Overdriving LEDs

oracacle

Senior Member
the difference is, I need only a single shot that lasts less than 1µs. currently getting around 1µs.

as it stand I am assembling equipment to start making PCBs, can't do a great deal more until that is done.
Wishing I had ordered more UV LEDs for my light box.
 

Jeremy Harris

Senior Member
the difference is, I need only a single shot that lasts less than 1µs. currently getting around 1µs.

as it stand I am assembling equipment to start making PCBs, can't do a great deal more until that is done.
Wishing I had ordered more UV LEDs for my light box.
How many UV LEDs have you got in your light box? I'm part-way through building a small one, using UV LED strip that has 120 off, 3528, 395nm LEDs per metre. I've stuck the strips side by side to the base of the box, linked at the ends so there is roughly an LED every 9mm across the whole base of the box, in both axes. These are wide-angle (120 deg) LEDs, so I'm hoping that they will be even enough to expose the board that will be around 30mm above the top of the UV LEDs.

I've no idea if this UV LED strip will work or not, but it seemed the cheapest and easiest way to wire up a lot of UV LEDs on a close grid. The biggest board I will ever need is a Eurocard, so 100m x 160mm, but I've made the box with a roughly A5 illuminated area, so could, at a pinch, make boards up to about 140mm x 200mm.
 

oracacle

Senior Member
I have 40 of these: http://www.ebay.co.uk/itm/301202861350?_trksid=p2060353.m2749.l2649&ssPageName=STRK:MEBIDX:IT

5 rows of six wide with 2 rows of 5 (one at each end) each spaced 2cm apart. I intentionally made my box quite deep, about 20cm. Once I have everything soldered together I will put a piece of paper in the glass and move the LEDs up and down until I get a uniform coverage. Think I am going to line the inside of the box with tin foil too to help it along a little.

Then hopefully tomorrow the ink I ordered for the printer will arrive and I can do some exposure tests to figure out how to expose for.

I am also trying to create a DIY drill press to hold my rotary multi too (dremel thing). If all goes according to plan I will try and post something in the finished projects, I suspect that it maybe handy for other out there - I had the wood left over from another project and other parts cost me £4, so no big loss if it doesn't all go well
 

hippy

Technical Support
Staff member
Rather late to the party but I have now finally put a logic Analyser on a pin to see what that produces -

Code:
#Picaxe 28X2
SetFreq EM32
Do
  PulsOut B.7, 1
  PulsOut B.7, 1
  Pause 10
Loop
Left attachment below.

That's a 28X2 using a resonator at 32MHz. That shows a 2us pulse and it would be logical to presume 64MHz would be 1us.

The difference with respect to the theoretical 1.25us and 0.625us likely comes down to software overhead as the pulse is bit-banged. Longer pulses are reasonably accurate 10us units at 4MHz but the overheads impact more with shorter pulses.

But can we do better ?

Code:
#Picaxe 28X2
SetFreq EM32
HSerSetup 1,%110
Do
  HSerOut 0,($FF)
Loop
That's a 0.25us pulse at 32MHz, which one would presume to become 0.125us at 64MHz.

Right attachment.
 

Attachments

binary1248

Senior Member
This thread fits perfect for another picaxe project that I have been thinking about.
I have been thinking of building a strobe (stroboscope ?) using some sort of LED. This thread is great as it points out various design problems and possible solutions. I had forgotten about white LED using phosphors. Although for my design I probably don't need to worry about persistence to much.
 

oracacle

Senior Member
Thanks hippy, at least it wasn't my measuring that was off.
Never would have thought to use hserout.
Need to decide on LED and order some smaller resistors for this to continue forward
 

Jeremy Harris

Senior Member
I love that work around by using hserout!

With a true output pulse width resolution as fine as 0.25µs, or, perhaps 0.125µs, then it makes this project look a lot more viable with a Picaxe. It also opens up the possibility of using short pulses for another application I've had in mind for some time.
 

hippy

Technical Support
Staff member
And with a small tweak; 125ns pulses at 32MHz, which should be 62.5ns pulses at 64MHz ...

Code:
#Picaxe 28X2
SetFreq EM32
HSerSetup 1,%110
PokeSfr $AF, 0
Do
  HSerOut 0,($FF)
Loop
That would set a 16Mbps baud rate at 64MHz. HSERSETUP cannot set that baud rate directly as using zero for the baud rate timer is taken as HSERSETUP OFF, but the fastest baud rate can be had by forcing a zero to the baud rate generator register using POKESFR after the HSERSETUP.
 
Top