DOUT + ULN2803

[jwillans]

It works!!  The problem was that I had not uploaded the midio128 "application" just the configuration file!  My fault, although it is not clear from the project page that applications are downloadable elsewhere.

Thanks to everyone for their help - particularly JimHenry who has guided me through the whole process - I owe you a driink.

James

[John_W._Couvillon]

Gentlemen,

On the same general subject.  I am developing the  12v power supply for the ULN2803 drivers driving magnets, from a 12.6 volt transformer with traditional bridge rectifier and Cap.  The transformer has a center tap, so I am deriving the 6 volts for power supply to the midibox cores from the center tap to one leg of the transformer (ac side of course). Is it possible for some switching spikes, surges to be developed with the magnets switching on and off rapidly, that could get back into the PIC with this arrangement?  Could a shunt capacitor on the dc side of the 12vdc supply supress the spikes?  What value cap would be needed if this is possible?

Thanks,

Johnc 

[jimhenry]

I believe the 2803s include suppression diodes.  Does the MIDIIO128 allow you to configure the output to send pulses?  You don't want to leave the SAMs continuously energized.

[John_W._Couvillon]

Hello Jim,

thanks for the reply. 

Actrually,  the magnets are no the pipes themselves, not on SAMS, and are only energized as long as you hold your finger on a key.  I always thought the diode on the ULN2803 was more to prevent interaction inbetween drivers.  Any way, whats to keep the spikes from coming back through the power supply, or the common ground?

Johnc

[smashtv]

I believe the 2803s include suppression diodes.

Yes they do.  Good enough to keep pretty hefty relay coils from glitching the rest of the system/spiking lots of inductive kickback current.  Back EMF can be huge sometimes even on smaller relays, I have seen a few that would kick back 10 times the rated drive current!

I know back in the day a 2.2mF 250v cap across the terminals of a flipper 'end of stroke' switch would reduce the spark across the switch contacts to almost nothing.....considering the 50v coils, the kickback was mean enough without the cap that the switches would burn and pit after only a few cycles.

You don't want to leave the SAMs continuously energized.

Sounds like you guys need a better coil design!  Anyone done a dual-winding coil for this use yet?  Something with a mean winding (short duty cycle-for moving from rest quickly) and wimpy winding (continuous duty-for holding the mech engaged) wound together on the same coil former perhaps?  -really- simple to control if an 'end of stroke' switch is possible....  ;)

Any way, whats to keep the spikes from coming back through the power supply, or the common ground?

Johnc

Not much, But with a little more info on the coils (#turns, gauge, bar core or moving plunger down the center, impedance, drive voltage, etc.) I can help you pick some complementary cap values to clean up most of the noise.

Nice to find my wasted years in the vending biz paying off a bit.  ;D

Best

Smash

[John_W._Couvillon]

Hi Tim,

First of, the coils are not energized continuously.  They open the pallet valves that let air in to the pipes.  The coil is energized as long as you hold your finger on a key.  No problem with this.

The magnets are anywhere from  40 ohms to 150 ohms, most common 90 ohms, and are rated for 12vdc.  These are standard designs that have been around for decades, nothing new.

I have no idea what the coil conductor is, probably # 30 awg or smaller magnet wire, nor do I know what the number of turns is.  I do know that I have one core receiving mididecoder, note on - note off messages with the douts (with ULN2803) and driving pipe magnets, and all of a sudden the midimerger doesn't work, and one note is missing, like the ini file has been corrupted deleting or erasing one entry.  I'm still suspicious that the power supply has something to do with it.

Regards,

Johnc