cromagn

LT Spice simulation of Loop Antenna preamplifier PA0FRI

2024-04-25T16:50:00.006+02:00

Just for fun I want to simulate the behavior of the loop preamplifier noted on PA0FRI web site (great job, guy !!!) version 25 jul 2023

This is the log of the activities:

Start with some useful links

- 2SC1908 BJT Datasheet

- Loop antenna LT Spice model

- Tran simulation LT Spice model(beta)

~~No real BJT Model.~~
Try to use a matched pair BJT: HFA3134 with this LTSpice model (pay attention to "CRLF" !!!)

.model HFA3134 NPN
+ (IS = 1.98E-16 VAF = 2.00E+01 BF = 1.20E+02 IKF = 2.10E+00
+ ISE = 7.70E-15 NE = 2.00E+00 VAR = 2.64E+00 BR = 1.78E+01
+ IKR = 6.10E-03 ISC = 2.26E-16 NC = 1.57E+00 NK = 3.00E+00
+ CJC = 6.00E-13 MJC = 3.80E-01 VJC = 6.00E-01 CJE = 7.00E-13
+ MJE = 5.10E-01 VJE = 8.72E-01 XCJC = 9.00E-01 CJS = 1.39E-13
+ VJS = 7.50E-01 MJS = 0.00E+00 FC = 5.00E-01 TR = 2.50E-09
+ TF = 1.30E-11 ITF = 2.80E-01 XTF = 2.01E+01 VTF = 2.37E+00
+ PTF = 3.70E+01 RC = 5.25E+00 RE = 8.70E-01 RB = 1.29E+01
+ RBM = 6.47E+00 KF = 0.00E+00 AF = 1.00E+00)

Now some other paired BJT to test with the LTSpice model:

MAT12: LTSpice Model (.cir) LTSpice Schematic (.asy)

SSM2122: LTSPICE Model (.cir) LTSpice Schematic (.asy)

Decoding a morse code message coming from a SatNOGS station

2019-08-16T12:05:00.000+02:00

aka...how to spend summer having lot of fun...

The idea is to decode a morse signal in a fully automated way. In the specific case, I want to decode a signal coming from a satellite CW transmitter heard by satNOGS receiver.

So, starting from a signal like this:

i want to achieve this:

The tested signal is part of the observation #910803 of a CUBESAT XI-5 (CO-58, OSCAR 58) CW transmitter.

How to proceed?

I want to give a chance to the following processing chain: Python with numpy & scipy, some open source library and a lot of patience...

Audio to Morse code processing chain

The step 1 is needed to remove the noise and keep only the tone (dit dee) of the signal.
The step 2 and 3 are the demodulator (mainly a rectifier and a integrator step)
The step 4 put the signal in a better shape
The last step convert the square wave signal in chars
Easy! Well...yes for simple signals. It works like a shame...but for more tricky signal...it is a different story
But...step by step.
First of all...the tools:

- Audacity to look into the signals and apply simple processing

- a Python IDE: pyCharm

- some Python snippet (here my GitHub page) to manage ogg and wav file, resample, visualizate, appy filter and so on

- good library like the clock recovery one (by Michael Ossmann)

Here the gist of the startup : load a wav file, see the shape and the spectrum and start reasoning...

First look at the signal

The signal seems sampled at 44 khz but no frequency after 3.5 kHz...strange (?)

Some noise but a good signal around 3khz (3147 Hz @ -46dB)

How to detect the frequency not using Audacity? Asking numpy to find the frequency of the max of the spectrum power

# identify freq

narr=2.0/N * np.abs(yf[0:N//2])

fmaxidx= np.where(narr == np.amax(narr))
fmax=int (xf[fmaxidx[0]])

It works ? Well...only for good signals. For others...it is not smart enougth...

In this example the signal frequency is not the maximum of the spectrum...so, to find signal we can leverage on a CW frequency satellite database and looking for a maximum near the observed one or try to discovery it using a local max of the PSD with some good properties.

Here the gist but following the idea is to select the max that is quite isolated (distance), height more than the mean of the signal/4 and with a good difference to other max (threshold = std)

# Plot the power spectrum with maximumplt.subplot(3,1,3)
data = data - np.mean(data)
f, Pxx = sp.signal.welch(data, fs, nperseg=1024)
#f_res, Pxx_res = sp.signal.welch(data, fs, nperseg=2048*2)plt.plot(f, Pxx)
plt.xlabel('frequency [Hz]')
plt.ylabel('PSD')

# Find good property to discover the real maxmPxx= np.max(Pxx)
vPxx=np.std(Pxx)
peaks,prop= sp.signal.find_peaks(Pxx, height=mPxx/4,threshold=vPxx, distance=10)
# peaks = sp.signal.find_peaks_cwt(yy,[50,100])   sexy but...
print("L Tresh:",prop["left_thresholds"])
print("R Tresh:",prop["right_thresholds"])
plt.plot(f[peaks.max()],Pxx[peaks.max()], "x")
plt.show()

Now the challange is to find the bandwidth of the passband filter and do a massive test to tune the script.

sciPy can help a lot with the function sp.signal.peak_widths

so...give them a try...

results_half= sp.signal.peak_widths(Pxx, peaks, rel_height=0.5)
results_half_int= np.rint(results_half).astype(int)
plt.hlines(results_half[1],f[results_half_int[2]] ,f[results_half_int[3]],color="green",

 linestyles="dotted")
plt.grid()
plt.show()

print(f[results_half_int[3]] -f[results_half[2]])

and it seems a good idea

Next step...have a look to see if something make sense... so, I apply a bandpass filter to the signal with different BP and i plot the shape...

Different BP around 3.273 Hz and 90 Hz BP

Not so bad at all !!! The dit and dat are now clearly visible just using a 4th order Butterworth BP filter.
Seems to me that the selected BP of 90 Hz is not enougth to see the first letter

Zooming on the first sounds, using a larger BP than 90 Hz, improve the signal s/n and we are able
to see .. ... _ :) this is the first part of the payload of the transmitter. The full message for a CubeSat OSCAR CO-58 (XI-V / SAI FIVE) is "XIV4 + bits" and we are able to decode "iv4"...Outstanding!

Now, I try to test the idea extensively and I got a new best friend...Memory Error.
Manage large (but not so large) array in Python on my PC is not good!
We can downsample the ogg file to 8k with ffmpeg (windows example)

for %%a in (*.ogg) do (
echo processing %%a
ffmpeg -i %%a -ar 16000 "%%a~conv.wav"
)

and chunk the signal in 30 s slot, just to see if everything sound good.
Now I'm ready to tune the filter function
A couple of gist with the code (main & tools) and what I v learned?
1. The noise profile is the same for every sample, so we can pre-process the file with a BP around 2kHz and 4 kHz

Typical NF from satNOGS sample

2. There are some lovely "stegosaurus like" signal (aka a the tone sequence)

Tone from CUBESAT (?)

