Fusion brain data questions

**AtomicMonkey** · 02-14-2008, 04:50 PM

Hi,

I got my Brain few weeks ago (actually even before I managed to send in the payment, wow thanks guys I hope you got the check alright) but after some reading and testing I'm a bit puzzled as to what's going on there.

First, the AI. The highest reported value I can get out of the 2 AI bytes is 510 ( the highest non-shifted value out of byte-1 is 127 and byte-2 oscillates between 0 and 192 ). So that works out to about a half of measurable voltage, and indeed if I connect a 3.5V source I still get 510. Also, anything greater than ~3.6-3.8V all the way to 5V completely freaks the unit out and it reports near 0 (3 max on odd byte and 192 on the even one) with intermittent flashes of full on values. ~1.5V and below is pretty stable, > 2V not as much and varies input by about 10% Sources I'm using for testing are very stable so that's troubling. Any insight or help would be appreciated. I hope my Brain's not broken

On the subject of digital outputs, I'm a bit confused. Pun intended. According to the doc I've read the Keep-Alive bit needs to be switched with every message to keep it alive I guess. Judging from my unit, I would say no, not really although I'm missing some consistency here. Port 0 works with or without the KA bit set. With KA bit set it will hold until bit-0 is reset to 0 and without it turn itself off almost immediately - sounds about right to me. Strangely Ports 1-11 will not close without the KA bit set to 1. Neither momentarily not constantly. KA bit has to be set to 1 but then of course the output needs to be manually set to 0 to disengage it. That is not a huge problem but I just want to get this straight.

Also, according to the doc 2k1Toaster posted in another thread, the D/O byte's bits 2-7 hold the auto-off timer value from instant to some time next day. My trouble is that a value of 1 (which ends up as 4 actually) holds the port closed for ~10s. A 2 lasts a whole minute and I got tired of waiting for others. The max value would probably end up being a month lol. The timer is a nice feature but at 10s min and exponentially apart pretty useless. Or am I missing something, I hope I am.

Lastly, just out of curiosity, D/O ports 0, 4, 8 and 10 I think set bits 0, 1, 2 and 8 on byte 32 (the one that holds the D/I state) when closed. Any meaning behind it?

TIA
A.M.

**2k1Toaster** · 02-14-2008, 06:12 PM

Originally Posted by AtomicMonkey

I got my Brain few weeks ago (actually even before I managed to send in the payment, wow thanks guys I hope you got the check alright)

Yup all good.

Originally Posted by AtomicMonkey

First, the AI. The highest reported value I can get out of the 2 AI bytes is 510 ( the highest non-shifted value out of byte-1 is 127 and byte-2 oscillates between 0 and 192 ). So that works out to about a half of measurable voltage, and indeed if I connect a 3.5V source I still get 510. Also, anything greater than ~3.6-3.8V all the way to 5V completely freaks the unit out and it reports near 0 (3 max on odd byte and 192 on the even one) with intermittent flashes of full on values. ~1.5V and below is pretty stable, > 2V not as much and varies input by about 10% Sources I'm using for testing are very stable so that's troubling. Any insight or help would be appreciated. I hope my Brain's not broken

It sounds like you are using a custom program. I would try with the default program to rule out a programming mistake. Open up the newest version, click about 1-2 pixels from the top bar, and type in "DEBUG" while holding shift+alt+control and the debug window should pop up. Type in "open raw debug" and you should be able to see the values for analogue input on the bottom and right (textual and visual).

Also what are you using to send these signals?

Originally Posted by AtomicMonkey

On the subject of digital outputs, I'm a bit confused. Pun intended. According to the doc I've read the Keep-Alive bit needs to be switched with every message to keep it alive I guess. Judging from my unit, I would say no, not really although I'm missing some consistency here. Port 0 works with or without the KA bit set. With KA bit set it will hold until bit-0 is reset to 0 and without it turn itself off almost immediately - sounds about right to me. Strangely Ports 1-11 will not close without the KA bit set to 1. Neither momentarily not constantly. KA bit has to be set to 1 but then of course the output needs to be manually set to 0 to disengage it. That is not a huge problem but I just want to get this straight.

This is not true as far as I can tell. I would assume you have flipped the KA bit and the state bit which seems to explain that behaviour.

Originally Posted by AtomicMonkey

Also, according to the doc 2k1Toaster posted in another thread, the D/O byte's bits 2-7 hold the auto-off timer value from instant to some time next day. My trouble is that a value of 1 (which ends up as 4 actually) holds the port closed for ~10s. A 2 lasts a whole minute and I got tired of waiting for others. The max value would probably end up being a month lol. The timer is a nice feature but at 10s min and exponentially apart pretty useless. Or am I missing something, I hope I am.

It is exponential as in the car, a couple secs and then a minute is good, an for home automation, long timers are needed. This is used when the PC loses communication.

For timers such as timed outputs, that is done all PC software wise.

Originally Posted by AtomicMonkey

Lastly, just out of curiosity, D/O ports 0, 4, 8 and 10 I think set bits 0, 1, 2 and 8 on byte 32 (the one that holds the D/I state) when closed. Any meaning behind it?

