logo elektroda
logo elektroda
X
logo elektroda
Advanced Search
logo elektroda
Discover miniature XIAO boards »
Discover miniature XIAO boards » ESP32-S3 • nRF52840 • RP2040
ADVERTISEMENT
Dostępna jest polska wersja

Czy wolisz polską wersję strony elektroda?

Nie, dziękuję Przekieruj mnie tam

HDDClock - clock from hard drive

mb1988 275342 243

TL;DR

  • Builds an RGB clock from a hard drive platter using a rotating slot, LEDs, and custom electronics.
  • Runs on an ATMEGA128, and the program uses under 10% of its flash memory.
  • Helps DIY electronics builders and hardware tinkerers who want to recreate the clock and see how the parts, control, backlight, and casing were assembled.
  • Shows how a reflection sensor with hysteresis, RGB SMD LEDs, and the original HDD housing can create a clean hidden-sensor clock effect.
  • Adds a later update with a BOOTLOAD.hex file for UART1 at 38400 baud, plus a Python script and extra packages for uploading the final application.
ADVERTISEMENT
Treść została przetłumaczona polish » english Zobacz oryginalną wersję tematu
Report a violation of the law
Reply Cool? Ranking DIY | New topic
Notify about new articles
📢 Listen (AI):
  • #121 7954387
    Matt_x89
    Level 12  
    After applying the 5V voltage to 8, the motor driver's foot starts without any problems. I checked all the legs you asked for and there is tension on them. The motor controller stand is well soldered.
  • ADVERTISEMENT
  • #122 7954415
    mlassota
    Level 18  
    Matt_x89 wrote:
    After applying the 5V voltage to 8 the motor driver's foot starts without any problems. I checked all the legs you asked for and there is tension on them. The motor controller base is well soldered.


    There remains the question of soldering the IDC10 socket - for programming - is it soldered on both sides of the PCB?

    If so, the problem lies in the processor ... After switching on the voltage on the leg 51 should be 5V immediately - no pilot signal is needed ...
  • #123 7954447
    Matt_x89
    Level 12  
    The IDC10 socket is well soldered. I will check all the guides tomorrow.
  • #124 7955173
    Matt_x89
    Level 12  
    Hello.
    I checked the signals on all the legs and this is what I received:
    (7) RCINT 4.89V (at a signal from the pilot 4.07V)
    (9) INT7 0.11V (with approx. 4,89V reflecting surface)
    (11) SCK 0.02V
    (15) RED 0.02V
    (16) GREEN 0.02V
    (17) BLUE 0.02V
    (20) RESET 5.09V
    (21) VCC 5,11V
    (22) GND approx
    (25) SCL 5.09V
    (26) SDA 5.09V
    (27) RDX 5.09V
    (28) TDX 4.70V
    (32) SQW 0.05V
    (51) MOTOR 0,02V
    (52) VCC 5,11V
    (53) GND approx
    (63) GND approx
    (64) VCC 5,11V

    I checked the connections between the IDC10 socket and the processor and everything is OK.
    on the quartz legs I observed the voltage of 0.07 on the MILI scale because in volts there was no reaction which I saw on the quartz of the DS1307 chip. Should it be like that?

    Here are the pictures of the layout.

    HDDClock - clock from hard drive HDDClock - clock from hard drive HDDClock - clock from hard drive HDDClock - clock from hard drive HDDClock - clock from hard drive
    HDDClock - clock from hard drive HDDClock - clock from hard drive HDDClock - clock from hard drive HDDClock - clock from hard drive
  • ADVERTISEMENT
  • #125 7955259
    mlassota
    Level 18  
    Hello Colleague

    I look at the tiles and intrigue me as you soldered from the TOP side - top stand under TDA because you have a simple stand and it can not be soldered on both sides of the plate and this is ABSOLUTELY necessary
    Take a photo of the side of the TDA stand and the IDC10 socket housing raised - it must be in February.
    greetings
    M
  • #126 7955316
    Matt_x89
    Level 12  
    Yes, I agree that you can not solder and I knew it immediately. I decided to circumvent this problem in such a way that before soldering the socket for the scalak I made guides from several wires in places where it is necessary and checked the meter if the circuit is closed. After that I soldered the stand and checked everything again. Besides yesterday, as you advised I gave the 8 TDA 5V outstretched and the engine moved without a problem. In the picture from above it should be seen that I solder guides under the stand in places where it is necessary.

