vreg output current not reaching spec???

Thread Starter

jaydag77

Joined May 28, 2007
17
Hi again all, I am still working on my vreg module design this afternoon- Some new reg's were delivered this morning so I have been trying them out, they are from TI and output 12V 500mA and 12V 1.5A. I have tried both in a test circuit and also tried another brand of similar spec- I am getting bad results when I test for current though- and am wondering if I did something wrong? They are putting out 12V dc just fine but the 500mA reg gives only ~200mA and the 1.5A only ~300mA !! The test circuit is typical- fed from 19vdc @ 2A or better, run through filter cap (470uF) after diode (4001), then direct into the reg. 0.1uF caps on both pins to gnd. Also a res/ LED indicator on the out tied to gnd.

When I test the current along various points, The LED cuts out. I am guessing that this is due to the meter testing/samples under a load? or is it a resistive thing? I think I may have been supposed to unconnect part of the circuit to take the test but cannot recall- plus it wouldn't make much sense cause what about if I wanted to test various portions of a circuit for current while under a load(like to make sure the circuit is running under loaded conditions)??

I suspect that I am just doing things wrong, but as far as the vreg output I just dunno, it is getting fed more than enough to run the reg,, and if there is that much a difference in the in to output then it must be very inefficient way to feed a circuit?? I'm disappointed. Tips, hackling welcomed, thanks in advance, jd
 

Thread Starter

jaydag77

Joined May 28, 2007
17
sorry 'bout that- part numbers are: UA78M12 (12vdc 500mA), UA7812 (12vdc 1.5A),, I just re-glanced at the sheet for the 78m12,, still cannot see what the issue is unless there may be a direct short?? I am using a breadboard for the circuit but am not reading any bridges with the multimeter, so...

really need to get this figured.....always somethin.
 

Gadget

Joined Jan 10, 2006
614
Obvious question, they are mounted on a heatsink...?
Also, does the 19 volts into the reg drop when the regs loose regulation..?
I'd also be increasing the 470 uf cap as well.
 

bloguetronica

Joined Apr 27, 2007
1,541
As far as I can see, these regulators may need heatsings to deliver their nominal current. An LM7812 will only deliver that order of amount of current without an heatsink, although they are capable of delivering 1A (thermal shutdown occurs). But I'm guessing without more information.

What is the type of regulators. I need to see part numbers/type numbers to see the specs.
 

bloguetronica

Joined Apr 27, 2007
1,541
sorry 'bout that- part numbers are: UA78M12 (12vdc 500mA), UA7812 (12vdc 1.5A),, I just re-glanced at the sheet for the 78m12,, still cannot see what the issue is unless there may be a direct short?? I am using a breadboard for the circuit but am not reading any bridges with the multimeter, so...

really need to get this figured.....always somethin.
Sorry, didn't noticed. The UA78M12 has close specs compared with the LM78M12. It also has short circuit protection and thermal shutdown. I can't compare the UA7812, but I see that it also has short circuit protection and thermal shutdown.

A short would cause the output to have no voltage. I think you really need an heatsink, since the thermal shutdown is limiting the output current (so the voltage drops once the thermal shutdown actuates). An heatsink will dissipate the heat, so the regulator can deliver more output.

You can do an experiment: just atach a clip to the metal tab of the regulator (TO-220 package), and you will see that you can have little more current (I expect more 50mA). If you blow at the regulator you can have even more current (done this experiment with my anmeter set to 10A with the leads connected to output and ground, and saw that the regulator really limits current)*.

*P.S.: Of course that is not the way to evaluate the nominal current of a PSU, so I don't advice it. Also, don't burn your fingers. The tab gets really hot.
 

Thread Starter

jaydag77

Joined May 28, 2007
17
Thanks all for the replies! alot more than I had expected!
First thing this morning I tried out the heatsink I am intending to use on the pcb, attached it to the MC7812 (12vdc 1A) and tested it shortly after starting it. The current reading is still about the same ~300mA-- I am going to try clipping some additional sinking to it right now and see if there is any difference. Strangely last night I was fooling around with the other part of the circuit (5V reg) and it showed readings at and well above spec @~1.5-3A on a L7805CV.(without sinking)

OK, I just tried playing around a bit more- attaching additional sinking, even tried blowing on the reg as suggested, no noticable difference at all, mind you my MM is set to the 10A current setting. I am thinking I must be doing something wrong in my measuring. I did a bit of a search to see about that and came upon a page which brought up additional question- "TO MEASURE AC OR DC CURRENT, YOU MUST BREAK THE CIRCUIT AND CONNECT
THE TEST LEADS TO TWO CIRCUIT CONNECTION POINTS. THE CONNECTION, MUST BE
IN SERIES WITH THE CIRCUIT UNDER TEST. CAUTIONS: Never connect the test
leads across the voltage source." -- The last part is what is unclear, what exactly is being conveyed here? As for the top portion, my circuit is being measured from the reg's common and Vout, which is in series as far as I can see? The only other thing in the circuit is the LED indicator which I have tested with it in circuit and pulled. I really must be doing something wrong cause there is no way that two brand new regs are that much off spec!

