Tuesday, December 30, 2014

Of(f) course...

As always, just when things are moving forward they turn to a screeching halt again.

For some time the MP3-triggers has behaved badly and "misfired" when playing MP3's. I thought this could be due to the solenoid-interrupt stepping in and messing with the communication. Should be easy to check, just remove the interrupts temporarily, right?

Wrong. 

What happened was that I've coded the solenoids to fire right away (for speed) and be turned off in the... that's right, the interrupt function. This lead to the solenoid being stuck in active and took me a while to notice since I was debugging code on the computer at the same time. Things that happen when you leave a project for some time, it's almost impossible to remember all decisions that were taken and for what reason. After that I heard what almost sounded like gravel dropping on the floor and thought "what the hell is that?".

Then came the smoke.

I moved faster than Mick Jagger on speed and shut off the machine, threw the glass front off and pulled the board up. I expected to see fire, but there were just one sizzling solenoid. I had to keep it cool by blowing air on it for a good long time. Phew.

Damage control.

1) Sizzling solenoids, anyone? I've heard they're all the rage...
The first thing was obvious. The drain kick solenoid was more fried than a bucket of KFC. Clocking in at almost 0Ω it was basically a short circuit.

Next thing - not a single solenoid worked after the incident. They seemed fine resistance-wise and didn't feel warm, so I suspected the power supply and check the voltage. It read 6V... not good. Should be 48V. After unplugging the short circuit solenoid the voltage went up to 48V again. Thank you, China manufacturer, for voltage limiting when shorting instead of blowing up!

2) New. Crispy.
I replaced the solenoid with a spare one I just happened to have in my "garbage" pile. (An exact match too. Seriously!?). Still no tango....

3) Spot the burnt MOSFET....

...because of course, the solenoid driver board MOSFET was blown as well.
I looked high and low for a replacement, but could only find IRF640's. I used these in the past and they worked, but kind of slow for the flippers so I scrapped them. They also became quite hot due to not opening fully when activated (not really a low-level voltage MOSFET).

The board seemed fine, so I soldered a IRF640 on there and included a heat shield. How about now?
4) Surely this has got to be it, right?

Nope.

Nothing.

Crap.

My theory is that the last time I used them was with the Ardunio Mega, which runs at 5V 40mA per output. The Chipkit has 3V and 20mA, if I'm not mistaken. This is simply to weak to open the gate at all. I raised the on-time with 10x and it still didn't work, so I once again scrapped the idea of using them.

I've now ordered a new power driver board (just in case) and a bunch of replacement MOSFET's (RFP30N06LE). It GOT to work.

On the plus side...
I've created the Atlas, Big Daddy and Upgrade game modes now, along with additional tweaks and things (like user selectable background music before launching the ball). Just got to get the MP3's to trigger reliably, as said, but Big Daddy is even now a treat to shoot at! :) 

5) Junkyard/graveyard. I've grown used to stuff not working, but this is getting ridiculous. 


No comments:

Post a Comment