3. Sometime the peak has been detected

Peak detected in unattended process

4. Sometime (most of time) no!

No luck...

So...retest after the prefilter action

Domoticz & blink(1)

2018-09-23T20:54:00.000+02:00

Domoticz: it's my favorites Home Automation System and I want to add a new hardware to it to enance the "notification" capability.

The chosen device is a blink(1) mk2 USB RGB led.There is a lot of scripts to control the blink(1) - mainly thanks to the "unix philosophy" used by the developers of this funny device - but I want to create a python plugin for Domoticz.

The infos on this device are located on github but mainly there are two way to interact with this device in the Linux space:
- blink1-tool : a powerful command line interface plenty of instructions (also in the nerd space)
- blink1-tiny-server: a web interface to the C library under the blink1-tool. Limited to one device a server and with fews instructions available (please, enhance it !!!) but remotly controllable with a simple http get like /blink1/fadeToRGB?rgb=%23ffffff&time=1.0. This interface is able to switch on and off the led, change the fade time and blink the led.

* localhost:8000/blink1/
* localhost:8000/blink1/on
* localhost:8000/blink1/off
* localhost:8000/blink1/red
* localhost:8000/blink1/green
* localhost:8000/blink1/blue
* localhost:8000/blink1/blink?rgb=%23ff0ff&time=1.0&count=3
* localhost:8000/blink1/fadeToRGB?rgb=%23ff00ff&time=1.0

On top at this interface the Domoticz plugin will be built. But...what kind of device is it? In the Domoticz space, i think that a Switch/ Lamp RGB is the best match to the blink(1).

Here my actual implementation of the Domoticz plug in. Not yet stable but a good MVP.

Obviusly the "on" and "off" command wake up the led handling a onCommand called for Unit 2: Parameter 'On', Level: 0 (the level remain the same at 0...) and the color is managed by an onCommand called for Unit 2: Parameter 'Set Color', Level: 88', Hue: {"b":97,"cw":0,"g":255,"m":3,"r":200,"t":0,"ww":0}
The blink capability is enabled if the Level is set under 95%, otherwise is a fixed led.

The Hardware device (and the configuration)

The new Blink device

RGB Blink configuration

KORG Monotron

2018-08-29T15:34:00.003+02:00

Here the plan:
1.Open it (And close also) --> done (Here the ifixit link)
2.Create a Spice simulation of the LFO & VCF --> working on it (ETA: sept 18)

First draft of VCF schematics

LTSpice file for VCF (super draft)

3.Create a version with a real power supply --> braining
...

Orange T-Six (or FsFly T-Six) and Micro Scisky Brushed Flight Control Board

2017-02-19T02:21:00.000+01:00

Some quick notes to remember to me how to set up a Drone (quadcopter) with a Orange T-Six transmitter and a Micro Scisky 32 bit Brushed Flight Control Board.

Set up the Transmitter:

1. "ACRO" mode : First and most important, put the transmitter in "plane" mode and not in "copter" mode. In copter mode there is a strange coupling between the AUX channel and the throttle control and this complicate things. Also, this prevent the channel under the switch "F-mode / Gear" to work in the correct way enabling the AUX. This switch will enable the drone to be armed.

2. Channel 5 (CH5) : Enable also channel 5 (CH5) controlled by "FLAP/GYRO" switch putting the FLAP setting as follow: FLAP NORM +100, FLAP LAND -100.

Set up the Flight Control board:
1. Upgrade: Upgrade the board with the last version of cleanflight. On my board (AFNA version 2) is mandatory to jump two pins with a cable before upgrade the firmware.
2. Cleanflight settings:
Remember to press "Save" every time you change something!
Dedicate the second serial port to the DSM2 receiver setting on the Serial RX. This enable the connection to the receiver.

CleanFlight Port setting for the Micro SciSky Brushed Control Board

Discover the min and the max value of the controller on the T-Six Orange transmitter.
Set the Stick Min and Stick Max value according to the min and max value (some point more). This setting enable the board to understand when the throttle stick is all low and arm the engine.

By the way, if we setup the Orange T-Six correctly (plane mode + Flap / Gyro enable) you can control the correct behaviour of the F-Mode / Clear switch used to arm / disarm the drone and the Flap switch. If you switch on or off this one, you can see the new value on the bar.

We must complete the setting of the receiver: RX_Serial is the correct value for the Receiver Mode tab, SPECTRUM1024 for the Serial Receiver Provider tab and AETR1234 for the Channel Map tab.

CleanFlight Receiver setting for the Micro SciSky Brushed Control Board

To enable the drone to be armed (aka "ready to fly") you must arrange the ARM control under the modes tab and select the value of the cursors near the value you see in the receiver tab for the AUX 1 channel.

So, if the cannel is off, no way to arm the drone, if this channel is on, you can arm the drone position the throttle stick on the lower value.

CleanFlight Modes setting for the Micro SciSky Brushed Control Board to arm the drone

The last action (before calibrate the drone) is to set the correct PWM frequency using the CLI with this command set motor_pwm_rate = 32000 and save

(---to be continued ...)

Arduino Digispark & Syma X5S / X5SW

2016-05-24T14:57:00.000+02:00

Now that the protocol used by Syma X5 drone use to communicate with the cam is (quite) clear, the next logical step is adding some other actions to this controller. The first way is to use a Digispark (by Digistump) or a smaller ATTINY85 clone to interpret the protocol and to do something great.

The first tests confirm that the supply voltage level (3.7 v from the drone) is enougth and the arduino board is able to run and to communicate with other Arduino's devices/sensor. Also, no problem with noise that can arrive from the motor to the power supply. Seems that the power supply is clean and good enougth.

To connect the Syma X5 drone to the Arduio board I use this 3 pins 1.27 mm connector. Keep max attention to the color of the cable and to be sure that you connect the correct pins to the arduino board. Using my connector, the black cable is the V+, the yellow cable is the GND and the red cable is the signal but please check it by yourself.

Some images & videos of the firsts test

Looking at the source code, the first thing to do is to be sure that we are able to read the signal pin and act according to it. This stupid example works an, if you push the left button on the drone controller, you see the Led lights up.

 int ledPin = 1; // LED connected to digital pin 1  
 int inPin = 0;  // dronce connected to digital pin 0  
 int val = 0;   // variable to store the read value  
 int stat=0;  
 void setup()  
 {  
  pinMode(ledPin, OUTPUT);     
  pinMode(inPin, INPUT);     
 }  
 void loop()  
 {  
  val = digitalRead(inPin);  // read the input pin  
  digitalWrite(ledPin, val);  // sets the LED to the button's value  
 }

Now a little more advanced example using the interrupt to detect and count the push of the button.

 #include ;  
 #define LED_PIN       1  
 #define FIRST_INPUT     0  
 static boolean State = HIGH;  
 volatile uint16_t FirstInputChangeCount = 0; /* Volatile since the variable will be updated in interruption */  
 uint8_t VirtualPortNb;  
 uint8_t VirtualPortNb_;  
 void setup()   
 {  
  TinyPinChange_Init();  
  VirtualPortNb = TinyPinChange_RegisterIsr(FIRST_INPUT, InterruptFunctionToCall);  
  /* Enable Pin Change for each pin */  
  TinyPinChange_EnablePin(FIRST_INPUT);  
  pinMode(LED_PIN, OUTPUT);  
 }  
 /* Function called in interruption in case of change on pins */  
 void InterruptFunctionToCall(void)  
 {  
   if(TinyPinChange_RisingEdge(VirtualPortNb, FIRST_INPUT)) /* Check FIRST_INPUT has changed (falling or rising edge) */  
   {  
     FirstInputChangeCount++; /* Rising AND Falling edges are counted */  
   }  
 }  
 void loop()  
 {  
 uint16_t LocalFirstInputChangeCount;  
  /* Blink the built-in LED */  
  if(FirstInputChangeCount> 10)  
  {  
   digitalWrite(LED_PIN, HIGH);  
  }else  
  {  
   digitalWrite(LED_PIN, LOW);  
  }

1. On the basement for Syma X5S Fixed Set I'v removed the plastic anchor to gain a bit more space for the Digispark

2. Using a 2 mm drill bit and the board as a dima, I make two little holes.

3.Using 2 M2 screw (15 mm and 6mm long) and 2 M2 bolt caming from mini4wd cars spare parts, i'v create the support for the Arduino board

2. Mount the Arduino ATTINY85 mini USB board and block it with other 2 M2 bolt. I add also a drop of threadlock to be sure that I don't lose this component during a flight.

3. A small test prove that the cables length is correct and that the lock fit well in the drone. Seems that this new board is ok.

4. Adding the led ring. This is a classic WS2812 device with 8 RGB leds controllable using the Adafruit_NeoPixel arduino library. To seal it to the rest, is necessary to do a hole near the border, use another M2 bolt and a bit of threadlock.

5. Only a side view of the assembly

6.And a front view of the same component

7. Final assembly and first tests.

Some modding to Syma X5

8.Preparing a good "case" for this mods.First of all Tamiya Primer, Tamiya Yellow TS-16 for the bottom of the drone (a lot of...) and Tamiya TS-42.
Copied some pattern from internet and trasformed in Vinyn sticker with a Silhuette Portrait cutter plotter. After this a good wave of TS-13 to protect the model.

Hacking drone Syma X5 camera protocol

2016-05-12T00:33:00.003+02:00

The idea is to hack and full understand the protocol used by Syma X5SW and Syma X5C drone to communicate with the cam or the wifi cam. As every Syma X5 serious drone users well know, you can control the cam using radiocomand by the switch at the right of the left controller, pushing it UP or DOWN. Here the pics of the controllers !

Hacking this protocol can enable the control of other payload attached to the drone so I connect the 3 pins cam connector of Syma X5 to a logic analyzer.

This connectors has a classic 1.27 mm pinout with Signal, Gnd and V+

Syma X5 cam connector

Syma X5 connected to the logic analyzer

So, looking at the result, I see that when I push "up" the button, a 2 seconds long impulse compairs on the screen and when I push "down" the button a 7 seconds long impulse arrive.

Here the screenshoot of the logic analyzer when the Syma controller button has been pushed.

Hacked Syma X5 cam protocol

Detail of the 7 sec long impulse when don button has been pressed

Domoticz Entity-Relationship database diagram (E-R)

2016-01-10T12:09:00.001+01:00

Domoticz is an oustanding multiplatform home automation software easily extendible, well documented and with a good integration with openzwave communication layer. Best of all...it's written in C language :).
Of course, Domoticz is the core of my Home Automation system as a consequence of the first sentence :)

As usual, i'm quite curious and I want to know a bit more about the internal structure of this software and today I want to see the design of the Domoticz's database. A quick google search did not give me useful result (but may be I miss some important results) so i decided to have a look on the database schema by myself. Create a back up of the database and start to analize the db SQLite file for Domoticz rev 2.3554

What tools? I already use for my day by day geek activities on Sqlite DB Browser for SQLite, SchemaCrawler to inspect the structure and, never miss, Graphwiz to plot the diagram.

What's happened? I prefer to generate the .viz file to manually adjust some aspect of the plot layout so, from my Windows box (a new entry...) we can type:
schemacrawler.cmd -server=sqlite -database=C:\domoticz.db -infolevel=maximum -command=graph -outputformat=xdot -outputfile=graph.viz
and, to generate an HTML output:
schemacrawler.cmd -server=sqlite -database=C:\domoticz.db -infolevel=maximum -command=brief -outputformat=html > db.html

Here the outpuf after some little adjust with gvedit:

E R Diagram for Domoticz

The tabular infos on HTML format are available on this link, the .viz file on this one and the image here

30 Sept 2018 : New ER diagram updated to Domoticz v.4.9700 generated with SchemaCrawler schemacrawler.cmd -server=sqlite -database=
C:\tmp\domoticz.db -infolevel=maximum -command=schema -fmt=scdot -outputfile=graph.viz

ER for Domoticz v.4.9700

Create a linux module for Jumping Sumo on a cross compiler environment

2015-12-25T01:25:00.000+01:00

I want to play a bit with the video interface (V4L2) in my Jumping Sumo Drone so, first of all, i need to learn how to add a kernel module to the kernel 2.6.36 embedded in the drone.
Really new experience for me. It's is a war and a bit a "try & error" path...now the process is working but still no working i have a working module! The main steps are:

Prepare the Kernel image:
1. Create a copy of the KERNEL downloading it from here
~~2.a Create the config for the Kernel running make ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- oldconfig~~
~~2.b Or make clean and make ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- menuconfig~~

~~3. Prepare the Kernel running make ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- modules_prepare~~

Important: Choose the Parrot arch family and accept all others default values (after...this sentence change a bit)
2. Copy the kernel.config file (from Parrot or from here) in the kernel source folder and rename it as .config
3.Execute make ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- modules_prepare

Prepare the compiler tool chain: Quite simple step: need to install the arm gcc with sudo apt-get install gcc-4.5-arm-linux-gnueabi.

