False triggering with Automat

[battlesmk]

Hi All,

I am getting some false triggering on the Automat. I have the Automat being triggered by Ableton Live. Ableton Live is outputed from my computer via USB and then is connected to four Polyend Percs, daisy chained together with 5 Pin MIDI DIN cables. At the end of this chain I have another 5 Pin MIDI DIN cable connected directly to the Automat MIDI in. This setup works perfectly and I’m able to trigger 16 separate solenoids to play very complex drum patterns on my acoustic drum set with pre-programmed MIDI signals.

I also have a handful of drum triggers (mostly ddrum acoustic Pro’s but one Yamaha DT-20) on my BD, Toms, HH (closed) and snare that I route through a DDRUM DDT interface into Ableton via USB to record my drum playing via MIDI signals and then loop these signals back to the solenoids so I basically am able to loop live drum patterns on my acoustic set on the fly. This also works perfectly when I am only using the Polyend Percs, with the Automat completely disconnected.

However, the Automat does not appear to play well in conjunction with the DDRM DDTI, as when I plug the USB cable from the DDTI into my computer while the USB output to the Polyend Perc is also plugged in to the same computer, I get some false triggering on certain outputs on the Automat. It’s clear the voltage causing the false triggering is constant and not intermittent since the solenoids that are false triggering have beaters attached to them; the beaters trigger, and then never release until I unplug the DDTI USB cable.

It is a strange issue that I thought could be caused by ground loops. However, I’ve controlled for ground loops in a number of ways, such as connecting all devices mentioned above into the same outlet through a 500 watt isolation transformer, and making sure no additional connections are possible, but the problem persists. The DDTI is ungrounded (2 prong), but the Polyend Perc and Automat are grounded.

Any help would be greatly appreciated. I’d like to add additional triggers to loop more complex patterns and utilize the Automat in this way for live performances, but I dont want rogue voltages to ruin any of this equipment.

Thanks in advance.

[johannes]

@battlesmk sent you an Email… will try to figure this out with you. J

[VincentP]

Please see the attached photo.

Note that the automat is powered on, but has no midi connection, and that solenoids 4 and 5 are energized, and stay that way without any external midi input.

Note also that the led’s for #4 and #5 are not on.

This is clearly not a software issue.

If I re-learn the automat, the solenoid or solenoids that stay energized may, or may not, change.

It does not seem to matter if I learn the automat from a usb cable or 5-pin midi din connector.

I had noticed that when using a program like Dadachron to drive the automat, that one solenoid might stay energized when I stopped the program running, and that the solenoid would stay that way until the automat was powered off.

The problem is intermittent, but frequent, and does not seem to do the same thing every time. That is why I got a new set of solenoids and eliminated the software to try to narrow down the problem.

Note that I am making an assumption that solenoids do not care about polarity, so the center pin on the connector might go to a red or a black wire on the solenoid. Since the solenoid is just an electromagnet, I assume that it is not at issue here. If I’m in wrong about that, please correct me.

If it matters, these are small 12v solenoids powered by a one amp 9v power supply. I get the same kinds of issues when using 5v solenoids with the same power supply.

Putting 9v in a solenoid rated at 5v for an instant isn’t a problem, but when the solenoid stays on, the magnet insulation starts to break down. That’s why I waited until I got 12v solenoids to continue troubleshooting this problem.

[johannes]

So as this is clearly a DIY setup - not using our tested components i have two guesses.

1. The solenoids are the problem. The automat has ESD & self protected N Channel Smart MosFETs to drive the outputs. If they get to hot, or you make shortcut they STAY ON until you take the power away and restart without the problematic part attached. If you work with some random solenoids you need to measure the resistance of the coil and calculate the current draw. The MosFETs are rated for 1.3A continuous draw.
2. If the MIDI signal gets lost on a channel / output which was on - it will stay on. The firmware isn’t counting to whatever number to shutdown the output. It’s quiet simple. If there is no Note OFF - it will stay on forever.

In your case my guess is that the solenoids are the problem. If you overdrive 5V solenoids at 24V that can be too much. Could as well be that your selfmade cables make shorts internally. It’s a guessing game. Can’t tell you more here. What is the PSU voltage?

[VincentP]

First, thanks for your quick reply, Johannes.

Second: I think it is my problem.

The problem seems to be the plastic screw-on shield on the cable connectors going to the solenoids. As Johannes noted, everything in my setup besides the automat is DIY. I used 20 gage solid core copper wire, which is stiff, it does not drape easily. These connectors are soldered to the copper wire and use a thread on plastic insulating cover. (Picture attached)

When the plastic sleeve is screwed on tight, it apparently keeps the connector from seating properly in the automat, and the false triggers happen. If I loosen the sleeve, which lets the connector sit deeper in the automat, I don’t see the triggering problem.

I will try reshaping the plastic to see if that works. Eliminating them entirely is a bad idea, but there may be some other option.

There are more expensive, and bulkier, solder-less screw type connectors that have a longer metal barrel and do not seem to have the same problem.

For the record: I am using a 9v 1a power supply. I was using 5v solenoids, because that is what I had on hand. The recent attempt was with 12v solenoids, but still using the 9v 1a power supply.

Thanks,

Vincent