F2624 V2 Help.

Discussion in 'Exclusively Sansui' started by cconger, Feb 10, 2018.

  1. cconger

    cconger New Member

    Messages:
    14
    Hi... I just finished soldering all the components on my new F2624 board... And plugged it in without using a dim bulb tester. Surprise suprise... Would not come out of protection and saw a spark at the bottom of a transistor. The led kn the left side kf the board also shortly went on after a few minutes. I was also unable to adjust the dc offset. It stayed around 12v on f04 and f05. Sorry if I come of very uneducated. I an still learning. Should I replace all the power transistors and try again with a dim bulb tester?
     

     

    Please register to disable this ad.

  2. Hipocrates

    Hipocrates Anti-Muppet

    Messages:
    1,494
    Location:
    Costa Rica
    Can you post a picture of the board?
     
    tnsilver likes this.
  3. tnsilver

    tnsilver AK Subscriber Subscriber

    Messages:
    3,246
    Location:
    Boston
    Sorry, I don't understand this sentence. The leds on V2 light up as a warning for way too much bias current. Whenever they light up, as dim as they might, you turn your bias settings down right away or shut down the unit immediately and speculate on what's wrong. These leds are warnings - they will not sustain rogue bias settings for long.

    F04 and F05 are used to measure bias current in miliamps - not voltage. You have to take the fuses out and measure current in series in miliamps. It typically requires shifting the probes on your meter to current measurement posts. Making sure the meter fuse is intact helps too. DC offset is measured at the speaker posts in milivolts.

    It's possible that if you were measuring voltage on the fuses, while adjusting bias and ignoring the leds, outside the safety umbrella of a DBT, you now have open resistors and fried semi-conductors on the F2624 driver amp. If you saw sparks - it's very likely. If you attempt to set DC offset voltage and not measure it, you may enter the range of voltage that kicks in the protection circuit. The relay opens, the red led flashes on and there's too much DC.

    I don't recommend you swap transistors blindly using a shot gun approach. First off check the other fuses on the unit and replace any blown ones, if any. Then, verify resistors integrity on F2624 (starting with R33/R34) and change any resistor that is shot. After that, meticulously verify all transistors on F2624 starting with the differentials (TR01...TR04) and moving on to the power transistors (TR11..TR14). Make sure you know how to test transistors with a DMM and look it up if you don't.

    Before you power up the unit again - verify your output transistors were not blown. Remove the heat sink at the back of the unit, take out all of the 8 transistors one at a time and check. Finally, never get off a DBT until your DC offset and bias current settings are good as per the service manual and are stable. You will need to readjust bias current and DC offset voltage after moving to mains (wall output) voltage.

    Please post some hi-res picture off the PCB so that it's possible to clearly inspect component orientation and mark the location where sparks were seen.
     
    Hyperion likes this.
  4. Hipocrates

    Hipocrates Anti-Muppet

    Messages:
    1,494
    Location:
    Costa Rica
    I would add that a cheap component tester is very helpful to learn, a good desoldering tool is vital, since that board is very delicate, when learning, desoldering can be tricky and can be bad for the board (broken traces, pads, too much heat, etc etc)
     
    tnsilver likes this.
  5. tnsilver

    tnsilver AK Subscriber Subscriber

    Messages:
    3,246
    Location:
    Boston
    Actually, the V2 board is quite resilient and will take an amazing level of soldering abuse without lifting a trace or damaging the pads. It will even recover from soldering iron scorching. The surface will self-heal and become smooth again (unless it's seriously burnt). Of course, a desoldering pump makes life easy regardless. Good point!

    +1 on the component tester. I just ordered this flimsy looking thing after watching an EEVBlog review of it. Haven't got it yet - but it appeared to have been doing a quick job (although not real lab type measurements) of sorting out components types and health. It's real cool for the money:

    https://www.ebay.com/itm/LCR-T4-ATm...2864-LCD-Capacity-ESR-Meter-Case/112244530499

    (no affiliation)
     
  6. Hipocrates

    Hipocrates Anti-Muppet

    Messages:
    1,494
    Location:
    Costa Rica
    man I had a brain fart and was thinking about the good old Sansui board with fusistors. My bad
     

     

    Please register to disable this ad.

