Maker Pro
Maker Pro

Make experiments safer, my workshop and concepts relevant in my opinion

Hi everybody. Upfront I want to let you know, that due to my serious health problems I am slow in progressing in my experiments.

29895094226_5a3f6de915_z.jpg


Its know quite a time ago that starting to experiment and learn how to do it I did use PyCharm IDE from Jetbrains to program my experiments on the RaspBerry Pi's. I had no experience in using Linux and never really went in depth into the topics around communication. I do have a PC with Windows and do use it as my development workstation. In the graph here that I did I show the IT infrastructure I am establish in in my private lab. I do sit on my PC and have the RaspBerry Pi's contacted via WiFi. The small antena symbols show this. My first hand goal is to develop y sheet control system for my model sailboat, to model critical parts of it using the modelling language Modelica and Mathematica to do HiL and SiL,to verify that the models actually reflect the true behaviour of the subsystems modeled.

9774113161_368d60fd87_c.jpg


The text in the graphic was targeted for spanish speaking audience, sorry. So deviating from the hardware modules shown in the block diagram I am using for my experiments a RaspBerry Pi board to control the LPCXpresso board listed.

So I had no problems to have the Raspberry Pi desktop to show up in a window on my PC, then using Windows 7 Ultimate 64 bits. I did use ssh, secure shell to establish the communication between the PC and multiple RaspBerry Pi boards. I was also able to have the PyCharm IDE running on my PC, that generates Python code to be executed on the RaspBerry Pi, short Raspi, by defining as the remote interpreter to be the one running on Raspian Jessie on my Raspi boards. I also do want to mention that I do have my Windows constantly updated and have powerful protection software suite in place. The starting point of my failure was to publish de Raspi boards on a DNS supplier in the internet, so that I was able to have the DNS to assign the proper IP ID to a communication.

Via the Internet and the Raspi communication link between the Raspis and my PC the attackers were able to render my PC inoperable! The good in the bad was that as all my intents to repair Windows on my PC failed, I did take advantage of Microsoft's offer to have my Windows OS updated to Windows 10 Pro , which then did fully operate again. Just for your entertainment here a picture of my electronics lab which is next to my office desk in my workshop:

28576480456_74de4e72df_z.jpg


I am working since many years in developing my workshop to get it to a degree of order so that I do know what I have and where it is.

Let's continue with the main story of this thread of mine. Having suffered this successful attack I did decide to learn what is involved to understand how to have my IT infrastructure to be as safe as reasonably posible. The link to the Internet has to be kept as I do want to have access to my sailboat model via TCP/IP. I am using the term TCP/IP and not i.e. WiFi or WLAN as with the technological advances around the IoT and IIoT or Industry 4.0 are making huge improvements and many offers how to establish a wireless communication. My phone provider i.e. offers for 1 Euro a month 100 MB of data volume, much more than a communication between my model and myself on a pond would need. Exception being if I want to have video streaming capabilities. Here I can book higher data volumes for just a single day. So this possibility to use celular data communication for my model is now realistic.

So I started learning Linux using books from a german university profesor seen as a guru and valuable source to learn facts about Linux. I did also start learning about the technologies to reduce potential attack surfaces in general and in embedded systems specifically. The MOOC's, university courses available for free are excellent sources to learn. Soon I run into realizing that profesional developers frequently do their development work using virtual machines on their workstations. Virtual machines, short VM's, do isolate the host, in my case a PC running Windows 10 Pro from the VM. Should an attacker be successful in attacking my experiment setup I just have to close the VM and reopen with a status previous to the attack. In the Windows environment 2 major suppliers of virtualization tools are Oracle with VirtualBox and Hyper-V from Microsoft. I started with VirtualBox to learn that VirtualBox does not work in Windows 10 if the BIOS setting of a host PC has the Hyper-V services activated. Hyper-V are services that take advantage of hardware functionality in the Intel CPU for virtualization, a key factor to establish an efficient VM! There was also another issue linked to have the Hyper-V services activated, but later more to it! So I tried Hyper-V and could generate a VM with Ubuntu 16.04 TLS running within it. I had also no problems in having the PyCharm IDE installed on this VM and the Jetbrains licensing scheme here is very open to have the multiple instances of PyCharm running in parallel or at least available both on the Host OS as on the Linux running within the VM. Here is the point where I am right now in my efforts related to this.

