Is bypassing PSU capacitors effective?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Jump to Latest
KSTR

KSTR

Member
Joined 2007
AndrewT said:
Now,
can I ask you to start a thread on building that pulse generator?
Click to expand...
Here an AppNote to a classic pulse generator design method, with an Avalanche Transistor:
http://www.zetex.com/3.0/appnotes/apps/an8.pdf
If you need higher power, build a spark gap*** ;)

My edge rate test voltage generator is simple: 74HCU04 oscillator --> 74HC541 octal bus driver, divided down to 0.5V and 50Ohm impedance or other values, depends on application. Powered from 4x1.5V dry cells. To get higher edge rates, a logic family with higher slew rate can be used (but check supply range).

**) or simply use a piezo oven lighter. I use this spark also for "hardcore" ESD-tests. Usually, only tube gear survives...

- Klaus
 
M

metalsculptor

Member
Joined 2010
fff0 said:
Hi guys, So can I assume, its not a bad idea to use no more than 1/50 the value of the main capacitor value to avoid the above said problems?

Example: 1000uF bypass with 20uF film
Ok for such arrangement even without doing stimulation or measurments?
Click to expand...

Why bother, I doubt the 20uf film cap will have an ESL much different from the Electrolytic capacitor so you might as well use a pair of 1000uF electrolytic capacitors. For critical systems, bypass capacitors are low ESL devices mounted near the load. Using twisted leads from the PSU to the amplifier board will reduce the ESL more effectively than adding capacitors to the PSU. Any bypass capacitor needs to be near the load.
 
Karl vd Berg

Karl vd Berg

Member
Joined 2012
Old power caps - Increased ESR with age

This thread has been "around" on other forums as well.

Just to ask further info: I understood that the use of bypassing small film caps with LARGE electrolytics could apply ONLY if the large electrolytic capacitor is too old, right? Let's say in a vintage amplifier from the 70s and early 80s? See the comment below:

The purpose of a film bypass (parallel cap) is to lower the impedance of the large electrolytic cap at high frequencies. When the practice of using film cap bypasses for large electrolytic caps was initially introduced in the late 70s/early 80s, the then current electrolytic capacitors had increasing ESR with increasing frequency and the film cap compensated for this. But with modern low ESR capacitors, this is no longer necessary since these have low impedance out to 100kHz. In fact, a small film capacitor in parallel with an electrolytic capacitor NOT having high ESR will actually result in higher overall output impedance and ringing.
Click to expand...
So, If we are recapping an old amplifier (vintage) and have to keep the main old caps (increased ESR due to age), what's the "general rule" (if there is any!) for the bypassing caps' value?

I have seen people using ~1uF for every 10k uF. For example, a 1uF film cap bypassing 10,000uF; or 1.5uF bypassing a 15,000uF; finally 2.2uF film cap bypassing a large 22,000uF??
 
Last edited:
Baldin

Baldin

Member
Joined 2005
sensibleaudio.dk/
Very good points Eva and others ..... we always would like to think it's real easy, and just adding some more capacitors will make it all better
When using a GND plane it still makes sense to use decoupling near the load .... smd preferably.
See also the 4 small videos here .... and notice how important it is not to mix too many different values:
https://www.youtube.com/watch?v=mk61DNz27FI
 
A

AndrewT

R.I.P.
Joined 2004
Karl vd Berg said:
.................Just to ask further info: I understood that the use of bypassing small film caps with LARGE electrolytics could apply ONLY if the large electrolytic capacitor is too old, right? Let's say in a vintage amplifier from the 70s and early 80s? See the comment below:

So, If we are recapping an old amplifier (vintage) and have to keep the main old caps (increased ESR due to age), what's the "general rule" (if there is any!) for the bypassing caps' value?.....................
Click to expand...
Don't.
Put the decoupling in a location where it can react to transient demand, i.e. at the output device leadouts.
 
G

gootee

Member
Joined 2006
fullnet.com/~tomg/index.html
Karl vd Berg said:
This thread has been "around" on other forums as well.

Just to ask further info: I understood that the use of bypassing small film caps with LARGE electrolytics could apply ONLY if the large electrolytic capacitor is too old, right? Let's say in a vintage amplifier from the 70s and early 80s? See the comment below:

So, If we are recapping an old amplifier (vintage) and have to keep the main old caps (increased ESR due to age), what's the "general rule" (if there is any!) for the bypassing caps' value?

