Shop OBEX P1 Docs P2 Docs Learn Events
Help with Darlingtons — Parallax Forums

Help with Darlingtons

BRBR Posts: 92
edited 2012-06-23 03:17 in General Discussion
I have a circuit that I believe is drawing far more current than it should. My belief is based on the fact that the 5v regulator gets warm enough that it's uncomfortable to leave your finger on it for more than a few seconds, even when the device is in an idle state.

I have a theory that perhaps it is due to the way I am using a ULN2803 Darlington array. There's nothing in the data sheet that would make it obvious (to me anyway) that my theory is or isn't correct, and I've looked around at other references such as this N&V article to no avail. So I'm hoping folks on the forum can point me in the right direction.

To be specific about it, I have a ULN2803APG Darlington array, with prop pins directly connected to the Darlington inputs. I have the output side connects the cathode side of 12V LEDs to ground. See attached schematic.

My theory is: perhaps the 3.3v applied by the prop pin on the Darlington input isn't enough to drive the gate to fully saturated "on" or "off" since the load side of the Darlington sees 12v. Is this plausible? Or to put it another way...is there anything fundamentally wrong with the way I'm using the Darlington in this example?

Comments

  • BRBR Posts: 92
    edited 2012-06-21 19:36
    Digging a bit more...this post from Mike Green would seem to suggest the answer to my question is "no".

    The plot thickens...
  • kwinnkwinn Posts: 8,697
    edited 2012-06-21 19:43
    First, you mention a +5V regulator but your schematic shows a +3.3V and 12V supply. Is something missing in the schematic?

    Without current limiting resistors between the ULN output and the leds they will draw too much current and be damaged (usually a very breif flash and smoke).

    Driving the ULN inputs directly from the prop should work. If your theory was correct the ULN would be getting hot.

    Try posting a more complete schematic, particularly of the regulators and the ULN outputs to the led.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2012-06-21 19:55
    kwinn wrote:
    Without current limiting resistors between the ULN output and the leds they will draw too much current and be damaged (usually a very breif flash and smoke).
    Since he specified "12V LEDs", I assume he means internally-limited automotive types.

    -Phil
  • lardomlardom Posts: 1,659
    edited 2012-06-21 21:53
    Forgive me if I don't have a clear understanding. The way I've used the ULN2803 is that the ground pin (#7) is common to both the input and the output side. You can limit the current on the 12V side with a series resistor. If you have a component that is warm to the touch in an idle state the input/output pin may very well be faulty. You can retest your circuit with a lower voltage LED (to be safe) on different pins. I hope this helps.
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2012-06-22 02:22
    I don't see how 30ma delivered at 3.3volts into a Darlington (with 1000x gain) would not be enough.

    Is the ULN2803 running hot as well? While the device can power 60ma per pin, the total output for the chip is nowhere near 8 x 60ma. Darlingtions tend to run hot by the nature of having two transistors in the same package.

    Is the 3.3v regulator also running hot?

    It is rather absurd to have 1000x gain when what one really requires is 60ma output at some elevated voltage. A 2N2222 will easily provide 500ma.

    What you need to consider is that the jump from 12v down to 5v is dumping a lot of heat. Powered at 7.5 volt, the 5v regulator runs much cooler. Is this a 'true' 12v or is it coming from a lead acid battery at 13.8v or higher? Maybe you have a wall wart that claims 12VDC, but is unregulated and running much higher.
  • BRBR Posts: 92
    edited 2012-06-22 20:45
    Thanks guys, those are good points. If the Darlington wasn't being driven into a saturated state, it'd get hot (it isn't).

    I was intentionally vague as to other details of my circuit because I wanted to first focus on verifying (or rejecting) my theory that the Darlington (or my use of it) was the culprit. Theory rejected.

    So let me fill in some additional details...an updated (more complete) schematic is attached, as is a picture of the physical hardware. I'm using a Propeller protoboard...12V switched PSU...Darlington used to drive 12V LEDs (yes, the type with built-in current limiting resistors)...Prop is also driving a Nokia 5110 LCD display, a DS1302 timekeeping chip (with supercap attached to TC pin)...and I have 3 normally open push-buttons wired to 3 prop pins pull-up fashion.

    Anticipating a few comments here...
    1) Yes, I'm aware that the protoboard is only spec'd for 9V in and that I'm living life dangerously by using 12V in
    2) Yes, I'm aware that the 5v regulator would be expected to dissipate significant heat due to high voltage drop it has to cope with...IF there's significant current draw on the device...more on this in a moment
    3) Yes, I'm aware that it is bad practice not to have a current limiting resistor on the prop button pins to protect them in case the pins are set to output and driven low
    4) The Darlington is a ULN2803A, which has built-in current limiting resistors on the inputs

    Getting back to the 5v regulator heat load...my expectation is that the regulator will get hot only if there's a current draw on it, but if there's little/no current draw, it won't get hot, even in spite of the fact that I'm using it outside the protoboard spec. The thing that's got me stumped is...I can't see where there should be much of any current draw on the 5v regulator.

    I did another experiment just to try to isolate the issue a bit more...I disconnected the LCD, and took the DS1302 & ULN2803A out of their sockets. If I turn it on, the 5v regulator still gets hot...meaning that within ~30s, it is warm enough that it's uncomfortable to leave your finger on it for more than a few seconds. The prop also gets slightly warm to the touch. I haven't noticed any other "warm spots" anywhere else on the board (well...the 3.3v regulator also gets slightly warm but its barely noticeable).

    One final piece of information...prop pins 1, 4, and 5 seem to be dead (these are attached to the ULN2803A). I have no idea why or how. All other pins seem fine. Device itself seems to work fine in terms of executing code. The Darlington gates attached to 1,4,5 also still work...I can manually drive them on by touching a 3.3v source to the gate input.

    Is it possible there could be a short internal to the prop, perhaps in one of the dead I/O pins? I thought that when a pin gets killed, it typically looks like an open, not a short.

    POST-EDIT: and another thing that's got me stumped...how do I change the thread title? I'm not seeing it...
    960 x 720 - 65K
    640 x 480 - 192K
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2012-06-22 21:08
    One possible scenario: all pads on the Prop Proto board are surrounded by a ground plane connected to Vss. Despite the solder mask, a solder bridge from a pin connection to the ground plane could cause the symptoms you observe. Double-check all of your solder joints.

    -Phil
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2012-06-23 03:17
    I am not so sure that a lot of heat is being dumped just being in a standby state. Read the PDF for the exact brand and part number you have. Parallax says '9 volt', but they usually provide 7.5 volt wall warts. You can 'pre-regulate', by adding a 7809 between the +12 and the board. And that can have a nice big heat sink.

    And, it may be that the lack of an adequate heat sink is a big contributor to the problem. These new voltage regulator packages without a mechanical mount do heat up quicker - some get so hot as to melt the solder and slide right off the board. (Try applying the technically acceptable limit of about 30volts to the input and you can watch the solder melt.)

    Solder bridges are very common and very insidious. Many that I have had, did not show up on visual examination. Only when I dragged an Xacto knife between all the suspected situations did the problem get resolved. And I did use a 20x loupe for visual inspection.
Sign In or Register to comment.