Announcement

Collapse
No announcement yet.

Question for those who have stared at a scope screen for decades

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • nsubulysses
    replied
    Thanks Tom

    and everyone

    I have only used a scope for about a year so lots to learn

    Leave a comment:


  • Tom Phillips
    replied
    Originally posted by nsubulysses View Post
    At the B+ cap with DMM I get 3.5VAC at idle and about 11VAC when cranked to clipping. At the bias supply cap the AC stays at about .1V when idle or cranked.... is this within the an acceptable or normal range of AC?...
    I noticed that this question has not yet been answered so Iíll chime in and say Yes, all of those measurements seem to be normal for your amp.
    Cheers,
    Tom

    Leave a comment:


  • R.G.
    replied
    Just as a comment.

    Getting nice, solid non-wiggling scope traces is an art. In addition to the causes of odd traces mentioned above, you can get a few others.
    - ground loops: the oscilloscope is an "amplifier", and is plugged into the wall itself; it can have ground leakage issues, as can the circuit you're looking at, so you can get the equivalent of hum in terms of the scope seeing the signal wiggle up and down relative to the scope's own ground reference. This can happen even if you're reading a transformer secondary if the amp you're working on happens to connect the chassis ground to the speaker jack, even by simply mounting the jack bushing on the chassis. Also many other ways. If you happen to be looking at line frequency stuff, setting the trigger to "line" generally helps.
    - magnetic loops: Magnetic fields can impress a voltage across conductor loops - like the loop from the signal being measured through the o'scope and back down to the tip of the probe. That's why they sell little ground adapters that slip onto the probe tip ring right beside the inner pin.
    - ground is not ground: It can be maddening to try to get a stable trace if you only have one ground location on board or chassis. It is easy to find differences between "ground" at different places by clipping on the ground lead, then probing "ground" somewhere else. Your scope can easily trigger on this little wobble, or can trigger on this wobble plus the signal you're measuring. Worse, this voltage is often some combination of power line hum, rectifier ripple, and some component of your signal. The solution is to find the nearest ground point that is part of the circuit fragment you're trying to measure.
    - RF triggering: Most modern scopes are really, really good at triggering, including triggering on fast, small signals. The problem comes when you get spurious RF pickup on top of your signal from one of the above mechanisms and your trigger picks out random peaks on the RF added to the signal. For this one, you can try to set the trigger to "low frequency" or go shield your setup to keep RF out, or go single trace. Wow. Modern scopes naturally do storage. Storage scopes used to cost a fortune.

    Leave a comment:


  • J M Fahey
    replied
    Agree.
    Both measurements useful but for different things.
    If you have too much (as in many volts) ripple while idle, your caps are shot. Period.

    Or the common problem, mainly in SS amps,of having one main filter cap leg torn or its solder cracked, by excessive vibration, where the symptom is loud hum but no DC present on output and you measure, say, 40V DC on one rail but "28V DC" on the other .... this last one is in fact the rectified but not filtered "DC component" in full wave rectified AC.


    Now high ripple at full power (clipping a little), to the point of being annoying, means the supply is not exactly "broken" but filtering is insufficient.

    Leave a comment:


  • teemuk
    replied
    Originally posted by g-one View Post
    Normally, when checking for excess ripple it is done idling with no signal.
    This.

    The more current is drawn and the more power supply voltage sags the higher the ripple amplitude will become.

    You would measure entirely different readings on "idle" vs. "overdrive".

    Leave a comment:


  • J M Fahey
    replied
    Also, meter reading (RMS) times roughly 2.8 to get peak to peak voltage.
    True, but for sinewaves only.

    Square, triangle, sawtooth all differ ... let alone narrow pulses, less than 50% duty cycle.

    There RMS voltage shown on meter display will be *much* smaller than peaks, for the very good reason that meters usually integrate/average measurements during a certain time.

    Old TV repairmen will remember that on schematics, usually peak to peak signal values were given, because they could be accurately measured with a dedicated VTVM probe .

    I'm thinking about reviving my old Eico VTVM.

    Leave a comment:


  • g1
    replied
    On that last scope pic you can see the ripple is being modulated by your signal (the little steps in the waveform). Normally, when checking for excess ripple it is done idling with no signal.
    Not sure about the difference between the meter and scope, maybe due to the difference in signal (check with no signal).
    Also, meter reading (RMS) times roughly 2.8 to get peak to peak voltage.