All of this crap has really made me re-evaluate the use of standard regulators- I have never had a good history using them either! If anyone here has seen my previous thread to this then they will already know that the target is a RC platform I have converted to a sort of 'bot'. It is supposed to have: ~12vdc 1A H-bridge, ~5vdc .5A H-bridge and a 5vdc ~150mA for uC mainboard. All powered from ~14V 1.2A nicad battery pack. At this point I am thinking to just cut out the 12V reg and power straight from packs the bulk of the analog circuits and use a reg for the uC only? I could even use a separate pack for the 5vdc 150mA and keep things separate?

lightningman- I cannot seem to open the pdf?? it just is timing out or something...may be my wifi but browsers are fine so.....I'll try again.


Thanks again everyone! Me thinks I need to order some better regs for future use! regards, jd
 

lightingman

Joined Apr 19, 2007
374
Ahhh.I think I can see what you are doing...You are connecting your DMM accross the output of the REG to measure the current available.... NO..You need a load of a known value in series with the DMM to measure it.A constant current load would be good for this, you can make one with another 7805.input to the 7805 input terminal, output from the 7805 to your DMM +,DMM - to your PSU ground.Now the common terminal of 7805 the needs to be connected (NOT TO GROUND) but through a resistor to the output terminal, the value of this resistor will determine the current through the circuit. I think 100R gives about 50 mA, so 50R will give 100mA and so on. If you use a large wire wound pot it can be varied.....The resistor will get quite hot at high currents, so will need to be large, and once again the 7805 will need a heatsink.Have fun.....Daniel.
 

Thread Starter

jaydag77

Joined May 28, 2007
17
Thanks man! I am going to give a go in a few minutes,, I had a feeling I was using the dmm wrong when the vreg was heating up so fast and much and the LED indicator was bogging out everytime! I have been looking around for info regarding proper measurement but found not much. The 'manual' for my dmm is less than useful. At least now I don't feel so lost and have hope that my design doesn't have to be scrapped completely. I'll post a schematic in a while- haven't drawn one up yet but for paper scribbles, but do have a rough layout done.

btw I ended up being able to open the pdf after a re-boot- seem to be having issues with reader in past few days, just nother reason to switch back to linux fulltime.

cheers, jd
 

lightingman

Joined Apr 19, 2007
374
Here is an updated PDF, with the constant current regulator.
Turn the pot to maximum resistance and just connect it accross your PSU, adjust the pot to the level on the DMM that you wish the PSU to be load tested at.......... OH ON !!! I have made a mistake on the constant current schematic....... The output should be from the "common" pin on the reg and NOT the "output" pin....SORRY.......

Daniel.
 

Attachments

Thread Starter

jaydag77

Joined May 28, 2007
17
Lightingman; well thanks again for the feedback, and especially for taking the time to submit the pdf with updated schematic, which came in very handy considering my interpretation of the text had left me in error. I have just finished doing the test and the result was more appropriate, although over spec for the part used (MC7812, 12vdc 1A) came out as follows: input ~19vdc @~3A DMM result when testing through about 3.9ohm gave current result of 1.87A ,11.52vdc from posted config.

I guess the main point of the testing is that the regulator be tested under a load? How can I use this info to design my proposed regulation design? I'm having a bit of a hard time figuring out how to ensure my regs are going to feed enough amps to both the h-bridges and the uC main? Do the regs automatically adjust the current supplied to meet the requirements of the rest of the circuit(s)- I would have thought it to be a constant (for standard regs), but I guess that isn't really implied in the datasheet is it...?

Agh, I am just getting more confused as I go... I guess above all I should ask how or if I am able to measure for current in an active circuit(like what if I need to make a measurement on a pcb??), do I just break into where I want to test and tie the dmm in series or can this only be done in certain cases? If I know how to test then at least I can get back to floundering.....

OK, I just wired up the reg, LED and everything in the 12v circuit that I wanted and re-tested BUT I kept the 10uF cap that was suggested in your test schematic, and lo and behold the current reading was about 1.5A!! Is it possible that the cap (or lack of rather) was the problem(plus I also use the heatsink now)? either way I am impressed and believe that the measurement must have been correct-ie break the circuit and tie dmm in series,, still unsure bout' testing pcb circuits but I'll get there later tonight...I am going to try a different value cap to see if I can get it down to 1A, spec for the part supposed-to-b.

Thanks lightingman, you saved the day.....:D regards, jd
 

Attachments

hgmjr

Joined Jan 28, 2005
9,027
Lightingman,

Is it possible that what you intended to recommend was this configuration for the constant current source?

This diagram is straight out of the National Data Sheet for the LM317.

This circuit only works, I believe, when built up from one of the adjustable three terminal regulators. I have not seen it built up from one of the fixed three terminal regulators.

Or did I miss something in your suggestion?

hgmjr
 

Attachments

Gadget

Joined Jan 10, 2006
614
Yep... that'll be it. If you put an ammeter directly across the output, you are effectively shorting the output of the regulator, and it will go into protection mode.

To check a 12 volt reg for 1 amp, connect a large wattage 12 ohm resistor and measure the voltage at the output of the reg for any voltage sag.... (before the resistor starts to smoke up).
5 ohm for 1 amp outta the 5 volt reg. (use Ohms law).
 

lightingman

Joined Apr 19, 2007
374
Hi hgmjr..Yes it does work with the standard 78xx series.I use this arangement for keeping my RAM back-up batteries charged at 50mA (3xAA's). They have been in the lighting desk now for 10 years backing up 3 x 1Mb and 3 x 16Kb RAM's... (not bad) and still as good as the day they went in.....Cheers.....Daniel.
 

bloguetronica

Joined Apr 27, 2007
1,541
Lightingman,

Is it possible that what you intended to recommend was this configuration for the constant current source?

This diagram is straight out of the National Data Sheet for the LM317.

This circuit only works, I believe, when built up from one of the adjustable three terminal regulators. I have not seen it built up from one of the fixed three terminal regulators.

Or did I miss something in your suggestion?

hgmjr
An LM78xx can be used where the LM317 applies, can be even used for variable voltage supplies (the output voltage won't be less than xxV). However, the LM317 allows better regulation in that task, since the minimum voltage of 1.25V allows better voltage adjustments through the resistors (less error).
Don't forget the LM317 is a regulator like any other, but is voltage is of only 1.25V.

Also, the regulator will limit the current, so the current never reaches big values. Of course this won't give the nominal current. The best way to ensure a good dissipation is to use the formulas:

Preg = (Vin - Vout) / Iacross
Dissipator (ºC/W) = (Treg - Text) / Preg (I use 70ºC for Treg and 40ºC for Text, being the last the temperature inside the case, or ambient temperature).

The smaller the value (in ºC/W) is, the bigger the heatsink is, so choose one with a value smaller than the one calculated. Also, this formula aplies if the junction between the heatsink and the case is neglectible, so you should use heat conductive paste. Using it makes a lot of difference.
 

lightingman

Joined Apr 19, 2007
374
Hi..I think every engineer has a favourite device, yours must be th LM317..Yes your rite they are a very good reliable regulator.You can get them down to 0 volts with a -ve ref votage of 1.25v...Cheers, and good luck..... Daniel.
 

bloguetronica

Joined Apr 27, 2007
1,541
Hi..I think every engineer has a favourite device, yours must be th LM317..Yes your rite they are a very good reliable regulator.You can get them down to 0 volts with a -ve ref votage of 1.25v...Cheers, and good luck..... Daniel.
The LM317 is not my favorite. I prefer to use the ones from the LM78xx series if I can. Anyway cheers.
 

hgmjr

Joined Jan 28, 2005
9,027
Hi hgmjr..Yes it does work with the standard 78xx series.I use this arangement for keeping my RAM back-up batteries charged at 50mA (3xAA's). They have been in the lighting desk now for 10 years backing up 3 x 1Mb and 3 x 16Kb RAM's... (not bad) and still as good as the day they went in.....Cheers.....Daniel.
That is cool.

I will have to try that technique the next time I have a need for a constant current regulator.

Thanks for the tip.

hgmjr
 

Thread Starter

jaydag77

Joined May 28, 2007
17
thanks all who have responded! I value any tips, even conflicting cause they allow me an opportunity to learn through experiment and advice....that being said, the last few posts have not alot of value for me due to my poor understanding of the foundations (which surely must be quite apparent at this point in the thread......) The formula- though may be the ticket to finding the solution or part of it is beyond my understanding in that I do not know what they express? so....

Its back to foundations for me I guess, cause the farther I go with disecting this design the more confused and lost I am getting! I mean if I cannot even make measurement properly that is a good indication that I am out of my range.

To allow continuing with my project I decided late last night to just go with using direct from battery pack for the h-bridges and other analog portions which require current over ~150mA and stick with 1, 5Vreg for the uC and supporting hardware. Then at least I am cutting out half of my existing problem in one shot (still need to learn the material and methods but at least can keep working...)

I want to thank everyone again for taking time to make a post here, and apologize for not being able to use the suggestions better than I have--regards, jd
 
Top