Author Topic: super-low frequency PSU oscillations + AC coupling with electrolytics  (Read 2083 times)

0 Members and 1 Guest are viewing this topic.

Offline pepelevampTopic starter

  • Regular Contributor
  • *
  • Posts: 101
  • Country: nz
Hi guys. Riddle me this one.
My buddy and I (noob #1 and noob #2) are trying to figure out how to fix this awful noise floor of this audio tube amplifier: There is low frequency oscillation of the signal all by itself when nothing's plugged in. We are seeing these wiggles when simply recording silence on the PC (we know its not the audio interface, that goes past -120dB).

We're seeing sub and super-harmonics of 50hz, and at around 4hz its @ -84dBFS. When we turn the gain up & down on the amplifier, the wiggles/oscillations remain at constant amplitude.
Thus we suspect power supply in the amp and we suspect it is _after_ signal-amplification somewhere. We suspect something like the mains rectification diodes.

Now we don't have much gear. We just got ourselves a wee pocket oscilloscope. We want to probe around & find the oscillations in the circuit. But here is the problem:

The power supply inside the amp is 400 volts by the time it touches anything. -84dB wiggles of 400V are 40mV. So we're like ok cool - bring on AC coupling. No wakkas.

BUT:

We are uncertain if the AC-coupling in our wee DSO-112A pocket oscilloscope is cutting off all the low-frequencies we wanna see. And with a 10x probe to bring the DC into the scope's 50v tolerance, we fear we've attenuated the wiggles to nothing. We're not even certain if -84dB is the true value since its so close to DC that we fear its being attenuated by the AC-coupling of both our audio interface, the AC-coupling of the amplifier, and the DSO-112A pocket scope. My mate is the one doing the legwork here & can't make heads or tails of what the scope shows.

So I came up with an idea. Lets make our own AC-coupling which reaches unity gain at 1hz and above. Put this on the nose of our oscilloscope & go probing. But to make this AC-coupling / low cut front-end I need large capacitor values, namely whats provided by an Electrolytic capacitor.

My questions to the audience are:

* Is it safe to do AC coupling like this with an Electrolytic capacitor (even if we put several in series to achieve 400V tolerance)?
* Is Star Trek Picard going to be good?
* Is there any credibility to my claims of attenuated dB from the low-cut?
* Does anyone have a better approach?

The amp is the Sebatron vmp 4000e.
Yes we have changed the tubes & it hasn't resolved the problem.
Attached is the signals we can see (along with a different amp for comparison. Everything is at unity gain)

Thank you guys I know this is a bit of a corky one because its not actually me doing the arm/leg work but my mate is with us via 'tired-after-work' online comms. We are looking at building audio gear ourselves so this is a great exercise handed to us for our learning.

Thanks again everyone.
« Last Edit: July 18, 2019, 10:05:21 am by pepelevamp »
 

Offline TimFox

  • Super Contributor
  • ***
  • Posts: 9003
  • Country: us
  • Retired, now restoring antique test equipment
During the vacuum-tube era, many power supplies used a relatively high-inductance filter choke between two filter capacitors (usually less than 100 uF).  I have seen resonant frequencies of such circuits below 10 Hz, with Q > 1, that can magnify power-line fluctuations or interact with the amplifier.  What values are found in your amp’s power supply?
 

Offline mike_doom

  • Newbie
  • Posts: 5
  • Country: nz
Hi Tim,

I'm the guy with the Sebatron VMP Mic preamp. *See pics below*. I don't believe there are any chokes in this psu that i can see. I'm wondering if I just start replacing parts in the psu like the rectifying diodes and so forth and so on.
 

Offline exe

  • Supporter
  • ****
  • Posts: 2647
  • Country: nl
  • self-educated hobbyist
Afaik, an RC-snubber reported to be able to cause this. If so, it's fine. The article was on Rod Elliot's website. Can't find it atm.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Motorboating is due to a low frequency feedback path in the amplifier circuit.  It has little to do with power supply, though the feedback can be introduced through accidentally poor value RC or LC filters between stages (much as a phase shift oscillator works).

Just because an amp cost something, or has some weight in its name, means absolutely nothing.  The highest end amps are just as often, just as trash as the lowest.  This is not my experience directly, but a close friend does tube amp repair (among other things) regularly, and this is his experience.

He often fixes design faults in these things, as well as incidental repairs (burned parts, etc.).  (Most of which are caused by design faults anyway, i.e., that could've been reduced or eliminated through proper design.  Electrolytics placed beside stinking hot vacuum tubes, for example.)

(The low level 100Hz ripple also suggests poor design, either in filtering, balance or shielding.)

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 
The following users thanked this post: exe

Offline pepelevampTopic starter

  • Regular Contributor
  • *
  • Posts: 101
  • Country: nz
Hello.

I appreciate the help - but can we answer the question at hand? About AC-coupling and frequency response attenuation.
So we can continue investigating.
 

Offline pepelevampTopic starter

  • Regular Contributor
  • *
  • Posts: 101
  • Country: nz
Motorboating is due to a low frequency feedback path in the amplifier circuit.  It has little to do with power supply, though the feedback can be introduced through accidentally poor value RC or LC filters between stages (much as a phase shift oscillator works).

Just because an amp cost something, or has some weight in its name, means absolutely nothing.  The highest end amps are just as often, just as trash as the lowest.  This is not my experience directly, but a close friend does tube amp repair (among other things) regularly, and this is his experience.

He often fixes design faults in these things, as well as incidental repairs (burned parts, etc.).  (Most of which are caused by design faults anyway, i.e., that could've been reduced or eliminated through proper design.  Electrolytics placed beside stinking hot vacuum tubes, for example.)

(The low level 100Hz ripple also suggests poor design, either in filtering, balance or shielding.)

