. I see for this one) then you are down to 27.75 volts. If you want to take in low line voltage you might be as low as 25 volts that is actually regulated.
Several things you can do to get a little more:
Schottky diodes in the bridge. .8 v
2X3300 Ufd. 1.25 v.
Lower drop out regulator 1 v. (Careful most LDOs are only good for 30 volts)
Ignore low line 2.75

OK, this makes sense. It seems I incorrectly assumed the ripple would be negligible but I see how at higher currents it makes a difference. Just for fun I went and ran the numbers myself and got answers pretty close to yours. Maybe we'll add another smoothing cap at the input. I also like the Schottky diode suggestion. I don't know how much current he would ever need at 30V but I don't imagine it would be huge. The supply will be used mostly for hobby tinkering and probably rarely used above 12V. Thanks for the explaination and suggestions.

With an unregulated DC input of 30V, internal IC circuitry will prevent you from getting an Amp of output current from an LM317 type device due to internal SOA (safe operating area) limiting.

Where in the data sheet does it talk about this? Looking at the LM317 data sheet **broken link removed**, I see a graph for output current limit as a function input to output voltage differential but not one that describes output current as a function of input volatge. NTE datasheets are generally useless so I'll just have to assume that this LM317 characteristic is the same as the NTE956

There are also major power dissipation problem operating at lower output voltages. The way such power supplies are designed is we use multiple power transistors in parallel to share current (and spread power around).

I agree there will be a significant power dissipation issue at low output voltage/high currents. We will be using a heatsink and a fan and I'm also pretty sure that power dissipation is limited by the IC.

FYI, LM317 type devices are useful for generating regulated voltages but pretty useless for designing a bench supply where you need a wide output voltage range at usable current values. I know the semi makers put such circuits out but they are pretty bogus. Look CLOSELY at the published data curve showing available current vs input-output voltage differential and you will see why.

I also agree there are better, more useful ways to make a bench power supply. One of my first projects when I started tinkering was to make an LM317 based power supply and since then I have had projects where my supply was inssuficient forcing me to canabalize an old ATX power supply I had. We are making this power supply for my friend to just introduce him to electronics and give him something to get started. If he sticks with it, I'm sure eventually he like myself will need to upgrade.

Thanks to both of you for the input.

 

[bountyhunter]

Another thing I forgot to mention: the unregulated DC voltage equation of it being:

VAC x 1.414 - diode drops

Where the VAC is the transformer secondary voltage is only a crude approximation. At the transformer's rated current it will be significantly less due to the distortion of the voltage waveform from loading. It's hard to explain in a single post, but the current flows only during narrow time intervals which loads the winding then and causes the peak voltage to drop significantly.

 

[bountyhunter]

I see a graph for output current limit as a function input to output voltage differential but not one that describes output current as a function of input volatge.

It is input-output voltage differential that determines maximum available current. Here are the actual guaranteed specs. You only have two guaranteed points:

1.5A at Vin-Vout < 15V.

0.15A at VIN-VOUT = 40V