You are here

Looking for help to fix power supply

Customising, building or repairing your own gear? Need help with acoustic treatment or soundproofing? Ask away…

Re: Looking for help to fix power supply

Postby jjlonbass » Thu Oct 22, 2020 10:15 am

ESR is definitely important in power supply smoothing capacitors, especially when they are expected to supply heavy currents.
Thinking about the shape of the ripple traces shown here, they do look odd. I'd expect to see more of a normal sawtooth shape with a steep rising slope as the rectifier diodes charge the capacitor then a more gentle trailing slope as the load discharges the capacitor. These traces show a sort of additional sinusoidal curve at the top. This could well be due to excessive ESR and as it's seen on every cycle it is unlikely to be due to a problem with the rectifier diodes as a different pair is used on each half cycle.
I'd advise replacing the 6800uF smoothing capacitor with a similar value type of the same sort of dimensions rated at 105 degrees C and specified with the highest ripple current / lowest ESR you can find. Choose a reputable make such as Panasonic, Nichicon or Rubycon if you can.

John
jjlonbass
Poster
Posts: 11
Joined: Wed Dec 19, 2018 4:52 pm

Re: Looking for help to fix power supply

Postby wireman » Thu Oct 22, 2020 12:11 pm

jjlonbass wrote:I'd advise replacing the 6800uF smoothing capacitor with a similar value type of the same sort of dimensions rated at 105 degrees C and specified with the highest ripple current / lowest ESR you can find. Choose a reputable make such as Panasonic, Nichicon or Rubycon if you can.

My capacitors just arrived although annoyingly the wrong product was sent for the 6800μF so awaiting a replacement.
So I will have a choice of:

2x 4700μF
6800μF
10000μF

Was going to try the last one.
Figured there was no harm in having some spares.
wireman
Frequent Poster
Posts: 559
Joined: Fri Dec 17, 2004 1:00 am

Re: Looking for help to fix power supply

Postby Folderol » Thu Oct 22, 2020 12:30 pm

Having the choice, I'd go for the 10,000μF one. As a matter of interest that could quite legitimately printed as 10mF, but nobody does! Maybe it's a 'look at all these zeros' thing :lol:

Odd really. Nobody writes 10mV as 10,000μV!
User avatar
Folderol
Jedi Poster
Posts: 11629
Joined: Sat Nov 15, 2008 1:00 am
Location: The Mudway Towns, UK
Yes. I am that Linux nut.
Onwards and... err... sideways!

Re: Looking for help to fix power supply

Postby merlyn » Thu Oct 22, 2020 2:18 pm

You do see this with lithium ion batteries. The battery capacity is usually given in mAh. So you see ridiculous looking capacities like 26800 mAh for a power bank.

The unit for lithium ion batteries has become mAh and the unit for electrolytic capacitors has become uF -- possibly saves confusion and avoids moving decimal points :)
merlyn
Regular
Posts: 149
Joined: Thu Nov 07, 2019 3:15 am

Re: Looking for help to fix power supply

Postby wireman » Thu Oct 22, 2020 4:54 pm

Project fix-supply is currently held up until I can get some cable ties < 2.5mm wide, hopefully by the weekend. I can't easily solder the capacitor and add cable ties later, the existing ones broke into pieces.

I have taken out the existing capacitor (6800µF) and my multimeter measures it at approx 1400µF.

By the way the resistor is in parallel with the output so the capacitor would be discharged.
wireman
Frequent Poster
Posts: 559
Joined: Fri Dec 17, 2004 1:00 am

Re: Looking for help to fix power supply

Postby wireman » Fri Oct 23, 2020 4:42 pm

Supply is fixed !

10,000uF capacitor installed and nothing getting hot. Scope trace below for a 10.1Ω load.

Image

A bit fuzzy but Ch1 (yellow) is DC coupled 5V/div, Ch2 AC coupled 500mV/div and traces are relative to arrows. Both channels connected to the output of the supply so are the same signal.

Ripple was much less (0.7Vpp) for the 20Ω load which would approximate to more normal usage for me.

The output voltage of the transformer was 15.4Vrms.
The output voltage is higher than I expeted even allowing for a possible load of 4A which for me is not going to happen.

Thanks for all the help.
wireman
Frequent Poster
Posts: 559
Joined: Fri Dec 17, 2004 1:00 am

Re: Looking for help to fix power supply

Postby Hugh Robjohns » Fri Oct 23, 2020 4:48 pm

That looks very much healthier. That top trace is exactly what ripple on an unregulated supply should look like!

Presumably the amplifier hum has gone away now, too?
User avatar
Hugh Robjohns
Moderator
Posts: 28956
Joined: Fri Jul 25, 2003 12:00 am
Location: Worcestershire, UK
Technical Editor, Sound On Sound

Re: Looking for help to fix power supply

Postby wireman » Fri Oct 23, 2020 4:58 pm

Yes, no hum.
wireman
Frequent Poster
Posts: 559
Joined: Fri Dec 17, 2004 1:00 am

Re: Looking for help to fix power supply

Postby Hugh Robjohns » Fri Oct 23, 2020 4:59 pm

:thumbup: :clap: :D

The DC regulators in the amp will be able to deal with the reduced ripple amplitude, whereas with the failed smoothing capacitor there was just way too much 100Hz rippled coming in and the regulators just didn't have the range to smooth it all out.
User avatar
Hugh Robjohns
Moderator
Posts: 28956
Joined: Fri Jul 25, 2003 12:00 am
Location: Worcestershire, UK
Technical Editor, Sound On Sound

Re: Looking for help to fix power supply

Postby Folderol » Fri Oct 23, 2020 6:19 pm

Yep. Now looks like an absolutely classic example :thumbup:
User avatar
Folderol
Jedi Poster
Posts: 11629
Joined: Sat Nov 15, 2008 1:00 am
Location: The Mudway Towns, UK
Yes. I am that Linux nut.
Onwards and... err... sideways!

Re: Looking for help to fix power supply

Postby jjlonbass » Fri Oct 23, 2020 6:32 pm

Well done wireman, an excellent result.

John
jjlonbass
Poster
Posts: 11
Joined: Wed Dec 19, 2018 4:52 pm

Re: Looking for help to fix power supply

Postby wireman » Sat Oct 24, 2020 12:59 pm

jjlonbass wrote:ESR is definitely important in power supply smoothing capacitors, especially when they are expected to supply heavy currents.
Thinking about the shape of the ripple traces shown here, they do look odd. I'd expect to see more of a normal sawtooth shape with a steep rising slope as the rectifier diodes charge the capacitor then a more gentle trailing slope as the load discharges the capacitor. These traces show a sort of additional sinusoidal curve at the top. This could well be due to excessive ESR and as it's seen on every cycle it is unlikely to be due to a problem with the rectifier diodes as a different pair is used on each half cycle.
John

It is definitely worth trying to understand this.

The more I think about this I'm sure what is happening is that for that period around the peak of the rectified AC we have the capacitor charging but it can't reach the expected voltage in time due to the ESR and the load. Voltage is being dropped across the ESR, only when the rectified AC drops below some threshold is the capacitor supplying the current.

Here is a transient response trace from the Partsim website for a bridge rectifier with the same circuit values I had and adding a 1.5ohm resistor in series with the filter capacitor acting as a simple model for a capacitor with ESR.

Image

The traces are:

  • Orange Filter Capacitor current
  • Purple AC Input to rectifier
  • Blue Output Voltage (across load resistor, across cap+ESR)
  • Green Voltage across ESR
  • Black Voltage across Capacitor

If I learn more from the failed capacitor I will update in case anyone is interested.

What amazes me is that this failure mode must be something very common yet I can't find any treatment of this in simple terms showing actual expected waveforms.
wireman
Frequent Poster
Posts: 559
Joined: Fri Dec 17, 2004 1:00 am

Re: Looking for help to fix power supply

Postby Folderol » Sat Oct 24, 2020 8:11 pm

It's common enough in 'proper' electrical/electronic training, but these days is considered so mundane that many don't bother to mention it at all, on the rationale that nobody builds PSUs these days :tongue:
User avatar
Folderol
Jedi Poster
Posts: 11629
Joined: Sat Nov 15, 2008 1:00 am
Location: The Mudway Towns, UK
Yes. I am that Linux nut.
Onwards and... err... sideways!

Re: Looking for help to fix power supply

Postby jjlonbass » Mon Oct 26, 2020 9:49 am

wireman wrote:
If I learn more from the failed capacitor I will update in case anyone is interested.

What amazes me is that this failure mode must be something very common yet I can't find any treatment of this in simple terms showing actual expected waveforms.

Failure of electrolytic capacitors due to excessive ESR is very common - even more so with equipment that uses switched-mode power supplies i.e. just about every mains powered device made in the past 15 years or so. It seems to be especially common in flat-panel televisions, probably because these run quite warm. I know of many people who've been given dead LCD TVs or rescued them from skips - replacement of a handful of electrolytics often brings these TVs back to life.

John
jjlonbass
Poster
Posts: 11
Joined: Wed Dec 19, 2018 4:52 pm

Re: Looking for help to fix power supply

Postby merlyn » Mon Oct 26, 2020 5:13 pm

Yes, pretty interesting. Did you use the actual value that you measured -- 1400 uF?

The impedance of a 1400 uF capacitor at 50 Hz is :

1/(2 * pi * 50 * 1400 * 10^-6) ~ 2.3 Ohms (forget about j :))

So now you've got a potential divider with the ESR -- the maximum voltage the capacitor can get to is :

2.3/(1.5 + 2.3) * 20 = 12.1 V

I think that's the general idea.
merlyn
Regular
Posts: 149
Joined: Thu Nov 07, 2019 3:15 am

PreviousNext