Wahooooo! Exams are over, term starts next week and I’m free other than a few nights out with me chums over the next few days.
In the meantime, today brought the arrival of my first ever DSO (Hantek DSO5072P) and a new LCD (the other LCD came a couple of days back but that comes soon!).
As a quick tester for my DSO, I hooked it up to my LED boost circuit to scope out a few traces.
As of yet, my impressions are that it is about as good as the one I use at university for my labs – some Gwinstek that looks pretty much the same! It cost me £175 of the good ol’ eBay (http://goo.gl/u6rLKE – Link will expire!), there were cheaper ones but this threw in free DHL shipping meaning it arrived to me 5 days after I purchased it.
I’ve recently been redesigning the Phobass LCD board to use a higher quality LCD than the standard PCD8544 84×48 Nokia equivalent LCDs that I’ve been using. The reason for this being I’m not a fan of the way the PCD8544 LCDs mate with the screen and how they’re also dependent on PCB thickness to ensure they’re completely secured to the PCB. They were a good step into the non-standardised world of LCDs but now I’m more confident, I’ve moved up to a tasty 128×64 LCD with an FPC connector. I’ve also been thinking of designing a bare minimum smart watch (as previously mentioned – https://hsel.co.uk/2015/01/09/eon-ultra-smartwatch-update-2/) using an 8×2 character LCD. As I needed to find a suitable display, I visited what I essentially find as the home of nearly any LCD/OLED that you could want, BuyDisplay (http://www.buydisplay.com/). Searching through their selection, I found a really nifty 8×2 character LCD. Cheap as chips and interfaced through I2C. Now if any of you have read my previous posts, you will know that I hate using I2C on the STM32 series of chips because its generally just a massive ballache to get working quickly. After a fair bit of faffing however, I got it working with both this LCD and a HMC5883L compass meaning my generic I2C driver is getting much more reliable, though not yet ready to be unleashed on the world! Writing up a quick sketch using the STM32F0 RTC has made me a low quality, poor resolution (both screen and time) clock.
As you can see, its a pretty simple arrangement. Two wires for I2C, 3 capacitors for VLCD, the internal charge pump and the power supply, a jumper from reset to VDD, the backlight through a 100Ohm resistor and of course VDD and VSS for power. Its going to be quite a challenge making a meaningful smartwatch when you can only display 16 characters at once with 8 of them being per line!
Talking of better LCDs, my actual arrival today was an LCD I found scouring through eBay’s US website (http://goo.gl/pgzucv – Link will expire!). A nice little 128×64 LCD with tiny dimensions, about the same as a PCD8544, just with better pixel density!
This is a really cool LCD and only requires 4 pins to interface it. SCK, SDA, CS and RS. The LCD integrated both power on reset and software reset. It also supports clock speeds up to 33MHz! Much faster than that pesky PCD8544 with its 2MHz.
As those of you who are regulars may know, I’m designing a 100W LED torch! I’d previously built and explained a pretty poor boost converter based around the good ol’ 555 timer. After much advice from the EEVBlog forum, I decided to design a boost converter around a dedicated controller. My choice of controller may not be a good one but after extensive spice simulation, it does the job! It is the… MC34063, with the design of converters covered by Dave Jones of EEVBlog himself. now obviously supplying 34V at 3A will require much more than the measly 1.5A switching current of the MC34063 so I’ll be using two IRFN44Z Mosfets in parallel (maybe even one depending on whether it can handle the load) to switch along with a beefy 30A inductor. Using the standard laws for conservation of energy and knowing my input voltage will be @7.4v (2s LiPo battery), At 80% efficiency (an achievable value!), I’d be looking at an average current of ~17.3A. Thats a fair bit of current and could produce a fair amount of magic smoke, especially from the LiPo so much care will be taken. Not forgetting too, if the LED is consuming 100W, only ~12% (Source: http://en.wikipedia.org/wiki/Luminous_efficacy) is converted into light meaning round 78W is wasted as heat. Obviously this requires some pretty drastic heatsinking methods so I’ve gone full out with a fan operated computer heatsink, made for the most power hungry of processors… the Pentium 4! I picked it up off eBay for <£5 and with my trusty Dremel (ahem Ferm, Dremel ripoff), I drilled mount holes and purchased a cheap (<£5 again!) tapping kit to allow me to screw directly into the holes. I’ve also purchased a parabolic reflector to focus the light for me too.
Driving the IRFN44Z with a 555 timer and a 100Ohm gate resistor leads to minimal heating of the Mosfet so I should hopefully be able to get away with a small heatsink for it. The inductor however feels pretty damn warm to touch! I don’t really blame it since I did salvage it from a power supply sometime in my past life… Fortunately, the fan runs fine at the slightly reduced voltage and is quite. If the LED is running at low powers, it shouldn’t need the assistance of the fan to cool it.