Cool LED Displays: Boost That LED!

RE: Cool LED Displays: Boost That LED!

jw0752 — Mon, 04 Dec 2017 06:24:28 GMT

Hi Jon,

I love the way you let us look at the technical beauty that lies behind the scenes. Just about anyone can make an LED flash but making one flash in a new way is pretty neat.

John

RE: Cool LED Displays: Boost That LED!

jc2048 — Sun, 03 Dec 2017 23:36:54 GMT

[View:https://players.brightcove.net/1362235890001/NkxiVJdjx_default/index.html?videoId=5666451443001:740:466]

Another one. This time it's twenty LEDs and a better shape - a star. I've increased the on time so that the current is higher to compensate for the increased number. You can see from the waveforms that the initial current through the LEDs is now 50mA and the voltage at the top of the string lifts to 40V to get the current to flow. (Although the LEDs will have a DC maximum rating of something like 30mA, they don't mind being pulsed like this - the average current is quite low, just a few milliamps.)

[View:/resized-image/__size/480x234/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-a86be8f2-f3c5-4434-bc34-40b2212ef6b2/contentimage_5F00_186047.png:480:234]

The ringing is quite entertaining - the first half cycle tries to go below ground but is stopped by the MOSFET body diode conducting and the subsequent cycles continue with the negative peak just above ground.

And I can report that the MOSFET didn't mind at all being avalanched and is still happily working (whoops!). Probably not too surprising - with the coil limiting the current and the duration quite short it's only doing what any small zener would be able to do.

RE: Cool LED Displays: Boost That LED!

jc2048 — Wed, 29 Nov 2017 09:23:11 GMT

[View:https://players.brightcove.net/1362235890001/NkxiVJdjx_default/index.html?videoId=5661927153001:740:466]

I couldn't resist doing a bit more with this, so I added two more strings of LEDs (with the LEDs all interleaved) and used MOSFETs to switch one string at a time into circuit giving a travelling pixel effect. The LEDs aren't very even and I needed a small piece of cardboard halfway round to stop them shorting, but it sort of works, if a bit slapdash.

RE: Cool LED Displays: Boost That LED!

jc2048 — Sat, 25 Nov 2017 15:41:54 GMT

Whilst I had this set up, I thought I'd do a follow-up. I'm going to look at the ringing and then try using it to measure something (more on that in a moment).

The oscillation at the end, when the LEDs stop illuminating because the coil can't power them any more, seems to appear from nowhere and is all a bit mysterious as there doesn't seem to be anything in the circuit to cause it. It's a nice sinewave, the frequency is fairly constant, and it diminishes gradually. It's also exactly the same on each and every cycle of the converter. Those details give us the clue that we're probably looking at resonance of a tuned circuit. A tuned circuit that is kicking back and forth the final dollop of energy left in the coil. The inductance is obvious, it's the 10mH coil (it actually measured 10.3mH), but where is the capacitance? Well, the capacitance is hidden in the other elements of the circuit. The MOSFET has some internal intrinsic capacitance at the output. For a small MOSFET like this that will be quite small, a few tens of pF (with a power MOSFET it can be quite considerable). The LEDs also have some capacitance when they're reverse biased, though having seven in series helps us here because the capacitance that the coil sees will be a seventh of what it would be for a single LED. There will also be a very small contribution from the coil itself - there's capacitance from turn to turn, but it's not very much and, as with the LEDs, the overall value from end to end is much less. Finally there's the capacitance we tend to overlook and that's the probe tip capacitance of the 'scope probe.

What I'm going to try and measure is the capacitance at the probe tip. I'll then use that to calculate what the actual frequency of oscillation is (ie without any probes attached) and we'll see by how much that differs from the reading that the scope gives.

Here's the waveform with two scope probes attached. Why two probes? You'll see in a moment.

[View:/resized-image/__size/480x234/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-a86be8f2-f3c5-4434-bc34-40b2212ef6b2/contentimage_5F00_186043.png:480:234]

The frequency is 210.4kHz. Using the formula for resonant frequency of a tuned LC circuit and plugging in the values for frequency and inductance I get a value of 55.5pF for the capacitance.

If I now remove one of the probes and do the same thing again

[View:/resized-image/__size/480x234/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-a86be8f2-f3c5-4434-bc34-40b2212ef6b2/contentimage_5F00_186044.png:480:234]

the frequency is now 234.4kHz. That gives a capacitance of 44.8pF.

The probe capacitance is in parallel with the circuit capacitance, so it adds to what is there. The capacitance for one probe is then simply the difference between the two figures, ie 55.5pF - 44.8pF = 10.7pF. That's a reasonably good figure for a passive probe (you'd expect a modern x10 probe to be in the range 10pF-20pF, though an old one might be anything up to 30pF) - there's a natural limit to how low it can go because of the screened cable to the oscilloscope.

So what would be the frequency with no probes at all attached. If I take the 44.8pF figure and subtract 10.7pF for the single probe I get 34.1pF. Plugging that into the formula gives me a frequency of 268.5kHz. So that means that when I probe the circuit, the frequency drops by about 13%.

It's easy to forget the effect that a small amount of capacitance can have on a circuit, particularly when you spend a lot of time looking at logic where the impedances are low and an output can happily throw 50pF of capacitance around with rise times of only a few nanoseconds. Even a x10 probe like the one I was using here, where the tip capacitance is much less than a x1 probe, doesn't necessarily save you from problems when you're looking at more sensitive analogue stuff.

RE: Cool LED Displays: Boost That LED!

Jan Cumps — Sat, 25 Nov 2017 10:40:17 GMT

Jon, what current probe did you use for the scope traces? a commercial one or self-built?

RE: Cool LED Displays: Boost That LED!

doug65 — Fri, 24 Nov 2017 19:42:29 GMT

I like the idea. I never thought of trying to get anymore than the 5 volts.

RE: Cool LED Displays: Boost That LED!

shabaz — Thu, 23 Nov 2017 18:30:04 GMT

Hi Jon,

Nice project and video! Driving 7 series LEDs from 5V with such few components is very cool : )

RE: Cool LED Displays: Boost That LED!

mcb1 — Thu, 23 Nov 2017 07:54:52 GMT

Nice.

RE: Cool LED Displays: Boost That LED!

rachaelp — Thu, 23 Nov 2017 00:14:06 GMT

I love what you’ve done here, nicely explained [View:/resized-image/__size/16x16/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-a86be8f2-f3c5-4434-bc34-40b2212ef6b2/contentimage_5F00_1.png:16:16]