With everything assembled, I have an issue. I set up a demo program to just turn on a channel of my high side driver and one of the SCRs. The motor begins to turn. After a delay of 1 second the SCR gate is brought back low. This allows for the second transition of the button which rides the D shaft to pass without shutting off. The motor completes its cycle and stops when the first transition of the switch occurs. The high side driver is then shut off. My problem is that about 1/4 of the time the motor stops when the SCR gate is brought low. Does anyone have an explanation for this? I have a reverse diode across the motor as should be and placed a small ceramic cap across the motor. This did not seem to help much. I then placed a .1u electrolytic across and it tested fine for about 10 cycles but I can't just assume this is the solution. (And I don't want to solder in 80 large caps) I am driving the SCRs with 74hc595 shift registers so the gate is brought to ground. I thought maybe the noisy motor could be pulsing the current low enough to shut off but it only occurs when the gate is brought low, never after. The motor draws about 120mA and the holding current is 5mA. Maybe bringing the shift registers to tristate would fix it but then I'm worried about stray current triggering the SCRs plus my pcb is not wired to do so.
Also I have 4.7k gate resistors as the trigger current is only 200uA. The output to these is 3.3V. I don't think the resistors are too big as it has no trouble turning on.
The motors currently have an internal reverse polarity protection diode but not a reverse parallel diode so it looks like I will have to solder in 80 diodes already but adding 80 caps is more work and more money. Is there an alternate solution. I'm not even sure if the caps are the solution without hours of testing.
Summation: What would cause a SCR to shut off when the gate is brought low other than current falling below holding?
Thanks
EDIT: I have three shift registers to control the 10 SCRs and 8 channel high side driver. They are cascaded but I do make use of the latch clock so the bits are not shifted through at the outputs. Is is possible that, regardless of this, that the shift register is toggling the high side driver even though the bits have not changed? ... I need to get to an oscilloscope.
EDIT 2: Sorry I'll stop after this. I placed a 10uF cap between the motor positive (high side driver output) and ground instead of across the motor so therefor I would only need 8 for the whole rig and it still stopped. I thought for sure that would fix it.