How to stress test a LM7805?

Discussion in 'General Electronics Chat' started by Grayham, May 18, 2010.

  1. Grayham

    Thread Starter Member

    May 18, 2010
    79
    2
    Hi, I have made a simple 12v to 5v circuit using the 7805 + heatsink and want to stress test the output for a few hours at 1A, then if that's stable another hour on 1.5A, etc.

    What is the best way to do this?

    I know what I technically need is a 5watt, 5ohm resistor from Vout to GND, but they don't exist as far as I know (only 1/4 watt, etc)

    Anyone know of any other devices or ideas how I can do this?
     
  2. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    Well, there are resistor with power rating more than 5W. For this test, especially for the 1.5A case, you will need a quite big heat sink to keep the regulator cool.
     
  3. kingdano

    Member

    Apr 14, 2010
    377
    19
    5 ohm 5W resistors certainly exist, and can be found on digikey in that rating and much higher.

    if i were you i would go above and beyond on the power rating - but how did you arrive to the 5W number?

    (Vi-Vo)*I would give you the heat dissipation - so (12V-5V)*Imax = Pd

    7V*1.5A = 10.5W


    im sure someone (sgtwookie...looking at you here) will correct me if im mistaken - i am not familiar with the device you are using.


    based on my math though - i would use at least a 20W resistor - like this:

    http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=KAL255F-ND

    i would also recommend mounting it to an aluminum block and using thermal compound between the resistor and the block - this will ensure enough heat dissipation (dont place the block on carpet or something silly like that - a nice solid work benchtop would be the right choice)

    use a beefy wire gauge as well (something like 14-16AWG is more than enough - just dont use 30 gauge wires)

    edit:

    we use this compound at work

    http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=KAL255F-ND


    also , make sure you selected a suitable heatsink - i did something similar to this recently with an LM317 - was regulating 36V down to 8V and needed a MASSIVE heatsink for the TO-220 package - the thing still gets smoking hot too. use sil pads to mount the heatsink to the device as well if possible - or that thermal compound
     
    Last edited: May 18, 2010
  4. Wendy

    Moderator

    Mar 24, 2008
    20,765
    2,536
    The resistors you ask for exist, but generally I parallel resistors.

    For example, if you want a 5Ω 5W resistor, and can get ¼W resistors, you will need 20 resistors total. 20 X 5Ω is 100Ω, so you need Qty 20 100Ω resistors in paralllel.

    NOTE: DO NOT try to make a resistance that will be a long term project at exactly the wattage needed. There is an informal but strong rule that says you should derate a part to 50% of its max level for reliability over the long haul. If this was the case you should use qty 40 200Ω resistors, which would be 10W.

    As long as you don't bunch them together, using a bunch of small wattage resistors also eliminates the need for a heat sink. They need good air flow (with air outside the box), but no one resistor will get that hot.

    Another advantage to this technique is you can tweak the values to much tighter tolerances. Most cheap resistors are 5%, you can get much better by mixing and matching odd values to bring it well under 1% tolerance (the DVM is basically the determining factor here).
     
  5. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    In addition to Bill's information, it would be good to leave a significant space between the resistors in order to have better airflow around each resistor and thus better cooling.
     
  6. kingdano

    Member

    Apr 14, 2010
    377
    19
    bill

    that is good advice, but assuming this is a home project budget may be an issue

    i think using one resistor is OK as long as proper preparation is done and care is taken to isolate the hot parts from anything flammable...children etc.

    ;)
     
  7. Wendy

    Moderator

    Mar 24, 2008
    20,765
    2,536
    It doesn't hurt that 5% ¼W are cheap. My personal cost is 2¢ each if bought in quantities of 100 or more.

    Project: Resistor Parts Storage

    ***********************

    I was posting this before I saw yours, but I answered your question anyhow.
     
  8. kingdano

    Member

    Apr 14, 2010
    377
    19
    good point.

    :)
     
  9. Grayham

    Thread Starter Member

    May 18, 2010
    79
    2
    I am no math expert, but I assume if Vout if +5.00v, then to stress test at 1A I would need a 5ohm, 5W+ resistor in series.

    Where if using a 10ohm resistor it would do 500ma, or 2.5ohm would be 2A, etc.
    Someone please correct me if my calculations are wrong, but I am sure they are correct.

    I did this test using a bunch of resistors in parallel (equaling 5ohm at the end in about 5W), but it's impractical to use a bunch of resistors like this.
     
  10. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    I think a single 5W resistor is more expensive than a hundred 1/4W resistors.
     
  11. kingdano

    Member

    Apr 14, 2010
    377
    19

    the total power dissipated will be the difference in voltage times the current - either the load or regulator will have to dissipate the power.

    using the resistor will draw a 1A load - but the device will still have to deal with the rest of the power you are trying to lose regulating 12V to 5V
     
  12. Wendy

    Moderator

    Mar 24, 2008
    20,765
    2,536
    Good point, generally linear regulators get hotter than the load they are providing.
     
  13. Grayham

    Thread Starter Member

    May 18, 2010
    79
    2
    Yeah I am aware the 7805 will dissipate much heat going from 1A 12v to 1A 5v.
    This is exactly what I want to test it doing though, but for a long period of time.
    Goal is to make sure it doesn't overheat in the box or what type of heatsink I will require.

    I am unsure of the best practices people use to stress test dissipating the 5W into thin air though.
    Like for example, using a bunch of 1/4w resistors, or using a 5W bulb, or using an old rusty 5ohm kitchen knife from the bottom of the cutlery draw, etc.
     
  14. kingdano

    Member

    Apr 14, 2010
    377
    19
    an overrated (power rating) load is what i use at work
     
  15. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Generally accepted practice is to double the actual power requirement for the resistor. So, you need a 5 Ohm 10W resistor, or a pair of 10 Ohm, 5 Watt resistors in parallel.

    We have no clue where you are, since you have not bothered to take 20 seconds to fill out your location in your profile.
    (hint: click the "User CP" link up top, then click the top link under that, and scroll down to where the "Location" box is, type your location in the box, then scroll to the bottom and click "Update")
    However, if a Radio Shack is near you, a pair of these would work just fine:
    http://www.radioshack.com/product/i...1&filterName=Type&filterValue=Power+resistors
     
Loading...