Not sure what you mean. We just happened to have room in that byte to squeeze 'em in!

**AtomicMonkey** · 02-15-2008, 10:12 PM

Hi, thx for getting back to me.

It sounds like you are using a custom program. I would try with the default program to rule out a programming mistake. Open up the newest version, click about 1-2 pixels from the top bar, and type in "DEBUG" while holding shift+alt+control and the debug window should pop up. Type in "open raw debug" and you should be able to see the values for analogue input on the bottom and right (textual and visual).

Also what are you using to send these signals?

Ok, i dl'ed what appears to be the latest app on your list (2.2.1 me thinks) and the AI voltage is reported exactly the same, that is < 2.5V shows fine but >2.5V still goes to 2.5V max (2.49V something or 511[int] TBE). This is not a software issue at least not outside of the either FusionUSB.dll or the mpusbapi.dll libraries depending which USB driver's loaded. Your app gets the same bytes I do. The odd-byte never gets the 128 bit set HI no matter the voltage applied except when you put >~4V it flashes full 1023 on every now and again and mostly reports close to zero. We tried just about anything, went over the board, ensured proper contacts, grounding etc all seems well, + like I said, < 2.5V is measured accurately albeit with occasional 5-10% fluctuation above ~2V. Something is definitely not right or I am seriously missing something.

As far as signal source they're various high grade batteries 1.3V ~ 4.6 verified stable to within 0.001V at any given time and a reference type 5V power supply.

This is not true as far as I can tell. I would assume you have flipped the KA bit and the state bit which seems to explain that behaviour.

Well maybe we're talking the same thing. I flip it to 1 when I want the D/O on (active). I naturally send the KA-bit 0 when the output is to be disengaged but only then. I noticed in your debug thingy that you basically send D/O state every half second or so and indeed those KA bits constantly change despite consistent ON state. Not sure what that's all about, though if it works. I only send the D/O state when I actually want it changed for one of them. It works so I'm not complaining, was just curious about that flip-flop thing. If I set the state of say D/O-1 to 00000011 it will stay on until I send a 00000000 or 00000010 value to it. No flipping in the meantime necessary. Not on my unit anyhow.

It is exponential as in the car, a couple secs and then a minute is good, an for home automation, long timers are needed. This is used when the PC loses communication.

For timers such as timed outputs, that is done all PC software wise.

I understand, I was just curious as the doc you posted specified

time----- x-time [s]
Immediate 0 Immediate
1 second. 1 1.20s
2 seconds 4 2.07s
3 seconds 6 2.99s

... and so on and was curious if it was supposed to be true. With the x = 1 mine stays on for anywhere from 12-18 seconds (not 1.2), 2 around 2 min and 3 is close to 10 minutes, so nowhere near those lower values that were in the doc.

Anyhoo, that's not at all important. My only real concern is the voltage which seems either choke or to reach a measuring roadblock at half it's range (exactly half to be precise). BTW firmware version reported by the PIC (63 & 62) is 4.2 and the thing is hooked up to XPSp2 box.

**greenman100** · 02-15-2008, 11:22 PM

Originally Posted by AtomicMonkey

Ok, i dl'ed what appears to be the latest app on your list (2.2.1 me thinks) and the AI voltage is reported exactly the same, that is < 2.5V shows fine but >2.5V still goes to 2.5V max (2.49V something or 511[int] TBE). This is not a software issue at least not outside of the either FusionUSB.dll or the mpusbapi.dll libraries depending which USB driver's loaded. Your app gets the same bytes I do. The odd-byte never gets the 128 bit set HI no matter the voltage applied except when you put >~4V it flashes full 1023 on every now and again and mostly reports close to zero. We tried just about anything, went over the board, ensured proper contacts, grounding etc all seems well, + like I said, < 2.5V is measured accurately albeit with occasional 5-10% fluctuation above ~2V. Something is definitely not right or I am seriously missing something.

Feed port 0 around 5vDC (not greater than 5.05vDC) and post a screenshot of the debug window. It should look like this:

**AtomicMonkey** · 02-16-2008, 01:40 PM

4.97V on A0x0

The port A0x0 flashes full on occasionally but it's was quick and infrequent for me to cap the screen.

Few more caps in case they mean anything:

2.52V on A0x0

2.52V on A0x0 (with A0x1 input pin grounded)

1.46V on A0x1 (with all other AI input pins grounded)

EDIT: one more thing, I forgot which may, may not be normal. With the 12VDC power removed from the Brain, and USB disconnected but 4.97V source still connected to AI0x0 the ext power LED on the board is still lit. Unhooked the source, went away, re-attached source, LED comes back on.

**greenman100** · 02-17-2008, 06:13 PM

It's normal for the LED to stay on. You're actually powering the PIC through the ADC in the PIC. It's a quirk that Microchip warns about, but to design around it you have to put buffers in, which we will do in the future.

However, it sounds like you have either a bad firmware or board. We can ship you a new brain, for free, if you agree to send back the old one. We apologize for the inconvenience. Please contact us at our email address, fusioncontrolcentre AT gmail DOT com