  7. Hyperion

    Hyperion Roobarb & Custard Subscriber

    Messages:
    43,381
    Location:
    Hertfordshire, UK
    That is a really nice feature. :thumbsup:
     
    tnsilver likes this.
  8. cconger

    cconger New Member

    Messages:
    14
    Here are some photos. Thanks for the help. I have a very good desoldering tool and I am able to desolder components. Sparks were seen on the Emitter of TR08. I immediately unplugged. Would it be a smart idea to test diodes as well? What gets me is on my other F2624 v2 board(yes I have two I peeled of a pad on the other board) the led of the same side lit up. Could my problem we with the F2656 I replaced r30 r31 r32 (way out of range) on it. However the unit leaves protection with the F2624 not installed so it makes me thing F2656 is not as issue. image.jpg image.jpg image.jpg image.jpg image.jpg
     
  9. DougBrewster

    DougBrewster AK Subscriber Subscriber

    Messages:
    467
    I see that the tails of TR07 and TR08 (and others) have not been cut off; is it possible one - or all - of them could have grounded or been pushed together? When you say you saw a spark at TR08 emitter which side of the board were you looking at? Maybe you just saw a reflection of a flash from somewhere else?
     
    tnsilver likes this.
  10. tnsilver

    tnsilver AK Subscriber Subscriber

    Messages:
    3,246
    Location:
    Boston
    Hi Carter. Your STV-3H diodes are suppose to compensate for temperature changes. These are to be mounted on the heat sink of TR-13/TR-14, right on the transistors body with a thin layer of thermal grease for better heat transfer. I am not surprised you cannot bias the unit in this configuration, there's simply no, or very insufficient temp/co. On a quick glance the transistors on your F2624 seem to have the correct orientation and so do the EC's. I can't tell about the diodes b/c there's not enough resolution so I include a couple of shots of a tested board for reference. Yes, it's a (very) good idea to test diodes too. Bias related accidents wreak havoc on the driver amp, so test anything you can, but start with resistors and transistors. You may want to trim the excess of the leads of the soldered transistors like Doug mentioned. No reason to leave them that long, they are just an accident waiting to happen. Other than that, I think your issues are related to the bias adjustment and DC offset measurements in conjunction with the lack of temp/co.

    Tested F2624 components orientation - for reference:

    [​IMG]

    [​IMG]

    [​IMG]
     
  11. tnsilver

    tnsilver AK Subscriber Subscriber

    Messages:
    3,246
    Location:
    Boston
    No big deal. Dig into the trace with a small flat screw driver to expose it and bridge (jumper) the missing conductor-trace to the pad with some left over capacitor leads and solder it in. I like to form a small loop in the cap lead so it looks like a soldering pad. I also cover the mess with some JBWeld compound. Ugly - but effective.
     

     

    Please register to disable this ad.

  12. cconger

    cconger New Member

    Messages:
    14
    Thanks, I am in the process of checking the components. I will report back after I finish checking. Here is a higher res photo of my board. https://imgur.com/gallery/nbNi7 I mounted the STV-3H diodes on the heatsinks. I ordered SV-03Y diodes from ebay. Supposably the work like the STV-3H however they dont mount on the heatsink. Would these work as a replacement? Thanks for the help!
     
  13. tnsilver

    tnsilver AK Subscriber Subscriber

    Messages:
    3,246
    Location:
    Boston
    The SV-03Y's are axial and not much use to us if they don't mount on the heat sink - are they? The only reliable stock solution that we ever tested, other than our own makeshift replacements, is this: https://www.ebay.com/itm/STV3H-Stabistor-Diode-SANKEN-STV3H-Y/321845661752 (no affiliation) but if you managed the heat sink mounting with whatever STV-3H equivalents, give it a shot with a DBT after correcting everything else. With good temp/co diodes you'll see bias current beginning to drop after about 4 minutes from a cold start. It will stabilize in time. Your hi-res pic on imgur looks good. I can't spot anything wrong with component orientation and it actually looks like good work. This makes me more confident that with the proper temp/co diodes mounted correctly and a meticulous following of the bias / DC offset adjustments, you will solve the issue. Just be careful! That biasing routine is a PITA and potentially risky.
     
    Hyperion likes this.
  14. cconger

    cconger New Member

    Messages:
    14
    My only worries are that I turned it on without a dim bulb tester and blew somthing. Currently the led does not turn on. Howwver the unit will not come out of protection. Ill report back after exchanging diodes. Should I wait 3-4mins fir it to come out of protection?
     
