Problem with Teabox expander board - urgent help advice needed.

[david.r...@ntlworld.com]

I've just recently started using the Teabox extender in earnest for a
project that's starting in a couple of weeks.

I've got 4 long range IR sensors on inputs 1-4 (xlr), and 2 close
range IR and 2 floor pressure pads (infusion) on inputs 5-8 (xlrs on
the extnder).

What's happening is that the data stream freezes from time to time -
all the input values drop to 0 and stay there for 2-3 seconds. (As
I've inverted most of these streams, this means that all values go to
full, and it get's very noisy). It's happened 3 times while I'm typing
this email!
This isn't just the usual occasional blips of "noise" from individual
sensors, this is all 8 of them at once.

I'm pretty sure that this is only happening with the expander plugged
in (though I _will try with it disconnected and see what happens).

Is this a faulty expander? Or have other people seen this happening?

thanks

David

[Timothy Place]

Hi David,

If all of the sensors are dropping out like that, then it suggests to
me that something is shorting out, or the sensors plugged into the
system are drawing too much power. I hope it isn't a short in the
expander, but you should be able to find out by trying with your
sensors plugged straight into the Teabox.

best,
Tim

Cycling '74 | http://cycling74.com
Electrotap | http://electrotap.com

[david.r...@ntlworld.com]

Well, first I tried removing the expander, and connecting the close
range IRs and pressure pads to inputs 1-4. All ok.

Then I added the expander - no sensors. Still seems ok.

Now I've connected the long range IRs to the expander. So far so good,
although the data from the long range IRs seems a _lot_ more jittery
than it was when they were connected to ins 1-4.

I'll leave it running a while and see what happens.

David

[david.r...@ntlworld.com]

Hmm.

So far so good on the drop out problem, but it seems that input 2 on
the expander (input 6 to the Teabox) is exceptionally noisey. I've
tried swapping the sensros around, and it's always in 6 that’s jumping
around all over the place. I've even tried connecting a short range IR
to that input - and it's still noisy (quick jumps of up to 0.2 with
nothing in front of sensor, as opposed to sensor connected to in 5,
which isn't jumping at all).

So there may be something wrong ith the board? If so, what can we do
about it? I need this working for the end of next week!
(We can take this off-group if you prefer - david.resonant (at)
ntlworld.com)

David

[david.r...@ntlworld.com]

ok, apart from the noisy input on 6, the freezing problem _isn't the
expander. I just used some xlr-3 pin adaptor leads to get the short
range IR and floorpads into inputs 5-8. Long range IRs into 1-4. Was
on the phone to someone when the inputs all dropped to zero again for
several seconds.

... later ...

to make sure that it wasn't my patch causing problems, I modified the
Teabox help patch to count instances of the input dropping to zero,
and had that as the only patch open. 4 long range IR on 1-4 (xlr
ins) , short range IR on 5,6, pressure sensors on 7,8. 3-pin ins)

I'm still getting intermittent drop outs! So i guess it"s either
something with the OS (could something be grabbing the processor??) or
the Teabox itself. I think it's only happening when I have 8 sensors
attached; or maybe it's the long range IRs drawing a lot of power?
(But it's an intermittent problem)

??????

David

[Timothy Place]

Jesse is the real expert for hardware -- I just try to fake it. But
one question I'm wondering about is if this only happens when you are
using the phone, or when the phone is near the system, or something
like that.

best,
Tim

Cycling '74 | http://cycling74.com
Electrotap | http://electrotap.com

[Jesse Allison]

This really sounds like a power drawing issue to me.

The Teabox can source 250mA for the sensors. The IR sensors are
spec'd to draw between 20-50mA a piece. I believe the way they work
is by cycling off measurements every 30ms or so - which means if they
line up right, you would draw more power than the Teabox could
support. We used to have bad problems with this with the 1.5m IR
sensors. The longer range ones don't seem to have the problem as much.

One thing to check. When the dropout occurs, does the little blinking
light inside the Teabox flip out and then come back into tempo? This
would indicate that the microcontroller has browned out due to the
power draw, then come back on line.

The solution is to add another power source to the mix. Try the setup
with only 3 sensors or so and see if the dropouts still happen.

If they stop, We can chat about adding another power regulator into
the mix.

-Jesse Allison

.

[david.r...@ntlworld.com]

> One thing to check.  When the dropout occurs, does the little blinking  
> light inside the Teabox flip out and then come back into tempo?  

I'll try and check that, but it may al be over too quickly for me to
see!

> The solution is to add another power source to the mix.  Try the setup  
> with only 3 sensors or so and see if the dropouts still happen.

It certainly doesn't happen with 4 sensors. I'll try with 6 and see
what happens.

[david.r...@ntlworld.com]

Jesse -

I left the Teabox running overnight with 8 sensors in. There were
"brown-outs" at intervals between 3-6 minutes, and occasionally up to
20 minutes.There were 80 occurrences between 11.30 last night and 7.30
this morning. Each brown-out lasted either 2800ms or 2700ms (+/- 2ms,
which might just be timing errors in my patch)

There didn't seem to be as many when I briefly tried with 7 sensors
yesterday evening, but I'll leave the patch running with 7 while I'm
out this morning and report back later.

David

[Jesse Allison]

Hi David,

This is almost definitely drawing more current than the Teabox can
supply. To fix this, you will need to supply more power. It's a
relatively simple circuit - as long as you can get the parts - a
regulator a couple capacitors and a power supply (9-12v typically).

Here are 2 posts which discuss making a simple regulated power circuit:

The generic overview:
http://idiarts.net/hci/show/1

and a specific 5v device:
http://idiarts.net/hci/show/18

The principle is like this. . .

The Teabox has a 5v power source that can supply 250mA to all
sensors. Since the power draw at times is above this range, you need
to supply the capability of drawing more current.

You can take another power source that has the same potential - 5v -
and connect the 5v and Ground to the Teabox's 5v and Ground. The
potential is the same, but if the circuit of sensors draws more than
250mA, the current can come from the other supply. In practice, they
both will pass current along proportionally.

The one issue you can run into with this method, is if your new power
supply is noisier than your first one. All of that noise will be
transfered into the sensor data.

This circuit is what we have on our Powered Expander - it uses the
same regulator as the Teabox and adds 500mA of current capacity. http://shop.electrotap.com/products/expander


One last thought. . .

Since the brownouts are intermittent, it seems that the cyclical power
draw from the IR sensors has to line up to go over the maximum. A
capacitor is used with regulation circuits to store a little bit of
energy and when there is a large draw, it dumps some current back into
the circuit which smoothes out the power line. If the time when the
Teabox is being overdrawn is very small. . . you may be able to put a
number of capacitors between 5v and Ground and have them supply the
momentary power that you need to keep it from browning out.

I'd try this first just in case this is all you need to do. (Don't
bet on it, but it just might work.) Just take a smattering of
capacitor sizes - 0.1mF to 10mF (you could go beyond this range if
that was all you had). Then connect one side to 5v and the other side
to Ground. Go ahead and connect a number of them to see if you can
get enough energy stored up to stop the brownouts. If it is still
browning out on occasion, you'll have to add a proper regulator as
described above.


Let me know if either of these work out.

-Jesse Allison

.

[david.r...@ntlworld.com]

Hi Jesse,

Did you get my direct reply? How quickly could you get a powered
expander to me - I don't want to start messing with the teabox at this
point!

David

[david.r...@ntlworld.com]

Oops. I see that you did!