SparkFun Forums

[geste]

I stumbled across some 4-inch tall red 7-segment LEDs in a local surplus shop and it wasn't long before I bought the SFE GPS Clock board. I'm no PIC programmer, but I figured I knew enought to piece this together.

I used the Bloader to install the provided sample code and that seemed to go OK

The LEDs have 10 pins. I didn't have a hookup sheet, but I used 9 volt leads and mapped out the common cathode and pins for segments a through g. Or at least I made each of a-g light up (I ignored the decimal point) with 9 volts with the ground/negative connected to pin 3. I cut a stock RJ45 patch cable in half and hooked up ground and segments a-g and then plugged that test LED into port 6. When I power it up, segment f blinks on and off, but that is it. I took it out on the deck and left it for 10 minutes so that GPS would have time to acquire (thinking that maybe the display would just blink something until acquisition. I also plugged that LED into RJ45 jacks 1 through 5 with the same result that segment f just blinks.

Before I went much further I figured it was time to stop and question my assumptions

I took it from the board's schematics that ground/cathode was common. Is that true?

Since I am lighting up at least 1 segment, I am assuming that my pin 1 to pin 8 RJ45 orientation/assignment (with Pin 1 white/orange connected to ground) is correct. Fair assumption? It looked like the traces from U16 to U17 to JP6 would line up just like in schematic, so I went with RJ45 pin 1 as ground.

Should I be assuming that the a-g outputs will drive a 7-segment display like this directly?

If everything else is hooked up correctly, will the sample code available for download actually display a HHMMSS clock off of the 6 RJ45 outputs? I looked at the code before I loaded it but I would have to struggle a lot more to make sense of it.

Should I be able to eyeball status/output if I hooked a serial console up again?

If everything comes together, I am going to build a wood and smoked lexan frame and install this in my office (I tested with a handheld and I should be able to keep plenty of satellites in view from my window).

I wondered if I should have left the unit outside for longer than 10 minutes, but it started to rain

Thanks!

Jim

[jasonharper]

I took a quick look at the schematic for the clock board, and it's clearly incapable of driving a common-cathode display - the ULN2003 driver chips are only capable of pulling down, not up.

You've also made some sort of wiring error, possibly as simple as counting the pins starting from the wrong side of the connector, as the expected outcome of the wrong display polarity would be no segments lit at all.

I'm afraid I don't see any easy way for you to get this to work. Would wiring up 42 transistors and 42 resistors (one per display segment) be within what you're willing to do to fix this project?

[geste]

First, thanks for such a quick reply (on a weekend no less) and to a pretty naive set of questions.

I took a quick look at the schematic for the clock board, and it's clearly incapable of driving a common-cathode display - the ULN2003 driver chips are only capable of pulling down, not up.

That started to dawn on me as I went back and re-read the last section of the project writeup....

You've also made some sort of wiring error, possibly as simple as counting the pins starting from the wrong side of the connector, as the expected outcome of the wrong display polarity would be no segments lit at all.

Well, I guess that helps explains things, too

I'm afraid I don't see any easy way for you to get this to work. Would wiring up 42 transistors and 42 resistors (one per display segment) be within what you're willing to do to fix this project?

These LEDs were cheap, like $3 each. If I could fix the problem by getting some common-anode replacement LEDs, they could be $12-$15 each from what I have seen. So, it might be worth trying the transistor/resistor fixup and use the LEDs I have on hand.. I have some perf board around and I could try it one just one LED for starters. I, guessing transistor price would be a big determinant. So fire away. How to wire up?

Thanks!

Jim

[inventore123]

Well the ULN2003 can only drive a common anode display, that's true, but before each ULN2003 there's a 74HC4511, if I'm reading the correct schematic. And the 74HC4511 drives common cathode displays
The only problem is the current that they can source. A quick look at their datasheet shows that they can drive nearly 10mA per each segment at a voltage of 5V. If the displays you have light up enough brightly with that voltage/current, just bypass all the ULN2003. Don't forget to add resistors to limit the current. Also consider that the display now draws current from the 5V, so the 5v power supply needs to be boosted to provide the extra current.

[jasonharper]

I doubt that a display of this size is going to light up adequately with only 5V 10mA per segment, but it certainly would be convenient if it did... Jim, in any case you're going to need some more exact specs on the displays to continue. Hopefully there's a standard part number on them, that can be Googled to find a datasheet. If not, you're going to need to hook one back up to your power supply, and measure the voltage across, and current through, a lit segment.