High current driver for 16 segment LED display

[Mitch]

I'm having trouble getting this to work. Everything appears to be good up to the gate of the NDP6020P, then it is not switching.That scope image is QuickPrint2.png. The other image is at the drain. The reason for this circuit is that the HK16K33 can't supply enough current to drive large displays, and they run on a higher voltage. I want to stay with that chip. Help is appreciated. Thanks in advance.

[n1ist]

Vgs(max) for the NDP6020P is 8v, so it may have been damaged by feeding it 12v.  I'd add a resistor between the NDP6020P gate and the 2N7000 (so the gate of the NDP gets pulled to 7V above ground (Vgs of 5V) when you want it on.

/mike

[Mitch]

Thank you.  I will try that tomorrow.

On Tuesday, August 13, 2019 at 9:14:16 AM UTC-4, Mitch wrote:

[gregebert]

The gate waveform is OK, but the duty-cycle is very low so it will be harder to see the LEDs on. Remember - This is a PMOS device, so it will be 'on' when the gate is low.

Now, for the drain, you can see that it's on for a short time, then decays in an RC-like manner. It's probably the RC time-constant of your scope+probe. It looks like the segment drivers are off at this time; if they were on, you would see a rapid decrease in drain voltage rather than a decay.

Check the sequence for your segment-drivers. You want those to go 'on' when the gate is driven low. And you probably want the 'on' time of your PMOS driver to be longer (gate=0v for more time).

Hopefully you have not damaged the gate-source; running it at 12V instead of 8V will probably impact it's long-term reliability somewhat, but probably did not destroy it. The resistor you want to add as noted by Mike will be roughly 910 ohms (standard value); you can go as low as 510 ohms.

[John Rehwinkel]

>
> I'm having trouble getting this to work. Everything appears to be good up to the gate of the NDP6020P, then it is not switching.That scope image is QuickPrint2.png. The other image is at the drain. The reason for this circuit is that the HK16K33 can't supply enough current to drive large displays, and they run on a higher voltage. I want to stay with that chip. Help is appreciated. Thanks in advance.

It's a really nice chip, I use it too.

I'm unclear what the scope photographs are, but I'm seeing what looks like an RC time constant, which may well be the 1k resistor charging the gate capacitance of the NDP6020. I'd suggest using a gate driver chip instead of the 74HC04 and 2N7000. A TSC426 or 472 (depending on whether you want the signal inverted) or similar should do the job. These chips can supply 2A of gate charging current, so they switch quickly.

- John

[Mitch]

Thanks, John. I was at a dead end with this circuit. I tried several values of resistors in series with the gate. That didn't work, but disconnecting the 2N7000 and grounding the gate on the NDP6020, did turn on the LED. 

I will try the TSC426. The only reason for the 74HC04 was to invert the signal because the 2N7000 also inverted it, so a non-inverting chip will simplify the circuit. The only issue is the cost. Eight are required to drive 16 segments and they are over $7 on digikey, but 27 cents on aliexpress. Does a series resistor go between the HK16K33 column driver and the gate of the TSC426, and is a resistor required between it's gate and source? Also, I think the NDP6020 is overkill to drive at most four LEDs on the largest segments, so I'll probably switch to a NTR1P02T1G or something similar.

Here is the first word clock based on the HT16K33 http://angryelectrons.co/wordclock-1.

If I can find a similar circuit using the HK16K33 that will drive IN-17 tubes, I'll  do a version for them, too.

[Bill Notfaded]

Interesting project Mitch. I too love alphanumeric displays!

Bill