Author Topic: Just lost my job. Finishing up unfinished personal projects before heading out.  (Read 2465 times)

0 Members and 1 Guest are viewing this topic.

Offline technix

  • Super Contributor
  • ***
  • Posts: 3223
  • Country: cn
  • From Shanghai With Love
    • My Untitled Blog
So I just lost my job. The compensation package is okay for me to live on for one month or two. I am thinking about finishing up some of my unfinished projects before heading out again, either to graduate school, or to the next job. This post is mostly a checklist for me, but I feel like sharing it. Since I have bought most of the parts already (sunk cost, at least the major parts) the only cost is my time and building the PCB.

Some of the projects that is waiting for me to finish:

The climate clock project.

This is from some home necessity as I need some kind of climate sensor in my room. The hardware is actually already built and the code is under progress.

Specifications:

* Measures: time, temperature, humidity.
* Timekeeping: Crystal oscillator with NTP synchronization
* Temperature and humidity: DHT12 sensor
* Main microcontroller: STM32F303CCT6 (or STM32F103CBT6, if I want to rewrite some drivers.)
* Display: HD44780-comaptible character display
* Connectivity: Wi-Fi 802.11b/g/n 2.4GHz through ESP8266 module, USB Full-Speed

Github Links:

* Hardware: https://github.com/SushiBits/LCDClock
* Firmware: https://github.com/SushiBits/LCDClock-Firmware

Apple EarPods to Bluetooth adapter

Apple removed the headphone jack since iPhone 7. I use my phone heavily while listening music on it, so battery life is a real concern. This project builds an adapter that converts Apple EarPods (ones with headphone plug) to Bluetooth, with full buttons support.

Specifications:

* SoC: CSRA64215 (Bluetooth audio with apt-X)
* Audio codec: either the one built into CSRA64215, or CS4272 over I2S. Which one is better?
* Audio analog chain: Either TLC2272/TLC2274, or TL072/TL074, or the built-in ones in STM32F303CCT6. Which one is better?
* Helper microcontroller: either STM32L031F4P6 (without built-in op amps) or STM32F303CCT6 (with built-in op amps)

The Allwinner V3s tablet project.

The linux-sunxi project have made some tremendous effort in getting the mainline Linux kernel ready for V3s, and I have sourced all parts except a case. This project will result in a low cost tablet that runs a custom, cut down Linux distribution, based on the linux-sunxi mainline branch.

Specifications:

* SoC: Alwinner V3s (Cortex-A7 MP1 @1.2GHz)
* RAM: 64MB DDR2-533 (built-in)
* Flash: 64MB SPI NOR FLash
* Display: 1024x600 7in with capacitive touch
* Audio: Headphone jack with microphone and Apple-compatible buttons, analog line-in
* Camera: 5MP rear facing camera through OV5640 on MIPI CSI2
* Connectivity: System USB OTG, Debug USB-Serial, Bluetooth 4.0, Wi-Fi 802.11b/g/n 2.4GHz, wired 10/100 Ethernet, microSD card slot

The Mess-of-RAM-wire project.

This is more of a design challenge than an actually useful project. The goal of this design challenge is to route the 32-bit wide memory bus to four 8-bit devices successfully, and secondary goals includes implementing DRAM bus-sharing peripherals like NAND and display controller, as well as writing proper framebuffer (not fbtft) and touch sensor driver for the display controller.

Specifications:

* SoC: AT91SAM9260 (ARM926EJ-S @180MHz)
* RAM: 256MB PC133 (4x MT48LC64M8A2-75)
* Flash: 2GB SLC (MT29F16G08ABABAWP)
* Display: 800x480 4.3in with resistive touch, through RA8875L3N on 16-bit system bus
* Audio: Line out and line in, through CS4272 on I2S
* Connectivity: 4x USB Full-Speed host port (integrated USB hub,) USB Full-Speed device port, Debug USB-Serial, Bluetooth 4.0, Wi-Fi 802.11b/g/n 2.4GHz, wired 10/100 Ethernet through DP83848 on RMII, microSD card slot

There are also a few less speced out but still planned projects I may or may not complete right now:

* Vintage Computer Project x86 (8088 + FreeDOS?)
* Vintage Computer Project 68k (68SEC000 + what OS software?)
* DIY network-connected test gear set (multimeter, bench power supply, soldering station etc, all with either wired Ethernet or Wi-Fi connectivity)
* Open Source GPS Car Tracker (I don't trust any service provider for the location of the family car. This system is fully open source and you can host your own tracking server.)
« Last Edit: October 09, 2017, 07:33:43 pm by technix »
 

Offline Corporate666

  • Supporter
  • ****
  • Posts: 2004
  • Country: us
  • Remember, you are unique, just like everybody else
Sorry to hear about losing your job - but I don't feel sorry for you, because based on the list of projects you are working on, you clearly know your stuff very well and I have no doubt it will be easy for you to find a new job with great pay.

I'm always looking for additional business for my company, and I sometimes peruse the job ads and contact the employers to pick up design/manufacturing work as it's a fertile hunting ground for such work.  I've seen a huge increase in jobs in automation, rapid design/development/prototyping, robotics, etc.  There's tons of money being thrown into these fields and I think that will continue for the foreseeable future.  My point is that if you're looking to earn big $$ and be in a growth industry with less chance of layoffs, these industries might be worth getting into.  You clearly have the talent for it!

Interested in seeing the progress of your projects too - they are interesting.
It's not always the most popular person who gets the job done.
 

Offline Bondguy

  • Contributor
  • Posts: 41
Interesting projects.


But not sure who or why should use:
* Vintage Computer Project x86 (8088 + FreeDOS?)
* Vintage Computer Project 68k (68SEC000 + what OS software?)
Do you know that?

* DIY network-connected test gear set (multimeter, bench power supply, soldering station etc, all with either wired Ethernet or Wi-Fi connectivity)
But why should I have/use soldering station, power supply  connected via Ethernet/Wi-FI?What's point in that?

Where do you plan to use The Mess-of-RAM-wire project?


 

Online NiHaoMike

  • Super Contributor
  • ***
  • Posts: 5638
  • Country: us
  • "Don't turn it on - Take it apart!"
    • Facebook Page
* DIY network-connected test gear set (multimeter, bench power supply, soldering station etc, all with either wired Ethernet or Wi-Fi connectivity)
But why should I have/use soldering station, power supply  connected via Ethernet/Wi-FI?What's point in that?
Think further and imagine needing a specific test jig for some specific use. No off the shelf solution exists and a custom solution must be made. I have worked at a few jobs where they used Arduino-based devices for testing, including one where I actually designed three different jigs all based on Arduino.
Cryptocurrency has taught me to love math and at the same time be baffled by it.

Cryptocurrency lesson 0: Altcoins and Bitcoin are not the same thing.
 

Offline technix

  • Super Contributor
  • ***
  • Posts: 3223
  • Country: cn
  • From Shanghai With Love
    • My Untitled Blog
Interesting projects.


But not sure who or why should use:
* Vintage Computer Project x86 (8088 + FreeDOS?)
* Vintage Computer Project 68k (68SEC000 + what OS software?)
Do you know that?
Educational purposes. Unlike modern computers those older architectures have clearly separated parts that I can point a finger to, and much simpler bus protocols that I can hook an oscilloscope to.
* DIY network-connected test gear set (multimeter, bench power supply, soldering station etc, all with either wired Ethernet or Wi-Fi connectivity)
But why should I have/use soldering station, power supply  connected via Ethernet/Wi-FI?What's point in that?
You get the benefit of fully automated testing with networked test gear (power supply, multimeter, etc.) The networked soldering station is just a curious side effect as it uses the same main circuit board as the power supply.
Where do you plan to use The Mess-of-RAM-wire project?
I am not sure yet. At least I can show it off to people...
 

Offline JacquesBBB

  • Frequent Contributor
  • **
  • Posts: 784
  • Country: fr
Nice projects,

But why use the DHT12 ?
Its accuracy is really awfull for humidity.

I recommend the BME280 which you can find at low cost on ebay and that gives pressure as well and is a much better humidity sensor.
 

Offline Brumby

  • Supporter
  • ****
  • Posts: 9386
  • Country: au
Skills demonstration on a project with no commercial implementation is not a waste.  It is a case of having a defined objective and overcoming the obstacles in achieving that objective.  If that project is unusual or unique in some way, that just adds to significance in completing it.

A potential employer will see that as a desirable skill ... one that they can use by defining an objective and handing it to someone who can take it and run with it.
 

Offline technix

  • Super Contributor
  • ***
  • Posts: 3223
  • Country: cn
  • From Shanghai With Love
    • My Untitled Blog
Nice projects,

But why use the DHT12 ?
Its accuracy is really awfull for humidity.

I recommend the BME280 which you can find at low cost on ebay and that gives pressure as well and is a much better humidity sensor.
DHT12 is just slightly cheaper in China, and it is easier to hand solder.
 

Online blueskull

  • Supporter
  • ****
  • Posts: 12479
  • Country: cn
  • Power Electronics Guy
Man, from your projects it seems like you really like DIY projects and you really need a lot of time to do these.
You won't get any meaningful and polished projects done in one or two months, so I recommend you to go to grad school and take your time to do your projects.

The CSR project, if DSP is involved, can be a good learning experience of DSP algorithm.
The RAM project can be a good practice platform for leaning advanced PCB routing.
The GPS project is a good idea and may attract VCs in the future.

If I were you, I would keep myself away from vintage computers. These are long long term projects that best suit those people with free time and won't give up in a few years, in other words, mature engineers with less stress to live and good amount of experience. I won't do them in my 20s.
 

Offline basinstreetdesign

  • Frequent Contributor
  • **
  • Posts: 324
  • Country: ca
Tough luck on losing your job.  As others have said, your ability speaks for itself so I have no doubt you will find a new situation soon.

However, having said that, IMHO you should consider all of these projects as spare time projects.  Your real full-time job, as of now, should be finding a job.  That should be your 8 hour/day task until you get one.  It may take 1 day or it may take 3 months.  You don't know.

And if you have some spare cash laying about, fill the fridge.

There was a time many years ago when I didn't follow this advice.  In fact, when I was laid off, after a month I went and spent my last $300 on a telescope so I could look at the stars on a country road on a cold February night.   ::)  A week later I got a phone call from a prospective employer saying that I got the job... but I didn't know how lucky I was.   :phew:

Anyway, good luck.
STAND BACK!  I'm going to try SCIENCE!
 

Offline technix

  • Super Contributor
  • ***
  • Posts: 3223
  • Country: cn
  • From Shanghai With Love
    • My Untitled Blog
Man, from your projects it seems like you really like DIY projects and you really need a lot of time to do these.
You won't get any meaningful and polished projects done in one or two months, so I recommend you to go to grad school and take your time to do your projects.

The CSR project, if DSP is involved, can be a good learning experience of DSP algorithm.
The RAM project can be a good practice platform for leaning advanced PCB routing.
The GPS project is a good idea and may attract VCs in the future.

If I were you, I would keep myself away from vintage computers. These are long long term projects that best suit those people with free time and won't give up in a few years, in other words, mature engineers with less stress to live and good amount of experience. I won't do them in my 20s.

I did get DSP involved in the CSR project - in fact I have changed the architecture to make the DSP the absolute center of it. For now it is an ADAU1701 that comes with preprogrammed algorithms (so I can have a knowledge of each tunable parameter of the algorithm) and I will create a revision with ADSP-BF592 later, using a proper programmable DSP, to learn about the actual DSP algorithms.

The RAM projects will probably also go through a few revisions, replacing the QFP chips with more and more BGA, as a way of not only exercise my advanced board routing skills but also BGA soldering/reworking skills. It also involves Linux kernel driver coding which can be useful for some embedded systems.

I don't think my moral compass will tolerate me to send the GPS project to a VC - it is intended to allow the user to have absolute control of the entire tracking system, hence the reason why it have to be open source and non commercial.

For the vintage computers, my intention was actually put those older but discrete CPU into a FPGA board as kind of a hard IP core (It would be interesting to implement the bulk of IBM PC/XT motherboard using a single EPM7128 CPLD.) Now it is more on the backburner anyway.

Tough luck on losing your job.  As others have said, your ability speaks for itself so I have no doubt you will find a new situation soon.

However, having said that, IMHO you should consider all of these projects as spare time projects.  Your real full-time job, as of now, should be finding a job.  That should be your 8 hour/day task until you get one.  It may take 1 day or it may take 3 months.  You don't know.

And if you have some spare cash laying about, fill the fridge.

There was a time many years ago when I didn't follow this advice.  In fact, when I was laid off, after a month I went and spent my last $300 on a telescope so I could look at the stars on a country road on a cold February night.   ::)  A week later I got a phone call from a prospective employer saying that I got the job... but I didn't know how lucky I was.   :phew:

Anyway, good luck.


For now I am actually more interested in getting into some graduate school. At least in China you can only get so far with only a Bachelor's. Thanks to how Chinese colleges work there is next to no boss domestic values actual ability - instead someone with good abilities but a bog-standard Bachelor's will be target of someone's green eyes. I lost my previous job exactly due to my team leader being jealous of me being more able at the job than he does - my boards cost less, perform better and consume less power, my code runs faster and integrates with the customer's code easier. I suspect corruption, the team leader has a Master's, and the boss prefers degree over ability.
 

Offline nctnico

  • Super Contributor
  • ***
  • Posts: 17971
  • Country: nl
    • NCT Developments
I'd go out and look for a job and tell what kind of projects you are working on. The interviewers won't care they aren't completed.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 
The following users thanked this post: NiHaoMike

Offline technix

  • Super Contributor
  • ***
  • Posts: 3223
  • Country: cn
  • From Shanghai With Love
    • My Untitled Blog
I'd go out and look for a job and tell what kind of projects you are working on. The interviewers won't care they aren't completed.
This does not solve my jealous peer problem. For some Chinese they have a strong belief in documents with an official stamp on them. If I managed to outperform those peers without an officially stamped document showing a higher level than them backing it they get jealous and try to get me kicked out.

I went through college in a fairly different way than other students. A lot of Chinese college students does not focus on actually learning stuff in college and instead treat university as a chance of socializing. I was never into socializing and spent my time either in labs or in my dorm room experimenting stuff. At least in college professors can recognize my accumulated abilities - I was filling in the role of a secondary TA in a few of the courses.
« Last Edit: October 12, 2017, 03:56:36 pm by technix »
 

Online NiHaoMike

  • Super Contributor
  • ***
  • Posts: 5638
  • Country: us
  • "Don't turn it on - Take it apart!"
    • Facebook Page
I went through college in a fairly different way than other students. A lot of Chinese college students does not focus on actually learning stuff in college and instead treat university as a chance of socializing. I was never into socializing and spent my time either in labs or in my dorm room experimenting stuff.
Where I went to (Texas A&M), they had a club called "Cepheid Variable" which is a little like a cross between a hackerspace (but mostly for software work - bring your own tools) and a study lounge. I found that it's a great place to hang out and get school or personal projects done at the same time. I spent a lot of time there from when I discovered it to the time I graduated. I think every university should have one. (Hint for those still in school: if such a club doesn't exist, look into starting one!)
Cryptocurrency has taught me to love math and at the same time be baffled by it.

Cryptocurrency lesson 0: Altcoins and Bitcoin are not the same thing.
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 13165
  • Country: gb
  • 0999
* Audio analog chain: Either TLC2272/TLC2274, or TL072/TL074, or the built-in ones in STM32F303CCT6. Which one is better?
Noise wise, f = 1kHz, 9nV/?Hz TLC2272 is better than TL072 18nV/?Hz, with the STM32F303CCT6 being the poorest at 109nV/?Hz. Both discrete op-amps you've listed have higher minimum operating voltages than the MCU, TLF2272 a minimum of 4.4V and TL072, which the data sheet specifies down to 10V but there are graphs showing various characteristics down to 7V.
http://www.ti.com/lit/ds/slos190h/slos190h.pdf
http://www.ti.com/lit/ds/symlink/tl072a.pdf

I'd probably go for a low noise, low voltage op-amp such as the LMV751,
http://www.ti.com/lit/ds/symlink/lmv751.pdf
 

Offline technix

  • Super Contributor
  • ***
  • Posts: 3223
  • Country: cn
  • From Shanghai With Love
    • My Untitled Blog
* Audio analog chain: Either TLC2272/TLC2274, or TL072/TL074, or the built-in ones in STM32F303CCT6. Which one is better?
Noise wise, f = 1kHz, 9nV/?Hz TLC2272 is better than TL072 18nV/?Hz, with the STM32F303CCT6 being the poorest at 109nV/?Hz. Both discrete op-amps you've listed have higher minimum operating voltages than the MCU, TLF2272 a minimum of 4.4V and TL072, which the data sheet specifies down to 10V but there are graphs showing various characteristics down to 7V.
http://www.ti.com/lit/ds/slos190h/slos190h.pdf
http://www.ti.com/lit/ds/symlink/tl072a.pdf

I'd probably go for a low noise, low voltage op-amp such as the LMV751,
http://www.ti.com/lit/ds/symlink/lmv751.pdf
This project have been significantly revised to include a DSP.
 