    Leave a comment:


  • nsubulysses
    replied
    At the B+ cap with DMM I get 3.5VAC at idle and about 11VAC when cranked to clipping. At the bias supply cap the AC stays at about .1V when idle or cranked.

    Here's the AC at the B+ cap at .5V/div with 10X probe


    I have never "viewed" a power supply this way before. Thank you for your input and directions.
    I know these old caps have to go anyway, but is this within the an acceptable or normal range of AC? I'll have to start checking this on all upcoming amps to get an idea or normal behavior.

    Edit: it seems these DMM (Fluke 27/FM) and scope voltages do not add up??

    Edit: also I was unfortunately inconsistent with my test signals. The original video is 600Hz as it says on the digital scope but I think for future tests I used 400Hz, 600Hz, 1K and 2K to see if ripple was more or less visible at different frequencies.
    Last edited by nsubulysses; 10-02-2014, 11:52 PM.

    Leave a comment:


  • teemuk
    replied
    Try larger division, something that actually catches the waveforms of low 60 - 120 Hz frequencies in the power supply.

    Probably then you will see the "ripple" superimposed on the clipped signal peaks.... similarly to this image:

    At "peak point" of ripple wave the B+ is highest so the clipping threshold is respectively highest it can be with given loading as well. At lowest valley point of the ripple waveform the B+ voltage is lower and the signal is in turn clipped much earlier. With larger divisions on scope you should see the "sawtoothy" effect of the ripple modulating the clipping threshold and the output waveform.

    When you merely examine a high frequency waveform you only see the magnitude of clipped peaks "bouncing around" as the duration of the ripple pulse far exceeds your usual signal frequencies. Your scope at current settings operates too fast to portray this.

    Leave a comment:


  • J M Fahey
    replied
    1) yes, that's ripple

    2) there is VERY LITTLE of it.
    So little , in fact, that if you were measuring it on a PSU cap I wouldn't believe it , but as mentioned above, push pull transformer coupled circuits cancel most of it, that justifies it being so low

    3) the up down bouncing and waving on top and bottom peaks comes from the beating (google it) between the test signal and the ripple one.

    If they were *exact* multiples you would have a static signal ... but that's almost impossible, unless both your oscillator and the line frequency are very stable, so you will practically always have a small signal modulating the peaks.
    Since they are close anyway (your scope shows signal wandering between 599 and 601 Hz , or almost 5 X 120Hz , 10X 60 Hz ) the end result of them mixing creates a small AM modulation at the difference frequency, in your case a few Hertz, so it's visible.

    At higher differences the modulation will still exist, but it will be too fast to follow by eye, so you'll see a "fuzzy top" or an apparently thicker trace on top, on the bottom peak as well.

    Leave a comment:


  • Enzo
    replied
    And what else is on those grids? The bias supply. How much ripple is on the bias suply?

    Leave a comment:


  • Tom Phillips
    replied
    Sorry Dean. I got you mixed up with someone in another thread.

    If the cap cases are bulgey then I think you should replace then now.

    Cheers,
    Tom

    Leave a comment:


  • nsubulysses
    replied
    You are correct the scope shots are taken from the speaker jack. But also, my name is Dean.

    The ripple is at the speaker jack, and the power tube grids, and NOT at the phase inverter or before. Which had me wondering if the two caps I didn't replace were problematic. But I guess they are doing ok. They are from the mid 70s I'd guess and are original. They are bulgey and need to go, but still work ok I suppose.

    Enzo likes to do it the easy way with the DMM or the scope on AC at the filter cap, which makes great sense.

    Leave a comment:


  • Tom Phillips
    replied
    Originally posted by imaradiostar View Post
    ...In a push-pull amp it doesn't matter so much as the ripple cancels itself out of the audio at the power amp. By the time power gets filtered down past the screens and phase inverter that stuff is usually so far down it's north worth worrying about.
    Jamie
    That is correct and I grant that it's probably not a problem. However, the scope images that Wes has been showing are of the output of the amp. Correct? So the ripple we have been talking about is what has made it to the speaker.

    Leave a comment:


  • Enzo
    replied
    The ripple is on top of the other waveform, so we ignore the other and look only at the rippled tips. Get it more or less stable on th screen, then yes, measure teh top and bottom limits of it against the V/Div setting for a p-p amount. Frankly for that I get out my Fluke and set it to AC volts. On top of the 1kHs test wave that wouldn't work, but if this is ripple, why not go to the filter cap instead and just scope the power supply?

    Leave a comment:

Working...
X