You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
299 lines
14 KiB
299 lines
14 KiB
|
|
\documentclass[11pt]{article}
|
|
%Gummi|065|=)
|
|
\usepackage{graphicx}
|
|
\usepackage{caption}
|
|
\usepackage{xcolor}
|
|
\usepackage[vcentering,dvips]{geometry}
|
|
\geometry{papersize={6in,9in},total={4.5in,6.8in}}
|
|
%\title{\textbf{Door Alarm}}
|
|
\author{Steak Electronics}
|
|
\date{}
|
|
\begin{document}
|
|
|
|
%\maketitle
|
|
|
|
%\tableofcontents
|
|
\textcolor{green!60!blue!70}{
|
|
\section{Attiny Solar Energy Harvest Tests}}
|
|
|
|
I have the following:
|
|
\begin{itemize}
|
|
\item Solar panels
|
|
\item Attiny 10
|
|
\end{itemize}
|
|
To this list, I will add a supercap, and an energy harvesting IC. The goal being to load the super cap during the day, and to run 24/7. I will need an exceptionally low power micro. The super cap will need to be about 3.3V or 5V.
|
|
\textcolor{green!60!blue!70}{
|
|
\subsection{Micro Considerations}}
|
|
The Arduino Atmega328P is not an option. I'm looking to have a current draw of only 1mA max, (ideally 500uA) when active. Moteino is also not an option for this. Those are made for batteries. I want to be battery free. A super cap, however can be used to store energy. I'll get to that shortly.
|
|
|
|
For micros, I have some Attiny10 on hand, and these have a reasonably low power pull in active mode. Let's build those up first. What will the micro do? No idea. I haven't a clue.
|
|
|
|
\textcolor{green!60!blue!70}{
|
|
\subsubsection{Micro Notes}}
|
|
Must run at 1.8V / 1MHz per front page of data sheet, for 200uA draw in active mode.
|
|
\\
|
|
\\
|
|
|
|
|
|
\textcolor{green!60!blue!70}{
|
|
\subsection{Energy Storage}}
|
|
I don't want a battery. Let's go with a super cap. The solar panels will only be active some of the time, so I will want to harvest energy with some kind of IC into the cap when the sun is out.\footnote{Reference: www.analog.com/media/en/technical-documentation/technical-articles/solarenergyharvesting.pdf is a start. I'll need to do more research.}
|
|
|
|
\textcolor{green!60!blue!70}{ \subsection{Make parts, not scrap}} I will
|
|
want to make sure that all parts I build are perf board parts, not
|
|
breadboard scrap (to be torn down and rebuilt again). This is an Attiny,
|
|
so no need to test much, yet.
|
|
|
|
\textcolor{green!60!blue!70}{
|
|
\subsection{Programming}}
|
|
To program the Attiny10, I'll use the Arduino adapter from the Junk + Arduino blog. I built it up\footnote{Had slight error where the Arduino + board wouldn't read - pins too short on headers, then the arduino wouldn't boot - due to bad connection on perf board shield. Thankfully, the USB port didn't try to run. Protection circuitry cut in on the laptop.}, and was able to Read the memory. In order to upload to the board, you will need a compiler setup. You can possibly do it in AVRGCC, but instead I opted for either Arduino IDE (via Attiny10Core which didn't work), and then went to Mplab. In order for mplab 5.25 to work, it will need XC8 compiler, and there is a pack that can be downloaded through the IDE to get Attiny10 support.
|
|
|
|
It appears the AVR Dragon (which I have) can not be used. However, other programmers can be used. Pickit 4, Mkavrii, stk600, I think.
|
|
|
|
|
|
\textcolor{green!60!blue!70}{ \subsubsection{Testing Arduino Loader}}
|
|
Tested this with the blink\_LED.c in code folder. The code is as simple as possible.
|
|
It is the following:
|
|
\begin{verbatim}
|
|
//#include <xc.h>
|
|
#include <avr/io.h>
|
|
#include <util/delay.h>
|
|
|
|
int main(void)
|
|
{
|
|
// PB2 output
|
|
DDRB = 1<<2;
|
|
|
|
while(1)
|
|
{
|
|
// Toggle PB2
|
|
PINB = 1<<2;
|
|
_delay_ms(500);
|
|
}
|
|
}
|
|
\end{verbatim}
|
|
When programmed in Mplab, with XC8 compiler, and Attiny10 support, I get the following
|
|
hex output:
|
|
\begin{verbatim}
|
|
:100000000AC020C01FC01EC01DC01CC01BC01AC01B
|
|
:1000100019C018C017C011271FBFCFE5D0E0DEBF41
|
|
:0A002000CDBF03D000C0F894FFCF5D
|
|
:10002A0044E041B940B95FE966E871E05150604087
|
|
:0A003A007040E1F700C00000F5CFB0
|
|
:02004400DDCF0E
|
|
:00000001FF
|
|
\end{verbatim}
|
|
The content of this hex isn't the focus of this passage. Instead, I want you to review the
|
|
results of a D for Dump Memory, by the Arduino Loader.
