Lights Are On, But Nobody's Home

[AetherSmith]

I wrapped up assembly of my DIY kit last night and, after appropriate continuity testing on the voltage regulators as recommended in Marqs’ tips, gave a try at powering on the unit. What I found was a little odd: it shows no signs of anything going spectacularly wrong (no magic smoke escaping, no sign of anything getting excessively hot), but simply doesn’t seem to do much more than light up.

Specifically, the LCD lights up when I switch the OSSC on, and the LEDs show red for a split second before switching to green. Upon switching, the screen briefly flickers off, then back on. I can detect some warmth coming from the TVP7002’s heatsink, and the HDMI port outputs a blank signal (I know because if I plug it in through my AVR, switching it on causes the AVR to go from a watermark of the receiver’s logo to said blank signal). Nothing appears on the front LCD at any point, and even if I try plugging in a source and cycling through with the PCB button, the HDMI output still shows nothing but the blank signal.

All this seems remarkably well-behaved for some sort of short or open circuit, and, after examining the board a few times with a 10x lens, I haven’t found anything that leaps out at me as a potential soldering issue. I’m not familiar enough with the ins and outs of the OSSC for this to be a particularly educated guess, but could the firmware be missing? I used the Advanced DIY Kit, so the firmware chip was supposed to come preprogrammed, but is it possible that the one I got was skipped by mistake?

[b0nz0beavis]

Is there any text on the LCD? Sounds like there’s not…

Which console, cable and OSSC input are you using? Even if the LCD is blank, the unit might still work. If you connect something up you can still use remote buttons or hardware buttons to cycle through the inputs.

I had two problems with my construction:

– A resistor in the wrong place (noticed by accident totally by and rectified before first power on of the unit)
– A missing cap. Despite going over the board a few times, I didn’t notice it as I’d added solder to both pads (one of them by mistake) so there was no copper visible. Marqs pointed me in the right direction based on the missing green signal on AV1.

That’s all the basic checks I can suggest (which you’ve probs already tried anyway). Good luck.

[AetherSmith]

Thanks, as I mentioned in my initial post, there is indeed nothing displayed on the front LCD, it just switches on and lights up. For the testing I mentioned, I had a PS2 connected via AV2 (YPbPr Component). I used that because I’m still waiting on a replacement SCART socket to arrive from overseas. Electrically, it should be the same as having it installed with nothing plugged into AV1, shouldn’t it? In any case, I did try cycling with the PCB button, but didn’t get any response.

I’ll go over a few more times, of course, but I’m doubtful at this point that a solder join is the issue.

[b0nz0beavis]

Yeah sorry, pretty hungover when I read that and I missed the part about the display 🙂

Before I installed the firmware on my DIY barebones assembly, only one LED lit up when the unit was powered on. I remember because I was wondering why they weren’t both illuminated and being concerned that I must have installed one LED with incorrect polarity or messed up something else. I don’t recall the colour, but I’m 100% that only one was lit up. So perhaps the source of the fault is something other than the firmware.

Anyway, sounds like you’ll need help from an expert rather than a dabbler.

[AetherSmith]

Hrmm… the red LED only shows for a split second, briefly enough that I missed it the first couple of tries powering on, and after that it’s just the green that stays on. I’m going to try re-flashing it via JTAG, if only to potentially rule out the possibility.

I think I’ll swing by my old university later in the week and see if I can use one of the lab’s JTAG programmers supplies rather than buy one. Funny thing, while I was studying there, we had a number of assignments on the same Cyclone IV FPGA and Nios II softprocessor used in the OSSC, even down to using the same Terasic development board Marqs originally prototyped it with!

[b0nz0beavis]

If firmware can be unflashed via JTAG, I’ll try that tonight and observe the LED behaviour at power on, to see if it matches the description you’ve given.

[AetherSmith]

Thanks, that would certainly be helpful to know, though I wouldn’t want to put you to too much trouble.

[marqs]

I’d first verify that FPGA works and has booted correctly. If you don’t see anything on the LCD and have no response to IR codes (green led blink), then I suggest measuring relevant regulator outputs (U12-14) and ensuring voltage is within +-5% of the spec. If that’s fine, then you could power off the board and check that CLK27 (pin25), MSEL pins (94,96,97) and other essential pins (supplies, nCONFIG, nCE and flash-related pins are correctly connected). If you have a scope, you might also want to check Y1 output as I’ve personally received a couple dead oscillators from Mouser.

[AetherSmith]

Fair enough. I haven’t tested its response to IR commands yet; I was intending to just add it to my Harmony remote, but the process on the wiki doesn’t work for the newer Harmony software. I’ll have another try tonight at finding the old desktop app. If it doesn’t respond, I’ll try taking the measurements you mentioned.

And of course, thanks a lot for all your work on developing and supporting this project, Marqs.

[b0nz0beavis]

I’m working on the ‘unflash’ of the firmware. Haven’t found an exact guide, but one thing I read was that I need some file from an unflashed chip.

However, given Marqs has suggested other trouble-shooting, his take is that this is a hardware problem so firmware checking might be a null activity. I’ll have another go tonight anyway and let you know if any progress.

[AetherSmith]

Okay, following up on what Marqs recommended, I have been able to program my Harmony with the NEC codes, and my OSSC does indeed respond to them by blinking the green LED as expected. If I understood Marqs correctly, that indicates that the FPGA may at least be booting, though the remote codes don’t yet appear to actually have any other effect and it still shows nothing on the LCD screen.

The replacement SCART socket I mentioned also arrived today, and I added and tested it to see if it produced any new behaviors. As with the component input, I got only blank video on the HDMI output. The SCART audio was correctly passed to the audio out jack, but unless I’m mistaken that’s just a passive passthrough, so it wouldn’t be touching anything relating to the unit’s main operation, anyway.

[b0nz0beavis]

Had another play in Quartus Lite last night, wasn’t able to remove the firmware. Good luck with the trouble shooting.

[AetherSmith]

That’s fine, thank you very much for taking the time.

[marqs]

If you get IR response, then FPGA has loaded firmware correctly and is at least partially working. You should next check all connections between FPGA and character LCD, and examine essential connections (HDMITX_PCLK, HDMITX_DE, HDMITX_HSYNC, HDMITX_VSYNC) between FPGA and HDMI transmitter IC with an oscilloscope and tiny probe.

[AetherSmith]

Found it!

There was the tiniest solder bridge between pins 49 and 50 of the IT6613, shorting SDA to ground. With that removed, everything is working beautifully. I found that SDA was grounded while examining the LCD’s interface, but it was only through going to every possible point it was connected to that I found the bridge; it was such a tiny bit of solder that I easily missed it in my previous examinations, and even knowing what I was looking for, I could only make it out under exactly the right light!

[Author]

Posts