  15. tnsilver

    tnsilver AK Subscriber Subscriber

    Messages:
    3,246
    Location:
    Boston
    The protection circuit "detects" DC voltage. It has nothing to do with the diodes or bias current or the leds. After fixing the damage, I suggest you stay on DBT, lower the bias trimmers (make sure you turn them the correct direction), measure DC offset at the gray and blue wires that come out the bottom of the PS board (each is a different channel) and lower it down using the F2624 DC trimmers - until you are out of protection mode. Then bias the unit - take out the fuses and use miliamps scale. Use this for reference:

    ---------Left Channel--------------------Right Channel------
    ------bias-------------DC--------------DC------------ bias------
    (-) VR03 (+) | (+) VR01 (-) | (-) VR02 (+) | (+) VR04 (-)
    --------------F05-------------------------------- F04---------------

    EDIT: With protection mode disabled and green led on, you can now probe for DC offset at the speaker terminals. When in protection mode, nothing is coming out the speaker terminals.
     
    Last edited: Feb 10, 2018
  16. cconger

    cconger New Member

    Messages:
    14
    I measured the DC offset at the grey and blue wires.... they both were .47V However. I put my multimeter on either aide of the STV3H diode and the one of the right measured 10v while the other was .6v. Ive sucessfully biased a 8080db but this 9090db is just getting me.
     

     

    Please register to disable this ad.

  17. tnsilver

    tnsilver AK Subscriber Subscriber

    Messages:
    3,246
    Location:
    Boston
    0.47V is 470mV which is probably enough to trip the relay on the protection circuit. IIRC it can tolerate voltages approaching 300mV - tops, but better research it.
    Is it coming out of protection? Neither 10V nor 0.6V seems right for the STV-3H diodes in idle bias current. The 10V is way too high. We tested 1.7Vf on the stock
    STV-3H diodes and about the same with the 3 1n4148 in series, and 2 X 1N4148 plus one UF1004 replacement diodes yielded about 2.2vF.
    The German Sanken STV-3HY measured about 2.5vF. Which temp/co diodes are you using for the measurement, what type do I see in your picture?
    Looks like something is still off with that driver amp. Do the differential pair transistors check out?

    Which components did you find faulty and replaced after the sparks incident?

    EDIT: please also solder the large heat sinks to the PCB using the two joints per each, in order to take the stress off the power transistors soldering points.
     
    Last edited: Feb 11, 2018
    Hyperion likes this.
  18. cconger

    cconger New Member

    Messages:
    14
    So far I haven't replaced anything because the sparks incident was 30minutes before I started this thread. However, I think I figured out the problem. The diff pair TR 01-04 all read differently on my multimeter. I will order some german stv3h diodes. The ones I am currently using are the stock factory ones. They where on the original pcb which burnt in the first fire after replacing R33.
     
  19. tnsilver

    tnsilver AK Subscriber Subscriber

    Messages:
    3,246
    Location:
    Boston
    Gotcha. Take your time. Look for shot components. First, low value resistors that are open and then transistors.

    Remember:
    • NPN transistors show continuity (low resistance, i.e. 'meter beeps') only on base-collector (red probe=b, black=c) and base-emitter (red probe=b, black=e). Any other combination that beeps means the NPN transistor is shot.
    • PNP transistors show continuity (low resistance, i.e. 'meter beeps') only on collector-base (red probe=c, black=b) and emitter-base (red probe=e, black=b). Any other combination that beeps means the PNP transistor is shot.

    [​IMG]

    The differential pairs (ZTX694B) are NPN's
     
    Sebastien and Hipocrates like this.
  20. cconger

    cconger New Member

    Messages:
    14
    After testing... I think I found the friend transistors. These are the results of putting C on positive and E on negative. TR01-02 go from 1000 on my meter to OL slowly. TR03-04 hover aroud 300v. TR05-06 hover around 600v and TR07-08-11-12-13-14-15-16 I assume are OK because they say OL and work with the other methods. TR17-18 I assume are fried because they hover around 900v on C and E when they are supposed to say OL. Are there any other components I should check before I Place an order on mouser? Should I get new VR01-02-03-04? I tested the all but zener and STV3H diodes and they all measure around 700 the same as used extras I have.
     

Share This Page