|
|
\begin{verbatim}
|
|
|
|
Current memory state:
|
|
|
|
registers, SRAM
|
|
+0 +1 +2 +3 +4 +5 +6 +7 +8 +9 +A +B +C +D +E +F
|
|
0000: 05 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
|
|
0010: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
|
|
0020: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
|
|
0030: 00 00 00 00 00 00 03 00 00 79 00 03 00 00 00 00
|
|
0040: B7 AD AE FA 58 70 63 6B FB 5A B4 1B FF FF 35 3F
|
|
0050: 67 D7 33 43 DF 5F FB 72 C9 7D FE E9 9D C5 00 12
|
|
NVM lock
|
|
+0 +1 +2 +3 +4 +5 +6 +7 +8 +9 +A +B +C +D +E +F
|
|
3F00: FF FF
|
|
configuration
|
|
+0 +1 +2 +3 +4 +5 +6 +7 +8 +9 +A +B +C +D +E +F
|
|
3F40: FF FF
|
|
calibration
|
|
+0 +1 +2 +3 +4 +5 +6 +7 +8 +9 +A +B +C +D +E +F
|
|
3F80: 79 FF
|
|
device ID
|
|
+0 +1 +2 +3 +4 +5 +6 +7 +8 +9 +A +B +C +D +E +F
|
|
3FC0: 1E 90 03 FF
|
|
program
|
|
+0 +1 +2 +3 +4 +5 +6 +7 +8 +9 +A +B +C +D +E +F
|
|
4000: 0A C0 20 C0 1F C0 1E C0 1D C0 1C C0 1B C0 1A C0
|
|
4010: 19 C0 18 C0 17 C0 11 27 1F BF CF E5 D0 E0 DE BF
|
|
4020: CD BF 03 D0 00 C0 F8 94 FF CF 44 E0 41 B9 40 B9
|
|
4030: 5F E9 66 E8 71 E0 51 50 60 40 70 40 E1 F7 00 C0
|
|
4040: 00 00 F5 CF DD CF FF FF FF FF FF FF FF FF FF FF
|
|
4050: FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF
|
|
4060: FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF
|
|
(...some memory omitted here for brevity...)
|
|
43E0: FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF
|
|
43F0: FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF FF
|
|
\end{verbatim}
|
|
|
|
Notice that the "AC020C01F" is set. That is from the hex. But the 01000...
|
|
before it seems to be missing. Some deciphering of how the Arduino programs
|
|
the Attiny is in order here. It also doesn't end the same.
|
|
|
|
Regardless, when programming, the Arduino reports 70 bytes written, and
|
|
likewise in the Mplab project memorymap.xml file, it also notes 70 bytes
|
|
for the sketch. This lines up.\footnote{Although for an unknown reason, every command registers twice on the Arduino serial monitor, but this appears to be harmless.}
|
|
|
|
The blinking LED works. Let's move on.
|
|
|
|
\textcolor{green!60!blue!70}{\subsubsection{Conclusion on Arduino Programming Attiny10}}
|
|
It's possible, but you have to make a dedicated jig (almost), so it might be easier to use the official programming tools. However, based on this https://www.avrfreaks.net/forum/pickit-4-and-avr-mcu I might not have a choice. So I will use the Arduino for now. But will have to devise what on board parts are req'd for programming, and incorporate into proto board layout.
|
|
|
|
\textcolor{green!60!blue!70}{ \subsubsection{IO Port Switching Speed}}
|
|
Using the above code without any delay\_ms, I get the following results from a default clock
|
|
speed, and a 128KHz clock speed. This test was done to confirm that I could change the clock with
|
|
\begin{verbatim}
|
|
//Write CCP
|
|
CCP = 0xD8;
|
|
//change CLK to 128KHz
|
|
CLKMSR = 0b01;
|
|
\end{verbatim}
|
|
There was no issue.
