Progress on Trackbot tachometery?

[George Rudolph]

Bruce,

Any idea yet when a firmware update that may include tachometry will be released?   I saw the email from February where you
talked about working on Trackbot again.
I too have been working on other things and now am back to
working with a student on my project that uses Trackbots.

George Rudolph

[Bruce Boyes]

We've taken a hit with the economic downturn, too, mostly last year.
And some of our customers just made bad decisions and that has
affected us (e.g., royalty income is therefore not what it was
forecast). So this has put TrackBot progress many months behind where
we wanted it to be.

On the other hand, since the start of 2009 we actually see a modest
uptick in orders of our mainline embedded products, and some TrackBot sales.

We've had some setbacks in TrackBot related things, such as:
1) incoming battery quality has led to our rejecting a huge
percentage of batteries. We have purchased some special battery test
equipment and found a new vendor. More on this soon when that work is complete.
2) We have encountered issues of repeatability with the tachometer
hardware and/or firmware. The bottom line is that some TrackBot
motors return reasonable data and some don't. Testing is tedious and
time-consuming which doesn't help. We don't have a solution so
decided to set this aside for the moment and work on beaconing which
will add so many power features (assuming it works as we intended).
We'll get back to this approx in June.
3) Bluetooth radios to link multiple bots to one PC, with an
autonomous control program per bot... pretty much a dismal failure on
the PC due to the way Bluetooth and virtual COM ports work. Some
layer under our application code opens and closes the serial
connection multiple times, when the application just wants a serial
stream opened once and left open. This breaks the application as you
might imagine. Very frustrating. Works fine on the MAC. We are
looking at other radio options. This would create sort of a
"virtualized" brain for TrackBots, at low cost (assuming the radios
are affordable).

We expect to be back on TrackBot code as a major priority in late May and June.

In the meantime there is some significant progress on the Greenfoot
simulation of beaconing in the SOB (Swarm of Brian) project which
will be a BOF at JavaOne in about 30 days from now:
https://swarm-of-brian.dev.java.net/

best regards

Bruce

------- WWW.SYSTRONIX.COM ----------
Real embedded Java and much more
+1-801-534-1017 Salt Lake City, USA

[grudolph]

>
> We've had some setbacks in TrackBot related things, such as:
> 1) incoming battery quality has led to our rejecting a huge
> percentage of batteries. We have purchased some special battery test
> equipment and found a new vendor. More on this soon when that work is complete.
> 2) We have encountered issues of repeatability with the tachometer
> hardware and/or firmware. The bottom line is that some TrackBot
> motors return reasonable data and some don't. Testing is tedious and
> time-consuming which doesn't help. We don't have a solution so

Thanks for the update.

I know you mentioned at one point using Weiner filters or Kalman
filters
for signal correction. Is that something you would like help on, or
are the motors just bad enough that even a filter is not worth
pursuing?

> decided to set this aside for the moment and work on beaconing which
> will add so many power features (assuming it works as we intended).
> We'll get back to this approx in June.

> 3) Bluetooth radios to link multiple bots to one PC, with an
> autonomous control program per bot... pretty much a dismal failure on
> the PC due to the way Bluetooth and virtual COM ports work. Some
> layer under our application code opens and closes the serial
> connection multiple times, when the application just wants a serial
> stream opened once and left open. This breaks the application as you
> might imagine. Very frustrating. Works fine on the MAC. We are
> looking at other radio options. This would create sort of a
> "virtualized" brain for TrackBots, at low cost (assuming the radios
> are affordable).

I still have problems with Bluetooth connections on Windows Machines
with Lego robots too. ;( I think the Windows drivers are buggy.

>
> We expect to be back on TrackBot code as a major priority in late May and June.
>
> In the meantime there is some significant progress on the Greenfoot
> simulation of beaconing in the SOB (Swarm of Brian) project which
> will be a BOF at JavaOne in about 30 days from now:https://swarm-of-brian.dev.java.net/

I'll look at it at some point.

[grudolph]

>
> We've had some setbacks in TrackBot related things, such as:
> 1) incoming battery quality has led to our rejecting a huge
> percentage of batteries. We have purchased some special battery test
> equipment and found a new vendor. More on this soon when that work is complete.
> 2) We have encountered issues of repeatability with the tachometer
> hardware and/or firmware. The bottom line is that some TrackBot
> motors return reasonable data and some don't. Testing is tedious and
> time-consuming which doesn't help. We don't have a solution so

Thanks for the update.

I know you mentioned at one point using Weiner filters or Kalman
filters
for signal correction. Is that something you would like help on, or
are the motors just bad enough that even a filter is not worth
pursuing?

> decided to set this aside for the moment and work on beaconing which
> will add so many power features (assuming it works as we intended).
> We'll get back to this approx in June.

> 3) Bluetooth radios to link multiple bots to one PC, with an
> autonomous control program per bot... pretty much a dismal failure on
> the PC due to the way Bluetooth and virtual COM ports work. Some
> layer under our application code opens and closes the serial
> connection multiple times, when the application just wants a serial
> stream opened once and left open. This breaks the application as you
> might imagine. Very frustrating. Works fine on the MAC. We are
> looking at other radio options. This would create sort of a
> "virtualized" brain for TrackBots, at low cost (assuming the radios
> are affordable).

I still have problems with Bluetooth connections on Windows Machines
with Lego robots too. ;( I think the Windows drivers are buggy.

>

> We expect to be back on TrackBot code as a major priority in late May and June.
>
> In the meantime there is some significant progress on the Greenfoot
> simulation of beaconing in the SOB (Swarm of Brian) project which
> will be a BOF at JavaOne in about 30 days from now:https://swarm-of-brian.dev.java.net/

I'll look at it at some point.

>

[Bruce Boyes]

At 08:56 05/05/2009, you wrote:

> >
> > We've had some setbacks in TrackBot related things, such as:
> > 1) incoming battery quality has led to our rejecting a huge
> > percentage of batteries. We have purchased some special battery test
> > equipment and found a new vendor. More on this soon when that
> work is complete.
> > 2) We have encountered issues of repeatability with the tachometer
> > hardware and/or firmware. The bottom line is that some TrackBot
> > motors return reasonable data and some don't. Testing is tedious and
> > time-consuming which doesn't help. We don't have a solution so
>
>
>Thanks for the update.
>
>I know you mentioned at one point using Weiner filters or Kalman
>filters
>for signal correction. Is that something you would like help on, or
>are the motors just bad enough that even a filter is not worth
>pursuing?

We are looking at Kalman filters, but they don't do well if the
sensor is noisy - they do well with a clean sensor but noisy data. In
our case it may be the sensor (the motor commutator) which is noisy.
Also the tread itself is not uniform in friction as the drive wheel
rotates, so it adds a periodic signature to the speed of the motor.
It's a more complicated and messy problem to solve than we first realized.

Thanks for the offer. If it makes sense to somehow make this an open
source project we will do so.