Tim

Don't give us this. This isn't a helpful contribution. This amp is in use because it has a purpose and we like what it does. We have our own reasons for fixing it - not for money. We also understand it could be a design not an introduced fault. We don't need to be told about something being great because of its brand name or whatever - that is condescending. Appreciate that we have our own reasons and offer us help - or not. But don't be demeaning like this.
 

Offline pepelevampTopic starter

  • Regular Contributor
  • *
  • Posts: 101
  • Country: nz
Afaik, an RC-snubber reported to be able to cause this. If so, it's fine. The article was on Rod Elliot's website. Can't find it atm.
Thank you this is awesome. Gonna go have a look at this. Thanks for your help
 

Offline Tomorokoshi

  • Super Contributor
  • ***
  • Posts: 1212
  • Country: us
Hi Tim,

I'm the guy with the Sebatron VMP Mic preamp. *See pics below*. I don't believe there are any chokes in this psu that i can see. I'm wondering if I just start replacing parts in the psu like the rectifying diodes and so forth and so on.

1. Need a larger overview picture.
2. Where does the Brown/Red pair of wires go to / come from?
3. Where does the Yellow/Green/Blue set of wires go to / come from?
4. What are the circuit traces on the bottom of that power supply board?
5. Are there any other power supply circuits?

Motorboating is due to a low frequency feedback path in the amplifier circuit.  It has little to do with power supply, though the feedback can be introduced through accidentally poor value RC or LC filters between stages (much as a phase shift oscillator works).

Just because an amp cost something, or has some weight in its name, means absolutely nothing.  The highest end amps are just as often, just as trash as the lowest.  This is not my experience directly, but a close friend does tube amp repair (among other things) regularly, and this is his experience.

He often fixes design faults in these things, as well as incidental repairs (burned parts, etc.).  (Most of which are caused by design faults anyway, i.e., that could've been reduced or eliminated through proper design.  Electrolytics placed beside stinking hot vacuum tubes, for example.)

(The low level 100Hz ripple also suggests poor design, either in filtering, balance or shielding.)

Tim

Don't give us this. This isn't a helpful contribution. This amp is in use because it has a purpose and we like what it does. We have our own reasons for fixing it - not for money. We also understand it could be a design not an introduced fault. We don't need to be told about something being great because of its brand name or whatever - that is condescending. Appreciate that we have our own reasons and offer us help - or not. But don't be demeaning like this.

6. Tim just saved you a few days of pain.
7. If this forum has a fault, it is a tendency to give more information than is requested, or less information than requested, but seldom just the information that is requested.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 22436
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
I would gladly give a more specific assessment, if a schematic can be provided. :)

mike_doom: you can, sure.  Diodes are the least likely culprit though.  More likely reactive components (like the capacitors in the possible RC filters I alluded to), or failings in the topology (too many capacitors, in a sense; but more correctly, in an unlucky combination), or effectively failings in the topology but caused by failed components (that is, an open or short circuited component effectively changes the circuit).

This is why a schematic, and preferably layout and wiring diagrams, are critical information to provide.  I guess it would be condescending of me to assume you knew that, and for that I apologize.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline pepelevampTopic starter

  • Regular Contributor
  • *
  • Posts: 101
  • Country: nz
Re: super-low frequency PSU oscillations + AC coupling with electrolytics
« Reply #10 on: July 20, 2019, 01:21:58 am »
6. Tim just saved you a few days of pain.
7. If this forum has a fault, it is a tendency to give more information than is requested, or less information than requested, but seldom just the information that is requested.
Its not the electronics tips, its the going on about it brand names of amps or not being worth anything etc. Its irrelevant & condescending.
Besides we asked about whether it was safe to construct an AC-coupling with an electrolytic so we could continue investigating, because its complicated trying to relay details over the forum like this.
My gut feeling is no. But it wont be till monday or so until we have an opportunity to get some help with it by someone who knows more about power electronics and has more capable gear. But we are still weary of even high-end scopes AC coupling frequency response which is why we asked.
 

Offline mike_doom

  • Newbie
  • Posts: 5
  • Country: nz
Re: super-low frequency PSU oscillations + AC coupling with electrolytics
« Reply #11 on: July 20, 2019, 04:41:51 am »
I wish I could provide a schematic but i have been unable to get one from the manufacturer or anywhere for that matter. *See a zoom out pic. I'll try get some more information  on the components.
 

Offline mike_doom

  • Newbie
  • Posts: 5
  • Country: nz
Re: super-low frequency PSU oscillations + AC coupling with electrolytics
« Reply #12 on: July 20, 2019, 04:53:11 am »
1*see new pic
2 led next to power switch on front panel
3 an input transformer I believe which is switchable 230v/120v etc
4 ill have to get back to you on that.
5 no just that one there

-The black molex socket provides power to different rails on the main mic pre-amp board.  *labeled voltages Ive currently unplugged the rainbow coloured cable going to the main board.

-Blue and brown are attached to power switch on front panel

-the small transformer *as shown in pic here which is just after the main input transformer is this guy here:
 https://www.altronics.com.au/p/m7064a-powertran-5va-12+12v-pcb-transformer/
 

Offline mike_doom

  • Newbie
  • Posts: 5
  • Country: nz
Re: super-low frequency PSU oscillations + AC coupling with electrolytics
« Reply #13 on: July 20, 2019, 04:55:05 am »
see labeled rails
 

Offline mike_doom

  • Newbie
  • Posts: 5
  • Country: nz
Re: super-low frequency PSU oscillations + AC coupling with electrolytics
« Reply #14 on: July 20, 2019, 08:02:25 am »
I Should also make note that those sub harmonics still show even when the 12At7s tubes are removed.

 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf