Max Voltage

[Overclocked]

I cant seem to find the max voltage for LM338K, but it does say the max differential between input and output is 40V. My input voltage (unloaded) from the caps is 58.8V DC, I just want to regulate the output. Im pulling 2 Amps of current and it brings the voltage down to about 10V! When loaded, the ripple voltage is .9VRMS.

Obviously, its really inefficient. I have 8 caps total. Each is 1000uF, 50V. I have 2 Caps in parallel with each other to give 500uF 100V. I have 4 Sets of those to give 2000uF 100V. Should I use more Caps before I connect the regulator? The transformer puts out 45VRMS.

 

[freeskier89]

I cant seem to find the max voltage for LM338K, but it does say the max differential between input and output is 40V. My input voltage (unloaded) from the caps is 58.8V DC, I just want to regulate the output. Im pulling 2 Amps of current and it brings the voltage down to about 10V! When loaded, the ripple voltage is .9VRMS.

In the National Semiconductor datasheet, it clearly says:

"Since
the regulator is “floating” and sees only the input-to-output
differential voltage, supplies of several hundred volts can be
regulated as long as the maximum input to output differential
is not exceeded,"

This means there is no absolute max voltage (Do not try to go regulating voltages like 10KV though lol). If you have an input voltage of 58.8V, then do not try to regulate it to anything below 18.8V.

Obviously, its really inefficient. I have 8 caps total. Each is 1000uF, 50V. I have 2 Caps in parallel with each other to give 500uF 100V. I have 4 Sets of those to give 2000uF 100V. Should I use more Caps before I connect the regulator? The transformer puts out 45VRMS.

You need to redo those figures. Caps in parallel add capacitance. ie. 2- 50V 1000uF caps in parallel will make 2000uF at 50V not 500uf at 100V.

Hope this helps

 

[Overclocked]

You need to redo those figures. Caps in parallel add capacitance. ie. 2- 50V 1000uF caps in parallel will make 2000uF at 50V not 500uf at 100V.

Hope this helps

Well, when caps are in series their voltage adds up while the capacitance in halved (or actually its reciprocal) I just divide by 2 since they are both the same .This is how they are connected.

 

[Hero999]

There's nothing wrong with the way you've connected the capacitors and what you've said is correct the capacitance does halve.

freeskier89 is also right, there is a limit on the voltage difference between the input and output but it's maximum rating is 40V. With 58.8V at the input and 10V at the output the difference is 48.8V that's 8.8V higher than it's designed for.

However you've said the input is 58.8V when unloaded and the ouput drops to 10V when loaded.

What's the input voltage and ripple from the smoothing capacitor when loaded?

If the peak voltage less than 50V you're alright.

What's the output voltage when unloaded?

If it's greater than 18.8V then you're alright.

 

[freeskier89]

When you said "I have two caps in parallel..." I assumed the caps were in parallel not series

 

[Overclocked]

When you said "I have two caps in parallel..." I assumed the caps were in parallel not series

Opps thats my fault, I meant "2 caps in Serial with each other"

There's nothing wrong with the way you've connected the capacitors and what you've said is correct the capacitance does halve.

freeskier89 is also right, there is a limit on the voltage difference between the input and output but it's maximum rating is 40V. With 58.8V at the input and 10V at the output the difference is 48.8V that's 8.8V higher than it's designed for.

However you've said the input is 58.8V when unloaded and the ouput drops to 10V when loaded.

What's the input voltage and ripple from the smoothing capacitor when loaded?

If the peak voltage less than 50V you're alright.

What's the output voltage when unloaded?

If it's greater than 18.8V then you're alright.

The specs I gave are not using a linear voltage regulator, its brute force (rectifier then filter caps)

Just to clearify:
The voltage output is 58.5V When Unloaded (this is from the caps- I should really add a 10k Resistor is parallel to give more of a accurate measurement). Once I load it, It goes all the way down to 10V. It also has a ripple of .9V RMS.

One thing I dont know is if This transformer Im using is really a 2Amp transformer. I estimated that number from its size.

 

[audioguru]

The LM338 can regulate 5A only with 10V or less across it. With 30V across it, it limits the current to only 1A to protect itself. That is why the output voltage drops and why there is so much ripple.

It can't dissipate heat very well so if the voltage across it and the current through it heat it too much then it shuts down.

 

[Overclocked]

The LM338 can regulate 5A only with 10V or less across it. With 30V across it, it limits the current to only 1A to protect itself. That is why the output voltage drops and why there is so much ripple.

It can't dissipate heat very well so if the voltage across it and the current through it heat it too much then it shuts down.

Im not using the regulator- yet. Im asking if it would be Ok to use it.

I dont plan to use it as a vairable regulator, I just want it to regulate. Im assuming if I use it, the voltage would be more stable (its the only thing I have)

 

[audioguru]

When operating within its ratings, an LM338 regulates extremely well and its output ripple is typically 1/1000th its input ripple. Adding a capacitor to it reduces the output ripple to be 1/5000th.
Its input voltage will be very high. How high do you need its regulated output voltage to be?

 

[audioguru]

Wait a minute!
The output voltage drops to 10V without the regulator? Then the transformer is not powerful enough and will get very hot. A regulator won't make the transformer work better.

 

[Overclocked]

When operating within its ratings, an LM338 regulates extremely well and its output ripple is typically 1/1000th its input ripple. Adding a capacitor to it reduces the output ripple to be 1/5000th.
Its input voltage will be very high. How high do you need its regulated output voltage to be?

Not far from 58.8V, So I'll say 50V to keep it within 10V. Its mounted on a heatsink.

For some reason though, and Ive always had this problem, when I do the formula to get Vout, I end up with a vary Large Number. For R2 I used 50K. For R1 I used 100Ohms.

1.25(1+(50000/100)+50e-6*5000=625.5625

50k is a pot for adjustment.

Wait a minute!
The output voltage drops to 10V without the regulator? Then the transformer is not powerful enough and will get very hot. A regulator won't make the transformer work better.

Thank you! I kinda figured this. Well Time to move on to a bigger transformer. Luckly, I have the same transformer, but Bigger. But its a multi-tap

 

[audioguru]

With 1.25V across the 100 ohm resistor, the resistor to ground must be about 39k for 50V out.

 

[xiongyw]

In the National Semiconductor datasheet, it clearly says: This means there is no absolute max voltage (Do not try to go regulating voltages like 10KV though lol). If you have an input voltage of 58.8V, then do not try to regulate it to anything below 18.8V.You need to redo those figures. Caps in parallel add capacitance. ie. 2- 50V 1000uF caps in parallel will make 2000uF at 50V not 500uf at 100V.

Hope this helps

About the NS datasheet at <http://www.national.com/ds/LM/LM138.pdf>, it says in the first sentence that:

"The LM138 series of adjustable 3-terminal positive voltage regulators is capable of supplying in excess of 5A over a 1.2V to 32V output range."

So I am wondering what "32V" here really mean, as it's "floating" as mentioned above. Does that mean when output voltage >32VDC, short cut protection will not working? Is that the only effect of outputing more than 32VDC?

For example, if I have a unregulated 42VDC, and I want to regulate it to 36VDC (which is higher than 32V), and typical current is about 2A, will LM338K do it well? I do not need shortcut protection.

Thanks,
/bruin