May be other version works also well, but I read somewere on internet that this is the correct version. Also you can download here the entire toolchain from Parrot, but no experience!

Download the target module to be compiled: My first goal is to create a loopback device for the V4L2 video dev, so I want to create this v4l2loopback module. This is a loopback video device that are able to create "virtual video devices" that normal (v4l2) applications will read without problem (i hope so) and that are feeded by other applications.

Prepare the Makefile for the module: I need to change a bit the Makefile to be sure that work with the already prepared Linux header and use the correct compiler.

Mostly add on top export and change the make file in this way.

$(MAKE) ARCH=arm CROSS_COMPILE=arm-linux-gnueabi- -C $(KERNEL_DIR) M=$(PWD) modules

Here is my Makefile for ARM platform to keep as example.

~~Some errors occour when I try to insert the module:~~
~~[JS] $ insmod v4l2loopback.ko~~
~~insmod: can't insert 'v4l2loopback.ko': invalid module format~~

~~and the reason is simple:~~
~~[ 99.922413] v4l2loopback: version magic '2.6.36 mod_unload modversions ARMv5 ' should be '2.6.36 preempt mod_unload ARMv5 '~~

~~Now I need to recompile the kernel (not all, only module prepare) and find the correct setting !!!~~

~~Ok, now better...adding in the preparation of the kernel the option:~~

~~Processor type and features~~

    Preemption Model
        ( ) No Forced Preemption (Server)

        ( ) Voluntary Kernel Preemption (Desktop)
        (x) Preemptible Kernel (Low-Latency Desktop)

~~and recompile...the insmod seems have a better aspect:~~
~~v4l2loopback: version magic '2.6.36 preempt mod_unload modversions ARMv5 ' should be '2.6.36 preempt mod_unload ARMv5 '~~

~~Find a way to remove the modversion flag~~
~~Enable loadable module support~~
~~[ ] Forced module loading~~
~~[*] Module unloading~~
~~[ ] Module versioning support~~

~~[*] Source checksum for all modules~~

~~Again no luck !~~
~~insmod: can't insert 'v4l2loopback.ko': unknown symbol in module, or unknown parameter~~

~~v4l2loopback: Unknown symbol __stack_chk_guard (err 0)~~
~~v4l2loopback: Unknown symbol __stack_chk_fail (err 0)~~
~~v4l2loopback: Unknown symbol add_preempt_count (err 0)~~
~~v4l2loopback: Unknown symbol kmem_cache_alloc_notrace (err 0)~~
~~v4l2loopback: Unknown symbol __gnu_mcount_nc (err 0)~~
~~v4l2loopback: Unknown symbol __tracepoint_kmalloc (err 0)~~
~~v4l2loopback: Unknown symbol sub_preempt_count (err 0)~~
~~v4l2loopback: Unknown symbol kmalloc_caches (err 0)~~

~~Removed the kernel flag Enable -fstack-protector buffer overflow detection (EXPERIMENTAL)" and the first 2 error disappair. The new list is:~~

~~v4l2loopback: Unknown symbol add_preempt_count (err 0)~~
~~v4l2loopback: Unknown symbol kmem_cache_alloc_notrace (err 0)~~
~~v4l2loopback: Unknown symbol __gnu_mcount_nc (err 0)~~
~~v4l2loopback: Unknown symbol __tracepoint_kmalloc (err 0)~~
~~v4l2loopback: Unknown symbol sub_preempt_count (err 0)~~
~~v4l2loopback: Unknown symbol kmalloc_caches (err 0)~~

~~Removed also CONFIG_FUNCTION_TRACER and CONFIG_DEBUG_PREEMPT~~
~~and remains only:~~
~~v4l2loopback: Unknown symbol kmalloc_caches (err 0)~~

~~To remove this, the config to change is the CONFIG_SLUB...and use SLAB~~

Using a simple hello.c module ... this work!
#include /* Needed by all modules */
#include /* Needed for KERN_INFO */
int init_module(void)
{
printk(KERN_INFO "Hello world 1.\n");

}

void cleanup_module(void)
{
printk(KERN_INFO "Goodbye world 1.\n");

}

See the result after an insmod command:

[ 23.314615] wifi_bcm: no IPv6 routers present
[ 493.993382] Hello world 1.

[JS] $

apt-get install results in Hash Sum Mismatch

2015-12-13T21:02:00.001+01:00

This resolve for me the "apt-get install" results in a Hash Sum Mismatch

sudo rm -fR /var/lib/apt/lists/* and sudo apt-get update

Domoticz & Nodon Soft Remote

2015-12-06T22:18:00.000+01:00

I'm using Domoticz to set up a Smart Home environment and I choose (i don't know why) Z-wave as connection protocol. To support Z-Wave, Domoticz need the bridge OpenZwave.

All of this is installed on a Toshiba AC100 netbook computer converted from Android to Linux. Quite simple to do following this instruction and also upgrade the distribution to the last available.

To compile Domoticz on the new AC100 box is importanto to growth the swap file size. I use the external SD for the new swap file.

Many Z-Wave device works really well but some trouble with Nodon Soft Remote. I found here the procedure to put it at work. I use the following definition:

adding it under the Domoticz/Config/Manufacturer_specific.xml

After this, only to add the device to the network again and under Hardware->Zwave, configured the Nodon to act as "Scene Activation" rather and "Central Scene" (press button on the Nodon to awaken it and let it receive its config)
under Hardware->Zwave->Groups & Network, add the controller (node 1) to each of the groups for the Nodon switch.

Parrot Jumping Sumo and Arduino Digispark

2015-09-27T19:32:00.000+02:00

Now all the pieces are aligned and we can start doing some interesting things!

We put all together our Jumping sumo, the smallest Arduino (Digispark), a 128*64 i2c display and a ultrasonic distance meter. All fitted on a small proto board under the jumping sumo and connected by a USB cable.

The idea is to add hardware feature to a solid rock land drone to improve the capability without reduce agility.

Till now I can control the OLED display and the distance meter and send this infos to Linux board. The distance in real time is available on the FTP server of the drone

Jumping Sumo with arduino pod
(Front view)

Jumping Sumo with Arduino Digispark, Oled display and ultrasonic range
(rear view)

Some pics of the pod

Jumping Sumo with Arduino explained (1/2)

Jumping Sumo with Arduino explained (2/2)

Only module coming from dx.com.