Virtualization is a mature technology in the server arena and is being used there for decades now! Having virtualization applied in the embedded arena, mainly for security reasons being mandated by governmental laws for devices connected to the Internet is a different story! System resources in embedded systems are scarce by definition compared to those in servers! This made me aware of the functionality of hypervisors, type 1 and type 2 exist.

32253679192_6949a3798f_o.jpg


As the picture illustrates, hypervisor type 1 execute "bare metal". The term meaning running directly on the hardware, same as any code we in electronics write for Arduino i.e.! Xen being one of those hypervisors. Type 2 hypervisors do run as an application on top of a host OS, in my case Windows 10 Pro and the program Hyper-V having the role of a Hypervisor type 2. As guests are meant VM'S! The picture makes evident that a Hypervisor Type 1 can be implemented demanding much less resources. For my experiments I will experiment with y type 2 kind of solution running on top of Raspian Jessie on my Raspis! If this is properly executing y am considering also running VM on my Raspi for safety purposes! But that is on my agenda for my hobby! Studying virtualization in depth under the perspective of implementing security made me aware of another technology that can realize even implementations with demand for less resources, is called Containerization and the tool is Docker.

32026995660_84ea354840_o.png


This picture illustrates the concept of Containers versus virtual machines! Containers generated using Docker run on top of the host OS and of the Docker engine, but each container does contain only those resources needed to execute while not replicating the entire OS as it takes place in a VM! But while VM's offer the best isolation and in consequence the higher safety compared with Containers whose isolation is more similar to that of one application to another application being executed within an OS. Well known to be much weaker. Due to the paramount importance of safety in devices connected to the Internet and how scarce hardware resources are in an IoT compared to a full fledged PC, a hybrid is being researched for to combine the benefits of both.

33729468036_0c4c84282c_z.jpg


So conceptually I am aiming to implement what is shown in this graphic! I will now go into the details as it would be even more extended is this contribution already is. But the goal is to benefit to the degree reasonable and performant enough for my objectives both in my electronic workshop and for the implementation of the electronics in my model sailboat! But I do want to mention that as not being an expert I was able within less than 30 minutes to generate a Container, write a PC of code in Python for it and generate the "Hello world" message on my PC!

Taking advantage of a special offer by one of my preferred publishing companies, Packt, I could by at a very low price 5 books that cover a wide area of information regarding Docker and Containers. A lot of safety can be achieved in a Container by being very strict in defining parameters that define who might engage a communication and what each task absolutely requires to be authorized to. This applies to the OS itself, to the VM's and to Docker. Well defined parameters allow to make penetrating an expensive task. Not really having anything worth the effort, attackers might decide to go for more fruitful targets!

Finally I want to include in this contribution that the requirement of safety is making research and implementation of technologies to achieve safety in devices connected to the Internet an activity with very "deep pockets"!

32458738165_8e45d16a38_o.png


All major providers driving the development of technologies have their way to it, but as devices in the IoT and IIoT dominantly are based on controllers based on ARM technology. The core of this technologies is to have 2 different environments, Trusted, here in green, and Non Trusted, here in red. Non Trusted is microcontroller functionality as we all are aware of. If we take the Raspis as an example. Linux runs in 2 privilege modes. One being the user mode as we know it from any operating system and one is a privileged mode in which Linux kernel is executed and which enables the OS to administer the use of system resources by an application and protecting the execution of the applications by preventing them to interfere.

For the Trusted Zone ARM has defined a new even higher priority level and its execution alone allows the use of special instructions which operate on the special hardware resources available within the "TrustedZone" has the concept is called by ARM. A hypervisor is the one that executes code in this highest privilege mode that allows for efficient operation of VM's! The first 2 ARM Cortex M controllers published and licensed by ARM are the ARM Cortex M22, which is equivalent in performance to the ARM Cortex M0 and M0+. The ARM Cortex M33 is the one equivalent to ARM Cortex M3 and M4! The one most visible to me is the implementation of this technology by NXP Freescale and its i.MX8 product family. NXP is mainly targeting the automotive marketplace with this products and as a consequence the i.MX8 that comes in the following flavours

