PSU Using The LM338

[MrClamperSir]

Hi, I am building a supply using the lm338 and I am wondering what input voltage to use. Currently I'm using a transformer with a 12V, 2.5A on the secondary side and I can't get the Vout (with load) above 8.4V.

I thinking I may need more current on the input but the datasheet for the lm338 doesn't say (or at least that I can find).

Any suggestions would be helpful.

Thanks

 

[davenn]

rectified and smoothed DC input to the reg chip should be around 9.5V and there will be around 1.2V drop across the voltage regulator
so your max of ~ 8.4 VDC is about right

to get more voltage out, you need more V in
what is the max voltage that you are interested in ?


I was going back to front, sorry ... trying to concentrate on too many things at once ... Its still my work day here

OK the 12V I assume is the 12VAC from the transformer ( no load)
if using a bridge rectifier ( 4 diodes ) there will be a 1.4V drop (0.7V for each of the 2 conducting diodes in each 1/2 cycle)

so 12VAC = 10.6 V of rectified DC
The smoothing capacitor, you said you were wanting the PSU capable of around 2 Amps, so 2200uF to 3300uF is required ( 1000uF / Amp is the rule of thumb)

after smoothing we get 10.6V x 1.414 = ~ 15V smoothed DC ( this is all WITHOUT load)

putting a load on it will drop it several volts, and there will also be a 1.2V drop across the regulator chip

trying to find a reference to working out the Vdrop with load


Dave

 

[MrClamperSir]

rectified and smoothed DC input to the reg chip should be around 9.5V and there will be around 1.2V drop across the voltage regulator
so your max of ~ 8.4 VDC is about right

to get more voltage out, you need more V in
what is the max voltage that you are interested in ?


Dave

Thank you for your reply. I'm looking to get 20V-25V out. I'm only using 3V-9V around 1A but I'm told that easing the ceiling on what it "can do" vs what it's doing is better for the circuit.

I also want to keep everything fairly low temp as I have had problems with circuits I've built getting really hot on the transistor.

Should a 24V 2A transformer give me at least 9V 1.5A out and stay reasonable with heat? That's the max load I will be drawing.

 

[davenn]

I was just about to do some editing ... was hoping to do it before you got back online
I was throwing you off course a little

let me edit my post OK ?

 

[davenn]

OK just learnt something
an interesting www page for you to read and me to read more fully ......

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

it states that the transformer rating is it's voltage when the stated current is drawn

ie. your 12VAC, 2.5A rated transformer is this .... it will show 12VAC across its terminals when 2.5A is being drawn by the load
therefore, if the load is less than 2.5A, its output will be a bit higher ( no load = approx. 12% increase in voltage)


Dave

 

[davenn]

sorry for the initial mis-info ... the sheet I was reading was for working out the needed secondary transformer voltage when you know what DC output you needed


hope the new info helps


D

 

[davenn]

Should a 24V 2A transformer give me at least 9V 1.5A out and stay reasonable with heat? That's the max load I will be drawing.

with that you should be able to get your 20 - 25V out if you ever require it

I assume you are using the metal cased LM338 ... preferably the steel cased one, not the aluminium one
The steel cased one is more rugged but will still need a decent heatsink
It will handle 5A without any need for a pass transistor


Dave

 

[MrClamperSir]

OK just learnt something
an interesting www page for you to read and me to read more fully ......

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

it states that the transformer rating is it's voltage when the stated current is drawn

ie. your 12VAC, 2.5A rated transformer is this .... it will show 12VAC across its terminals when 2.5A is being drawn by the load
therefore, if the load is less than 2.5A, its output will be a bit higher ( no load = approx. 12% increase in voltage)


Dave

This is REALLY good info. in other words I DON'T want to overload my circuit with Amps in because If I'm not using them all ( which I won't be) that equals out to be an increase in voltage (which I won't be using either so that's just more heat to dissipate!?).

Am I understanding this correctly? Just get a transformer rated for the Amps out I need?

 

[davenn]

the 2.5 A rating of that transformer is OK for many hobby uses

you just need one with a higher voltage output so you can adjust the LM338 to a higher voltage out if needed

power in Watts to dissipate = voltage drop across the regulator x the current drawn by the load

eg 18V into the regulator and 12V out, that's a 6V drop and say 1.5A of current being drawn by the load

Watts dissipated = 6V x 1.5A = 9W of heat to get rid of


Dave

 

[BobK]

The heat wasted in a linear regulator is:

(Vout - Vin) * Iout

If you need only 9V out, using a 24V transformer will produce 5 times the heat of using a 12V transformer.

Bob

 

[davenn]

The heat wasted in a linear regulator is:

(Vout - Vin) * Iout

If you need only 9V out, using a 24V transformer will produce 5 times the heat of using a 12V transformer.

Bob

yup as stated

 

[MrClamperSir]

The heat wasted in a linear regulator is:

(Vout - Vin) * Iout

If you need only 9V out, using a 24V transformer will produce 5 times the heat of using a 12V transformer.

Bob

Yes that makes sense. The 12V input doesn't give me enough headroom though. I'm going to try an 18V. The 338 metal case regulator is rated for 50W and at most using an 18Vin at 3.5Vout drawing 1.5A I will be at 21.75W which is under half the rated value.

That's not counting the voltage drop with the diodes and regulator itself.

Does that sound about right?

 

[davenn]

The 338 metal case regulator is rated for 50W and at most using an 18Vin at 3.5Vout drawing 1.5A I will be at 21.75W which is under half the rated value.

It still MUST be heatsunk


Dave

 

[MrClamperSir]

It still MUST be heatsunk


Dave

It's heat sinked on the outside of the aluminum enclosure. So far it's staying warm at most.

Everything else sounds right to you?

 

[davenn]

cool

 

[eetech00]

Hi

The LM338 can drop as much a 2.5-3v across it. Plus 2v for rectifier..so maybe an 18v transformer output would be better.

 

[BobK]

A 12V transformer will produce about 15VDC. This is plenty to regulate down to 9V. If you are not getting 9V there is something wrong.

Bob

 

[MrClamperSir]

A 12V transformer will produce about 15VDC. This is plenty to regulate down to 9V. If you are not getting 9V there is something wrong.

Bob

Interesting! I think I may know why.

I'm using a 12VDC adapter I modified as the input because it was easier than ordering a transformer just to see if my circuit was working. I've got it hooked up at the beginning of my circuit board (before the rectifying diodes). Could this be the problem?

 

[BobK]

Yep. That will get you only about 10.6V and you need 12 to get 9 out of the regulator. Try hooking it up after the rectifier bridge.

Bob

 

[Petkan]

Linear regulators (like the one you are using) are limited by the difference between Input and output voltage, especially under heavy current.Their heat dissipation is defined by this voltage difference times current.
Assuming you want to stick to the transformer you have, a best approach is to use a switching mode regulator like LTC3115,which will have noт just way better efficiency.but also the ability to produce output voltage lower or higher than the available input voltage. Go to www.linear.com for details and LTSpice illustration.
This Buck Boost regulator can tolerate up to 45V at the input and can produce up to 40V at the output with max load of 2A.
You can regard this regulator as DC transformer. The product of output voltage and current (output power) will almost match the respective number at the input (minus around 15% losses). This implies that at lower than input voltages you can get higher than input current (as step down transformer implies) As other voltage regulators, the output voltage is defined by a resistive divider, scaling down the output voltage for comparison with internal reference voltage. This implies волтаге control range above the reference as the formula is Vref*(1+R1/R2). If you need more than 2A have a look at LT8705 - again Buck Boost circuit but with external MOSFETs. The LT demo board for LT8705 illustrates up to 200W at input voltage as high as 80V.