Hi guys,

I'm making a variable power supply where I want the current to be limited as well as the output voltage.

I'm wondering whether the crowbar (U3) is a good way of limiting current? Also how do you know how much the current will be limited? I remember seeing somewhere that the limit for a lm317 in crowbar configuration will be 1.2/R = Imax. Can anyone explain why this is?

Also, I'm wondering if the Zener and resistor (D13, R33) is a good way of limiting output voltage. So when the voltage is set high than 22V, the resistor will begin conduction 22/100 = 0.22A. Since the crowbar limits the current to 1.2 / 2.5 = 0.48A, the most current on the output will be 0.48 - 0.22 = 0.26A at 22V.

Does this look correct? Also is there any other problem with the circuit?

Thanks.

 

I think you need to place a resistor before the zener (D12) to dissipate power coming off the bridge that would potentially overload the zener. This resistor will also help in limiting your power supply's current: (Vrect-Vzener)/R.

The way you have U3 connected, I believe, uses the LM317 as a current source, continually outputting whatever current you set it to with R28. The resistor added after the rectifier should eliminate the need for this block of your schematic. Also, using the LM317 as a voltage regulator like you are at U4, should provide a built in current limit inherent to the device.

I'm not sure that you need the zener at the end (D13) since the LM317 is outputting the voltage you control with the variable resistor (V1).

I'm just a student though and far from experienced in this area.

 

Thank you for your thoughts.

To clarify:

D12 is there to eliminate voltage spikes. Since the input transformer is of good quality and has a fuse, it should not be too much of a problem, I don't think a resistor is needed there.

Although the LM317 has current limiting, I want to make sure it is not possible to have more than 0.5A on the output, because it may damage the circuitry it is connected to.

The zener on the end, D13 is there so that the output can never be higher than 22V. The variable regulator with 5K pot will be able to output up to ~26V, I want to make sure it is not possible to have it go beyond 22V because again, more than that could damage the circuitry its powering.

 

The resistor in series with D13 will allow the voltage to go higher than 22V. If you tap the output across the zener, then you'll get 22V out. Any voltage off the regulator >22V will be burned off on the resistor (R33). If you set the LM317 to output 22V, if should be pretty stable around that voltage due to the cap (C18) and the zener is unnecessary.

What's the RMS voltage coming off the transformer? If it's below the LM317's max input (34V?), the zener before the regulator is also unnecessary.

You can also find DC-DC converters other than the LM317 that will allow you to control the maximum output current.

 

Hi, thanks for the reply.


Do you mean this (see attached), where the output is taken across the zener?


You're right, D12 is not really needed, since the transformer won't output more that 34V.

Also, I have to use the 317 cause thats what I have already.

Thanks.

 

Yeah, that's what I meant. I still think the zener is unnecessary there though since the LM317 will provide steady DC output of 22V.

 

Thanks for the repply.

The purpose of the zener is to make sure the voltage never exceeds 22V. I need a variable regulator that can go from 10 - 22V, but not exceed 22V. In the current configuration, when there is more than 22V it will work fine since the zener will start conducting current and R33 will bleed off the excess current so there is no short. The problem is that when there is less than 22V, there will be a huge voltage drop accross R33 since it is 100ohms (v = 0.5A * 100 = 50V).

Is there no way to limit the maximum output voltage of a variable regulator?

 

Thanks for the repply.

The purpose of the zener is to make sure the voltage never exceeds 22V. I need a variable regulator that can go from 10 - 22V, but not exceed 22V. In the current configuration, when there is more than 22V it will work fine since the zener will start conducting current and R33 will bleed off the excess current so there is no short. The problem is that when there is less than 22V, there will be a huge voltage drop accross R33 since it is 100ohms (v = 0.5A * 100 = 50V).

Is there no way to limit the maximum output voltage of a variable regulator?

In theory, you could add a 20.75V zener in parallel with your pot. Good luck in finding one.
Below is an adjustable limiter that should work.

 

If using the regulator as shown in the datasheet --> Vo=Vref(1+R2/R1). You should be able to select a value for R1 that controls the range of voltages you can output. For Vo=22V you want R2/R1 = 16.6 (using Vref=1.25V). If R1 is 300Ohm, and R2 is your 5kOhm Variable resistor you get about 22V. 1.25*(1+5000/300)=22V.

That drop across R33 won't be possible. The max output of the LM317 is 34V. The zener tops out at 22V. R33 will at most see 12V. 12V/100Ohms = 120mA. I think that R33 and the zener should come out of the circuit. R33 will limit the current to your load. Especially if the regulator isn't outputting way more than 22V [for example: (25V-22V)/100Ohms = 30mA to the load].

 

Thanks, Ron.

 

To limit the output voltage, all that is required is an internal trim pot in series with the main voltage control. If you are working for NASA and am really worried about the output voltage going too high, a good way is to use a NPN power transistor connected across the output (C -> +, E -> -), then connect the base to the anode of a zener, with its cathode going to the + line. Vz = Vmax -.8V. If the output goes over voltage the zener turns on and injects base current into the power transistor which drags 50-100 that current from the PSU, and in the event of a catastrophic failure will blow the supply fuse. The alternative that manufacturers often employ is a whole circuit board that "fires" an SCR to short out the output terminals. This sort of circuit is often prone to "firing" on obscure noise pulses. Ya pays ya money. . .
Frank

 

Just a couple of other comments. For 22 Volts of output your will need at least a 24VAC input. Therefor the capacitor is not rated high enough, should be 63 Volt. You might want to increase the voltage on the transorb across the capacitor to somthing over 35Volts.

 

@Ron: Thanks for the design, thats exactly what I was looking for. I like the idea of using a transistor that turns on when there is too much voltage, I will keep that idea in mind in the future.

@mlundgren: Thank you very much for your help. Setting R1 to 300ohms is a very good idea.

@chuckey: That your for your suggestion, it is very similar to Ron's. And lets be glad I don't work for NASA...yet.

 

@Bailey45:Good catch, that cap is actually rated at 50V, that's just a typo. Out of curiosity, is there a convention for how far over your caps should be rated. If the voltage from the transformer is 24VAC is a 30V cap good enough?