I have seen people using ~1uF for every 10k uF. For example, a 1uF film cap bypassing 10,000uF; or 1.5uF bypassing a 15,000uF; finally 2.2uF film cap bypassing a large 22,000uF??
Click to expand...

If we are recapping an old piece of electronic gear, then we DO NOT keep the old main caps. Those would be the first and most-important caps to replace.

If we are recapping a beautiful old ANTIQUE piece of electronic gear, the closest we would ever come to not replacing the old main reservoir/filter caps would be to hollow them out and install the modern electrolytics inside the old caps' cases.

All small-value bypass/decoupling caps can only be beneficial if they are within a millimeter or two of the point of LOAD, where the transient current demands need to be met.

Also, do not put small caps alone in parallel with rectifier diodes. At best, they will only lower the frequency of any ringing or resonance.

A "snubber" is a resistor. Usually, a snubber resistor has a small capacitor placed in series with it, so that only the unwanted frequencies have access to the resistor, which is only done so that the snubber resistor will not dissipate too much power, or affect other frequencies. Do it the best way: Put an RC snubber across each transformer secondary, not at each diode. If it occasionally does prove necessary to snub each diode, use an RC snubber, not just a C.

"Large companies" do not place single caps as snubbers across rectifier diodes. An individual who works at a large company might do that. But the fact that it happens does not imply that it is good practice. (Theoretically, there might be a "perfect cap" scenario where the ESR of a cap is the optimal R needed as the snubber. But I've never seen it in practice.)
 
D

DF96

R.I.P.
Joined 2007
myweb.tiscali.co.uk/g8hqp/
Large companies might not place single caps as snubbers across rectifier diodes; I don't know. They might place single caps as RF bypasses across rectifier diodes - to avoid modulation hum in equipment which contains an RF oscillator or is used in strong external RF fields.

The relevant R for a single cap used as a snubber is not cap ESR but the resistance of the entire secondary circuit. This may or may not be sufficient to damp the resonance by the required amount, which may or may not be equivalent to critical damping. Simply adding a cap will drop the resonant frequency and lower circuit Q. This may be enough. At the very least it will reduce induction to nearby circuits.
 
G

gootee

Member
Joined 2006
fullnet.com/~tomg/index.html
Hi DF. I am sure that you are correct. My post wasn't meant for engineers or anyone who knows all about this stuff. It was probably more like when people tell very young children that they should never lie, even though the person telling them that knows that there will eventually be hopefully-very-rare times when the child is older that lying might be the best course to take, but trying to explain that to the small child would just confuse the issue.

I didn't even know about that type of RF application. So I'll have to claim ignorance on that. But I probably wouldn't have mentioned it, in my post, even if I had known about it. Maybe I've gone overboard. I don't know. Maybe I should stop worrying about it so much.
 
D

DF96

R.I.P.
Joined 2007
myweb.tiscali.co.uk/g8hqp/
As in all engineering, you have to decide what you want to do before you decide how to do it.

Putting small caps across rectifier diodes will prevent the sudden change in current caused by charge storage - or at least it will confine it to the small loop consisting of the diode and cap. By reducing the area of this loop it will reduce induction and radiation. It will also reduce the effect in the wider secondary loop. Using a full snubber across a diode might reduce the effectiveness of the cap because in a resistive circuit the voltage and current can, in principle, have instantaneous changes (or nearly so in real life).

There may still be a slower change in current caused by normal diode switch-off. A snubber across the secondary may help reduce the ringing effect of this.
 
S

syklab

Member
Joined 2009
DF96 said:
As in all engineering, you have to decide what you want to do before you decide how to do it.

Putting small caps across rectifier diodes will prevent the sudden change in current caused by charge storage - or at least it will confine it to the small loop consisting of the diode and cap. By reducing the area of this loop it will reduce induction and radiation. It will also reduce the effect in the wider secondary loop. Using a full snubber across a diode might reduce the effectiveness of the cap because in a resistive circuit the voltage and current can, in principle, have instantaneous changes (or nearly so in real life).

There may still be a slower change in current caused by normal diode switch-off. A snubber across the secondary may help reduce the ringing effect of this.
Click to expand...

Hi DF,
How do you calculate the RC value for the snubber applying across the secondary to prevent ringing?
 
audiostrat

audiostrat

Member
Joined 2012
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom