> Now I'd like to change all the leds to a 5 watts 220 volt bulb, > so the matrix become an 8 x 128 5watt 220 volt. Here are the last relevant posts of what I think was the only other time this was discussed ==================================================== From: Lance Allen To: Subject: Re: 220volt lamp matrix Date: Wednesday, 10 November 1999 08:32 Gabriel Callao wrote... > Hi all, > > I have to make a 220volt lamp matrix of 7x5 lamps=35 lamps. With it, I > will be able to "draw" different letters. If they were leds, I would only > need to control a row (or a column) and do refreshment in every other > column (or row). Does anyone know how to control all these lamps > using the fewer lines of a Pic micro, and also fewer triacs ? > I would consider rectifying the 220v AC to DC and then switching in a matrix using IGBTs.The High Side switches would be N channel as would the Low Side switches. The High Side gate drive will float and obtain their power from small high frequency transformers, the gates being triggered by opto couplers. Thats 12 IGBTs and driver ccts. I am presently in the final design stages of a variable freq 230vac source and this is the technique I have chosen. Note that the supply to the lamps will at best be muxed at 5:1 so brightness may be a problem. There are some seriously dangerous voltages and high instantaneous currents present so beware. (Standard warning for newbies who might not fully understand the dangers involved). Kids... dont try this at home. Lance Allen Technical Officer Uni of Auckland Psych Dept New Zealand www.psych.auckland.ac.nz ===================================================== From: Paul B. Webster VK2BZC To: Subject: Re: 220volt lamp matrix Date: Wednesday, 10 November 1999 09:07 Lance Allen wrote: > I would consider rectifying the 220v AC to DC and then switching in a > matrix using IGBTs. > I was going to suggest Thyristors, but bit my tongue. Why? Well, > you must multiplex so that each lamp is pulsed at least 50 times > per second. For the same reason as you always requite a minimum > multiplex frequency of this order. This means that for five rows, each > row is driven for 4 milliseconds. Note that your rectified DC must > also be smoothed so all rows get the same voltage! IGBTs it is! Also note - one diode in series with each lamp - this is the secret to multiplexing incandescent lamps! > Note that the supply to the lamps will at best be muxed at 5:1 so > brightness may be a problem. Indeed. Let's look at this. At full brightness, the lamp has a certain resistance. Using rectified DC however, we are driving at 1.4 times the voltage, thus twice the power. However, it is on at a 1 in 5 duty cycle, so receiving overall 40% full power. If this is acceptable, (it may be, to extend lamp life) then go ahead. Otherwise consider - using 110V lamps on the 220V supply, using the PWM control to limit their full brilliance (and along the way, compensating for actual power line voltage by measuring same). Also, you can consider "pre-emptive" filament warming of lamps at least 1 second *before* they are switched to full brilliance, using about 10% of full PWM time. In other words, perform your animation one second in advance! In case it's not obvious, this is a full-fledged PWM application and also, if the micro "hangs" in mid-multiplex there's going to be considerable smoke released unless you put in place a few protective measures. The WDT at *no* prescale is an obvious one to start, as is using pull-ups to default to lamp drivers *off*. > There are some seriously dangerous voltages and high insantaneous > currents present so beware. (Standard warning for newbies who might > not fully understand the dangers involved). Kids... dont try this at > home. Yep. That's pretty obvious too! It's a pretty exciting design -- http://www.piclist.com hint: To leave the PICList mailto:piclist-unsubscribe-request@mitvma.mit.edu