    HDDClock - clock from hard drive HDDClock - clock from hard drive
  • #127 7955413
    mlassota
    Level 18  
    My friend
    If you are sure about the correctness of soldering, do so
    Put the primitive program into the processor's memory:

    Code: text Expand Select all Copy to clipboard
    $regfile = "m128def.dat"


$crystal = 16000000
Config Pina.0 = Output
Config Pinb.5 = Output
Config Pinb.6 = Output
Config Pinb.7 = Output
porta.0=0
portb.5=0
portb.6=0
portb.7=0

Do
Set    Portb.5
Wait 1
Set Portb.6
Wait 1
Set Portb.7
Wait 1
Set Porta.0
Wait 3
Reset Porta.0
Wait 1
Reset Portb.5
Wait 1
Reset Portb.6
Wait 1
Reset Portb.7
Loop


    Below compiled hex
    He has to turn on the diodes and turn off the engine and then perform the reverse sequence. Let me know what the effect is.

    To upload Hex, use the ISPPROG program available here
    Remember to configure your programmer.
    Attachments:
    • noname1.zip (479 Bytes) You must be logged in to download this attachment.
  • #128 7955527
    Matt_x89
    Level 12  
    Thank you for the testing program.
    After uploading it to the processor, the LEDs are always red. It looks like the processor does not make a loop, it just stays in place, it concludes that the remaining colors do not turn on and the motor does not react. Could it show damaged quartz and lack of synchronization for prock?
    Once again, thank you for the test program :)
  • #129 7955598
    mlassota
    Level 18  
    Try to change the fusebits on the internal 8Mhz oscillator (bits w / g image)

    HDDClock - clock from hard drive

    Below the same program the same but at 8Mhz ... maybe now it will start :)
    Attachments:
    • 8mhz.zip (474 Bytes) You must be logged in to download this attachment.
  • ADVERTISEMENT
  • #130 7955649
    Matt_x89
    Level 12  
    I came up with the idea of checking if everything works, but unfortunately the environment has gone crazy for me to program. Could a friend give me this program first on 16Mhz only in 3 versions, let's say it starts with a green color. The next program would start with the blue color and the last one would start with the engine being switched on.
    If this is not a problem, I would ask for such versions of this program. Then it would be known that the percentages start the system and the wine can be on the quartz side
  • #131 7955725
    mlassota
    Level 18  
    In the appendix programs starting from different pins - the program name is the name of the port from which the program starts - I hope I did not crash any error in the program :) - in the end is so complicated :)
    Attachments:
    • wersje.zip (1.77 KB) You must be logged in to download this attachment.
  • #132 7955769
    Matt_x89
    Level 12  
    Thank you very much for the programs. After uploading b5, the diodes are red, b6 green, b7 blue, and the engine turns only (after about 40 seconds of turning for a while it slows down as if it was turned off and continues to spin). Now the question remains why the loop does not work and the ports are not activated in turn. Is this momentary turning off the motor and the lack of port switching indicates the damaged quartz?
  • #133 7955832
    mlassota
    Level 18  
    Try to really switch the processor to an internal oscillator at 8 mhz and then upload the programs from the file below. Exclude quartz failures.
    Solder out capacitors with quartz - maybe one has a puncture and pulls to the mass

    The upside is that we are sure that the ports are operational and did not burn during the soldering
    Attachments:
    • Wersje 8Mhz.zip (1.77 KB) You must be logged in to download this attachment.
  • #134 7955949
    Matt_x89
    Level 12  
    I changed the source signal to the internal 8MHz and uploaded the program on this version. Earlier I was soldering the quartz and the system was behaving unchanged. then I switched to external 16Mhz clock and the system refuses to cooperate here:

    HDDClock - clock from hard drive
  • #135 7956030
    mlassota
    Level 18  
    And porcek has connected quartz?
    If not, he has the right to behave like that
  • #136 7956548
    Matt_x89
    Level 12  
    I re-soldered the quartz to the system before I changed the bits and uploaded the program to 8MHz. The system can not be recognized by any program at all.

    Added after 1 [hours] 53 [ minutes]:

     I decided to desolder the slingshot and ordered a new system and quartz. Thank you again for the test programs. I wanted to ask if you guessed what could be the cause of the damage to the prock, or what could I do wrong? Could the temperature of the soldering tip damage the processor?
  • #137 7958155
    mlassota
    Level 18  
    Hello, Kolegów

    I undressed my watch and uploaded the test programs.
    On a properly folded (working) watch they work the most.
    The LEDs light up one by one and the engine starts up ...
    So the problem lies in something else.
    Perhaps you've overheated the chip during soldering and maybe the quartz is guilty, capacitors ...
    I thought that when I was at work and writing test programs, I screwed something up, but I'm sure it does not work on my ass.
    Before desoldering the processor, if you did not do it try to insert a different quartz - new - not even in the smd enclosure - maybe it overheated ... There is also the possibility that the processor fell from ESD ... although I solded a few and did not spoil in this way.
    What are you programming - what programmer? If under LPT, check if the computer is connected to ground.
  • #138 7958261
    Matt_x89
    Level 12  
    I've desoldered the slingshot now. The processor was soldered to a soldering station with a 0.5mm tip and each soldered separately, I was also thinking about inserting a 1mm tip and applying one stop at one stroke.

     I have installed AVR Studio 4 on my computer and loaded the hex file from this program and set fuse. Could the other systems damage the slingshot, walk on the signals that I put in the morning it looks like everything is correct? Which download program do you recommend to install? I would like to avoid damage to the next prock so I'm open to any suggestions ;)
  • #139 7965177
    mb1988
    Level 12  
    I've always uploaded the program using avrdude. I have it configured so that I call it from the Programmers Notepad level from the WinAVR package, it handles it very easily.

    As for your problems with the watch, you wrote that the target program did not start, but with test programs written by a colleague mlassota that's right. When we look into the main.c file of the watch source, we notice that before switching on the engine it first communicates with the RTC system. Therefore, there is a small chance (although anything is possible) that it hangs when you try to make contact with this system. It is easy to check if the RTC system has been correctly initialized - on its seventh leg a rectangular wave with a frequency of 4096Hz should appear.
  • #140 7967045
    utak3r
    Level 25  
    After watching the movie, I was about to write that it would be good to add 12 dots on the edge of the dial and slightly shorter hints for increasing the readability of the clock, but only later I read the principle of operation ... it is rather unfeasible in the solution with the aperture.

    And if you do a typical POV, with a ruler of diodes on the plate ... the problem with the aesthetic performance of such a thing, it would be best to cut the gap and place the ruler in it, or under the plate to make it shine up. With this solution, it would be possible to create a clearer image of the dial.

    Creation of the author's design: for a high five, everything elegant and pleasant, zero chalay. :) The decoration is also very interesting. Could you explain how your friend put the pattern on the surface? Directly on plexiglass she drew, or how?
  • #141 7967511
    mb1988
    Level 12  
    The idea with the led line is worth considering, but it also has its drawbacks - you will not get such a smooth image in this way, it will be divided into concentric circles. But it's a matter of taste.

    The graphics you see were made by hand with a pen on a piece of paper. This card was then cut and placed under 2mm plexiglass.
  • #142 7967543
    mlassota
    Level 18  
    Hello
    Problem 2 - feeding the ruler on a rotating plate ...
    Problem 3 Balance
  • ADVERTISEMENT
  • #143 7967703
    utak3r
    Level 25  
    In fact, the image will be clearly composed of circles ... You can still do all around, but over let the plate release the permanent lighting (possibly changing, but at any speed), which would slightly offset this effect. The question is, what would it look like eventually ;)

    As for the problems reported above:
    - power supply - fact, it is a mechanical challenge, you can try to plug in the rotor, but in the case of the main drive motor, it is rather difficult
    - balance - in the design of the colleague also had this meaning, so he attached a suitable counterbalance under the plate. Here, it would have to be the same.

    As for the drawing - I do not think I would have dropped in to put a simple card under the underside ... it looks really good.
  • #144 7970486
    Mavrodi
    Level 1  
    Good day. but how to open files from the HDDClock archive - schematic and PCB.rar
  • #146 7977147
    maliiva
    Level 1  
    Can someone explain to me step by step how to program a blank atmmega. whether to use an ISP or RS232? which program to use? please explain in detail with screenshots because I'm new in this field. Thanks in advance. maliiva(_at_)yahoo.com
  • #147 8016662
    michail_w
    Level 12  
    Hello,
    I want to build this watch, but I have a lot of problems buying TDA5140. Can you replace this element with ULN2003?
  • #149 8016923
    michail_w
    Level 12  
    Terribly much :) Earlier someone gave a link, where it is available for purchase for PLN 11. This is a more affordable price. But 30 PLN is a bit much :)
    Especially with the price of ULN2003, which costs max. 2 zlotys per piece, depending on the store.

