Testing ULN2003 Darlington Array

Recently my ULN2003 got hot and subsequently seems to not drive my stepper motor so I pulled the ULN2003 chip and put it on a breadboard by itself.

I tied E to ground. I added on Pin 16 (Output 1C) a 1K Resistor and an LED in series. If I apply 5V to any pin of the ULN2003 my LED is lighting up.

Is my chip fried or am I misunderstanding how the chip works?

You need to draw the circuits you are using, verbal description is too vague.

That chip will only act as a switch to ground, so your LED has to be wired to a positive voltage with a resistor to the chip’s input to get it to light.

Thee_Captain:
Is my chip fried or am I misunderstanding how the chip works?

Certainly sounds like it.

Why would you be using a ULN2003 in any case? :astonished:

The basic "is it fried" test, put a positive voltage on each input pin 1-7 and test continuity between each corresponding output pin 16-10 and pin 8. This only tests the "switch" function of the chip and doesn't stress test it. The chip costs less than 25 cents. If your use is playing around, keep using it as long as it does what you want it to do. If the use is more critical, replace it when in doubt. Also, consider replacing it with a TBD622003APG, which uses n-channel FETs instead of darlington BJTs, which runs cooler, and is pin compatible.

Assuming you have it wired correctly - The common emitter pin (physical pin 8) would go to ground, the negative side of an LED to one of the outputs, positive side of said LED to a resistor and then +5v - if the LED turns on when you connect a positive voltage to any input, instead of just the one specific to that output, yeah, the chip is trashed.

Did you have physical pin 9 (the one that all the freewheel diodes are connected to) tied to the supply to the motors? Or if not, did you have a freewheel diode on each motor? If the answer to both of these is no, that's why you fried it - when you turn off a motor or other inductive load, the current wants to keep flowing, and this will generate a large voltage spike, which will damage the switching element if you don't have a freewheel diode to clamp that to the motor supply voltage.

Paul__B:
Certainly sounds like it.

Why would you be using a ULN2003 in any case? :astonished:

Probably because they're dirt cheap, abundant, and there are thousands upon thousands of tutorials that use them - they're pretty poor switches compared to modern MOSFETs, but they're perfectly serviceable for non-demanding use cases.

Perehama:
TBD622003APG, which uses n-channel FETs instead of darlington BJTs, which runs cooler, and is pin compatible.

I have looked for a chip like that for YEARS! Can you please post the correct part number? Google search for that part number gives no results (even without the APG on the end).

Here 'tis:
https://www.digikey.com/product-detail/en/toshiba-semiconductor-and-storage/TBD62003APG/TBD62003APG-ND/5514087

Aaah! TBD62003, not TBD622003

THANK YOU!

TBD63083 is the version of the ULN2803 (the 8 channel one) and TBD62783 is the 8-channel high side one.

DrAzzy:
Aaah! TBD62003, not TBD622003

THANK YOU!

TBD63083 is the version of the ULN2803 (the 8 channel one) and TBD62783 is the 8-channel high side one.

Sorry about that. There are others, for example, the TI TPL7407L, but Toshiba is the only one making them in a PDIP package to my knowledge.

I'd seen the TPL7507L - but that needs 8.5v on COM, which is limiting.

Just ordered 5 each of those parts in DIP and SOIC :stuck_out_tongue:

Seems that a member of the TPIC6xxxx family can have the same current capability.
Don't forget the thermal limitations when driving more than one output.
And shift registers require less drive pins.
Leo..

The TPIC6-series (which are excellent parts, highly recommend) is more of a replacement for a '595 than a replacement for a '2003.

You can't exactly PWM an output of a TPIC6-series :wink:

DrAzzy:
You can't exactly PWM an output of a TPIC6-series :wink:

Depends on your coding ability.
SPI control (optional) is very fast.
And the TPIC chips have an output-enable pin that can be PWM-ed.
But yes, sometimes it's easier to just use a darlington mosfet array.
Leo..