All led on at power on?

Discuss Nava construction and related issues
  • [EDITED} PS. Power up with ENTER or START/STOP combo does nothing.

    However, 1,3,5 does give me the double flash.
    I tried re-applying 1.28 firmware and it goes through and completes, restarts and back to just showing the blocks on line1 of the display.

    So maybe the bootloader not competely dead, but def no running the firmware properly.

    Ordered an ICSP programmer anyway...

    All voltages present and correct, wondering if I've just damaged the Atmega...
  • So after several weeks of debugging its all finally working again.

    Adding my findings here to aid anyone else.

    Symptoms.

    Display with top line boxes only.
    Power on all leds on, sometimes none, sometimes random.
    No voltage on 5v trim on sequencer, ENTER ignored on startup.
    1,3,5 double flash and successful sysex update.


    So first off, the top row boxes on display. This seems to be a sign that somewhere between the Atmega or other circuits leading to display there is a short or a bad connection. Searching the forum, lots of different fixes have resolved, like a bad crystal, bad solder to display etc. In my case I think it was pin1 on the display didn't have a connection to ground, or at least not a clean ground. Seems the sequencer section is susceptible to damage easily. I may have damaged the pac when removing the display in my debug efforts. Added a wire to pin 16 as it was good. Also had a bad connection on one of the links to the Atmega. Added a wire on solder side of board. Again likely damaged when removing the display.

    The Atmega.

    I got a usbasp and avrdude. The 1284p that came with the kit was working, and success with the lock, fuse, firmware and bootloader. But still no joy. I orderded a couple from mouser. Got 1284 instead of 1284p. Not recognized when inserted. Ordered a 1284p and same issue. Turns out by defailt they ship new with very slow clock settings. Add "-B4" to the avrdude command and it sorts until you get the right fuses.

    The one I got from mouser looked used... pins already bent... failed to verify firmware. Then tried original and it failed to verify too... messed about with fuse settings and bricked them both! Beware!

    So, ordered yet another. After going over everything in sequencer and adding those display wires I figured I'd stick one of the 1284 in (non P variant) and check if i could still speak to it. Worked, so ran through unlock, fuses, firmware, bootloader, lock. All good.

    Turned boards over and powered on to check voltages etc before shelvinv until next 1284p turns up and what do uou know... it frigging booted and all normal!!!!!!

    So, the 1284 works just fine (8Mhz vs 16Mhz) and everything wss good. Retrieved 5v to 5.00v, refreshed EEPROM (original chip) and all good.

    Everything seemed fine, but no accent! Figuring the mux boards wires, as they worked before) checked them all and voktages good. Some more forum searching and started replacing things. Started with the MCP, et voila! Accent back and everything working. So it could have been the start of all my woes! Or whatever fried it.
  • A few things I wasted time on.

    You will measure continuity (400 and 1800 in each direction) between 15v and -15v if you have ANY M5218L soldered in the analog section. I soldered the ones in the noise section before I started socketing them. I discovered this by cutting all the RP links and tracking it back to them. If you try a continuity test on a chip on its own you get the above measurements.

    In the sequencer section you will measure continuity between gnd and 5v if nothing connected. This seems to be the flow of the continuity test current across the capacitors in the section. If you test between and of them you get the same.

    Not being an expert, I spent time chasing these and it seems normal. When powered up, your 5v, and 15v will measure perfect. So this all seems normal.
  • Hello there.
    I have a question about the Atmel 1284.
    I built a Nava with a 1.01a PCB but I made a mistake with the CPU, I take a Atmel 1284 but not P with firmware 1.028.
    But my problems are:
    -screen is like this topic: viewtopic.php?f=24&t=1093&hilit=TP1
    -I have 13V on TP1 and no change when I turn TM2
    -nothing is happening when I push start/stop...
    -at start up some LED randomly stay lit on, other lit on and switch off

    Crystal and welds seems good (I still have some tests to do).

    So have I to change the CPU or change something to run the 1284 or my problem is elsewhere ?
  • The system runs fine on mine with the 1284 (non P) version.

    The voltage suggests something shorting or a bad connection in your cables. What do you read on the power pin on IC1 (the TL072) should be 15v

    The rest of the IC in the sequencer should be 5v
  • Ok thanks for the info, I'm going to make better voltage and continuity tests.
    Does the firmware need to be changed for the 1284 non P ?
  • No, just download 028beta as normal.
  • I've had trouble with both of mine with the display showing just the top line of boxes. I eventually worked out I'd fried one of the displays, or at least done some damage to the pads on the display, as if I connected via the other set of pads on the bottom edge it worked.

    In both cases I damaged some of the traces / via's when swapping / removing the connector strip.

    In the thread you link to there is an image that shows where you should have continuity, BUT its missing one vital connection to ground, the 5th pin needs to be GND too - there is a very small connection on the bottom side of the board with just one thin trace to the ground plane, I added a wire between pin 5 and pin 16 on the display and it sprang into life. I've damaged pin 5 trace on both 1.01a and 1.02 boards so its possible this is your problem too. The top line of boxes on the display means its not getting reset by the atmega, and it seems missing ground on either of the 3 (pin1,5,16) will cause this.

    As for your 13v on TP1, try a different TL074 in IC110, and be sure to be in "ENTER" boot mode, it only seems to adjust correctly in that mode - if its not moving at all with TM2 (which you do need to turn a lot, mine started at 11v) then maybe you have a short somewhere, but most likely the TL074.
  • Hello there.
    I make a complete contiunity test for the 1284 pin out, they are all good.
    So I think it's my CPU. The man who upload the firmware doesn't use Arduino, he uses a CPU/EEPROM programmer.
    I made tests with others CPU (Atmega 162) from him on my X0xb0x and these CPU don't work normaly (very low start-up, midi sync doesn't work, impossible to change the tempo...). I sent to him the 1284 of my Yocto to change the firmware because I couldn't make it with the midiOX technic and this CPU now make perfectly the job (the CPU came from E-Licktronic with an original firmware).
    So my question is: without an arduino programmer, do we have to change or program something on these CPU before upload the firmware (like frequency or other)?