|
|
\begin{verbatim}
|
|
Default CLK (8MHz? or 1MHz?): 160KHz IO Switch
|
|
128KHz CLK: 2.5KHz IO Switch
|
|
\end{verbatim}
|
|
I am going to pursue 128KHz for starters, for lower current dissipation. Note that with the Arduino loader, it is cumbersome to test and change code as you move along. It is
|
|
therefore going to be necessary to use a programmer, with a dedicated header on board.
|
|
\textcolor{green!60!blue!70}{ \subsubsection{VCC 1.8V}}
|
|
The lowest power supported: 1.8V can be applied, without any configuration
|
|
needed. It does not affect IO switching speed (although obviously amplitude is affected).
|
|
\begin{verbatim}
|
|
128KHz CLK (5.0V): 2.5256 KHz IO switch
|
|
128KHz CLK (3.3V): 2.5477 KHz IO switch
|
|
128KHz CLK (1.8V): 2.5849 KHz IO switch
|
|
\end{verbatim}
|
|
As voltage drops, IO increases.
|
|
\\
|
|
\\
|
|
\textbf{VCC Dropout voltage:}
|
|
\\
|
|
From 1.5, it drops out at 1.248V or so. Comes back at about 1.34V
|
|
\\
|
|
\\
|
|
Test size of 1.
|
|
\\
|
|
\\
|
|
Can't run this with one (AA) battery, but you could with 2.
|
|
\\
|
|
\\
|
|
Current Draw: 128KHz - IO test, 1.8V, 0.08mA (~78uA) (tested w/3478A)
|
|
\\
|
|
\\
|
|
\textcolor{green!60!blue!70}{\subsection{Application}}
|
|
First, I need a board for these and a programmer, to quickly program. Second, I need an application. I want extremely low power. Hopefully, solar with no batteries, to start. This is extremely low - that is the point. Let's keep this ridiculous.
|
|
|
|
Given the power requirements put me under 1mA (with my current panels), I'm considering the following: EEPROMs would require SPI protocol. Doable, but overcomplicated for now.
|
|
\\
|
|
\\
|
|
Eink (need to find a small and cheap enough option. So far, they have either too many pins, and/or use too much current. Something like what stores use to display prices would work, but that doesn't get the data out, only makes it readable.),
|
|
\\
|
|
\\
|
|
Third option would be RF. That is a viable path, but not today. Let's skip that for now.
|
|
\\
|
|
\\
|
|
Fourth option that comes to mind is IR. IR diodes, as in TV remotes, would work well here. I am choosing this as the first project. I will have dumb clients, that consist of - Attiny / IR / Sensor powered by solar. I will have a BBB that receives the IR data, and does all intelligent data gathering. To keep things simple, the IR will be binary ADC data, or otherwise sensor numbers. No SPI, no protocol complexity. That would require space on the Attiny.
|
|
\\
|
|
\\
|
|
Let's build some boards based on the above.
|
|
\\
|
|
\\
|
|
For sensors:
|
|
While building, I came across an option. Hall effect sensors. I think also capacitive sensors can be used. This may find a use in a gate sensor, for when a driveway gate is opened or closed. With a small battery, it would work for years.
|
|
|
|
Footprints: I had to make a footprint for this module on board package for one sensor. The solution to get footprints right? copy graphic image and make it into silkscreen on the board. Easy.
|
|
|
|
The sensor I looked at was a temp and humidity sensor SHT11 (SHT10 is obsolete). It is low power enough. However, it's \$20. So not in my price range. Otherwise, it would work here. Looks like communication is a shift register, or SPI.
|
|
|
|
\textcolor{green!60!blue!70}{
|
|
\subsubsection{Magnetic Current Sensor}}
|
|
There is this:
|
|
BM14270AMUV-LB
|
|
Which is low enough current here (<1mA). But \$7 in qty, and req's I2C. Not today.
|
|
|
|
\textcolor{green!60!blue!70}{
|
|
\subsubsection{Accelerometers}}
|
|
These are an option.
|
|
|
|
Best pinout (for deadbug) is LIS344ALHTR (but lacks vcc down to 1.8)
|
|
2nd Best pinout with full 1.8 -3.6 vcc is ADXL337BCPZ-RL7
|
|
(Keep in mind, these are low end options only)
|
|
(Analog output only. keep it simple for now.)
|
|
Runner up to all above, is KXTC9-2050-FR . But has worse pinout.
|
|
|
|
Going with AD part. \$5 in single qty.
|
|
Digital output Accelmeters are cheaper.
|
|
|
|
All have tiny package sizes.
|
|
|
|
Since I am grabbing 1 output only, will need to orient or choose correct output.
|
|
|
|
\textcolor{green!60!blue!70}{
|
|
\subsubsection{Temperature Sensors}}
|
|
Temperature can be boring, but why not. Let's throw one of these on: LMT84LP . Pin compatible with LM35. Supply current is maybe 8uA. Extremely low.
|
|
LM84 (1.5V starts, to 5.5), LM85 (1.8V to 5.5)
|
|
|
|
\textcolor{green!60!blue!70}{
|
|
\subsubsection{Gas Sensors}}
|
|
Lowest is 5mV as of writing on dkey. Skipping. The SHT would work, but its too expensive.
|
|
|
|
\textcolor{green!60!blue!70}{
|
|
\subsubsection{Supercap}}
|
|
For now, trying this:
|
|
FG0V155ZF
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
\textcolor{green!60!blue!70}{
|
|
\subsubsection{Hall Effect Sensors - Push Pull vs Open Drain Outputs}}
|
|
\begin{verbatim}9.1.1OutputTypeTradeoffsThe push-pulloutputallowsfor the lowestsystempowerconsumptionbecausethereis no currentleakagepathwhenthe outputdriveshighor low. The open-drainoutputinvolvesa leakagepaththroughthe externalpullupresistorwhenthe outputdriveslow.The open-drainoutputsof multipledevicescan be tied togetherto forma logicalAND.In this setup,if any sensordriveslow, the voltageon the sharednodebecomeslow. Thiscan allowa singleGPIOto measurean arrayofsensors
|
|
\end{verbatim} From DRV5032 data sheet.
|
|
|
|
\textcolor{green!60!blue!70}{
|
|
\subsection{Starting Sensors}}
|
|
So as a recap, to start with affordable, low power sensors for my project, I have the following types:
|
|
\begin{itemize}
|
|
\item Temp sensor (cheapest)
|
|
\item Magnetic Sensor (hall effect)
|
|
\item Movement Sensor (accellerometer) (analog output) (tiny package)
|
|
\item capacitive sensor (azoteq)(may only be short range)
|
|
\end{itemize}
|
|
|
|
Output, I have not determined yet. IR will not work, as its too high power. Unless I dedicate a battery just for the IR diode... Or make it battery powered. I'll start with batteries, but for solar panel and supercap, it will likely not be viable, unless I transmit extremely rarely. That is also an option, however.
|
|
|
|
\subsubsection{Farad to mA}
|
|
1.5F supercap can supply 1.5A for 1 second. That is 0.025A for 60 seconds, or 25mA for a minute.
|
|
Let's stay I use half that, so 12mA for 60 seconds is my supply. If I transmit once every other hour (when in sunlight)...
|
|
|
|
\subsection{PCB}
|
|
I have built rev2 of the board today. Using an Uno, user must remember to include VCC and GND. So the programming takes up all 6 pins. The 0.1'' headers were slightly close to the resistors, and the top row of VCC and GND headers are separated, so I labeled them V+2, and GND2.\footnote{Forgetting to plug in GND, and or 5v+ (for Uno), while still plugging in the TPI pins, did not break the Attiny10, in my tests today.}
|
|
|
|
The 0.05'' pin headers work perfect for scope probes. The extra breakout I made demonstrates this.
|
|
|
|
\subsection{PCB Programming}
|
|
|
|
By default, the Attiny10 idles around 1.5mA (5V), before programming.
|
|
|
|
First thing to do, is to program it into a low power mode.
|
|
|
|
In order to use low power, whilst using the Uno, I will need to add a jumper to the PCB,
|
|
so that you can switch between V+2, and V+1. Three pin jumper. Outside pins are each V+, internal goes to Attiny V+. I'll need to remove the trace on gerber rev2, that goes from 4 to VCC.
|
|
|
|
\subsubsection{PCB programming and use}
|
|
It's not possible to leave the Arduino plugged in, while testing the Attiny10.
|
|
example: Even if you power down the Uno, and use a jumper to change V+ rails,
|
|
the power dissipation through the TPI pins (10,11,12,13) will cause a draw of about 545uA, and the Attiny10 will not toggle its led.
|
|
Therefore, not only a power jumper is required to use the Attiny, but also
|
|
a 4 or 5 pin 0.1" cable for the TPI pins, must be disconnected before using the Attiny.
|
|
\footnote{In code section, mplab - tests2, the power draw of the 128KHz internal CLK with PB2 flipping at about 400us a cycle, at 3V VCC is about 115uA. At 2V it is ~95uA.}
|
|
|
|
|
|
|
|
|
|
\end{document}
|
|
|