    I am asking you again if you can replace the TDA5140 on ULN2030
  • #150 8021356
    mlassota
    Level 18  
    Hello

    Here probably is quite cheap :)
    You can not use ULN2003 - how do you see it? - processor control? .... It is quite difficult to get high revolutions from other controllers than those specialized ...
    TDA is the best solution in this and similar cases ...
Reply Cool? Ranking DIY | New topic
Notify about new articles
📢 Listen (AI):
Report a violation of the law

Topic summary

✨ The discussion revolves around the creation of a clock using a hard drive (HDDClock), where participants share their experiences, challenges, and solutions related to the project. The clock operates by utilizing a rotating plate with RGB LEDs that illuminate through a cut-out slot, creating a visual display. Key topics include the design and assembly of the clock, the selection of components such as the TDA5140A motor driver, programming the ATmega microcontroller, and troubleshooting issues like motor control and sensor calibration. Participants also discuss the importance of balancing the disk, the impact of power supply on performance, and the use of various sensors for detecting the slot in the rotating plate. The conversation highlights the collaborative nature of DIY electronics projects, with users providing insights and assistance to one another.

FAQ

TL;DR: A DIY HDDClock spins a hard-disk platter at ~60 Hz and draws ≈1 A @ 5 V to project a full-colour, 240-step clock face; “vibration is almost missing” [Elektroda, mb1988, post #6985675] Why it matters: this FAQ helps makers reboot or improve the viral RGB HDD clock with fewer pitfalls.

Quick Facts

• Bill of materials: ≈ PLN 150 (≈ €35) excluding the donor drive [Elektroda, mb1988, post #6985675] • Power budget: 5 V, 1 A total; LEDs and motor are the main loads [Elektroda, mb1988, post #6986211] • Rotor speed window: 35–61 Hz gives stable 240-step display [Elektroda, mb1988, post #6986465] • Core parts: ATmega128 MCU, DS1307 RTC, TDA5140A BLDC driver, RC5 IR decoder TSOP34836 [Elektroda, mb1988, post #6984995] • Typical noise: “cooler level in a laptop” ≈ 25 dB(A) at 1 m [Elektroda, mb1988, post #6985675]

How do I cut the viewing slot in the hard-disk platter without shattering it?

Use an angle-grinder with a 1 mm cutting disc; keep the cut to about 0.6 mm for sharper hands. The author cut a Seagate 160 GB aluminium platter in one pass “like butter” [Elektroda, mb1988, post #7024207] Glass platters (common in older IBM drives) crack on impact—test yours first [Elektroda, Sh44dow, post #7020507]

What capacitors control RPM on the TDA5140A driver?

Cap-ST (pin 9-10) sets base speed. Start with 18–22 nF from the datasheet; smaller values raise RPM, larger slow it [Elektroda, mlassota, post #9289906] Cap-TI must be ≥ 2 nF for kick-start stability NXP TDA5140A Datasheet.

My disk stops after one minute when using the original PCB—why?

Drive firmware parks the motor if heads aren’t detected. Either transplant a standalone BLDC driver (e.g., TDA5140A) or cut the power trace and feed coils directly; leaving stock electronics risks auto-shutdown and excess current [Elektroda, whd, post #7362488]

Which fuse bits should I set if I skip the UART bootloader?

  1. Disable ATmega103 compatibility (M103C=1).
  2. Keep BOOTRST=0.
  3. Set CKSEL to external 16 MHz quartz. Example avrdude line: -U lfuse:w:0xFF:m -U hfuse:w:0x9F:m -U efuse:w:0xFF:m [Elektroda, mnemonic, post #17683505]

How can I balance the platter after cutting?

Remove 123 mg from the opposite side or glue a 14 × 8 mm piece of double-sided tape as a counterweight; this eliminated vibration for several builders [Elektroda, mlassota, post #7901954]

What causes flickering when the background turns red or green?

Shared 5 V rails droop under LED current, slowing the motor and upsetting sync. Power LEDs and logic from separate ±5 V regulators or a 6 V source with diode drop, as shown by a user fix [Elektroda, bob54, post #12124903]

Can I use a slotted opto instead of the reflection sensor?

Yes. Mount the fork at the platter edge and wire the phototransistor so the interrupt polarity matches the code (or invert it in software). Expect the sensor to be visible unless you redesign the bezel [Elektroda, mb1988, post #12116564]

How do I flash new firmware with MegaLoad after 2019 update?

Follow this 3-step script the author provided:
  1. python -m pip install pyserial bincopy.
  2. Run megaload.py --port COMx --file app.hex --reset at 38 400 baud (UART1).
  3. Wait for Page #n -> success messages; finish when “Flashing completed!” appears [Elektroda, mb1988, post #6984995]

Edge case: my motor jitters but won’t spin—solutions?

Jitter means the driver can’t overcome static friction. Verify you wired all three (or four plus star) coils correctly, use 6–7 V on VMOT for startup, and confirm CAP-ST ≥ 18 nF; replacing a weak TDA5140A fixed the issue for one builder [Elektroda, PanBo, post #7369702]
ADVERTISEMENT