2014. március 29., szombat

3D Printing - The struggling continues

After two successfully printed but quite small object I started to work on something bigger. I faced  the leveling problem of the Z axis again. It is aligned well on one side but not on the other.
After some fighting it finished this object. Anyhow I didn't liked the brown burned small spheres what appeared on the print. I presumed that these are some plastic from the previous prints.
On Tuesday evening I started to work on the next piece.
At the setup I realized something about the Z axis:
If I not hold tide the threaded rod on one side when I turn it on the other, they turning together and wryness will be left there. (It is weird a little bit, because the two sides has two independent stepper motors. Maybe an additional timing belt is there)
After this I started to print. It forecast the project to 9 hours. Even in the evening I seen, that there is some problem with the project, because the brown spheres are continuously generated. 
The head leaks somewhere. And I seen the same in the morning but the print was not finished.
While I was waiting to a meeting with one of my clients I searched for the replacement part list (exchange the head). And I found even better thing:
http://diy3dprinting.blogspot.hu/2014/01/best-robo-3d-printer-fixes-tips-tricks.htmlFor not to much money I can upgrade the machine.
When I arrived to home, checked the print. The last few layer was missing and the layers were there were slipped away a little bit. The extruder was clogged. I tried to clean it in an easy way but the filament broke at half way. This means I should disassemble the whole extruder sometime during the night.
First of all, I ordered the new metal extruder.
After this I started to clean up the old one. Removed the head from the holder, but I still wasn't able to remove the filnament from it. I wanted to take it apart anyhow, to eliminate the leakage. This ony can be done when it is hot.
Switched on the heating and started to screw it with two pliers. I grabbed the metal part at a wrong spot. There was a little spark.
The head taken apart and cleaned.


The result:
I can't screw it together, I suspect the plastic should be heated for this to expand.
The head doesn't heat at all.
I found out what is in there (RAMPS 1.4 board), grabbed drawings for it. I was measuring around a little bit. It looks like I killed one of the MOSFETs (I think, that a 60V 55Amps rated FET should survive this and the fuse should blow, but it makes no difference).
The ST made FETs are in there according to the drawing was not available at the Lomex, so I bought some with similar parameters from IRF.
Yesterday evening finally removed the control board from the enclosure, removed the RAMPS board from the top of the Arduino and...
"a 60V 55Amps rated FET should survive this" - Yes, it did, and the fuse also. But the trace on the board did not:






My opinion, this is a shitty design. At least they should use thicker foil on this type of board. It is power electronics admit it or not. I soldered it together and the head started to heat.

Now comes some acetone and screwing the head together.

2014. március 26., szerda

Refigirator 3. - Money thrown away

The refrigerator I repaired it seems working but I'm not satisfied with it. At twenty-something degree external temperature (don't forget to bring a thermoeter there) it is able to cool the internal space to 16-17 degrees.
From the beginning I felt that the original cooling fan of the peltier module is not sufficeient.
I wanted to exchange it for a while. I found a shop near to my workplace called "ventilatorház" (house of the fans).
On Wednesday morning finally poped up the thought to measure the size of the fan (92mm).
After this I went into the shop and asked for a size 92 fan with 12V DC supply. They said that only 230V AC available, at a high cost. I bought it. At that time I had the feeling that it is a wrong and costly decision, but I wanted to finish this task.
On the second day I connected the cable of the new fan and tried place into its place.
It won't fit in. This thing has only 3 mounting holes. At the 4th place the power connection and the control unit located. This is not a problem by itself, but the refrigerator has long molded pins and short screws instead of holes and long screws. To place the fan into its place I need to cut off one of the pins what I didn't wanted.
Afternoon I bought in the Lomex (one of my regular local electronic part store) a 12V fan on the 6th of the 230V ones price. In addition according to the datasheet it has twice as much airflow.
On Friday morning I built this one in. It generates even less noise.
I'll try to bring back the inappropriate 230V one, but I've no confidence if it will work out because I even soldered in the cable.

