+- cool tech zone zone (https://forum.cooltech.zone)
+-- Forum: Tangara (https://forum.cooltech.zone/forumdisplay.php?fid=3)
+--- Forum: Hardware (https://forum.cooltech.zone/forumdisplay.php?fid=8)
+--- Thread: Building one's own tangara (/showthread.php?tid=63)
Building one's own tangara - Eve - 2024-12-03
Hello, I am making a thread to share our experiences building our own tangara's from the sources.
First, thank cooltech crew for sharing this great project. And thanks a lot Jacqueline for your support so far.
I have assembled two mainboards, and one faceplate.
2 mainboards, because I thought the first one was failed and hope I'd have more luck with the second.
Truth is, I don't : )
Both boards power on, and I can run commands from the console using minicom. But except that everything seems frozen. I can change the display brightness and see the pwm duty cycle change on the ESP32 corresponding pin ; so something runs, but except that the clock signals (SPI clk for example) are 0V. I have a few clock cycles at the power on and then it stops (and the screen is stuck saying it can't read sd card)
I'll continue debugging my connections, but it the meantime : has someone seen the same problem ? to me it is weird that the esp32 stops the clock on SPI. Is there a command to (re)start SPI from console ? has someone experienced similar problem when building ?
Cheers
Eve
RE: Building one's own tangara - jacqueline - 2024-12-04
Hi!
That does sound very odd. Especially the bit where you do see a couple of clock cycles before it stops. The fact that the display initially works also makes me very suspicious of the SD card part of the SPI bus... maybe there's an issue with the SPI mux that sits between the ESP32, SAMD21, and SD card? Although I'm not sure I can think of any kinds of soldering issues off the top of my head that it would cause it to fail like this.
RE: Building one's own tangara - Eve - 2024-12-04
Hi Jacqueline, thanks for your reply.
Indeed, I think there is ssomething weird on the SPI mux, the S pin is high and if I'm not mistaken it means that the SPI signals are taken from the SAMD.
Maybe there is something wrong with the PCA8575 that drives this S pin then, I'll check around this one.
Also I don't have a jack plug yet (out of stock when I ordered on Mouser...) and the signal "3.5mm_DETECT" is high.
Cheers
Eve
RE: Building one's own tangara - Eve - 2024-12-09
Hi ! Progress
!I disconnected the R32 pullup, so that S goes to 0. Now the SD card (and lua ls works) is detected and screen refreshes !
It is still not in the 3D case and the "naked" touchwheel doesn't react. But soon.
Do you Jacqueline or someone reading there have an idea why S stayed high ?
Eve
RE: Building one's own tangara - jacqueline - 2024-12-10
Hi!
Have you double checked the soldering of the PCA8575 pin that drives this signal? There may be a soldering fault there. Another thing you could check is whether the volume buttons work (they should show a little popup on the screen with a volume meter). If the volume buttons are also not working, this could indicate a more severe issue with your PCA8575.
Regarding the touchwheel, the touchwheel *should* respond (albeit with some sensitivity issues) even without a cover installed. The touchwheel sensor IC is on the same I2C bus as the haptic motor. Are you able to get any feedback from your haptics at all? (Assuming you've soldered on a motor).
RE: Building one's own tangara - Eve - 2024-12-10
Hi !
I doubled check for shorts to ground / vdd /adjacent pins on PCA8575 and observed no issues. It's not easy to probe though, so maybe there is an issue anyway. The volume buttons work.
I can measure traisn of pulses when I run the command "haptic" in the console, but then I can't stop them until I unplug USB. I think the pulses are not long enough for the motor to vibrate, but there are pulses (and the motor works). But I measure no pulses at all on the motor pins when I use the touchwheel, and also the TOUCH_INT signal never changes (stays 0).
(also, just in case you think it's relevant to add this, the haptic motor is not part of the BOM generated by kicad).
About I2C bus, I could observe clock pulses on SCL when triggering i2c communication with haptic command for example.
RE: Building one's own tangara - Eve - 2024-12-11
Hi !
I assembled a second faceplate, and in the process I finished breaking the first. But the second now has the touchwheel working! (I made a video thinking I could share it there)
Now I still have to fix the issue(s) with the mainboard, because as we could expect there are still issues, maybe related to PCA8575: for example, audio doesn't work... Yet!
RE: Building one's own tangara - Eve - 2024-12-11
Also, I re, re, re tripled check all the solders and of course you were absolutely right. I Re did the pca8575 and it's fine. Now I need to Re solder the amp and the tps too. PS: I did re solder everything and it works!! I could play music! Just in time for a gift this week end. I need to assemble the case tomorrow. And afterwards, I'll make one for me too. Thanks for your support!
RE: Building one's own tangara - jacqueline - 2024-12-11
Hi!
Can I ask what your process is for soldering the QFN components on these boards? They can certainly be very challenging to solder correctly, and it does sound like most of your issues so far have stemmed from soldering faults on the QFN packages. e.g. even for a lack of audio output, it could be either a fault on the mute pin of the PCA8575, or a separate fault on the amplifier (U1).
RE: Building one's own tangara - Eve - 2024-12-12
Yes, all qfn components were not correctly solder.
I used a hot plate (MHP50 and 30). I placed solder wire (with lead) on all the pads with a small tip iron and flu. It may have better to use solder paste but I didn't have access to it when I mounted the boards. I'm novice with heat plates and temp curves for reflow so I just set it to heat at 250°C. Then I put the pc with the placed components on the plate and waited for solder to melt, helping components to get in with the tip of a thin plier.
Then when I saw that all components were bad, I used the soldering iron and binocular magnifier to add solder wire on the pins and pads.
I think the whole process would have been more successful if :
-I was less excited / doing things fast
-took the time to learn about reflow curves
-I used solder paste instead of wire, but not sure
-I'd put less solder (I was careful not to put too much, but it was still too much) on the big exposed gnd pad, because it seemed that the components were too high to touch the solder on the other pads
Your inputs on these issues are welcome too. I assembled a working one that I'll offer to my partner but will make one for me too and a coworker is mounting one too for himself!
RE: Building one's own tangara - jacqueline - 2024-12-13
It sounds like you've worked out a decent approach then! All the advice I would give is advice you've already worked out yourself
They can be a very frustrating package, but I'm sure you'll get them all sorted soon!!
RE: Building one's own tangara - Eve - 2024-12-13
Yes I did on the first tangara that now works well! (and is so cute!)
I'll fix the other board I assembled too.
When mounting the tangara in the case, I have problems with the volume buttons that stay on sometimes, but if I solder the button more on the inside of the footprint, it will be better. (next week, though)
Now the tangara is in my previous phone box, wrapped in gift paper!
Eve