Power supply project: Constantly changing AC ripple after installing filter caps?

Discussion in 'The Projects Forum' started by Blatboy, Sep 22, 2014.

  1. Blatboy

    Thread Starter Member

    Jun 11, 2012
    46
    0
    Greetings. I'm building a power supply, using the book "TAB Electronics Guide to Understanding Electricity and Electronics" by G. Randy Slone. Feel free to check my earlier posts on this if you want to know more.

    I'm moving along here... have now installed the bridge rectifier, which went without a hitch, and the filter caps.

    According to the book, after I've installed the caps, I should get a reading of around 34v DC across them. All good there...checks out great.

    Then, it says to measure AC Volts across them to check the ripple, and I should get a reading of 5-20mV. When I check AC volts across them, I get a repeatedly changing value... seems to be oscillating, over a wide range.



    The book calls for me to use two 4400uF caps. I didn't have that, but I did have four 2200uF ones, so I wired them in parallel. I've double checked all the connections (after discharging the caps of course) and everything seems to check out ok.

    I noticed the behavior happens after the power has been turned off from the supply as well and the caps are charged.

    Is this a common error or a red flag to something specific?

    I may just go and get 2 4400uF caps anyway to make it more neat inside. Admittedly squishing all those caps in there makes it difficult to work on... made me glad I got my chops up for working in tight places from that classic old board game, "Operation" as a kid... haha.

    ahem.

    Thanks!
     
    Last edited: Sep 22, 2014
  2. Blatboy

    Thread Starter Member

    Jun 11, 2012
    46
    0
    Actually, here's this portion of the supply. Instead of C1 and C2... it's C1a, C1b and C2a, C2b wired in parallel. And... I used one center tapped transformer, not two transformers.

    [​IMG]
     
    Last edited: Sep 22, 2014
  3. Blatboy

    Thread Starter Member

    Jun 11, 2012
    46
    0
    and here, in my most amazing penmanship, complete with missing dots for connections around the caps (should be obvious what I meant, I hope) is how I wired the caps.

    filtercapsinparallel.jpg
     
  4. Wendy

    Moderator

    Mar 24, 2008
    20,764
    2,534
    There may be nothing wrong, or it could be a wiring error. Any chance you can get access to an oscope? My first suspicion is the diode bridge is either connected wrong or bad.

    It might help if you take a picture of the project itself.
     
    Blatboy likes this.
  5. #12

    Expert

    Nov 30, 2010
    16,247
    6,739
    That reading is so wrong that you must have a mistake in there. Your probes should be on the terminals of a single capacitor while the power is on. Period. Nothing else. You don't have any load connected, so the AC component should be really small...like 2mv.
    AC component = current/ (1.414 x C x F).
    If current is zero, AC should be zero, give or take some leakage in the rectifier, and that's going to be tiny.
     
    Blatboy likes this.
  6. Blatboy

    Thread Starter Member

    Jun 11, 2012
    46
    0
    I double checked the connections...tested them all. I hadn't actually been checking each cap individually...and I did. I got the same result. It seems the value sweeps between something very low to something around 28V.

    I have a Tek 2465 I just picked up. I'm no expert with it, but I do know the basics. Perhaps I should take a look at the AC ripple on the scope?

    I did test the rectifier before I used it and it seemed ok. (Doing diode tests across all pins in all combinations.) If it was wired incorrectly, wouldn't the DC voltage across the caps also have something wonky?

    thanks! I really appreciate your input!

    I have photos, but it's really hard to take a picture that shows the wiring, being that they're all so squished in there...

    IMG_0827.jpg IMG_0828.JPG IMG_0829.jpg IMG_0831.jpg
     
    Last edited: Sep 22, 2014
  7. MrChips

    Moderator

    Oct 2, 2009
    12,414
    3,353
    I agree with #12. Something is wrong. You should get no ripple with no load connected.
    Maybe one of your capacitors is bad.

    For initial testing, you don't need to put two capacitors in parallel. One will do for now. See if you can find a bad capacitor.
     
  8. Blatboy

    Thread Starter Member

    Jun 11, 2012
    46
    0
    So I should pull the caps and test them? I did test them individually, but they were connected (as they are in the pics.)
     
  9. MrChips

    Moderator

    Oct 2, 2009
    12,414
    3,353
    It looks like a lot of work to pull the caps and besides they look new.
    Do you have an extra bridge rectifier and cap?

    It appears that it is easier to unsolder the wiring at the secondary of the transformer. I would try to find out what is wrong by starting afresh with the output of the transformer to another bridge and one cap just hanging loose on the table for testing from centre-tap (GND) to one side of the bridge.

    MATCH THE POLARITY OF THE CAP CAREFULLY TO THE POLARITY OF THE BRIDGE.

    Edit: There is a chance that there is nothing wrong with your components and wiring. It may have something to do with your meter and metering. An oscilloscope test will reveal it all.
     
    Last edited: Sep 22, 2014
    Blatboy likes this.
  10. MrChips

    Moderator

    Oct 2, 2009
    12,414
    3,353
    If you are going to buy more capacitors, you don't have to have exact values.
    4700uF/63V would be more common than 4400uF.
     
    Blatboy likes this.
  11. KL7AJ

    AAC Fanatic!

    Nov 4, 2008
    2,039
    287
    Sounds like your meter is trying to change scale. Hit the range hold button. :)
     
    Blatboy and Wendy like this.
  12. #12

    Expert

    Nov 30, 2010
    16,247
    6,739
    That's true. An old, analog meter with a moving needle is a lot easier to interpret than a jumpy digital readout. The last one I bought was $20. Still have it after 25 years because it only gets used for jumpy readings.
     
    Blatboy likes this.
  13. Blatboy

    Thread Starter Member

    Jun 11, 2012
    46
    0
    I spent all evening in the basement working on this and I'm now seeing the last comments. I don't have an analog meter, but I do have a 4 channel scope. It took me a while, and I measured just about everything around those caps... individually and otherwise. By the time I really figured out how to use the scope smartly, I compared the signal coming out of the rectifier (top signal on picture) to what happens after the caps (bottom.) Check out the photo. Maybe I'm ok? I've never tried a range hold feature, but I know I have it on my meter.

    I feel like I really learned a lot on that scope tonight.

    IMG_0843.JPG
     
  14. #12

    Expert

    Nov 30, 2010
    16,247
    6,739
    Your time scale is too fast. 10 microseconds will distort the answer beyond recognition. You should be around 5 milliseconds per division.
     
    Blatboy likes this.
  15. Blatboy

    Thread Starter Member

    Jun 11, 2012
    46
    0
    guess I didn't learn enough though. ha! I'll do that tomorrow and get back to ya. Thanks!!
     
  16. The Electrician

    AAC Fanatic!

    Oct 9, 2007
    2,281
    326
    KL7AJ may have the reason.

    Use your meter to measure a flashlight battery with the meter set to measure AC. Measure with both orientations of the leads--first with red connected to the positive tip, then with black connected to the positive tip. Does your meter indicate anything?
     
  17. #12

    Expert

    Nov 30, 2010
    16,247
    6,739
    I hope you read post #5 well. You can't expect the voltage to hold still after you turn the power off!
    The natural discharge of the capacitors will look like AC to a digital meter.
     
  18. Blatboy

    Thread Starter Member

    Jun 11, 2012
    46
    0
    I've been tinkering at this stuff, trying to find the issue.
    I wasn't able to get an analog meter yet, but I did borrow a much nicer Fluke.

    I'm still confused at my findings.

    Just as a reminder, spec would be that the ripple on the caps should be in the 5-20mV range.

    The fluke, which certainly behaves better on this than my other two meters, seems to start at about 5mVAC then goes down and hang out around 1mVAC or just below it. Maybe, again, I'm somehow using the instrument incorrectly. Or, maybe even a really nice digital meter isn't the right tool.

    IMG_0853.JPG

    The scope, now that I think I've got it figured out, seems to imply the caps aren't doing anything.

    IMG_0862.JPG

    The top signal is from the rectifier. The bottom one is from one of the caps. (And I get the same result with all four.) The top one is using an old tek 10X probe that automatically updates the volts/div display on the screen. The bottom one is using a newer probe, set on 10X, but doesn't have the extra little nub to tell the scope to update the display... so it seems that one signal is at 20mV/div and one is 2mV/div, but they are actually both at 20mV/div. At a 5ms scale, the screen blinked too much... the trace wasn't fast enough (maybe because it's one beam drawing 2 signals?) so 2ms is still a good ballpark for this, yes?

    And, since obviously I don't really know what I'm doing, I figured I'd show how the scope was set in case I'm still doing something that would keep this from being accurate.

    IMG_0865.JPG

    So.... assuming this is correct... I'm leaning towards giving up on my idea of using four caps. I may just order two 4400uF caps and an analog meter... and hopefully be done with this part of it.

    Whaddya think?

    Thanks!
     
    Last edited: Sep 24, 2014
  19. MrChips

    Moderator

    Oct 2, 2009
    12,414
    3,353
    No, no, no. You are using your scope improperly.

    You should set your scope and probes to read 10V/div, DC.

    Start afreash. Remove all the capacitors and look at the +ve output of the bridge
     
    Blatboy likes this.
  20. Blatboy

    Thread Starter Member

    Jun 11, 2012
    46
    0
    set for DC? But I'm looking for AC ripple, yes? I am confused...
     
Loading...