high side driver failure

Discussion in 'General Electronics Chat' started by drobe011, Mar 1, 2013.

  1. drobe011

    Thread Starter Member

    Dec 28, 2012
    33
    2
    Needing a couple of SSRs, I cheaped out and bought a dual SSR board that had snubbers built in. It has the following connectors (Vcc, GND, CH1, CH2). Vcc supply and CH1&2 are 3-32VDC. The SSR board works fine. I will only use 1 channel (SSR) at a time and it consumes 18-25mA to activate one SSR.

    The MCU I am using is a 3.3v system with a max of 15mA output.

    I built the attached driver just using a single driver on a bread board and worked perfectly. I transferred to perf board with the dual and the SSR stayed on just with Vcc and GND (no input to the CH pins).

    I figured it was my sloppy perf board that was the problem, so I etched a PCB, soldered it up, ensured I had not shorts/bridges and I have the same problem. I then realized I put in all the transistors in backwards. Flipped them all around and still have the same problem.

    Writing this I wonder if I pulled the bases of the NPNs to ground through a R would help..

    Anyone see anything I screwed up on?

    My last resort is just to desolder the SSRs and activate them directly by sinking the current.

    R1 and R8 just simulate the load to activate the SSR.

    Thanks for your help.

    -Dave
     
  2. #12

    Expert

    Nov 30, 2010
    16,355
    6,852
    A 10k at each base wouldn't hurt. Change R2 and R6 to 1200 ohms. Check for smoked transistors.
     
  3. thatoneguy

    AAC Fanatic!

    Feb 19, 2009
    6,357
    718
    I'm a bit unsure why you are using the NPN-PNP pair to drive the SSR, why not a single NPN to switch the ground side on/off?

    Your PNP transistor isn't getting enough drive current to be in saturation for 18-25mA (Ib=Ic/10). Best case is maybe 10mA. What voltage is measured at the collectors of Q2 and Q4 SSR in the built circuit? If undreadable, what voltage is read if a 220 ohm resistor is substituted for the SSR from the collector of each to ground (current can then be determined)?
     
  4. drobe011

    Thread Starter Member

    Dec 28, 2012
    33
    2
    A single NPN switching the low side would work, if the SSRs weren't built on this module. I can switch the ground on the module, but then I still would have to feed the channels of the SSR module through something. (One ground connector).

    Thanks for the advice. I will verify my base currents and change resistor values as necessary.

    Hope the transistors are not smoked. I'm starting my own little transistor grave yard over here.
     
  5. drobe011

    Thread Starter Member

    Dec 28, 2012
    33
    2
    Got it working. The datasheet for the relay board was a little ambiguous. After measuring what the board actually consumes on all pins, I found that the 18-25mA was just on the Vcc line. The input signal for each channel only pulls .24mA at 3.3V. I can easily source that from my MCU.

    I'm pretty sure I smoked the transistors when putting them in, because after recalculating all base currents, I had the same problem. I think all the transistors I have commit suicide when I switch my soldering iron on.

    -Dave
     
  6. drobe011

    Thread Starter Member

    Dec 28, 2012
    33
    2
    Just an update in case others have similar problems.

    Bottom line, the flux I was using (Oatey All-Purpose Liquid Soldering Flux) stayed conductive even after cleaning with Acetone and 91% Alcohol.

    More details: Almost everything I soldered, especially if it was what I call delicate (like sending small currents to transistor bases) failed. My usual procedure:

    0. Have good idea
    1. test concept on simulator; works
    2. build prototype on solderless breadboard; works
    3. build same on perf board or PCB; now gets tricky...
    4. clean pads/components
    5. flux pads/components
    6. solder
    7. clean with acetone, then with alcohol
    8. 50/50 shot of working
    9. gouge in between pads and traces with exacto knife; works sometimes
    10. add board to junk pile
    11. go to step 1

    Testing between solder joints, I would get between 5k - 5M Ohm resistance. Some things would work fine, but if transistors were involved it was a no go.

    After testing and confirming my flux was the problem (the solder did flow good though) I bought some good solder (Multicore) and some decent flux (MG Chemicals rosin flux). Everything I have soldered since just works; crazy.

    I was able to salvage some of my old projects by soaking them in Xylene for a little while, scrubbing them, then the same with Acetone and Alcohol. Then reheating the solder joints with either the new flux or just a little new solder. Although on one board, the Xylene (I guess) somehow must have penetrated a LED and caused it to allow bidirectional current flow.

    I feel pretty retarded, especially after about 30 years of not using any type of flux when soldering to not immediately identifying that with the problem.

    Dave
     
    #12 likes this.
  7. #12

    Expert

    Nov 30, 2010
    16,355
    6,852
    THANK YOU!
    Immediately ran to the shop and checked which kind of flux I have.
     
  8. praondevou

    AAC Fanatic!

    Jul 9, 2011
    2,936
    488
    Retarded is the one who invented this flux.
     
Loading...