Cuda's SX-850 Refurb Thread

Discussion in 'Pioneer Audio' started by cuda440, Jan 15, 2019.

  1. cuda440

    cuda440 AK Subscriber Subscriber

    Messages:
    229
    Location:
    San Jose, CA
    Thanks for the analysis. The base of Q10 comes in at 6.4v.

    But wait, there's more.

    The function selection indicators (there are 5 choices) are a decent brightness when no speakers are selected, but select either speakers (lighting one of the indicator bulbs) and the brightness dims to, well, about half brightness by my eye. Measuring the power supply (pin 3) it comes in at 6.1v with no speakers selected and 5.2v when one is selected.

    The function and speaker indicators share this power supply (in parallel) and are the only users of it.

    I have a SX-1050 to compare with and it has a similar design, 5 function indicators and 3 speaker indicators. Measuring the same supply shows 5.23v with no speakers selected, 5.19v with one, and 5.17v with two selected. No discernible difference in brightness.

    Sounds like the 850is having trouble sourcing enough current. What might cause that? What should I look at?
     

     

    Please register to disable this ad.

  2. kdris10

    kdris10 Active Member

    Messages:
    116
    Last edited: Jan 19, 2019
  3. cuda440

    cuda440 AK Subscriber Subscriber

    Messages:
    229
    Location:
    San Jose, CA
    The BOM I have specifies a 512-KSC2383YTA.
     
  4. Watthour

    Watthour Electron Rancher - JS3600

    Based on your measurements (see what I did there?) the base of Q9 should be around 5.8-5.9V, with the resultant regulated voltage being about 5.2V or so. At lower than the rated voltage, those lamp filaments are probably going to suck almost as much current to try to achieve the temperature needed to maintain filament resistance. Thus, that 8V/50mA lamp may be close to 42 mA at 5.4 filament voltage.

    Thus Q9 needs to pass at least 145mA without losing voltage - One function lamp and two speaker lamps, all 8V/50mA rated. The marking on the diagram shows 100mA, which is obviously only when one speaker set is selected.

    A KSD526Y would get it done if the original 2SD313 is gone, but the original 2SD313 should be more than sufficient. R20 should also be of little concern unless it is weakened, since 150mA through that should not appreciably affect it, and you can verify this by measuring the collector of Q9 when all the lamps are lit. I'm guessing it still shows at least 15V.

    Out of curiosity, when you load down the supply with three lamps lit, what happens to the 13.5V supply on Pins 4/5? If they change more than a few mV, there may be something happening upstream.
     
  5. tsd71

    tsd71 RIP Tom Petty

    Messages:
    2,617
    Location:
    New Jersey
    Understood, this has been bandied about for awhile now, but my experience building 850 and 950's (lots of them) when KSC2383 is used at Q10 it will dim the function lights, the 2N5551 works perfectly, no dim lights. I have a bunch if you wanna try. Also I use KSC2073 at Q9 with no issues.
     
  6. cuda440

    cuda440 AK Subscriber Subscriber

    Messages:
    229
    Location:
    San Jose, CA
    I swapped in a KSC2073 for Q9. No difference. So I put the KSC2383 back in for Q10 and then put back the two oddball transistors (Q8 and Q9) that I found when I got here. They are not 2SD313, they are labeled T D234 O 66. See the picture.

    That fixed it.

    So what are these? And why are these here instead of the 2SD313?
     

    Attached Files:

     

    Please register to disable this ad.

  7. Watthour

    Watthour Electron Rancher - JS3600

    2SD234s are rated for 3A. That should certainly do it. The specified 2SD313s are also rated for 3A. Perhaps Pioneer depleted its stock of 313s in production and subbed in 234s, which is never a complete shock when coursing through these old units.

    It'll be fine...
     
  8. cuda440

    cuda440 AK Subscriber Subscriber

    Messages:
    229
    Location:
    San Jose, CA
    Okay, I see it now. The SX-850 BOM I found specified the 863-MJE15032G for 2SD313 (Q8, Q9), which I didn't have on hand. Somehow I found the KSC2073 as a suitable replacement (I haven't quite figured out how I came up with the KSC2073) and I had them on hand so that's what I used. The 2073 doesn't have nearly the current capacity (1.5A) of the MJE15032 (8A).

    Am I on the right track?
     
  9. tsd71

    tsd71 RIP Tom Petty

    Messages:
    2,617
    Location:
    New Jersey
    Where are the heatsinks for those two transistors?
     
  10. Watthour

    Watthour Electron Rancher - JS3600

    That's a valid question. I presumed they were not yet installed after components were swapped, for testing only. Perhaps they were left out for modification, like these:

    [​IMG]
     
  11. cuda440

    cuda440 AK Subscriber Subscriber

    Messages:
    229
    Location:
    San Jose, CA
    Watthour has it right, not installed for the quick test. I'm going to order some 863-MJE15032G and install them.
     

     

    Please register to disable this ad.

  12. cuda440

    cuda440 AK Subscriber Subscriber

    Messages:
    229
    Location:
    San Jose, CA
    Now that this problem has been figured out, there was another problem associated with it that I didn't mention - a (60Hz?) buzz in the speakers. It was heard with the volume off and didn't vary when the volume was turned on/up. It was only present when one of the function lamps or the speaker lamps was lit - i.e. there was a load on the circuit in question. Since the SP-A lamp is burned out (and presented no load when selected), the buzz was pretty loud when A was selected (since B wasn't selected, no buzz was heard here). When I added speaker B, the brightness of the function was dimmed as the B lamp came into play, the buzz became half as loud, and present in all the speakers.

    So what was happening here?

    And thanks again for the detailed analysis.
     
  13. Watthour

    Watthour Electron Rancher - JS3600

    It sounds as if one of the filter caps in the power supply is failing or is completely gone. Take a look at C13 and C12 on the PS board, as well as D9 and D10. You may find a problem among those.
     
  14. cuda440

    cuda440 AK Subscriber Subscriber

    Messages:
    229
    Location:
    San Jose, CA
    All are new parts, Nichicon caps and On Semi 1N4004 diodes.
     
  15. cuda440

    cuda440 AK Subscriber Subscriber

    Messages:
    229
    Location:
    San Jose, CA
    When I put the old transistors back in, the brightness and the buzz problems went away. I'll update on this once I get the new parts from Mouser.

    In the mean time, I need some advice. So as I reported earlier, there are some rust spots on the chassis. It's going to take some grinding and sanding to remove them and not only does the rust eat right through the cadmium plating, so does the removal process. I'd like to dress it up once the rust is removed and was thinking I'd paint it. I can't really reproduce the cadmium look and so thought maybe I'd go satin (semi-gloss) black or perhaps silver.

    Has anyone else done something like this? Maybe just clear coat?
     
  16. bluehills5

    bluehills5 Active Member

    Messages:
    257
    Location:
    Brooklyn NY
    Personally I would not change the color. I'd just sand off the rust spots and perhaps clear coat or just coat the spots themselves - perhaps even a spot of grease to halt the corrosion and leave it at that. The transformer I'd paint with matte black.
     

     

    Please register to disable this ad.

Share This Page