• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Alleged Y cap issue attributed to Bruno Putzeys

DIN audio (5 pin) had voltage signals for playback (pin 3 and 5) and current for recording (1 and 4), 2 is common.

Biggest issue usually in audio is leakage currents trying to equalize between 2 devices via the shield (audio ground) adding a small (unwanted) voltage drop over the common wire which adds to the audio signal.
As the shield of the RCA is usually connected to common and mostly also to the chassis (which may or may not be connected to safety ground) it will be impossible to make that connection balanced (and thus have high CMRR). Cable then also would not have to be shielded but twisted or the 'common wire' should not be connected to the chassis and the only the chassis would be connected to shield (and thus have 2 signal wires, one to common).
Yes, that is the standard in audio gear. I was pointing out that it _need not_ be from an engineering standpoint.
In other applications there are two-conductor balanced/differential signals paths. 3 conductors are not inherently required.
That is what I was pointing out.
 
. X caps are designed to fail "open", with Y failing to a short.

It's actually the other way around. ;)
X caps have to short and are used between L and N and when they short must blow the obligatory fuse.
Y caps are used between input and output for instance and in case of failure must open to ensure mains is never connected to the isolated output voltage.
They are also used between safety ground and N in filters so in case of failure the chassis is never connected to mains (N or L).

Y caps between input and output of SMPS (for EMC emission compliance) are usually the reason for the leakage currents.
 
Last edited:
It's actually the other way around.
X caps have to short and are used between L and N and when they short must blow the obligatory fuse.
Y caps are used between input and output for instance and in case of failure must open to ensure mains is never connected to the isolated output voltage.
They are also used between safety ground and N in filters.
Doh!
Yes, I got them 180 degrees out of phase!
 
However, a capacitor across the transformer barrier sometimes is installed (primarily to provide a current path for the induced common mode leakage current across it) which may reduce downstream emissions. There, either a Y or X may be used. It comes down to a design decision on which failure mode is most appropriate for the application.
These are extremely common due to the reason you mention (passing conducted emissions). I remember trying to help a company find one without the Y cap and it was darn near impossible. Even the "medical grade" supplies had such even though specs implied lack thereof.
 
In the by @HaveMeterWillTravel mentioned application only Y caps may be used.

They are essential in switched mode power supplies to create a path for HF switching noise from the secondary to primary if they want to comply to certain standards for RF emission. This path could also be safety ground but for double isolated stuff it can't go through the safety ground and will flow though either N or L.
The higher the switching frequency the lower the capacitance can be and the lower the 50/60Hz leakage current will be (assuming there are no other paths like the transformer itself for unwanted HF power) which also should not be underestimated.
 
The higher the switching frequency the lower the capacitance can be and the lower the 50/60Hz leakage current will be (assuming there are no other paths like the transformer itself for unwanted HF power) which also should not be underestimated.
That is definitely another path for mains leakage.
 
A third path (for HF) is wire/PCB routing and capacitance of ground-planes for instance.
At least .... when done incorrectly at least.
There are designs where pre- and post filter mains wires run really close together (design mistake) which can be the difference between a pass and fail of EMC testing (emission mostly but sometimes even immunity or ESD)
 
In the by @HaveMeterWillTravel mentioned application only Y caps may be used.

They are essential in switched mode power supplies to create a path for HF switching noise from the secondary to primary if they want to comply to certain standards for RF emission. This path could also be safety ground but for double isolated stuff it can't go through the safety ground and will flow though either N or L.
The higher the switching frequency the lower the capacitance can be and the lower the 50/60Hz leakage current will be (assuming there are no other paths like the transformer itself for unwanted HF power) which also should not be underestimated.
One will be surprised by how many devices do not use the proper specified capacitors internally.

Having conducted various EMC testing for over a decade, I know how it often happens. For trouble shooting purposes you tend to have a small kit of mitigation components. If you are trying different value capacitors you're only worried about the value and voltage. I don't think I even had X and Y caps in the kit, let along X1,2,3 Y1,2,3.
Once it passes at that frequency range you continue until the system passes all tests. During this process you are reporting the status and results to the customer.
While it is generally their responsibility to properly evaluate and integrate the changes in the final design, many forget that safety caps are not solely limited to the EMI power filters and the power supplies. Often the project is behind schedule and over-budget and they just implement what ever random cap was used.

Or, mitigation components were necessary to pass EMC testing but are omitted in the shipping product. This is one of the main reasons why I do not blindly trust certifications.
 
In the by @HaveMeterWillTravel mentioned application only Y caps may be used.

They are essential in switched mode power supplies to create a path for HF switching noise from the secondary to primary if they want to comply to certain standards for RF emission. This path could also be safety ground but for double isolated stuff it can't go through the safety ground and will flow though either N or L.
The higher the switching frequency the lower the capacitance can be and the lower the 50/60Hz leakage current will be (assuming there are no other paths like the transformer itself for unwanted HF power) which also should not be underestimated.
For low noise applications one common technique is the use of a single AC/DC SMPS. at a common voltage, say 24 or 48 Vdc.

The DC return of the P.S. is then bonded to chassis and physical earth, and sufficient capacitance is incorporated. Essentially now you have a DC link bus with extremely low mains frequency common mode voltage.

Isolated DC:DC converters are hung off the bus to provide voltage/power for the various parts off the system where as you mentioned, the higher switching frequency also helps. The technology has really progressed over the last decade or so in this area.

Then the designer can choose how to reference that power plane to chassis so there is more design flexibility.

Not that this is necessary for common audio equipment! :)
 
A third path (for HF) is wire/PCB routing and capacitance of ground-planes for instance.
At least .... when done incorrectly at least.
There are designs where pre- and post filter mains wires run really close together (design mistake) which can be the difference between a pass and fail of EMC testing (emission mostly but sometimes even immunity or ESD)

Another is when A/C power is run up to the front for a mechanical power switch and then back to the power supplies.
Before the concept of 'standby power' many benchtop measurement devices (such as an HP meter) used a mechanical linkage from the front button to the back to avoid it.
 
One will be surprised by how many devices do not use the proper specified capacitors internally.

Having conducted various EMC testing for over a decade, I know how it often happens. For trouble shooting purposes you tend to have a small kit of mitigation components. If you are trying different value capacitors you're only worried about the value and voltage. I don't think I even had X and Y caps in the kit, let along X1,2,3 Y1,2,3.ds
Once it passes at that frequency range you continue until the system passes all tests. During this process you are reporting the status and results to the customer.
While it is generally their responsibility to properly evaluate and integrate the changes in the final design, many forget that safety caps are not solely limited to the EMI power filters and the power supplies. Often the project is behind schedule and over-budget and they just implement what ever random cap was used.

Or, mitigation components were necessary to pass EMC testing but are omitted in the shipping product. This is one of the main reasons why I do not blindly trust certifications.
Sounds like real world. Not for audio only.
 
Back
Top Bottom