2014. március 23., vasárnap

3D Printing - Tricycle

No, no, I haven't printed a tricycle.
My daughters has one. The thing has a pipe on it. This pipe leads to the closed chain chamber. They used to throw things into it, what can be only removed with the disassemly of the thing.
As the 3D printer started to work, let's print a plug to it.
Here is the design:


And the finished piece:

3D Printing - Success

I didn't wrote about my 3D experience. Unfortunately I had nothing to write about. To be honest there were subject but I not happily write about my faliures.
The whole thing was a catastrophe.  Nothing was working, it slipped away, even the head was clogged once, what required a total disassembly.
Finally I've lost my zest and put the whole thing onto the side.
a  two-three weeks ago I seen my friend on the facebook with his siny new 3D printer. And he started to print. On THE FIRST DAY.
Wow.
I wrote a letter to him for asking help. I wrote down my concerns. I received some setup tip.
Yesterday finally I cleaned the messy baseplate (3D glue, hairspray, stacked PLA, some cotton). A bit of acetone and some water (the acetone doesn't dissolve the hairspray) was able to clean it.
My UV Lightbox is waiting for me for a while. The blocks tighten the bottom to the frame are planed from aluminum, but the plastic will be sufficient.
I designed it:


And printed.
The first trial was a failure because of the old plastic on the head.
For the second trial I was able to create all of the four pieces:



Adventures in CADLand - The solution

Here (this was not translated to English up to now):
http://it-pro-hu.blogspot.hu/2013/11/kalandozasok-cad-orszagban.html
I wrote about the OpenCAD's circle rendering problems. I don't know if they already corrected it, or just I wasn't find the solution.
Now I started to work with OpenSCAD and faced this problem once more. I started to search about it. Finally I found out that circle and the cylinder function has a $fn argument. Setting to $fn=100 gives a very nice result.

2014. március 19., szerda

Entrepreneur Course 1. - 0th impressions

In early morning when I woke up looked into the site. I don't see what will be with this for a while. Deliverable only scheduled to the end of the course, but the 60% of it is in text format.
I've concerns with it:
  • I don't understand how will it be evaluated
  • I'm not so confident with my English language skills, to be able to write this.
  • There are some kind of interviews needed. Frankly I don't have a clue who can be the subject and what is the context
So I start to read and watch the material and we will se.
May it wasn't a good idea to sign into the ID Verified Certification on the first place?

2014. március 18., kedd

Clock with analog seconds - Eager follower

I'm  corresponding with Csaba for a while in the hobbielektronika.hu. First he contacted me, to ask for a help building the thermometer I designed. He read this blog.
It looks like he likes my analog second clock design idea and he was drawing in his free time.
Here is the result:



I like it.
Thank you Csaba.

Course - Be Entrepreneurs




In theory MIT's Entrepreneurship 101 course starts today. This is a nontechnical course what I need for my business. So I registered to it. Unfortunately the course material is still not available. I wanted to check the schedule to know the deadlines.

In detail:
https://www.edx.org/course/mitx/mitx-15-390x-entrepreneurship-101-who-1312

Project Idea - Frequency counter acessories

I was planing the frequency counter basically to be modular.
So I had some idea in my head for accessories:
1. Precision timebase Actually I'm using a calibrated but not thermally compensated simple 4.194MHz XTAL oscillator. This needs some uplift in frequency anyhow (I tried it but I've failed on the breadboard), and as an option I would need an OCXO timebase. I bought OCXO for this already:



In addition to the OCXO this accessory requires a frequency divider to 1Hz. It came to my mind to create the divider from an MCU (an MSP430G2230 or an NXP LPC810) instead of discrete logic. This solution has an additional advantage, that I can put a digital thermal sensor IC under the OCXO case (if I choose the MSP430G2230, it can be put under the OCXO itself because of the integrated thermometer) and an additional LED to check if the OCXO achieved the operating temperature.
2.Prescaler
A few month ago I bought a Xilinx CPLD development board.



