[FOLLOW UP] Adjustable LM317HVT High current supply design

Thread Starter

Marcus2012

Joined Feb 22, 2015
425
Adding a current limiting resistor at the base of the PNP seems to have solved the problem of the blowing regulator. I will now try to combine the NPN output and regulator output and see if it works :)
 

ronv

Joined Nov 12, 2008
3,770
Adding a current limiting resistor at the base of the PNP seems to have solved the problem of the blowing regulator. I will now try to combine the NPN output and regulator output and see if it works :)
It could be the pnp is shorted or backwards. The 20 ohm resistor may need to be larger (watts) depending on the base resistor.
 

Thread Starter

Marcus2012

Joined Feb 22, 2015
425
It could be the pnp is shorted or backwards. The 20 ohm resistor may need to be larger (watts) depending on the base resistor.
I'm going to quadruple check the transistors again just to be super sure they are ok :) The base resistor I added was 1K btw :) The new design is below.

Linear PSU 24-48V adjustable Final.png
 

Thread Starter

Marcus2012

Joined Feb 22, 2015
425
IT'S ALIVE!!!! Thanks guys it seems to be adjusting voltage across the transistors now, seems that base current was destroying the regulator. I'll get the load set up and seem what happens :) It does make the lights dim though when I turn it on lol. I guess that's just the inrush current for the filter capacitor.
 

ScottWang

Joined Aug 23, 2012
7,413
I suggest that you should use 4 or 5 power transistors, I bought a power kit long time ago, it used 2 piece of 2N3055 (15A) to output 4A current, 15A rating current only used for 2A application, and it never damaged, what do you expect 20A can be afford in your application? (6.67A for each power bjt)
 

Thread Starter

Marcus2012

Joined Feb 22, 2015
425
Well I hooked her up to the induction heater and well what can I say.......it's a vast improvement over a 12V supply in one of these voltage fed heaters. Thanks for all the help everyone :) Top range of the supply seems to be around the 56V mark which was actually higher than I expected to be honest but I guess I can just chalk that up to tolerances. One problem does seem to be this initial capacitive load, it's tripping the 16A breaker in the garage so I have to hold it up so the capacitor charges and the current drops (this is without load). Should I have a resistor in series with the capacitor to soften the start-up?
 

ronv

Joined Nov 12, 2008
3,770
Something like this is sometimes used:

In your case a 48 volt relay. It puts the resistor in series with the primary until the cap charges up then the relay shorts out the resistor.
 

Thread Starter

Marcus2012

Joined Feb 22, 2015
425
Ahh awesome thank ronv, this looks like it fits the bill :). Any idea on how I go about deciding how much to limit the current on the primary winding and therefore my resistor value? Or is it a case of just keeping the secondary current below the surge current limit of my FWB rectifier?
 

ronv

Joined Nov 12, 2008
3,770
I'd just try the 10 ohms so the current will be below the trip point of the breaker.
You have a toroid transformer?
 

ronv

Joined Nov 12, 2008
3,770

Thread Starter

Marcus2012

Joined Feb 22, 2015
425
I'd just try the 10 ohms so the current will be below the trip point of the breaker.
You have a toroid transformer?
Yeah it's a toroidal but it's 240V here unfortunately and I think that circuit was designed for 120V, if I used a 10 Ohm that would be 24 Amps which would trip my breaker and have me worried about my rectifier:oops: I could go with 48 Ohm 10W which would give me 5A and limit my secondary to 30A.


I found this relay too and seems to fit nicely :)
TE CONNECTIVITY / OEG OJ-SS-148HM,000 General Purpose Relay, OJ Series, Power, Non Latching, SPST-NO, 48 VDC, 10 A
 

Thread Starter

Marcus2012

Joined Feb 22, 2015
425
Hey guys, Time for me to do a little follow up on this. Thanks for the inrush limiter circuit Ronv that work a treat and stopped the breaker going every time but the huge cap keeps it on for about 5 minutes after I switch it off, but that's no big deal :)

It works quite well with the inductive load I have designed it for but unfortunately I haven't got a resistive load I can use to test it with at the moment but I am assuming if it can handle an inductive load it should be ok with a resistive or capacitive. As expected the high voltage provided to the royer oscillator decreased heating time significantly.
One thing I am noticing though is that before a load is switched on the highest voltage measurable would be closer to 60V however once load is attached the maximum voltage decreases to the expected 52V approx. I can only assume that the voltmeter is measuring a peak value here before a load is applied. In addition to this when an inductive is switched on the voltage will drop very low and quickly increase to near the previously set value. I say near because it seems and inductive load slight skews my range of voltages, not by much but by some (5V max). I'm guessing this forced voltage drop is just due to the fact I can't change the current instantaneously in an inductor, but no damage to LM317 seems to result from this.
I chose some rather small emitter resistor values to balance my current load across the NPNs before realise the cable resistance is 4 milliohm per foot so I had to make the cable lengths match before attaching them to my resistor array.
Air flow in the unit is maintained by a LM383 PWM fan controller mounted on the top, inside the case with the respective thermistor behind the filter capacitor as this is the most likely area to encounter stagnant airflow. Four 40mm 0.12A axial fans mounted in opposite pairs maintain the airflow. The rear heat sink for the Power BJTs also has active forced air cooling (two 80mm axial fans) which are maintained by an additional fan controller with it's respective thermistor behind the heat sink and set low enough so ambient temperature increase is directly proportional to heat sink temperature.

So BIG thanks to the following members for helping me troubleshoot this and getting it going. I really appreciate it thanks :)

Ronv
Dodgedave
Lestraveled
Hp1729
Scottwang
Wbahn


I've attached the final schematic and some photos in case anyone is interested. As you can probably see I fell into my usual trap of barely leaving myself any room to fit it all in :oops: Also if anyone has any advice, suggestions or criticisms on the finished unit please feel free to comment, I would also appreciate this as I always make a list of refinements for future implementation (FYI that burned resistor did get replaced lol). :)

Thanks again everyone
Linear PSU 24-48V adjustable Final (+inrush limiter).png 20160308_225810.jpg20160330_180021.jpg 20160330_175712.jpg


20160330_184216_resized.jpg
20160330_184236_resized.jpg
20160308_172331.jpg

20160308_172318.jpg
20160330_180040.jpg
 
Last edited:

ronv

Joined Nov 12, 2008
3,770
Hey guys, Time for me to do a little follow up on this. Thanks for the inrush limiter circuit Ronv that work a treat and stopped the breaker going every time but the huge cap keeps it on for about 5 minutes after I switch it off, but that's no big deal :)

It works quite well with the inductive load I have designed it for but unfortunately I haven't got a resistive load I can use to test it with at the moment but I am assuming if it can handle an inductive load it should be ok with a resistive or capacitive. As expected the high voltage provided to the royer oscillator decreased heating time significantly.
One thing I am noticing though is that before a load is switched on the highest voltage measurable would be closer to 60V however once load is attached the maximum voltage decreases to the expected 52V approx. I can only assume that the voltmeter is measuring a peak value here before a load is applied. In addition to this when an inductive is switched on the voltage will drop very low and quickly increase to near the previously set value. I say near because it seems and inductive load slight skews my range of voltages, not by much but by some (5V max). I'm guessing this forced voltage drop is just due to the fact I can't change the current instantaneously in an inductor, but no damage to LM317 seems to result from this.
I chose some rather small emitter resistor values to balance my current load across the NPNs before realise the cable resistance is 4 milliohm per foot so I had to make the cable lengths match before attaching them to my resistor array.
Air flow in the unit is maintained by a LM383 PWM fan controller mounted on the top, inside the case with the respective thermistor behind the filter capacitor as this is the most likely area to encounter stagnant airflow. Four 40mm 0.12A axial fans mounted in opposite pairs maintain the airflow. The rear heat sink for the Power BJTs also has active forced air cooling (two 80mm axial fans) which are maintained by an additional fan controller with it's respective thermistor behind the heat sink and set low enough so ambient temperature increase is directly proportional to heat sink temperature.

So BIG thanks to the following members for helping me troubleshoot this and getting it going. I really appreciate it thanks :)

Ronv
Dodgedave
Lestraveled
Hp1729
Scottwang
Wbahn


I've attached the final schematic and some photos in case anyone is interested. As you can probably see I fell into my usual trap of barely leaving myself any room to fit it all in :oops: Also if anyone has any advice, suggestions or criticisms on the finished unit please feel free to comment, I would also appreciate this as I always make a list of refinements for future implementation (FYI that burned resistor did get replaced lol). :)

Thanks again everyone
View attachment 103410 View attachment 103414View attachment 103415 View attachment 103416


View attachment 103417
View attachment 103418
View attachment 103419

View attachment 103421
View attachment 103422
http://forum.allaboutcircuits.com/threads/value-of-a-capacitor.122450/#post-982688
Very nice!
Out of curiosity how much current does the heater draw with and without the bolt in it?
 

Thread Starter

Marcus2012

Joined Feb 22, 2015
425

Thread Starter

Marcus2012

Joined Feb 22, 2015
425
That will generate tons of heat.

I would encourage you to think of a switching mode pre- or tracking regulator before the linear regulator.
I did consider this but I had trouble understanding how I could use a pre-tracking regulator with the pass transistors without all of the current passing through the pre-tracking regulator. The main function of the unit is to operate at maximum voltage and I just accepted at the time that the voltage differentiation would be indirectly proportional to the current available without burning out my transistors. :(
 
Top