Offline JacquesBBB

  • Frequent Contributor
  • **
  • Posts: 784
  • Country: fr
Nice projects,

But why use the DHT12 ?
Its accuracy is really awfull for humidity.

I recommend the BME280 which you can find at low cost on ebay and that gives pressure as well and is a much better humidity sensor.
DHT12 is just slightly cheaper in China, and it is easier to hand solder.

I made some tests a while ago
https://www.eevblog.com/forum/beginners/problems-with-the-new-dht12-humidity-sensor-!/msg992744/#msg992744

With DHT12, for the same location, you can gethuge differences.
One gives 48%, another 72%.

So it is totally useless,
It claim 5% accuracy, but it is very far from this.

DHT22 is better.

BME280 is much better, and to avoid soldering problems, you can use
A breakout board that it not much larger than a DHT12 or DHT22.
« Last Edit: October 13, 2017, 05:00:25 am by JacquesBBB »
 

Offline chickenHeadKnob

  • Frequent Contributor
  • **
  • Posts: 823
  • Country: ca
  • doofus programus semi-retiredae
I imagine that you were drawn to the AllWinner V3 for the same reasons that grabbed my attention, price, easy to handle QFP package and on board RAM. I just checked for prices and availability of bare chips yesterday and it seems to have decreased in the last 6 months significantly as if the supply is drying up. I would be hesitant in spending any effort on a design with them. Obviously you are closer to the market and perhaps not worried.

As far as vintage computing goes, I am baffled by the attraction, especially in a young person. I am from that era and have no nostalgia for it at all, absolutely none. My first home computer was an Ohio Scientific superboard. A 6502 processor, micky-soft 8K basic in ROM and 4k RAM  with cassette tape for program store and a video modulator for TV output. I had to design, build and source the video modulator and power supply which forced me to learn electronics. I suppose that aspect was a benefit as  I was more into math and programming then. It was in truth a horrible and limited little machine. I can distinctly  remember dreamily anticipating and lusting for the features and power of the computers we have today, particularly full color flat screens and large RAM.
 

Offline technix

  • Super Contributor
  • ***
  • Posts: 3223
  • Country: cn
  • From Shanghai With Love
    • My Untitled Blog
I imagine that you were drawn to the AllWinner V3 for the same reasons that grabbed my attention, price, easy to handle QFP package and on board RAM. I just checked for prices and availability of bare chips yesterday and it seems to have decreased in the last 6 months significantly as if the supply is drying up. I would be hesitant in spending any effort on a design with them. Obviously you are closer to the market and perhaps not worried.
The only reason why I am still pondering about V3s is because of its ease of use in a low cost low complexity project, but being that underpowered it seem to me that even Allwinner is giving it the cold treatment. They are focusing on the better selling H3/H5/A64 and the only reason for driver improvement was H3 drivers trickling down to V3s.

As far as vintage computing goes, I am baffled by the attraction, especially in a young person. I am from that era and have no nostalgia for it at all, absolutely none. My first home computer was an Ohio Scientific superboard. A 6502 processor, micky-soft 8K basic in ROM and 4k RAM  with cassette tape for program store and a video modulator for TV output. I had to design, build and source the video modulator and power supply which forced me to learn electronics. I suppose that aspect was a benefit as  I was more into math and programming then. It was in truth a horrible and limited little machine. I can distinctly  remember dreamily anticipating and lusting for the features and power of the computers we have today, particularly full color flat screens and large RAM.
I am attracted to vintage stuff because of the exposed system bus that I can stick oscilloscope probes to. Modern chips embed everything, making them a lot less interesting to play around with.

There is a special mention to MC68SEC000 as it is planned in a much different way than the i80C88 system. It is being used on a dev kit along with an EP4CE22 Cyclone IV FPGA in kind of a ZYNQ-in-TQFP arrangement.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf