Very faint hum

Jimb0 · 4245

0 Members and 1 Guest are viewing this topic.

Offline Paul Birkeland

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 19316
Reply #45 on: February 02, 2020, 06:05:30 AM
Well, that's 13,000,000Hz, you can't hear that.  For low level measurements like that, it helps to be on the x1 mode on the probe.  x10 divides the signal to reduce the voltage substantially, so if you want to probe a 500V power supply and your scope has a 100V max input, then a 10x probe is helpful.

Since you have a good modern scope, there is a math function built in I'm sure.  You can set a lowpass filter at 5kHz and that will filter out all the nebulous grunge you're picking up.

Paul "PB" Birkeland

Bottlehead Grunt & The Repro Man


Offline Jimb0

  • Sr. Member
  • ****
    • Posts: 289
Reply #46 on: February 02, 2020, 06:24:42 AM
I was actually on the DC coupled before so I couldn't lower my scale. I went to AC coupled and the 275v unregulated has about 200mv ripple.

https://imgur.com/a/9LexdGf

The regulated is fairly clean
https://imgur.com/a/lfULDOp

I do have a math menu but I don't see a lowpass filter
https://imgur.com/a/lfULDOp



Offline Jimb0

  • Sr. Member
  • ****
    • Posts: 289
Reply #47 on: February 02, 2020, 06:37:00 AM
I found the lowpass filter but could only only limit 20 mhz



Offline Doc B.

  • Administrator
  • Hero Member
  • *****
    • Posts: 9539
    • Bottlehead
Reply #48 on: February 02, 2020, 06:54:42 AM
This is just a sidebar since the ripple before the regulator doesn't say too much about what it is after the reg. That first image is a fair amount less than 200mV. You will have to mentally subtract that hash your leads are picking up. It helps to adjust the y position of the trace so that either the peaks or troughs are touching a horizontal grid line. If you do I think you will see it's 150mV pk-pk or less. Which is about 50mV rms. Ripple is typically expressed in Vrms.

To figure out how to get the noise down so you can see what is going on with the regulated output ripple, take the leads off the circuit. You will probably see some very large sine wave, which is just pickup of EMI in the room. But see if the hash is still riding on the sine wave. That will help to determine if it is being picked up in the circuit or something that is being picked up from the environment.

Picking up this kind of hashy stuff riding on the main signal is pretty common. We had to give up doing low level tests at the old Bottleheadquarters because there was so much EMI/RFI being generated in the environment by our neighbors, and probably by that cell tower across the street. Luckily PB has a pretty quiet environment at his lab and we now use numbers measured there for our noise figures. Are you using a scope probe, or just test leads? If test leads, try twisting them together. That can help reduce some types of interference.

Dan "Doc B." Schmalle
President For Life
Bottlehead Corp.


Deke609

  • Guest
Reply #49 on: February 02, 2020, 10:55:53 AM
What Doc B and PB said - your scope is triggering on non-audible, super high frequency stuff from computer, routers, wireless phones, etc.

It's been a few months since I played around with my scope, and I'm barely past the stage of just knowing how to turn it on/off and hook up the probes, so I might be a bit off about what follows. But it may help.

i think you need to increase the horizontal time divisions A LOT.  From  your pics it looks like you had the time division set to 4 ms - I think that's too small of a window to show and trigger on a 60 Hz waveform.  Try 100 or 50 Ms instead.  Or, to get some practice dialing it in visually/manually, here's what I suggest for searching for 60 Hz hum.  Use both probes/channels.  Set the switch on probe/channel 1 to 10X and hook it up to the hot wire of your incoming AC voltage on the IEC inlet.  Make sure you have probe 1 set for 10X on the scope. Dial in the horizontal divisions until you see a nice 60 Hz waveform. If it's moving all over the place, that means the scope isn't auto-triggering or triggering on the wrong thing. Either hit auto-trigger, manually select a trigger (e.g., falling or rising slope), or manually adjust the trigger line so that it falls within the signal image - any of these should give you a stable waveform on the screen. Once you have that, you know how to dial in the horizontal time divisions for probe/channel 2.  Set the switch on probe 2 to 1 X and do the same on the scope.  Hook up probe 2 to either the left or right rca output jack (NOT the power supply, b/c that might damage the probe/scope depending on how much voltage they can withstand).  Dial in the horizontal divisions to match the figure you used for probe 1.  At the this point the easiest way to see what's going on at the rca output is to turn off the display of probe 1, so only probe 2 is visible. Adjust the vertical divisions until you see a waveform (assuming there is a 60 Hz waveform).  Then you can measure it (and I suspect your scope can do rms measurements for you - on my scope it's found under a menu intuitively named "measure").  There's also a way to keep both probe traces on screen by shifting them away from each other using the vertical position adjustment- but in your case that doesn't serve any testing purpose (although it looks kinda cool and is worth figuring out how to do for when you want to compare simultaneous waveforms).

Hope this helps, Derek



Deke609

  • Guest
Reply #50 on: February 02, 2020, 03:06:38 PM
One last suggestion - if the 60 Hz hum on the outputs is really faint/weak, it might be difficult for your scope to isolate it from all the other rfi emi and other noise. In that case, you should try what PB suggested above and suggested to me a while back - use the math functions. If I recall correctly, FFT (but possibly some other function - I can't remember) enables you to focus in on a frequency or a band of frequencies. A few months back it helped me to focus in on a hum waveform in my overly modded BeePre.

cheers, Derek



Offline Doc B.

  • Administrator
  • Hero Member
  • *****
    • Posts: 9539
    • Bottlehead
Reply #51 on: February 02, 2020, 05:10:27 PM
Quote
i think you need to increase the horizontal time divisions A LOT.  From  your pics it looks like you had the time division set to 4 ms - I think that's too small of a window to show and trigger on a 60 Hz waveform.  Try 100 or 50 Ms instead.

The horizontal scale is 4ms/div. The magnitude of one cycle of the sine waveform is approx two 4ms divs wide, or about 8ms. 1cycle/.008s is 125Hz. It's a 120 Hz wave.

Dan "Doc B." Schmalle
President For Life
Bottlehead Corp.


Deke609

  • Guest
Reply #52 on: February 02, 2020, 06:29:20 PM
Oops. My bad. Glad I prefaced my first post with "may be a bit off".  :-[  I was focused on the frequencies the scope was reporting in the 100s of kHz and 10s of MHz. I thought I recalled that increasing the time per division knocks down the sampling rate and can prevent the scope from triggering on higher frequency repeating signals.



Offline Doc B.

  • Administrator
  • Hero Member
  • *****
    • Posts: 9539
    • Bottlehead
Reply #53 on: February 02, 2020, 06:45:49 PM
I seem to infuriate some people when I use a scope, because I tend to actually use the scales and interpolate measurements from them. Probably has to do with being so old I learned how to use one before they had all of the built in numerical readouts.

Dan "Doc B." Schmalle
President For Life
Bottlehead Corp.


Offline Paul Birkeland

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 19316
Reply #54 on: February 03, 2020, 04:33:39 AM
Probably has to do with being so old I learned how to use one before they had all of the built in numerical readouts.
Yes, it's this. I am completely useless with an analog scope!

For your scope, page 49 in the manual has some stuff you can try for a noisy signal. 

« Last Edit: February 03, 2020, 04:45:40 AM by Paul Birkeland »

Paul "PB" Birkeland

Bottlehead Grunt & The Repro Man


Offline Doc B.

  • Administrator
  • Hero Member
  • *****
    • Posts: 9539
    • Bottlehead
Reply #55 on: February 03, 2020, 08:16:25 AM
FWIW I was the guy at Berkeley who couldn't get thru the five page long homework problems in Classical Dynamics 101, but could show the guys who got As in that class where the power switch was on a scope and how to set it up to get useful measurements in senior physics lab. I think we were equally shocked by the other's lack of intelligence.

Dan "Doc B." Schmalle
President For Life
Bottlehead Corp.


Offline Jimb0

  • Sr. Member
  • ****
    • Posts: 289
Reply #56 on: February 06, 2020, 10:09:29 AM
Thank you guys for all your responses. I need to sit down and really try to figure it all out so I printed out this page for guidance. Will check back in soon!



Offline Jimb0

  • Sr. Member
  • ****
    • Posts: 289
Reply #57 on: February 08, 2020, 03:06:53 PM
I'm not sure if I mentioned this before or if this means anything at all but I noticed when I first turn it on the hum is fairly loud for a few seconds, followed by a static noise, then into the 60 hz faint hum shortly after that



Offline Paul Birkeland

  • Global Moderator
  • Hero Member
  • *****
    • Posts: 19316
Reply #58 on: February 08, 2020, 03:12:03 PM
The amp shouldn't be quiet when you first turn it on.  With nothing conducting, the high voltage regulators aren't doing their jobs to clean up the high voltage supply. 

Paul "PB" Birkeland

Bottlehead Grunt & The Repro Man


Deke609

  • Guest
Reply #59 on: February 08, 2020, 03:39:37 PM
For your scope, page 49 in the manual has some stuff you can try for a noisy signal.

I haven't tried this, but it might work to strip out some EMI/RFI noise.  But it will only potentially work if your scope can invert a signal and sum the results of two probes. I got the idea from reading on the Emission Labs website about the following way of visualizing signal distortion on a scope: (1) use one probe to scope the incoming (undistorted) signal and a second probe to measure the outgoing signal; (2) invert one of the signals; and (3) sum the two signals. Whatever is common to both should cancel out (because one is inverted) and leave only the distortion.

I figure the same approach might work for stripping out emi/rfi: (1) attach one probe to measure the signal you're interested in; (2) lay out the second probe close to the first probe, but don't connect it to the amp (and make sure the probes tips aren't contacting the amp - leave the plastic probe tip protectors on if you have them); (3) invert the signal of the second probe and sum the two signals.  Assuming both probes are picking up the same emi/rfi, the noise should cancel out. If anyone tries this, please post about whether it works or not. I plan to try it the next time I'm fooling around with my scope.

[Edit: It just occurred to me that you probably need to connect the ground clip of the second probe to the amp - preferably to the same spot you connected the ground clip of the first probe]

cheers, Derek

« Last Edit: February 08, 2020, 03:54:29 PM by Deke609 »