Subject:Hysteresis Solar Engine (HYSE)
Date:Fri, 30 Apr 1999 10:02:31 -0700
From:Sean Rigter
To:beam , Chiu-Yuan Fang
Hello all,
After a few false starts I'm back on track with a new tested SE design.
This was build to prove that the hysteresis of a supervisory chip can be
increased and used for SE latching. It was tested using a MC34164 (5V)
voltage supervisor since I have no 1381 to play with. It worked fine
with a range of input currents ma to 30 ma using a simulated 9V OCR
solar cell.
The design was breadboarded and tested with 3 different motors including
the relatively common Mac ejector motor. With the 0.047F cap shown and 5
mA supply current @ 5V, the Omron motor runs for about .5 sec every 5
sec and the output shaft turns about 90 degrees for each pulse (which,
incidentally, is exactly what is needed for a walker application but
that's another design)
The way that the hysteresis is increased is worth mentioning. In this
circuit with V+ = 5.2V, the supervisor is still operating below the
threshold (!) This happens because the 34164 output is low and the 50 uA
current flowing through the 100K resistor also flows through the 10K
resistor raising the trigger threshold by 500mV. When the supervisor
trips at about 5.3V, the output changes from 500mV to 5.3V and the
current through the 100K and 10K resistors drops to zero while the motor
current "snaps" on. When the voltage across the 10K resistor drops to
zero, this also drops the lower threshold voltage by about 500mV and
provides the required. hysteresis.
When the voltage drops down to about 4.8V, the supervisor output turns
back on and the process starts again. Since the supervisor output
voltage rises to about 500mV before the threshold is reached, the FETs
used must have a Vgs (@0.1mA) > 1.0V or else the circuit will not
work!
I have not yet build a complete working "photopopper" type application
with this circuit, so you take your chances. The design should be
scalable to other chips and voltages as long as the supervisor chip is
the non inverting active low/open collector type and the Vgs of the FETs
is higher than the Vhys at the lowest supply current of interest.
Subject:74xx139 (Re: H-bridge)
Date:Fri, 30 Apr 1999 18:35:47 -0700
From:Sean Rigter
To:NSX -
CC:beam@corp.sgi.com
You asked for it: an explanation of the famous 74xx139 (Z bridge) chip
which should answer your questions. This is not much different from
other explanations out there but a bit more oriented to the current
discussion. Anyway I found it a useful exercise to try to keep the
explanation simple but complete.
Here is how the 74xx139 (Zoelen bridge) is drawn in layout format:
+---------- Reverse
Motor | +------- Forward
GND | | | | +---- Engage (connected to GND = YES)
| | | | | | | |
.-+--+--+--+--+--+--+--+-.
| 8 7 6 5 4 3 2 1 |
| [|
| 9 10 11 12 13 14 15 16 |
`-+--+--+--+--+--+--+--+-'
| | | | | | | |
| | | | | Vcc
Motor | | +------- Engage
| +---------- Forward
+------------- Reverse
74xx139 ASCII DRAWING (c) AA van Zoelen
http://www.xs4all.nl/~vsim/z-bridge.html
It consists of 2 identical independent sections called dual 1 of 4
decoders. In this example, each section controls one motor. Let's look
at section 1 (the top pins) in schematic format showing the pin numbers
and a truth table for reference.
Motor 74AC139 truth table
| |
+--+--+--+--+--+ Combination A B E Y0 Y1 Y2 Y3
| 4 5 6 7 | ----------- ----- -----------
| Y0 Y1 Y2 Y3 | 0 0 0 0 0 1 1 1
| Section 1 | 1 1 0 0 1 0 1 1
| A B E | 2 0 1 0 1 1 0 1
| 2 3 1 | 3 1 1 0 1 1 1 0
+----+--+--+---+ 4 x x 1 1 1 1 1
| | |
| | |
| | +------- Engage
| +---------- Forward
+------------- Reverse
The 3 lines marked E, A and B are inputs which can be connected to Vcc
(1) or GND (0). These 3 inputs control the 4 outputs Y0,Y1,Y2 and Y3.
The outputs are all normally 1. The A and B inputs can be connected to
Vcc (1) and GND (0) in 4 combinations (00,10,01,11). These input
combinations cause 4 combinations to happen on the Y outputs
(0111,1011,1101,1110). The E input is normally called the ENABLE input
and it must be connected to GND (0) for anything to happen to the 4
outputs. When the ENABLE pin is connected to Vcc (1), the output
combination is always 1111 regardless of the A and B input combination
(xx).
In some BEAM applications, the A, B, and E inputs are connected to
MicroCore outputs (ie A=Nv1 and B=Nv3 and E=PNC output). The motor is
connected to the Y1 and Y2 outputs and will turn only if either output
Y1 or Y2 is GND (0) while the other is Vcc (1). From the TRUTH TABLE it
is clear that when the PNC output is 1 (ie right after power is
connected), the motor cannot turn since both Y1 and Y2 are 1.
The MOTOR is STOPPED if the PNC is still Vcc (PNC=E=1) which is
combination 4 in the table
After the PNC output settles down (0) and the process is circulating
through the MicroCore then two active input and output combinations will
occur that cause the motor to turn.
REVERSE if the process is active in Nv1 (Nv1=A=0) which is combination 1
in the table.
FORWARD if the process is active in Nv3 (Nv3=B=0) which is combination 2
in the table.
From the table it is clear that no other combination of A and B can
generate active (01 or 10) outputs on Y1 and Y2 and that outputs Y1 and
Y2 can never both be 00.
The 74AC139 may also be useful in other applications for example in
controlling a HEAD (instead of the usual BiCore circuit). In that case
the output of left and right photo sensor circuits are connected to the
A and B inputs. Then think of turning FORWARD and REVERSE as turning
LEFT and RIGHT or UP and DOWN. When a head photo sensor detects a light
source input A or B will turn low (0) and the motor will start turning
in that direction until it receives equal light on both sensors and both
A and B are low (0). This is combination 1 from the table and therefore
Y1 and Y2 are both 1 and the motor is turned off saving power! This is
an improvement over a simple BiCore HEAD circuit which consumes power
even if the head is pointed directly at the light source. It should be
pointed out that the BiCore circuit outputs could also be turned off by
decoding the balanced or "locked on" condition.
The 74AC139 or equal can be used by itself to drive small motors. Two or
more chips can be stacked vertically with corresponding pins soldered
together. The Y1 and Y2 outputs can also be connected to a current
amplifier such as an H-bridge. Since the illegal output combination 00
cannot be generated by the 1 of 4 decoder, the Y1 and Y2 can be safely
connected to an H(bomb)Bridge without a possibility of generating a
mushroom cloud.
This information holds equally for section 2 (pins 9 to 15) of the chip.
These few BEAM examples are just the tip of the iceberg of many other
applications for which the 74xx139 Dual 1 of 4 decoder can be used.
BEAM, MicroCore and BiCore are (c) Mark Tilden
Z bridge is (c) AA van Zoelen
wilf
NSX - wrote:
>
> Hie all....i am wondering out where should i connect the pin labeled forward
> and reverse in the chip 74xx139 in beam tek's motor driver website.....more
> over the are forward and reverse on each side of the chip on on the left and
> the other on the right...any hints?and why does it have two engage/enable
> pins....wilf i know u know this one...care to help...and about the
> chat///mail me ASAP
Subject:step by step FLED SE analysis ( was RE: F#$@ beam)
Date:Wed, 5 May 1999 18:52:22 -0700
From:Wilf Rigter
To:"'Bumper314@aol.com'" , beam@corp.sgi.com
Hi Steve,
perhaps you can check one part at a time as you assemble the whole SE and
report what you find.
On the attached schematic, what values for C1 and R1, what motor and solar
cell? : so we can compare notes.
Here are some suggestions for checking each part/step along the way.
1. check the (illuminated) solar cell output (unconnected): 3-5V?
2. add the cap, charge and measure the voltage: 3-5V?
3. touch FLED across the cap: does it flash?
4. charge cap and touch motor across cap: does it turn?
5. connect 2N3904 and the motor across cap, charge cap, then touch a 330 ohm
resistor between 2N3904 base and +V : motor turns/moves?
6. add the 2N3906 and resistor, then touch a 1K resistor between the 2N3906
base and 0V : motor turns/moves?
7. add FLED, if circuit not working: touch 1K resistor between between the
2N3906 base and 0V : motor turns/moves?
let us know where you get stuck.
(try to) enjoy
Wilf Rigter mailto:wilf.rigter@powertech.bc.ca
tel: (604)590-7493
fax: (604)590-3411
> -----Original Message-----
> From: Bumper314@aol.com [SMTP:Bumper314@aol.com]
> Sent: Wednesday, May 05, 1999 5:02 PM
> To: swilliam@cadvision.com; beam@corp.sgi.com
> Subject: Re: F#$@ beam
>
> In a message dated 5/5/99 4:58:02 PM Pacific Daylight Time,
> swilliam@cadvision.com writes:
> > maybe it has to do with that little problem that seems to come up a lot,
> > just having enough voltage to keep the latch on, so it can not
> recharge, or
> > something like that, try your SE's under different light conditions.
> done that too, I have tried 55W and 65W incondesent, i tried a 40w halogen
> lamp, a halogen 4 D battery mag light, 75W halogen flood light, and of
> course, the sun...
> STeve
Subject:RE: Bicore head targeting?
Date:Wed, 5 May 1999 20:55:30 -0700
From:Wilf Rigter
To:"'Dennison'" , "John A. deVries II" ,
Chiu-Yuan Fang , Dave Hrynkiw ,
"Beam List (E-mail)"
As discussed at the chat today here is a (new?) idea.
With a phototropic BiCore head mounted on a uCore walker, the head should be
free to turn towards a light target and send an error signal to influence
the walker uCore to start turning toward the target. From some
experimenting, I found that integrating both BiCore phase signals produces
two average Vdc error signals proportional to each phase duty cycle. Since
the duty cycle of each phase is complementary to the other phase, this
produces two complementary error signals which can be applied to the two
corresponding uCore bias points to lengthen the process in one Nv and
shorten it in the other Nv, in effect causing the walker to turn. (ie <
50% = turn left, 50%= straight ahead, >50%=turn right) Given a relatively
high oscillation frequency for the Head BiCore, it is simply a matter of
connecting both BiCore outputs through 2 suitable resistors to the uCore
bias points using the uCore caps themselves to integrate the average DC
component of the BiCore outputs.
The problem is that the head circuit stops sending out an error signal (50%
duty cycle) when fully turned and locked on the light source. So once the
head is turned into position: no more error signal and no more influence.
Solution ? add a centering spring(s) between the head and the walker,
tensioning the head to face to the front of the walker. Now when the head
turns, the BiCore has to "work" to keep it turned towards the light and
therefore keeps sending an error signal to the uCore until the head and
walker both face straight towards the light. 8^)
What do you think?
Wilf Rigter mailto:wilf.rigter@powertech.bc.ca
tel: (604)590-7493
fax: (604)590-3411
> -----Original Message-----
> From: Dennison [SMTP:dennlill@buffnet.net]
> Sent: Tuesday, April 27, 1999 12:01 PM
> To: John A. deVries II; Sean Rigter; Wilf Rigter; van Zoelen, Bram
> SSI-TSEA-352; Chiu-Yuan Fang; Dave Hrynkiw; Beam List (E-mail)
> Subject: Bicore head targeting?
>
> I found out from dave that the controll system for his and Mark T's Heads
> are not Bicores. They use some other "poppernets" or something like that
> to control things. So that brings me back to bicores, how does one tell if
> the bicore has 'targeted' something? You should be able to compare pulse
> durrations, when the pulse durrations are equal in length you know that
> the head has 'locked on' and is no longer in motion. But how to compare
> these pulse's is the question. My first Idea was to connect both outputs
> in some fashion to the +v or gnd side of a sizable cap. The idea goes
> that, while one output charges the cap, the other output discarges. When
> the pulse durations are equal, then in theory the average net change in
> voltage across the cap should be zero. That would mean that at any other
> point, when the avereage net change was a positive value, the head was
> still in motion. I know it's confusing, but think about what would happen
> if the cap was being charged and discharged.
>
> Dennison
Subject:RE: LC circuits
Date:Wed, 5 May 1999 22:41:46 -0700
From:Wilf Rigter
To:"'afarley@sas.upenn.edu'"
CC:"Beam List (E-mail)"
Hello Alex,
No, not stupid at all ! Here is a fine example of an electromagnetic Beam
circuit. This Relay Core has a definite L 'ement" of retro about. The
triangles are meant to Show the similarity with the MicroCore. Relays are of
course natural Schmitt triggers. This low tech design could have been build
a century ago! Who knows, if someone had invented it then perhaps today we
would be driving walkers instead of cars.
enjoy
Wilf Rigter mailto:wilf.rigter@powertech.bc.ca
tel: (604)590-7493
fax: (604)590-3411
> -----Original Message-----
> From: afarley@sas.upenn.edu [SMTP:afarley@sas.upenn.edu]
> Sent: Wednesday, May 05, 1999 9:49 AM
> To: beam@corp.sgi.com
> Subject: LC circuits
>
> This may seem like a stupid question (it may just be so obvious
> that it does not immediately occur to me why not), but why aren't LC
> (inductor-capacitor) circuits used in any of the higher level circuits?
> Can't LC's be used for central pattern generation? Or is there a reason
> why inductors do not mesh well with the desired outputs? Are the outputs
> too regular (if they are why not have some type of LRC circuit)?
> Just wondering.
> Alex Farley
enjoy
Wilf Rigter mailto:wilf.rigter@powertech.bc.ca
tel: (604)590-7493
fax: (604)590-3411
enjoy
wilf