BEAM Bots

[Aaron]

I ordered some parts for some BEAM bots.
I'd like to try and collect some references for parts suppliers.
I got some solar cells, and 1381 voltage triggers from solarbotics.com.
I'd like to support these guys. They seem earnest... but $1.75 for
a very simple IC that looks like a transistor seems steep.
These triggers seem to be at the core of most solarengine designs.
(A Solarengine stores energy from a solar cell, usually in a capacitor, then releases it in a burst. It allows small cells to pulse-drive a motor)

I often get some electronic junk at Electronic Goldmine, www.goldmine-elec.com.
They have a limited selection of common components, like transistors, capacitors, and solar cells.

Another favorite was Marlin Jones, but they seem to be more about more integrated electronics. Not too many simple, base components.

Of note... one of the big guys of the BEAM bot community is from Los Alamos.

I intend to build a timmet/symmet or two. They are supposed to scutter about in pulses, and use mechanical means (tipping over) to avoid obsticals. ingeneous.

If I want to do more, the next would be a "popper"?. Probobally something with two solar engines, plus light-seeking circuitry. Something that might chase the sun
coming through a window. I'm looking for work on one of these with some collision-avoidal behavior.

[Aaron]

Here's a supplier of fairly priced FLEXIBLE solar cells:

http://www.siliconsolar.com/flexible-solar-panels-3v-p-16159.php

I think you can cut them up for differeing power/voltage/current characteristics
I'm also seeing some sellers on eBay...

[Aaron]

I got some of those flexible solar cells, and started working on some Soar Engines.
I don't have any 1381's (low-voltage checkers) but I have been playing with several varieties
of diode based SE's.

I have not been able to get any of the zener or regular diode versions to oscillate reliably.
The Flashing LED version will work, but often flashes visibly several times before the engine
latches. Like 10-12 times in low light. It only flashes visibly about 1 times out of 2 in direct sun
(faster charging, higher power). This wastes a lot of solar power. The energy to flash a LED
is significant when powering from a small cell.

I'm no where near as clever or experienced as many of these solar engine designers, but
I have some ideas.

These low voltage side of the motor only gets down to about 1v (not ground) when these guys
latch.

? Might a Solid State Relay have less loss across the junction ?

Do we tend to see transistors because relays need high "base" voltage... like 3V (not .7)?

Even if so... could we get some sort of histerisis (or time delay) that would keep
the relay latched for a bit... until the capacitor got quite low?

How about a hysterisis inverting buffer (shottkey?)?
I guess that might be tricky.
You might need a secondary power circuit to keep Vcc high for the buffer.
Perhaps another capacitor, but with a very slow discharge... just enough to
keep the buffer powered.

It seems that the big secret of these circuits is hysterisis in the latch.
So... starting out with a device that is mosty intended JUST to provide hysterisis
seems like it might be a good start... as opposed to a voltage switcher which is not
intended to have large hysterisis.

[Aaron]

I also built my understanding of this "Fred" circuit.
It seems to be the same as the base FLED design, but with the 2.2K pull-"up"?
resistor replaced by two resistors and a cap.

I don't have a small cap like that, so I used a bigger one (10uf)... and it didn't work.
I also wonder of the 3.3K and 33K resstors should be flipped.

I can see where some clever use of a cap could/should help this latch on
the first (significant) flash, instead of wasting several small flashes before latching.

If the FLED designs can be made to be as robust as the 1381, I'd like them.
It does help to be able watch the flashes in debugging.

Here is the original Fred circuit, and my adaptation, which you can see is EXACTLY
the same as the base FLED design, but with the 2.2k relaced by the little triangle of cap
and two resistors.

[Aaron]

I took off work early today, and picked up a 555 at Radio Shack on the way home.
I built this SE:

It works great! I dunno why they are interested in so many designs with higher part counts
(hard to free-form) and/or odd chips like the 1381, Maxim...

It seems that you can control the trigger voltage(s) with those two diodes.
It now fires at about 3.2v, and drains down to about 1.5v ish.
I bet we can lower this with a lower reference voltage. Unfortunately,
diodes all drop about .75v. I'm not as hip with this... but I think we can make
more like a .3v reference by tieing some leads of a transistor together (help?).
.75 was not enough voltage for the Radio Shack 555 to fire. 1.5v (3.2v trigger)
is a bit high for my 3.5v pannel. I'd like something in-between.

I have ordered more 555's from DigiKey. They seem to list some (more expensive... like $1)
CMOS 555's that should run down to 1v! They might work on a single diode voltage reference.
I just wish I had ordered some 556's too. (Dual 555 chips). I could use those for a popper.

It was odd, but even though they show photosensitive versions like these:

This one uses two caps and some diodes to balance the motors more, so you can have a larger, longer pop:

the first circuit does not work without the resistor.
I was thinking that some power leaks from the 555 biasing those voltage reference diodes.
So... I am a bit concerned about how/if these will work with the photodiode (no tie-up to Vcc).

[Aaron]

In reviewing the last photopopper circuit, it seems that the two caps are functionally in parallel. You could do it with just one big cap over the photocell like usual.

[Aaron]

On this last circuit, I'm thinking that one transistor sinks current for both motors.
The fact that the non-popped side goes through a diode on the way lowers
it's motor voltage by .75v w.r.t. the side that did pop. It would be that voltage difference that steers the bot.

The next question would be... how can I add a bump-sensor and back-up circuit to this? Some sort of DPDT relay?

[Aaron]

Here's my idea for a simple SE with a schmidt trigger.
This one should trigger around 4.5-5v, and continue to dump down to about
2.2v. One diode might change the trigger to about 2.2v, and drain down
to 1.2v... if your CMOS will operate at that low voltage.

[Aaron]

I have a SE design using 555's that I'm happy with.
It uses a single diode bias to set the "hi" trigger voltage.
At uA currents, this tends to be near .5v. This makes the 555 discharge until

Vlo-level ~= 2*thresh = 1-1.2v

This is a pretty good level. I think that lower discharges do not help much.
It is not enough voltage to drive the load.

Use small photocells, capable of producing a few micro-amps to bias the diodes for the trigger (high) voltage reference. I chose two. This has several excellent characteristics.
The brighter the light, the higher the hi trip voltage. Not only useful for photovores, but also allows the circuit to have higher power discharges under bright conditions. I find that parasitic losses shut down the SE under low light. Reducing the firing voltage in low light allows these circuits to continue firing, even if at lower power.

For a lower voltage photocell array, use just one photodetector (.2-.6v reference, depending on light). For a higher voltage pannel array, go up to three photodetectors (.6-1.8 v reference).

I think I might even like this better than the popular Miller engine, and it uses common parts. On DigiKey, I found 555's that operate down to 1v at 200uA, and low-power ones that work down to 2v, but at 100uA. I just ordered some "low power" 2v 556's (pair of 555's on a chip) to do BEAM bots that will follow light.

The Hi-voltage threshold would be at Vcc ~= 3* (voltage across ALL diodes)
With two photodetectors, as shown, it triggers at about 3.1 v in low light,
4.6v at higher light, which is pretty good for my 5v pannel.

Perhaps this is not a very popular SE since it needs some tweaking w.r.t. each solar array.

I am also hooking up the "thresh" and "trigger" voltage kind of backwards, in an attempt to get my desired hysterisis, as opposed to the standard 1/3 Vcc and 2/3 Vcc trigger points. So far, it seems to be OK.