Bom

1. MCU Attiny85

2.Display 128*64

3. Distance sensor SH04

4. Proto board

Hacking USB port of Parrot Jumping Sumo

2015-09-05T23:28:00.000+02:00

If you want to keep in contact our great friend Parrot Jumping Sumo with the rest of the word exist a lot of possibilities. Use direct connection with I2C but...you need to open the drone and sold some wires on the board or use the native USB port... how? Well...if the device is a real USB device that support CDC interface (like the GPS or a real arduino) no problem at all. Only follow this instruction to add CDC_ACM driver at boot time but...if you want to do more? Or use a ATTINY85 device? In this case you cannot use on linux board the CDC_ACM drivers because the Jumping Sumo linux kernel version dont like the low speed usb interface. The system recognize the device, you can also see it using dmesg but...nothing more than this. Many thread about this problem but I dont want to recompile nothing.
So, use the old classic approach...looking at what kind of libraries are already on board in the drone and try to use it. In this case the libusb-0.1 are in!
Not my preferred lib (is old and no more maintened) but...is already there so why not give it a chance? Try the test code sample in the package. Open Code::blocks for the cross compiler dev env and:
1.Create the project for a console prj and copy the code of testlibusb.c example.
2. Adding the reference in the linker to -lusb (remember to install libusb-0.1 and libusb-0.1 dev in your Linux cross compiler PC) for the DEBUG build
3. Adding the reference in the linker for the RELEASE build to the libusb-0.1 ARM lib using the full path. Here a copy if you don t want to download it from the drone. Remember also to create also a link to this lib.
4. Build the Debug copy and test it...Run like a breeze (but remember...sudo it is a must)
5. Build the RELEASE copy, copy the exec to the drone, run it and...no device found!!!
6. Rebuild it with usb_set_debug(3); after usb_init(); and we can discover the real problem!!!
Detecting USB devices...
usb_set_debug: Setting debugging level to 3 (on)

USB error: couldn't opendir(): No such file or directory
7. The usbfs device dont mount the usb bus standard directory but in /tmp so:
8. Set the env variable USB_DEVFS_PATH=/tmp/udev/dev/bus/usb
9. Export it using export USB_DEVFS_PATH
10. Now re run the .test program and...bingo!
Detecting USB devices...
usb_set_debug: Setting debugging level to 3 (on)
usb_os_find_busses: Found 001

usb_os_find_devices: Found 004 on 001

Now we can talk really at low level to the USB device from the Jumping Sumo !!!

USB CDC for Digispark Serial on WIndows 7 : Now working

2015-08-16T01:07:00.003+02:00

I have had many problem installing the USB Serial Driver for Digispark on my Windows 7.
When I install the driver to test the CDC_ACM library for the Digispark i got a "A Service installation section in this INF is invalid" and no way to solve this problem. After a bit (not so bit honestly speaking...) I resolve this issue adding a simple row in the DigiSerial.inf file.

From this:
[DriverInstall.NTamd64.CopyFiles]
lowcdc.sys

To this:
[DriverInstall.NTamd64.CopyFiles]
usbser.sys

lowcdc.sys

Here my DigiSerial.inf file.Now i can use the digispark mini arduino on windows...next stop, use it on drone :)

Adding a GPS to a Parrot Jumping Sumo Drone

2015-06-20T00:15:00.002+02:00

Part one: Finding device and load the driver

I try to attach to the drone a GPS device an read the GPS position. I use a GLONASS VK-172 USB Gps device coming from ebay but also aliexpress (and please note the price difference).

The interface specs for this device are available here but mainly is a cdc_acm device.

The Parrot Sumo has not already loaded this driver BUT is already availabe in the JS. Looking into cd /lib/modules/2.6.36/kernel/drivers/usb/class there is a nice cdc-acm.ko library.

Now is necessary to load the module using insmod cdc-acm.ko.

Testing with a lsmod you can see the magic:

[JS] $ lsmod

Module Size Used by Tainted: P

cdc_acm 13187 0

uvcvideo 61143 2

asix 11525 0

usbnet 12328 1 asix

wl 935621 0

bcm_dbus 47768 1 wl

ipv6 220695 16

ultra_snd 8432 0

aai 50643 1

snd_hwdep 4617 1 aai

snd_pcm 52836 1 aai

snd_page_alloc 3392 1 snd_pcm

snd_timer 16662 1 snd_pcm

usb_storage 33608 0

ehci_hcd 31168 0

usbcore 105021 8 cdc_acm,uvcvideo,asix,usbnet,bcm_dbus,usb_storage,ehci_hcd

2. Connecting the device...

Connecting the USB GLONASS GPS key using a USB OTG cable (male to female)

Parrot Jumping Sumo with USB GPS Glonass

Waiting some minutes that the GPS lock some satellite (the led must blink) and type a

cat /dev/ttyACM0...

$GPGSV,3,1,10,01,55,300,20,03,29,236,28,04,80,355,07,11,70,297,17*7A

$GPGSV,3,2,10,14,38,071,18,19,54,167,22,27,23,158,10,28,22,301,*70

$GPGSV,3,3,10,32,73,224,18,33,32,213,*71

$GPGGA,200906.00,4513.zzzz,N,00921.zzzz,E,1,07,1.36,57.6,M,47.1,M,,*65

$GPGSA,A,3,19,01,11,03,32,27,14,,,,,,2.65,1.36,2.27*0B

3. Load cdc_acm module at startup

Creating under /etc/init.d/ a file named init_gps

Write in this insmod /lib/modules/2.6.36/kernel/drivers/usb/class/cdc-acm.ko

Make it exec using chmod 777

Modify file init_manager adding

# load gps modules

/etc/init.d/init_gps

Jumping Sumo - Using SDK to create a control program with the JSPilotingNewAPI

2015-05-31T01:32:00.003+02:00

Step by step mini tutorial for Linux operating system (the only one) to run the example demo for Jumping Sumo using the ARDrone3 SDK

I want to test the new sample of ARDrone3 for the Jumping Sumo Drone starting from the git of the Parrot providing in the same time a silly tutorial to avoid many many hours on google to resolve build problems.

1. Set up the SDK library (follow my post here)

2. Download & set up IDE Code::Block from here. Other IDE works /editors of course...

3. Start a new C project and call as you prefer!

4. Add the reference to ALL the libraries of the SDK in "Setting|compliler|linker|link libraries" setting. (Eg: /home/user/DroneSDK/ARSDKBuildUtils/Targets/Unix/Install/lib/libarcontroller.a)

5. Add the reference to pthread lib in "Setting|compliler|linker|link libraries" setting. (Eg: /usr/lib/i386-linux-gnu/libpthread.a)

6. Add the reference to ncurse lib in "Setting|compliler|linker|link libraries" setting. (Eg: /usr/lib/i386-linux-gnu/libncurses.a)
Add also the reference of libc.a

9. Add all the 4 file to the project coming from JSPilotingNewAPI example (JumpingSumoPiloting.c, JumpingSumoPiloting.h, ihm.c, ihm.h)
The code of JSPilotingNewAPI is really simple to read

10. Save it! Here my project file with setting as reference.

11. Try to compile it and fix the error you see (i hope no one)

12. Open the file mplayer.conf in /etc/mplayer and add lirc=no. This remove only a silly warning telling "mplayer: could not connect to socket"

13. Push "run" button and enjoy :)...but on my PC dont work...yet. I v Ask help to Parrot!

UPDATE: Not a code problem in my opinion but...network firewall somewhere in my pc chain...Lost incoming UDP traffic.... Stay tuned...tomorrow I try to address!

UPDATE (AGAIN): Confirmed network issue on vmware. Using a NAT connection from the virtual machine to the host dont work because the UDP has some problem with NAT :(.
Use instead a bridged connection to the host and everything run smooth!

CH3NO2

2015-04-21T21:58:00.001+02:00

20/4/2015 - CH3NO2 - RIP - My unforgettable friend

Sharp PC-1350 & CE 126 P - Found in cleaning of my basement

2015-03-31T22:41:00.001+02:00

I spent a few hours of the weekend to clean my basement and I found a little treasure.

Nothing really valuable but I found the important piece that created my IT skills.
A Sharp PC 1350 (pocket computer using a 8 bit CPU SC61860A13, 40 k ROM, 4 k RAM and a great graphical display), a thermal printer CE 126 P for the pocket computer and, most of all, a copy of the book PC-1350 Machine Language Reference Manual. This book, together with the C K&R, create the basis for all my computer skills...and yes, im a IT professional...:(

Here the wikipedia link

Ok, first of all i need to scan the book and create a pdf. I dont find it on internet and i need to share with the rest of the world this great book.

Full scan of the copy of the 30 years old manual of the Sharp PC 1350 Machine Language Reference Manual. Each page contains good memories and interesting information even now.

1. Introduction
2. System Configuration
3. Machine Language Programming
4. PC 1350 I/O
5. PC 1350 Instruction List (1/2)
6. PC 1350 Instruction List (2/2)
7. APPENDIXES (with Sharp PC 1350 schematics & expansions)

Also found in internet the PC 1350 Service Manual and the printer CE 126 Service Manual

For the lovers of programming, found "only" the PC 1360 Operational Manual, but I'm sure that the basic command are really similar.

I need to repair the calc...no idea why but i need to isolate (look at the red scotch) the door closed contact...but now the calc seems that works!

PC 1350 internal repair

Now... the first great BASIC program to print something...but the printer print only the lower rows of the char...ok...tomorrow open also the printer and try to see why dont print every rows... again...no idea about this problem

Opened...and discoved also that someone sell the seiko mtp201-24ba printer !!!

CE 126 P

After this...i want to ack the printer. Im thinking some like RF commanded printer.

I attach my USBee clone at the printer but...simply dont work. Never mix old tech (>5 V) with new but low cost china tech (3.3 V). I suspect that the 10€ logic analyzer sink too much current from the pc 1350 and...dont work. If i connect all the cable...the printer dont print but i see some like this

If i connect only the Din pins (coming from printer) i see some nice (but useless) pict BUT the printer...print!

Controlling a 0.96'' 128x64 I2C display (SSD1306) for Jumping sumo

2015-03-01T00:14:00.002+01:00

I purchased from dx.com this 0,96'' little little OLED display. There are 128*64 blue dots controlled by a I2C SSD1306, powered by 3.3 V and addressed at 0x78.

I want to test the basic funct to be sure that, when I put it on top of the Jumping Sumo drone, the only problem (if any, i hope so...) come from the electrical connection layer and not on the software side. So, i use the "USB 2 I2C Converter" to try to find how this device work. For sure the chip is well documentated in the datasheet and there are a huge number of good library for arduino & c to take as example...but, anyway...start to got infos.

Here the basic i2c command to control the display

8D --> Power supply to the display (1/2) (needed)

14 --> Power supply to the display (2/2) (needed also)

A5 --> All display pixel ON

A4 --> RAM map will be displayed

AF --> Turn on display

AE --> Turn off display

A6 --> Display standard

A7 --> Display inverted

So, if I send some like 78 (Address), 00 (Command follow), 8D (Start supply the display) followed by 780014 and 7800A5 and 7800AF...well, something appears on the display.

Remember to write "3" on the "data to write" length field and put a "0" in "data to read" length field.

I2C Oled display connected to the I2C emulator

For the command follow value....if is 00 then the following byte are a command, if is 40 the following value is data to be stored on display RAM.

So, if you want to write some data in RAM, use 78(address)40(data follow) and FF(the value for example). Try this sequence....

78008D

780014

7800A4

7840AA

7840FF

7840AA

7840FF

7840AA

7840FF

7840AA

7840FF

7800AF

Now I have to learn to use the tools necessary to generate the bitmap from the drawing.It's il famous bitmap_converter and find a way to push the data to the display...may be tomorrow....

Jumping Sumo - Hardware specs: The beginning of all reverse engineering hacks

2015-02-07T23:15:00.000+01:00

Have a look at the internals of the Jumping Sumo boards. Quite good hardware quality board. Here also every datasheet that i have finded. This pics are realized using a 20 euros "USB microscope" got on Amazon.it, take a huge number of pics an sending to the really nice program "Image Composite Editor" created by Microsoft Research for free !

1. CPU & Driver Board

CPU: Sip6 - Is a classic ARM model (ARM926EJ-S)

Motor Driver DRV8833 - A Texas Instruments (TI) 2A Low Voltage Dual Brushed DC controller capable to control two motors.

Mosfet driver C3021LD - To control in real the motor, the Jumping Sumo use a pair of complementary enhancement mode Mosfet. Really low Ron resistance

2. Sensor Board

Another Mosfet Driver C3021LD

Sensor MPU-6050: This is a classic chip from the Arduino land :) Here we have a motion processor born for the smartphone with a 3-axis gyroscope, 3-axis accelerometer, temperature sensor and a Digital Motion Processor talking I2C language.

DM53801 : Really no info about this large chip. Please help me. etmoimeme report that this chip is in real a PCA9633 4 bit PWM LED driver. Sound really good for me, so tks etmoimeme! I will check and test a bit to add some interesting things to our drone!!!

3. RF board

BCM43526: this is a 5G WiFi 2-Stream 802.11ac Transceiver. Not many more public data but seems really interesting...as an example, this transceiver talk directly to USB
A couple of SKY85803 :WLAN Front End module for 802.11a/b/g/n/ac. Here I need to find the time to got an external antenna to provide a better sensitivity to the drone.

4. USB & Web Cam

USB2513B: This is a quite standard USB 2.0 hub controllers with many many feature inside.
eSP668P: This is the cam controller for the drone,. Is a china chip and I have not found nothing about docs.

Licorice liqueur

2015-01-25T20:02:00.000+01:00

I try to make the "licorice liqueur". Drank last week in the mountains (Presolana) now I try to do it.

Easy, 2.5 h

The recipt is simple:
1. 1 lt of water
2. 80 /100 gr of licorice
3. 1 kg of sugar
4. 1 lt of alcool (95° food use).

Step:
1. Pulverize liquorice (i 'v used a minipimer)
2. Dissolve the sugar in 500 gr of water using a slow slow flame (the water should never boil...it takes patience)
3. Dissolve the licorice powder in 500 gr of water using a slow slow flame (also this should never boil...it takes patience)
4 Wait until the two liquids are cooled
5. Mix together 2, 3 and the alcool.
6. Mix mix mix
7. Put it in a glass bottle
8. Mix mix mix again
9. Wait a couple of months
10 Serve chilled

Work in Progress

Set up & test the "USB to I2C IIC SPI UART TTL Converter" for Jumping Sumo

2015-01-25T14:21:00.001+01:00

To test the software to communicate with the I2C bus for my Parrot expansion, I bought a USB to I2C / SPI / UART converter (aka USB to I2C IIC SPI UART TTL Converter)

This one is available virtually anywhere (ebay, dx.com, aliexpress) for 10-15 euros and seems ok for my purpose. I bought it using Aliexpress

The board

Capability of the USB converter...great!

Some pain to find the sw and the driver, so i copied it on my Google Drive space

What occurred is the USB2I2C program so: need to download also MSSTDFMT.DLL to create the application (in VB !!!), register it with regsvr32 MSSTDFMT.DLL and register also tabctl32.ocx using regsvr32 tabctl32.ocx

Now, I can exec the program to talk to I2C device.Cool! it's work now. But how it works?

Find the spec of the bus converter chip (CH341A) but no real useful information. The real way to understand how to use this adapter is to understand how I2C works:
1.send Address info with read or write bit
2.send offset byte: from which address start the read or the write
3. define how many byte read or write.

Look at the following pic finded on the doc package...everything now is clear...or I hope so!

I2C read operation using the USB adapter

I2C write operation using the USB adapter

Try now to read some and write some info from the simplest I2C device on my collection, a 24c02 eeprom device (buyed from dx.com). This one:

Atmel 24c02 connected to USB Converter

Using the dedicated panel of the control sw works smooth. I can read and write the data on my flash 2 k memory. Now start to read the data using the other panel (most general one) to be sure that I have understud how it works.
Set the length field with 2 byte (all is hex notation), the data with the address of the device. Here start the problem. You need to use always the write addressing (LSB to 0) and use the "data to read" field to choose if you are writing (set to 0) or reading (set to the correct length)

For this example,i want to write "FF" at the first cell of the memory.

1. Put "3" in the "length" field in "write panel"
2. Put "A000FF" in the "data" field (the address of the chip is A0 (10100000), 00 is the offset address and FF is the data to write)
3. Put "0" in the "length" field in data to read field
4. Push "read/write" button...remember..the Data field in "data to read" dont change. You need to read it !

To read some data:
1. Set "2" in the "length" field in "write panel"
2. Set "A000" in the "data" field in "write panel" ( Address in write mode is "A0", offset is "00")
3. Put "A" in the length field of the "data to read" panel

Now some more funny test.
Connect a Clock I2C module based on Atmel DS3231. I need to persist the current data to the Jumping Sumo, so... use this module bought on dx.com than also got a 24 k eeprom based on a 24c32 device and a battery. This is the pic of the module:

DS3231 High Precision Real-Time Clock Module

Try to read hours, minute and second from DS3231 chip:

1. Set "2" in the "length" field in "data to write" panel

2. Set "D000" in the "data" field on "data to wite" panel (D0 is the write address of the clock device)

3. Set "3" in the "length" field in "data to read" panel

4. Got some like "38 15 10" --> looking at the specs of the clock chip, the first byte are second, the second byte is the hours and the last is the day...as usual in hex format.

Try to read the temperature:

1. Set "2" in the "length" field in "data to write" panel

2. Set "D011" in the "data" field on "data to wite" panel (D0 is the write address of the clock device, 11 is the offset's registry)

3. Set "2" in the "length" field in "data to read" panel

4. Got some like "1A 80" looking again at the specs of the clock chip, the first byte are the MSB of the temp and the second...well, got the first 2 bit, convert to dec and divide by 4!
So:1A == 26d, 80==10000000,convert to dec the first 2 bit(10 == 2d) and divide for 4 =0,5...at the end the temperature is 26+0,5 °C

Here the full memory map of the ds3231 device

DS3231 memory map

Adding pod for Jumping Sumo - The Laser one

2015-01-18T19:35:00.005+01:00

Not able to connect something to I2C port (grr...I have not so much time...yet) but I'm trying to add some basic hardware functionality to take the measure with the possibility of my Sumo. The basic idea is not to add many weight but to build a base to add some useful things (well .. nice) .
I start with a stand alone pod. The nexts one will be based on a common hw base.

1. Laser Pod for Jumping Sumo: a 5 mW red dot laser module, a 3.7 V 300mA LiPo battery and an inclinometer as safety switch. Works great...

Laser pod in action : a pic take from Sumo

The laser pod mounted on Jumping Sumo

Side view of the Laser Pod

First trip of the Laser Pod

Some other pics

Jumping Sumo SDK - First contact on Linux platform

2015-01-11T01:25:00.001+01:00

Try to compile the library and create some example code. My end game is to create a Python wrapper for this SDK.

Here the general instruction, but i put here my step by step instructions

1. Update my Lubuntu 13.10 distribution. Missing some needed tools:

sudo apt-get install automake
sudo apt-get install libtools
sudo apt-get install libtool

2. Download the SDK from Git-Hub:

mkdir SumoSDK
cd SumoSDK
git clone https://github.com/ARDroneSDK3/ARSDKBuildUtils.git

3. Make the library:

cd ARBuildUtils
Test if everything is ok using ./CheckEnv.py
then build ./SDK3Build.py -t Unix

The library will be created in /ARSDKBuildUtils/Targets/Unix/Install/lib and the .h header in /home/user/DroneSDK/ARSDKBuildUtils/Targets/Unix/Install/include

4. Start my preferred IDE (code::blocks), create a C project and:

reference the .h file in the compiler setting
reference the SDK library in the linker setting (adding also libpthread)
adding -static option in linker options
copy the example file Unix/JumpingSumoChangePosture and build & run !

Bingo for now...

Running sample for Jumping Sumo

Update: creating a new sample. Great libs and docs but a puzzling problem sending more command in sequence...i v no no idea why, so ... opened an issue. Also the sample

"JumpingSumoChangePosture" in my PC dont work correctly for the same reason

Simple Class A Amplifier JLH 69 (J. Linsley Hood 1969 amplifier) got on ebay

2015-01-04T20:25:00.001+01:00

XMas give me a really good blueray player and my LG TV works well but I realize that audio is not so...edge.

Other pic here for the work in progress...

So, i start to play with a really good Class A schema to fulfill the gap! Here a lot of info about this schema. I draw a LTSPICE model for the Linsley Hood 1969 Class A amplifier.

Using as LTSPICE model for 2N3906 this one:

.MODEL 2N3906 PNP(IS=4E-14 BF=400 VAF=50 IKF=0.02 ISE=7E-15 NE=1.16

+ BR=7.5 RC=2.4 CJE=6.3E-12 VJE=0.75 TF=5E-10 CJC=5.8E-12 VJC=0.75 TR=2.3E-8

+ VJS=0.75 XTB=1.5 KF=6E-16 )

and for 2N1613 this LTSPICE model:

.model 2N1613 NPN(Is=14.1f Xti=3 Eg=1.11 Vaf=100 Bf=88.85 Ne=1.5 Ise=0

+ Ikf=.75 Xtb=1.5 Br=5.591 Nc=2 Isc=0 Ikr=0 Rc=.7 Cjc=15.69p

+ Mjc=.3603 Vjc=.75 Fc=.5 Cje=55.06p Mje=.1553 Vje=.75 Tr=854.5p

+ Tf=1.332n Itf=1.6 Vtf=5 Xtf=55 Rb=10)

and, of course, for 2N3055 this model:

.MODEL 2N3055 NPN(IS=4.66E-12 BF=360 VAF=100 IKF=0.25 ISE=3.339E-11

+ BR=2 ISC=5E-9 RB=3 IRB=0.001 RBM=0.4 RC=0.04 CJE=5.802E-10 VJE=1.2

+ MJE=0.45 TF=8E-8 XTF=1 ITF=3 PTF=120 CJC=2.121E-10 MJC=0.4 TR=2.55E-6

+ XTB=1 )

Some adjust for the R5 value (to 130 k ohm from 100 k ohm to balance the offset value and...ok, start a simulation. I need to discover the real current on TO-3 transistor and...see the FFT of the amplifier.

First of all, please read this amazing post about good configuration of LTSPICE for FFT analysis. Is necessary to shorten Maximum Timestamp (I use 0.000001s) on Simulate | Edit Simulation Cmd | Transient and delete all compression on Simulate|Control Panel|Compression option to got a good graph.

Here what I got.

Really amazing...so, now start with the build! I got a Kit on ebay for a really low cost. The PCB are really nice but the transistor 2n3055 seems like not new...but work and the other 2 transistor are not the one recommended on the original schema.The pnp bjt is an S9015 and the npn bjt is an H669A. Also there is a RC circuit on the output.

The idea is to assemble the ampli, hear a bit and...after this start to modify the component using Hi End one.

First pict of the board

Other pic here for the work in progress...

15.2.2015: First audio check! Only one amplifier module is working (one of the strange transistor is broken) but the other channel works well! I did not have the patience to finalize the installation but...i need to hear the sound. So, really excited. No noise, a LOT of heat...beginning from the rectifier bridge, to the regulator and, for sure...the 2n3055 transistor. There is not enough cooling but I got many ideas.
The sound? I attach to my Ditton Celestion and...paradise, paradise, paradise... Need to hear a bit more, but finally, the amp sounds good. For sure not too much power, but enough! This schema works. I m planning about three more months of additional work...

07.2.15: Hi End Audio Components for JLH69 Class A Amplifier
Try some nice components...
Resistor: Vishay Metal Film Mil spec or High Power for really low noise
Capacitor: Nichicon KC, KA and KG (Type II ...so sorry), Panasonic Metallized Polyester ECQUL / ECWFE and TDK Metallized Polyester for..i hope so, a great sound
Transistor: Trying CentralSemi 2N3906 PNP and 2N1613

Audio grade components for JLH69 Class A amplifier

07.03.15: sound check amazing !!! Not so much power but....quality ! Jethro Tull, Pink Floyd, Status Quo and...Vivaldi and Beethoven.

Class A amplifier

03.01.17: a not so standard case. Just 18 month after my last post on this great subject, here the first pic of "gunny". Yes, this is the name of the JLH69 amplifier in a unconvenctional case for a Class A amplifier. Everything here is really hand made, so...be merciful !!!

Gunny - A JLH69 class A amplifier in a strange case

If some one is so crazy to try to do the same that i did, here the stl files for the 3D printing and some other notes on the capabilitis and the components i use to finalize the class a amplifier.

JLH69 - some detail of the case

Stl model: fan.stl, front.stl, top.stl e base.stl . I use a Sharebot 3D printer without any problema
The temperature sensor is a DC18B20 from Banggood and the heatsink for the TO-03 transistor arrive from ebay from a very good Italian supplier
The wood used as an anti dumper material is a 6 mm diameter "balsa" stuck with Vinilic glue: this simple addon can (hopefully) help a lot the mechanical characteristics of the chassies.

JLH69 - detail on the base with 4 temp controlled fan

Here an in-deep view of the base of the amplifier. There are 4 fan of 40mm used to cool the power transistor and controlled by an arduino controller located on the power module (not ready now). The fan is protected by a standard grid.

JLH 69 - Heatsink for 2n3055

A view on the heatsink part of the amplifier with the TO-03 socket (i want to test different NPN power transistor other than 2n3055 like KD503 Tesla or MJ15003 on semi in a really short time).The secondary little heatsink directly mounted on the body of the 2n3055 is really effective and can lower a bit the temperature of the transistor.
)
... tomorrow other pict and infos...