33549503331_7e5f241e06_o.jpg


represents a device of a complexity far beyond what we are used to. This is on one side reflected by the hardware functionality shown in the table and can be, to my believe. better visualized seeing their Multisensory Enablement Kit offering:

IMX8-MEK-BD.jpg


This complexity is due to the fact that it enables to create one central Unit for multiple functionality present in current top of the notch cars and those to come in a relatively close future. Development requires different engineering skills for different functionalities offered within the controller! Former Freescale announced this controller a couple of years ago and is closely working with selected tier 1 customers to use their device in the automotive marketplace. Just very recently a company that offers inexpensive boards using the previous generation controllers, i.MX6 and i.MX7 has announced to have a board in its pipe with the i.MX8.

But this complexity also demands from NXP to deliver IDE and libraries and application notes that make the implementation of this multiple functionalities possible to smaller engineering groups. While NXP had announced the availability of its i.MX8 devices sometime in 1Q17, this has not yet taking place. No wonder as to be ready to help interested customers of tier 2 and beyond demands a different class of support that has to be equivalent to the level of support we have been familiar with other products in the market. I believe this to be the reason for the still continuing delay of the general market to have access to this product!
 
Finally yesterday I was able again to continue my work on this that my health problems have prevented me to do.

I did create a VM using Hyper-V within my host os being Windows 10 Pro and install another license I own for Windows 10 Pro in this VM. Here I was able to define a standard desktop size which is equivalent to my screen with a lower resolution and activate the support for using the whole desktop available by the combination of 2 displays I have available on my host W10Pro. The benefit of this is that when I select the option on the right top bar for expanding the desktop to the whole available desktop the desktop of W10Pro running on the VM, the screen expands to cover both screens. When I do select the opposite the window size of the VM fits to the lower resolution display. i had tried to accomplish this on a VM with Ubuntu, but have not yet succeeded to achieve this. The reason is that the VM with Ubuntu is not that well supported by the hyper-V VM!

But during this exercise I found the response to my question of what were the differences between Docker on Linux versus Docker on a Microsoft OS. Docker requires the availability of a Linux Kernel and so Docker on Linux uses the Kernel of Linux just similar to the way normal programs under Linux do. In Windows 10 Pro and in Windows Server 2016 the installation of Docker does generate a VM called "MobyLinuxVM" which you see listed when opening the Hyper-V Manager. Docker than uses the Linux Kernel in this VM when it runs i.e. in Windows 10 Pro.

To recall. The reason for my efforts regarding virtualization and containerization was to implement a working environment for my future experiments that make the repetition of the successful attack I suffered when starting to experiment the use of Raspberry Pi boards. attackers rendered my the Windows 7 Ultimate 64 bits OS inoperable and I luckily could recover by updating the OS to Windows 10 Pro.

Now what I have learned yesterday afternoon and night was that using as my development workstation my PC running W10Pro and Docker actually benefits by using W10Pro also as the native OS in my Hyper-V VM. The key term is "nested VM"!

My goal is to run my IDE PyCharm for Python software development and CLion for C and C++, both IDE from Jetbrains, from within a VM on my development workstation and within a Container made using Docker. Using W10Pro, both on my PC and within a VM. The Container generated using Docker on the W10Pro within the VM runs the container within a so called nested VM within the VM!

I did find yesterday while investigating how to have the VM "MobylinuxVM" running within the VM the places in the Internet that instruct how to do it! But yesterday, and even today, I have not been able to continue the implementation. My damn health problems! But writing this I a way to develop my understanding of the topic better. I recently read, that Alzheimer's early effects are a loosing of the motivation to work on hobbies. So perseverance in pursuing this gives me the good feeling I am not yet in Danger of Dementia.

I believe I will have to activate the Hyper-V related services within W10Pro running within the VM, as I had to do in the W10Pro copy running on my workstation. To be able to do this I will have to activate the services related to nested VM support within W10Pro. If then I succeed to install Docker on the W10Pro within the VM the Hyper-V Manager of the W10Pro within the VM will then list the nested VM "MobyLinuxVM" as running proving that Docker was successfully installed on W10Pro within my VM.

While installing W10Pro on my VM running on my W10Pro workstation, both W10Pro copies were updated, which made the process very long. But in the process of installing the updates of W10Pro the wizard already offered me to make decisions that effect the security of my W10Pro copies. Microsoft has this support of Docker and Containerization a "work-in-progress", so that support for this functionality is constantly improving!