My goal was to learn using the programmable logic ecosystem (CPLD, FPGA, etc.). My second goal with it to create a prescaler for the frequency counter.
If I understand correctly this series of CPLD is able to handle signals up to 400MHz (maybe wrong) and the smallest version with 32 macrocells cost around €1.
3. Input circuit
I continuously feel the absence of a good input circuit. I was searching around the internet but haven't found solution (maybe just because I didn't built any of the available schematics). It is still fuzzy to me what would be the requirements.
The only thought what I have about it to pick a high speed integrated comparator and add some input protection, a shortable DC decoupling capacitor and a rail-to-ral variable threshold level to it.
4.PSUI need a PSU (may a directly mains connected one) for all of above. Not all of the listed things (the OCXO for sure, the input circuit may) are able to work from a single 3.3V. This makes it necessary to have a multi-output PSU.

I was thinking to put all of the accessories one by one on PCBs equally sized to the frequency counter itself. This can be modularly connected on top of each other to the counter.

Project Idea - Rubidium frequency standard

I've this in my drawer for a while:


The thing is a high precision rubidium frequency standard. It is like an "atomic clock" just it uses rubidium instead of cesium.
The DB-9 connector on it not just for the power source and the 10MHz output, but if I'm lucky and this model is capable (according to the ebay seller, it is) to program it between 1Hz and 20MHz.
I think, instead of just using this as a 10MHz standard, I'll put an MCU next to it with a display and some control switches/buttons. As the result I'll get precise variable frequency standard what can provide output in the frequency range mentioned above.

2014. március 17., hétfő

LED Blinker 4. - Errors corrected, software finished

As I mentioned in the previous post, the writing about the LED Blinker will continue.
Here it is.
In picture:


And in video:


And today morning (instead of the frequency counter) to try out the TI toolchain on my shiny new desktop PC, I wrote the software for the MCU version:


These can be places on necklace or  keyring.

2014. március 16., vasárnap

Back on track

Tonight, I started to take time with my unfinished projects.
These came out from the ultrasonic cleaner:



From left to right:
1. MSP430G2553IP28 Breakout board. I had hard times with it in the past, then I gave up and a little bit later I soldered in finally to a professional PCB. Today I just thrown into the ultrasonic cleaner because it was lying on my desk. I can't foresee If I'll use it anytime, because I don't plan new project on MSP430.
2. In my opinion this is the possible smallest self current valued Step-Up SMPS power regulator on the market. I built it for a while, but never tried it out. I had a plan to employ it in the next generation of my thermometer to enhance the battery life.
3. The LED Blinker's keyring version. This was soldered together today, it even works, but I'll write about it later.
4. Same as number 3., but with RGB LED and MCU. This doesn't work yet, because the code is missing. This was not planed to blinking rather than displaying the full light spectrum with PWM drive.

The thing what left out from tonight's work is the software of the frequency counter. Hopefully I'll have time for it tomorrow morning, because the first customer is expecting it within a few days.

Clock with analog seconds - I can't keep still

For this project idea (http://pakahuszar.blogspot.hu/2014/03/project-idea-clock-with-analog-seconds.html) I've been ordering LEDs for a while.
These are arrived on Friday. Here they are (60pcs, naturally I ordered more):


I has a fistful of almost ready projects (frequency counter, LED Blinkers, UV lightbox, GainClone, etc.), what I need a few days on the whole to finish.
I can't keep still. If I can try something new, the things wait to finish just left unfinished longer.
No, no, I not started to work on the clock. Just played around with trigonometry and 3D.
Created a component design for the arrived LEDs in KiCAD:


These need to placed on a PCB in a circle. This need some calculation. Excel+Trigonometry:


Put 60 LED to a schematic, after that netlist generation, pcb editor, write in the values from Excel:



Now dream a big. Let be 3D. For the LED we need to draw a cube in a software called Wings 3D. This software what I not really like, but I found an article, how can I work from OpenSCAD (http://happyrobotlabs.com/posts/tutorials/tutorial-3d-kicad-parts-using-openscad-and-wings3d/).
Let see:


Conversion to Wings 3D:


At the component the scaling factor is a question. Finally in a forum post I found the right value. If we work in the OpenSCAD in milimeters, than we need to multply it with 0,393700 (1/2,54).
Finally some settings, and here is the result:


The transparency of the LEDs -requires some additional work, but today I solved the 3D model problem in KiCAD, what I miss for a while.




2014. március 11., kedd

3D - Printing, miling, etc.

I still strugle with my 3D printer. I was not trying to work with it for a while and I not planed a next round until I'm not able to put it into it's final place.
But this is not the only way of 3D manufacturing, even it is not my only desire in the 3D world. I'm thinking about several things for a while.
The story begins sometime ago when I bought a drill stand for my Bosch hobby drilling machine.


I needed this because without it is almost impossible to drill PCBs. Shortly after this I started to think about a compound table. Even it was available for my drill stand.

Finally I not bought it.
Struggling with drilling and cutting with various enclosures are still a continuous issue.
Then came in my mind to step forward and purchase something what is able to mill - let say - aluminum.
I fount a machine to name it a Proxxon MF-70 miller.


It may suitable to part of my tasks, I was planing and looking around the available purchase sources, but it not came to the reality.

I had the possibility to build a CNC miller in my head for a while. But I was unwilling to start this project because I don't have even a bit of the required knowledge.
Once I collected my requirements against a CNC miller.
Here they are:
Drilling and milling PCBs:
If I can do this largely simplify and accelerate my life.
I think about something like this:
  • I need to work with less chemicals.
  • At least partially solve my problems with the two sided PCBs.
  • Eliminate the slow inaccurate hand drilling
  • The result will be a more precise PCB (with my current prototyping process I faced already that I'm unable to solder a DIP-40 package because of the missing accuracy)
Drilling and milling aluminum enclosure faceplate:
Old desire of me to create professional face/back plates. It would be nice if I could create the standard 44cm Hi-Fi size. If it smaller a little bit, it is not a real problem, just I will not be able to drill the whole area.

Drilling and milling wooden parts:
  • I was struggling a lot with the wooden frame of the UV lightbox and still not finished it.
  • There is the "GainClone" what needs wooden parts.
  • I have additional countless ideas in my head what needs wooden parts.
Everything else:
For example some acrylic engraving come in my mind.

If I collected this list I started to look around to find a machine fulfill my needs. I found a good looking one called ShapeOko.



The site where it is (not)orderable, provide a handful of options. I collected what I need. Before ordering I went to the Hungarian HobbyCNC forum with my needs and this machine to ask the opinion, if this fits to my needs. So mainly they laugh on me. To be fair not laugh on me, but correctly explained to me  why to forget it. The conclusion: The structure is not reinforced enough for the size I want, the belt drive is not precise enough. I need a trapezoidal spindle, but more over ballscrew drive.
I got recomendation, what to build. I started to collect the knowledge, the material sources, even I almost started to draw a mechanical plan.
Finally I looked inside: What am I doing right now?
MAN! YOU DON'T KNOW ANYTHING ABOUT THIS!
I fall back to the ShapeOko with a thought in my mind "the guys who designed and built this can't be that stupid" it should be sufficient for my personal needs.
I would order it, but the "MakerSlide" not available for several month, and I just see the continuous promise "the next week".
On this weekend I come across this Chinese machine:


Ballscrew drive, serious milling tool (not a hacked in Dremel like something), PC based control (instead of the beloved magnificent Arduino) 4th axis option, etc.
If I put together the price of the (not)orderable ShapeOko it become around the price of this machine. The only drawback I found that instead of the desired 44cm length it can handle only 39, but this compromise I willing to take.
One last question left: Where to order from? eBay/AliExpress form Hong-Kong or from the Hungarian reseller. Because, yes I found a such (but far above the Chinese price):
http://www.variometrum.com/index.php?route=product/product&path=78&product_id=137



2014. március 8., szombat

Solved - Schmitt Trigger Oscillator

A parallel capacitor in the feedback loop solved the instability of the circuit. But instead of the recommended 22pF I needed 470pF to be stable.


2014. március 7., péntek

Struggling - Schmitt Trigger Oscillator

In the morning I commenced the calibration of the frequency counter. It works now so it time to make it work. I picked a breadboard an put together a 120kHz RC oscillator based on a HCMOS Smitt Trigger:


It started, I measured it, the frequency was fluctuating in both directions with 20%. I know that a base RC oscillator is not the champion of the precision and stability, but I can't believe that the result can be such bad.
I was changing the ratio of the R and C values, even lowered the frequency to around 30kHz, it not stabilized.
Reorganized the power supply by putting the oscillator and the counter into a common USB rail. The result is a fluctuating 600kHz. This really stinks - Oscilloscope:


This is the rising edge of the signal.
With 50ns timebase.
It was known to me that the high frequency behavior of the breadboard is a catastrophic, but it was a surprise to me, that I can't put together a 120kHz oscillator on it.
It was in my head that a parallel ceramic capacitor in the feedback loop may slow down the circuit a little bit. My colleague who is a radio frequency expert confirmed it. But this comes tomorrow...

2014. március 5., szerda

Frequency Counter V2 13. - Breath to live

I was not working on this project for a long time but I should finish it soon. Today morning I connected and started the whole thing:



On the software I made only two modifications for the time being compared to the prototype:
- Changed the MCU to the one what finally went to the board - from MSP430G2252 to MSP430G2312
- Reordered the function selector switches (in the prototype they wasn't in the right order to have less jumper wire and the software handled it)
During process I found a small "bug" on the PCB. The header at the left side has the designation text on the outer edge. So if I solder in a 90° header it will not be readable.
It can be corrected at the assembly of the later pieces with soldering the header to the other side of the PCB. This doesn't interfere with the usability.
The other thing what can be seen that new programming "connector" works correctly.
Here comes the calibration and the bugfixing. In addition the software will get a new feature: Displaying a message and the version number at boot time, to make the further updates identifiable.

2014. március 4., kedd

Project idea - clock with analog seconds

This is just a variation to the desktop clock theme. Put a 4 digit 7 segment LED display to the center of a PCB what show the time. Maybe put a second smaller display under it for the temperature. It is just a normal digital clock until now. Naturally add an MCU, some LED driver and maybe some buttons.
Here come the twist. Pick 60 LEDs (these are rectangular blue ones in my head) and put them in a circle around the display. In every second increment the number of LEDs switched on and when reach the end, start it over.
The clock can be expanded with some precise timebase: DCF-77, SNTP, GPS, distributed time source (homebrew), or something precise (more precise than the regular 32768Hz XTAL) but not atomic clock based one like a TXCO. Maybe something from here:
The whole thing goes ta a "Sick of  Beige" enclosure together with the timebase what has a same size PCB like the clock itself. The whole thing screwed together with necessary spacers like a "sandwich".

2014. március 3., hétfő

ARM Course 9. - Lab 9.

This also finished.
I'm getting sick of LED blinking.
I'm sure that at this Lab the description and the required behavior is not the same. So I found only the solution in the forum.
In the description was "record PortF bits 4,1,0 every time the input changes or the output changes", but the solution:
record PortF bits 4,1,0 every time the input changes or on and the output changes
In addition the information was missing about the storage of the initial state value - it should not be stored.