Strange sequencer/power issue

Discuss Yocto construction and related issues
  • my nose tells me it might be a short or semi-short on one of the pins on the shift registers which drive the LEDs
    when the sequencer wants to light that led up, that causes the shift register to put 5V on the pin, but then that is shorted (or semi-shorted) to ground and consumes a lot of current, dragging the whole 5V supply rail down which may cause the cpu to misbehave
    so.. one thing you can do is, after it has been off for some time (so that it's kinda cooled down) turn it on and reproduce the error, while it's frozen and the 5V rail is lower than 5V - check the shift registers (the 74hc595s) if one of them is hotter than the rest

    a more complicated way to test them is to (with the yocto switched off) measure resistance from ground to each pin on the shift registers
  • hi antto,

    thanks so much for the advice.

    i really like this theory.

    unfortunately i do not feel any of the 595's or 165's being warm when the sequencer/led's are frozen after powering on. even after waiting some minutes.

    so i measured the resistance from each pin of the 595's to ground. unfortunately i get the feeling the readings are normal...?

    IC3
    1- 0 / 9- 0
    2- 0 / 10- 3.5k
    3- 0 / 11- inf
    4- 0 / 12- inf
    5- 0 / 13- 1 ohm
    6- 0 / 14- inf
    7- 0 / 15- 0
    8- 3 ohm / 16- 3.5k

    IC4
    1- 0 / 9- 0
    2- 0 / 10- 3.5k
    3- 0 / 11- inf
    4- 0 / 12- inf
    5- 0 / 13- 0.6 ohm
    6- 0 / 14- 0
    7- 0 / 15- 0
    8- 2 ohm / 16- 3.5k

    IC5
    1- 31.5k / 9- 0
    2- " / 10- ~3.5K
    3- " / 11- inf
    4- " / 12- inf
    5- " / 13- 1.5ohm
    6- " / 14- 0
    7- " / 15- 31.5k
    8- <1 ohm / 16- 3.5k

    IC9
    1- 31.5k / 9- 0
    2- " / 10- ~3.5K
    3- " / 11- inf
    4- 0 / 12- inf
    5- 0/ 13- 1 ohm
    6- 0 / 14- 0
    7- 0 / 15- 31.5k
    8- 3 ohm / 16- 3.5k

    previously i reported that jiggling wires or pushing certain ic's allowed the yocto to turn on after a few tries.

    specifically, if you touch some pins of the 595 chips with your finger, the sequencer unfreezes immediately. wires seem fine and jiggling them is not helpful as previously reported :P

    stranger still, just turning the yocto on/off several times in a row, without touching a thing on the pcb's, and after a few tries the sequencer starts normally. but after turning it off again and letting the yocto cool down/discharge for just some seconds, turning it on again produces frozen sequencer/led's.

    i do have a normally working yocto to compare readings to. perhaps i should compare the voltages of the 595 chips, perhaps one has become defective?

    any more advice more than welcome, aside from tediously measuring and comparing voltages i feel a bit directionless..
  • i replaced all the 595 and 165 chips, symptoms remain the same.

    ...

    i suppose i should continuity test all the sequencer parts according to the schematic, and if that seems ok, compare voltages between my dysfunctional yocto and good yocto? any troubleshooting tips massively appreciated...

    edit -- tried swapping the 40-pin os chip with another one, the behavior is unchanged

    EDIT ohhhh duhhhhh. after a day of rest the troubleshooting became obvious, though i still haven't done it...i've got 5v from the IO pcb, it should be fairly straightforward to then use the .brd file to follow the 5v on the main pcb and keep measuring the voltage until i see where things go awry...
  • still totally stuck. tracing 5v through the circuit on the main pcb doesn't work because the whole 5v rail is pulled down to about 3.5 volts.

    just to confirm, i disconnected the yocto pcb's and measured 5v (4.93v technically) coming from the power/io pcb.

    i reconnected the 15v, -15v and GND lines between the pcb's, and it was then i noticed that I measure 2v on the main pcb 5v rail without it even being connected.

    does that mean that somehow, somewhere, 2v is bleeding from the 15v onto the 5v line, I suppose do to a small bridge or short somwhere? where do i go from here to figure out where?

    could really use some tips here...somebody, anybody...thanks for any help!
  • i strongly think that the "5V" line shouldn't have anything to do with the +15 or -15V lines
    if you get 2V in that case, it smells bad

    check the resistance (on the main board) between the 5V line and the +15V line
    the fact that the PCB is black doesn't help much, but look around the PCB for ugly solder joints, brush the PCB with a toothbrush to get rid of anything potentially stuck (cold blobs of solder, fractions of component legs, etc..)

    basically if you put 2V and 5V together, you'll get some voltage in the middle, exactly how much depends on the effective series resistance, i think

    the 2V i'd guess is probably coming from a voltage divider on the +15V, such divider must have a smaller resistor towards ground
    but, i'm not fully sure if this is a problem (my supply cable is on a 4-pin molex connector, so i can't easily disconnect just the 5V to perform the same measurement as you)
  • thanks for the tips antto!

    yeah it doesn't seem right to me, the 2v on the dis-connected 5v..

    i measured the resistance between the 5v and +15v rails, and get about 5.6k ohms...

    i've cleaned and brushed the pcb many times to no avail...

    i had most recently suspected a fault in the accent section, since the 5v comes into the main board and then goes to one leg of the accent pot and the emitter of q34, as well as 6 or 7 spots in the sequencer section..i replaced q34 and the 4011 in the accent section for good measure, but i'm still getting the same results...
  • Hey Guys, just chiming in here as I have a similar problem.

    I didn't build my Yocto but after a successful RE-303 build I finished the Yocto off and even started a Nava.

    My Yocto was working fine until I put in in its enclosure. I even successfully did the internal Transformer Mod.

    But when I went to power it on last night (After wiring it up in the studio) it just didn't turn on. All my tests out of the case with the internal transformer went fine. I suspect it may be similar to pistrix's problem as the IO board is a tight fit with the IEC port near the PSU section and getting the IO board into the right position I may have bumped something.

    I was going to start by pulling the whole thing out of the case and replacing all the wires and header cables as the ones installed are short and are ribbon type.

    I did notice the master volume pot is quiet loose..

    I get no lights when I power on. Ive tested +5v, +15v -15v successfully at the IO and the power cable connection to the main board.

    Any suggestions? :)
  • Sorted.

    Changed all those horrible short cables for some longer more robust ones. Cleaned up some things, trimmed up some legs and reflowed a heap of joints. Happy dayz. ;)
  • Glad you got it fixed, but that was definitely a very different issue. My IO/power board is fine, and I have power to the mainboard. There is simply something mysterious pulling down the 5v rail on the main board that I have yet to discover..I have yet to systematically trace through the sequencer circuit or do a thorough retouching of the solder joints in the sequencer section, so perhaps it will end up being a simple soldering issue as well...