What I will have to do after having been able to generate the Container with Docker on the W10Pro VM is to make a list of all the parameters that affect the security. I had purchased a series of books related to Docker from the publisher Packt that addressed the aspects that impact security in the context of Docker. It has a lot to do with defining strict parameters about ports being used, what container is entitled to do just what it explicitly requires and to limit interactions between containers by explicitly defining source and target container of any communication and what they are entitled to do. It adds to defining none standard ports through which this communication takes place and to apply strict rules as to who is authorized to scan ports or do communication. It goes on defining explicitly what file systems and what part of the who is authorized to address and with which rights.

As I am talking about nested VM's and containers I have to get a very complete understanding of what is involved and who and how a participating software is operating. Once I have defined a list that lists all the parameters and their required settings I will have defined a workflow for the workstation part of my development environment. Then and just the I think I will have the understanding and knowledge to define a properly configured firewall for my workstation. Really what I have been and still have a long way to do is to get a detailed understanding of the "big picture" of my pursued safe-as-possible working environment. As the target of this object is a moving target due to the importance of security in devices connected to the Internet in an ever increasing hostility of the Internet environment, I will have to keep track of those developments and apply modifications to the "big picture" I am pursuing. Unfortunately I will not be able to publish either the list I have been mentioning to develop nor to many details about the parameters and their properties not to make available valuable data to potential attackers that might be attracted through this thread. I do repeat. This is just for personal entertainment and their is no valuable data behind all of this. So the cost of investing resources to crack my safety architecture does not provide any benefits that would justify the effort.

But as part of my studies and investigations regarding this topics is also the study of "penetration testing"! So monitoring my environment to identify intents to attack me are also part of what I am developing for my setup.

But once the above has been accomplished enough to start to reengage into my experiments with the RaspBerry Pi boards I also have to address the implementation of proper safety setups to those Raspi's. I have obviously already engaged into studying the involved topics as you can read from my earlier published contributions to this thread. I do expect that while implementing a secure as possible setup with Raspi's, those will migrate to a board using the new i.MX8 device mentioned in my previous posting in this thread. I will hide my embedded experiments setup behind a central device using the i.MX8 that will probably be using Yocto as the Linux OS. This way as of today I will have a state-of-the-art setup that I later will be using in the electronics for my model sailboat for which I am pursuing all this activities!

I believe that it is my obligation to this and similar forums to repeat that this intensive dedication to more data processing topics are a fundamental part of my work to model and design the electronics for my model sailboat, the "Carina" I am building from scratch! Security is becoming an essential aspect in designing electronics in systems connected to the internet in general and in embedded systems even more. IoT, IIoT and Industry 4.0 are application areas where embedded systems connected to the Internet are demanding best possible security at the design phase already. Autonomous cars and drones are an example of this!
 
No one is going to read you 5 pages or so of text and pictures.

If you have a question you want answered, please post it in a concise form.

bob
 
Dear BobK. I am not publishing to get an answer to a question. My intention is to share with those interested in the topic I am dealing with my activities and maybe share experiences. With no intention to offend you, but your response is the prove that you have not dealt with the subject! The subject of security in embedded systems and the need to take security into the design phase of a project of embedded systems connected to the Internet is a very important and critical demand that impacts any IoT or similar device project.

My thread addresses my efforts to implement my electronic workshop that implements means to achieve the maximum level of security to prevent to happen what happened initially to me and that rendered my PC inoperable! As a side effect, the topics do fascinate me and so I am embarking in it with a lot of fascination, curiosity and perseverance. I hope this reply addresses your posting and explains why I am doing it without any intention to offend you!
 

davenn

Moderator
probably was in the epic category
where a 3 part novel would have been better

But if anyone wants to read it all, good for them ;)


Dave
 
Sad that, different to other electronic forums I am just meeting as of now people that spend the time lecturing while obviously not having nothing to contribute! I hope there are adequate people in this forum! There are never many but those i find help me to learn a lot!
 
Well even sailboats sink when they get holes in the hull...

Your studies are supposed to belong to you, not the board.
I am still proud of you. I think that is a lot to share.I have never used linux.

With TCP-IP you are talking about building a dialup into your sailboat, and then having a set of arduino codes that operate it, and that means you will also need an ActiveX control pad that is browser enabled to control the boat.

That is a tall order. It has been done. but it is not as stable as it seems, sometimes the servers shut down for maintenance and after say 6hrs turn back on and stream all the commands they have recieved. Sell imagine your boat in the mifddle of the lake going round and round until finally it straightens up and runs right straight into a rock, rupturing the hull and slowly sinking into the lake near the shore, where some kid finds it about 30 years later and becomes "the bomb" of wired-to-wireless drone and prone control?
 
As I wrote, one is enough to continue, thanks! You are right, my studies and experiments belong to me and it is my hope that by sharing them openly I can return the help I have received a little bit!

As you addressed a kind of Top View" I would like to respond as I do believe electronics has a lot to contribute to naval modeling, much easier then with building planes models. The latter are very sensitive to malfunction and timing, naval models are more forgiving and can contain an infinite number of functionalities that electronics allow to build the way the modeller has in his mind. Benefitting for the possibilities of communication development this was a relatively early topic I have dealt with. I warn you, as always for me, I am lengthy!

I started and still own a R/C system from the german company "robbe" that operates as required by law in Germany at 40 MHz:

RH1-F4039.jpg


This R/C equipment, the F-14 does exist in a special naval version that has one of the control sticks replaced by:

RH1-F4039_1.jpg


For sailboat models this is perfect as you can control the position of 2 sails independently, putting the 2 sticks next to each other you can activate a switsh that connects both sticks to operate as one.

RH1-8084.jpg


and for proportional controls:

prt1512.jpg


I do own the one on the bottom that has 8 rotating knobs. To say it directly: a total mess! One of the first things I did when I build the first device to digitize the PWM signals coming out of the 8 channels of the receiver was to analyze it. What it did in a unit that you plugged into the receptacle of the receiver and which in turn offered 8 receptacles to connect i.e. servos it, and believe me the price was outrageous, was to send in 8 consecutive cycles of 20ms each, as usual for R/C receivers, the control PWM for the individual knobs and repeating it. The unit connected to the receiver just generated a PWM extended to 160ms. the result was bad operation of the servos connected and nearly no torque. All I had to do was to leave the even more expensive unit in a box and make a small module with a mega8 controller that read in the signal of the receiver and so manage to generate 8 PWMs at his outputs, each with the correct 20 ms cycle PWM and just updating the PWMs individually every 160 ms.

So having wasted a lot of money buying those expensive units from robbe was the starting point to make my own electronics for my naval modelling.

So in my R/C transmitter I did install the 8 proportional control knobs and the 8 multiswitches in the 2 areas left and right of the power indication of the transmitter.

The the 2,4 GHz R/C systems changed the technological platform of R/C systems making it possible for chinese companies to offer R/C systems and ridiculously low prices on one side and on the other side the vendors of R/C systems to offer ridiculously high priced computer transmitters and expensive units to interface with actuators and sensors and to display the data generated on their small black/white screens, still characterizing their R/C systems by defining how many "channels" their transmitter were now able to offer! What a marketing scheme to continue charging premium prices to their customers, when the 2.4 GHZ technology is a serial bi-directional data stream"

So my concept was to combine a 10" tablet placed in landscaped position, today that could be a smart phone with electronics, so that still such a transmitter has the familiar control sticks, 1 to his left and 1 to his right. In a row above that display 5 switches, in a row below 5 rotating knobs. The display would best have an Android OS and the display device a USB port to connect the external circuitry to the display. Android is better than other OSs as right from its provider it offers the tools to have an app running on it communicate via the USB port to external circuitry. Android when I started looking into it was still in a 3.x version but it was foreseeable that later versions would be even better suited for my purposes. Now comes the great thing about it:

The user would have HMI with the functionality he is used to from smartphones and tablets and prices for tablets would be falling dramatically, as it actually happened. Now the more limited number of switches and rotating knops would be assigned their functionality depending of the actual window being active. So, as a single motor plane pilot, multifunction displays in the cockpit was being discussed in those day. The screen has a "main window" that allows full control and small windows that just present key information. Should one of the auxiliary windows require the attention of the pilot it would either signal it by blinking or would take the place of the "main window". So those "soft switches and knobs" make it possible to assign them unlimited number of functions, depending on the "main window" being active. As tablets were still too expensive in those days I did change the focus of my activities to electronics within the sail boat model.

To your reference to Arduino and those kind of boards. In those days there still was no Arduino and much less RaspBerry Pis. So I build myself in a tiny space left next to my workbench an electronics lab:

4144940900_ca71e9b857_z.jpg


As you can see there are 7 power outlets and a switch. I do believe you have never enough outlets!

4163215980_7a301e7682_z.jpg


Here you can see "my experimental board" expanded. A friend of mine developed a tutorial to make it easy for naval modelists to take electronics as another technology offering infinite possibilities in their hobby and to overcome the intimidation that electronics has on many of them. Interestingly the modelists of planes did jump in much bigger number on this tutorial.

5013795345_80138b4113_b.jpg


Here addition I made to my workbench, today called my old workbench to store parts.

Todays look of my electronics lab was already shown in a picture earlier in this thread!
 
A lot has happened since3 my last contribution. Bad is that I have by now survived my 4th stroke. Thanks to medicine I take to make my blood thinner, the last 3 strokes were light. But often being tired and difficulties to get concentrated are delaying my work on my projects around my sailboat model Carina and the electronics and informatics topics around it.

As a word of warning: Nobody is obliged to read my lengthy contributions! If the content interests you, great I am here to exchange opinions, suggestions, and critics.

I am continuing starting from what I have already presented:

28576480456_74de4e72df_z.jpg


Wandering around the internet I found this item:

41536434242_466a4a8b6a_z.jpg


In the first photo, you see my electronics workbench and on the left side of the desk what I call my panel 2nd generation. This panel is fed from my modified pc power supply offering me a total of 8 different voltages. Seeing the above device I decided to have one of these devices above the receptacles for each of the 8 voltages supplied. My first design idea is shown in this photo:

41499568682_0b80aa54a6_z.jpg


But moving on with the project of a 3rd generation panel I decided I did not like a solution the way shown here! But even more, the serious reason was that the display with the current and voltage monitoring did measure the current on the negative pole. My 2nd generation panel placed the receptacles of the positive poles and had the 2x 6 black receptacles used jointly for all voltages.

41374099972_ec15782c72_z.jpg


This means I have to make ground receptacles for each of the voltages of my panel. Something like this:

27670963478_3ddfb2d486_z.jpg


My panel 2nd generation had 1 colored receptacle below a switch that was always available, while the receptacle above the switch was turned ON/OFF by the switch. Additionally, I had 5 screwable connectors on top of the panel and added a bunch of negative pole connectors on the left side.

43075699401_3a496f03b0_z.jpg


So I decided to make the panels 3rd generation as shown in the work in progress above. I did now have 2 connector pairs below the switch and 5 above controlled by the switch as it was done in the panel 2nd generation. As you can see in the photo now there is enough space on the panel to have a display above each of the voltages. Also, you see that now there is always a positive and a negative receptacle next to each other. To the right of the switch, there is a 5 mm diameter hole in which I will place an RGB LED. The screwable connectors are now placed below the panel. 6 screwable connector pairs, above the positive pole, below the negative pole. The RGB LED will get on when the switch in its "ON" position.

41560543904_30c14e8fa9_z.jpg


serveimage


Each RGB LED will display the color coding of its respective voltage. I have chosen the color coding from the color coding for PC power supplies. 8 RGB LED has each 3 pins that need to be driven by adequate PWM values so the proper color is achieved. The above photo is a board that offers 16 PM outputs. So with 2 of these boards, I can generate the PWM 24 signals required. These boards are controlled by a single RaspBerry Pi ZERO W. I take this Raspi because I Want to have a window on my Windows 10 Pro PC desktop to control the RGB LED, not just the color, bad also be able to dim them. So the Linux Desktop running on the Raspi will be in this windows on the PC with an application running on Linux to control the RGB LED remotely via the WiFi functionality. Additionally, this Raspi will control and monitor the electronic fuses I am building using MOSFETs. They will be monitoring the currents flowing from the power supply to the panel using the resistance of the MOSFET and switch the MOSFET OFF if a value I define an upper limit for the current is reached.

Sadly my health problems are delaying my progress on the making of this panel.
 
Keep fighting. It may be paranormal whats happening to you. Im also getting attacked, its a sign that u are a powerful person, that god or whatever monkey upstairs whats to put you away!!!
 
For someone with limited workshop space your approach towards a power supply seems rather confusing.

A power unit with dual VARIABLE outlets would be as much as you might need for all your experimentation. Many experimenters workshops will have such a supply AND NO MORE.

Individual systems (that you may be building) would normally have a single source to power them anyway and any sub-supplies would be derived internally (on board) from that single source.

It's called 'simplification' and/or 'minimalising' - something your work seems to be missing! I see (and read!) that you have over engineered the potential solution by a long way. All you're going to do is make construction and fault finding exponentially more complicated and time consuming.
 
@ratstar: I do not give up. The different topics I am dealing with do fascinate me and this is the way I do my hobby. Thx.
@kellys_eye: You are absolutely right... if my purpose was finishing something and do it efficiently! Let me respond to what you write in your last paragraph: What characterizes a German engineer? Why do it simple when you can do it complicated. And so is the issue with all my projects: It makes fun and it gets me in touch with topics I would have never met if I would not do it my way! And you know what? I find my panel 3rd generation nice. Just take as an example my dealing with security in embedded systems. 3 years ago I went to a trade show called "Embedded world" in Nurenberg. Due to my researches and studies and experiences in my own workshop, I knew that security would become a major issue. In the show, I just met one engineer that knew what I was talking about. A year later this topics was in everybody's mind at the same trade show. For more than 3 years I am studying and experimenting with the topic of security. Sadly my health problems do slow my efforts down more and more.
By the way, you are the first to name my workshop as having limited space. Surely because I have not presented it here yet.

Here a photo of my electronic corner of my workshop
21452150814_f971940121_z.jpg


Here nor with the boxes, I made to have my stuff stored.

35182074166_67aab24403_z.jpg


Here a picture of my tower milling machine. It is, together with my lathe what I call my new workbench:

8616405238_624cd960d3_z.jpg


Here a picture of the drawers in construction below the tower milling machine.

44170897224_2e3d3ee7d6_z.jpg


here the drawers below the lathe in my new workbench:

29954460447_db8ff32b7f_z.jpg


Here my lathe:

34398424444_41b49d5ee4_z.jpg


Here a photo from the center part of my so-called old workbench:
5013795345_80138b4113_z.jpg


Here space below my old workbench where I made a "hangar" for my sailboat hulls:

5013794883_de9e84fae8_z.jpg


Here the door closed:

5013794557_b1b1d56a26_z.jpg


Here a photo from the right side of my old workbench. By now you know I love drawers as much as I love power outlets:

4144943100_581687d269_z.jpg


Here the left side plus to top of my old workbench. On the left of this photo, you see de predecessor of my electronics workbench:

4144941988_5fd01bdcfe_z.jpg


Here the shelf to store screws and other stuff in a visible manner to know what choices I do have due to my existing inventory.:

41525492682_cb33534e1c_z.jpg


Here my "office bench" that is to the right of my electronics workbench:

26416442868_fed85e98db_z.jpg


Motivation to get an order in my workshop is that due to the fact I cannot earn money anymore I have to change my habit. I use to buy something that I knew I had because I could not find it in my chaos.

As you have recognized by now, I am proud of my workbench. And due to my health problems, I do switch from one of my projects to another one that I feel I can do. I also know that when I pass away, not too far from now I fear, most of the content of my workshop will be put into a trash container. My wife asked me the other day why I was putting so much effort into developing my workshop? I said that I do so as if my life would go on infinitely. or should I sit in a corner and wait to die? Also writing this responce is part of my activities.
 
Of course we must all fill our lives to make our presence here worthwhile but a load of half finished items will never succeed you as a fully completed item would. A legacy if you will.

Regardless, we each do what we want in order to make us happy while we are here - I certainly do! and my workshop is, for much of the time, as hectic and in disarray as yours looks to be. Suits me. Suits you. Great result!
 
Seems you are having fun and enjoying life. Who cares how long it takes and whether a project ever gets finished.

Not a fan of John Lennon normally, but I did like his quote: "The only time that is truly wasted, is time you do not enjoy" (or something close to that)

Sounds like the perfect life to me (except for the serious health problem bit of course)
 
Top