Advice Sought for Triac light controller

[hal9000]

hi all.

I've been developing a very specific random light controller using a picaxe 40X1 and i'm at the stage now where the programming part is almost done and i need to drive the intended lights from the picaxe..

I'm not an analogue electronics expert, I troubleshoot microprocessor based digital microelectronics in my usual job, so i'm a bit vague on analogue stuff and wondered if someone could comment on my selected circuit...

I am driving a number of miniature 11w fluorescent tubes, these use OSRAM electronic ballasts (QTFM 1x11's).

Shamelessly found on the internet, i modified the opto-driven TRIAC circuit found here :
http://www.mitedu.freeserve.co.uk/Circuits/Switching/triacctl.htm

I just removed the 470R and LED and substituted the 3040 with a Motorola MOC3020 (more plentiful). I'm Also using a BTA10 TRIAC, because i have a drawer full more than any other reason I'm also using 20mm fuseholders/fuses to aid in servicability. The fuse selection i was unsure of, but i figured to try as low as possible, my rudimentary calculations say 0.04 amps at 240v is required for the 11watt tube and electronic ballast, but i thought i'd scale up slightly, i have 100ma and 500ma fuses in stock, but currently i have only run the test circuit with a 1amp@240v fuse.

I built one example circuit on a stripboard and left it running the lamp with one of the 14 outputs i'm using on the 40X1.. for a few hours... The PICAXE turns the light on and off randomly, on and off times range from 2 seconds to 3 minutes. The triac didn't even get hot, nor did the opto driver, so i'm taking that as a good sign, the cooler this runs, the better as it will be enclosed with 13 other identical circuits..

my main questions for you all is:
- What rating of fuse should i go with? 100ma, 500ma ? or am i off?
- Do you think the circuit is suitable, could it be improved? have i made any dangerous assumptions?
- The BTA10 (400B) triacs are isolated tab, i was thinking of bolting these to an aluminium bar heatsink to share the heat dissipation (if any). Is this reccomended?


I thankyou in advance for your input, i love the forums here, and i love the picaxe, i have so many projects in my head it will take me a while to bring them all to fruition... Like i say, my background is much larger CPU based systems troubleshooting, small microcontrollers have always grabbed my attention, but i just havn't had the time until recently to get stuck in, and i don't come across them often (i work on stuff from 1971 to 1984 mainly)..

 

[sedeap]

Reliability

Hi:
I use the same concept for many of my designs and after several years, still work as intended.

you can see in the user projects some of them Here

This concept look like bullet-proof design.

The heatsink is recomended, and the isolation too.
and one small fan like CPU's fan in the end of the box is good too.

Bye.

 

[inglewoodpete]

During startup, a fluorescent lamp can draw an average current of about 10 times the operational current. So for about 1 - 2 seconds at turn-on, expect a current of ~400mA. Due to the thermal lag of a fuse, a 100mA fuse is unlikely to blow but in the long term it possibly will.

The BTA10 has a dynamic resistance of 40milliohms and a junction-to-ambient thermal resistance of 60C/W. If the fluorescent draws ~40mA when operational then the triac will dissipate 64uW. That equates to a temperature rise of about 0.004C.

Now the 'worst case' scenerio. At end-of-life the fluorescent will fail to start and so it could draw an average current of 400mA. This will cause the triac to dissipate about 6.4mW, with a temperaure rise of 0.4 degrees Celcius. Since you don't want the fuse to blow when the 'tube goes on the blink, a 500mA fuse would be more appropriate.

A heat sink on the triac would be overkill, although you may want to bolt it to something (eg the chassis) to keep it secure.