Alleged Y cap issue attributed to Bruno Putzeys

[RickS]

As promised from another thread, am starting this thread to fact check a contention that Y caps cause audio problems (notably as applied in AC inlet filters) and that noted designer Bruno Putzeys had stated as such. Have seen this contention perpetuated here and on other sites and so would like to see if ASR experts agree or not (about Y cap application).

Here is some background from other sites on this contention...

Here is one notable Bruno P quote from diyaudio.com:

“Y caps: as have been noted, they capacitively couple the mains to chassis. In AES48 compliant balanced systems this is no problem because they keep signal and ground well separated. In an unbalanced system the leakage current will also flow through the ground wire or shield of your $$$$$ RCA interlink and thus add noise to the audio.
As such, Y caps are difficult to avoid, they serve a purpose after all (to keep rubbish generated inside the box from getting out), but what it means is that it pays to try minimizing them. There are small Y caps in the SMPS600. Since they wholly suffice, there is no reason to compound them with external ones.”

If you look around, he talks about Y caps in many contexts. Notably both in a SMPS and in an AC inlet filter.

Here is another (even older) quote from Audioholics and appears to be the source of the confusion…

Bruno continues: "The difficulty lies in the fact that life isn't getting easier in the future. Switch-mode power supplies suffer an aggravated version of interwinding coupling across the transformer. The circulating current between the primary and secondary sides is quenched by placing a so-called "Y" capacitor between primary "ground" and secondary ground. This reduces the voltage imposed by the transformer between the two sides by a factor equal to the ratio between Y capacitance and parasitic capacitance between the windings. Increasing the Y cap reduces the problem, but never solves it. Worse still, the Y cap effectively connects the AC line to the audio ground! Similarly, other EMI problems are "fixed" by connecting caps to chassis everywhere. Y caps are also found sometimes in audio equipment that has a linear supply and a good reputation. They are also part of mains filters, devices often used as a selling argument in the audio world (cleaner mains! sure! but the dirt is now in your audio).

Such practices are of course of no consequence in computers and copiers, VCRs and TVs (when not connected to the audio system). Unfortunately the same people who have been smart enough to design these are now selling power supplies for use in DVD players etc.

Designing a Switch-mode power supply that does not create circulating currents and that does not need Y caps is certainly possible. I am afraid, however, that the number of designers worldwide capable of pulling this off can be counted on one hand, and they might not even actually be at it."

So, at the end of the first paragraph, Bruno is quoted as saying “They (Y caps) are also part of mains filters, devices often used as a selling argument in the audio world (cleaner mains! sure! but the dirt is now in your audio).” So this states exactly what has been circulating around regarding EMI inlet filters. Given the earlier quote, not sure how this “dirt” gets in the audio single or maybe he was misquoted? He is not hanging around diyaudio anymore and that, for unbalanced system, Bruno does mention leakage current as source of the noise.

Agree or disagree? Does seem clear that Bruno (on 2 different occasions) contends that Y caps can cause noise. I have measured a few Hypex that used inlet filter with Y caps and did not find any major issue but perhaps I missed something?

Please help!

 

[Speedskater]

That must be from a 2012 diyaudio forum 'Hypex Ncore' thread.

Hypex Ncore

here is the proper way: I see, thanks pos!

www.diyaudio.com

 

[DonH56]

Isolating the switching node and careful attention to the ground path is a critical part of any SMPS design. I fought and lost the Y-cap battle for one of our eval boards, and then testing showed how it coupled noise, so was de-populated for the final version. I would not use the word "transformer" in describing the switched inductor, though there are multi-phase versions I suppose could be considered transformers (not in the sense I normally think of them, but that's me), but mixing line and audio (in my case, inlet, digital, and analog) grounds was a Bad Thing and the y-caps coupled switching current spikes into the ground plane.

It likely to be highly dependent upon what the system ground is like (safety ground over here in the U.S.A.) and ground interaction with other components as to whether it is audible. I would expect it to be measurable, however, at least it was for me. It was a PITA to measure, however, since o'scope "ground" is rarely as good as the manufacturers would like you to think... I used differential probes for the 'scopes, and a differential scheme for our spectrum analyzer to look at the problem.

 

[RickS]

Thanks Don!

Is this a case of theory vs. reality? If so, gather it is some aspect of the capacitor properties? Is there some cost/benefit tradeoff (like can a lower leakage Y cap be had to avoid coupling the noise)?

 

[DonH56]

Thanks Don!

Is this a case of theory vs. reality? If so, gather it is some aspect of the capacitor properties? Is there some cost/benefit tradeoff (like can a lower leakage Y cap be had to avoid coupling the noise)?

Theory and reality will predict what happens pretty well. We (the company for which I worked) would usually have the SMPS manufacturer check the design and layout of our boards as we did not have any real SMPS experts on staff (myself included). I found it interesting that some SMPS manufacturers recommended using Y caps and others explicitly said to not use them. Power supplies are not really my thing, but I got sucked into helping design the decoupling networks and board layout for the analog sections, and of course that trickled into everything.

Nothing specific about the capacitors, and "lower leakage capacitor" is not the issue -- it is having capacitors there that causes the leakage. In the right place, Y-caps are great, as they provide a small (physically) wideband solution that works well when you have something like two signals that benefit from being tightly coupled through the ground path. That can cause offsetting currents that lower the overall noise. The problem with using Y capacitors "across" the SMPS terminals is that you couple input to output, a no-no when you are trying to isolate the two sides. It can make the input look better but couples more noise to the common ground and thus increases leakage into the signal path.

For me, it was another case of needing to watch very carefully the current paths, keep reminding everyone and myself that current is a loop and that the return path (ground) is part of that loop, and think (and analyze) carefully to ensure currents flow where desired without unintended consequences.

 

[Speedskater]

For other readers, Class 'Y' and 'X' capacitors are safety rated for use in AC power line circuits.

 

[RickS]

My primary edit as others have mentioned is the use of Y caps in EMI/RFI inlet filters. Notably one Class D amp supplier uses them, and Amir has tested those amps (EDIT) and found power supply noise. without finding any negative consequence. When I first found out about the noise assertion, I only found one limited quote from Bruno. Admittedly, this may be because the original discussion is quite old (the Audioholics quote in my OP is from 2006). The issue resurfaced recently in another amp supplier’s thread here so I decided to research again. Not finding a conclusive thread on ASR, I started this one.

Did not want to bog the OP with technical detail as it is simple to find info on Y caps online. For those who want to learn more, here is a link and here is a pic:

As for the specific use in an inlet filter, here is what I found in one vendor’s amp. It is a Delta and it uses Y caps as shown in this data sheet: https://filecenter.deltaww.com/Products/download/04/0406/iec_inlet_filter/BE_SWITCH.pdf

In this case, as designated by Cy the Y cap is coupling both the hot and neutral to earth ground.

So, why is this not an issue? If, as noted earlier in proper design application, the vendor keeps the signal grounds and earth grounds separate, Y caps do not present a noise issue. On the other hand, if there is no major benefit to the inlet filter, why go to the added expense?

 

[RickS]

Thanks again!

Theory and reality will predict what happens pretty well. We (the company for which I worked) would usually have the SMPS manufacturer check the design and layout of our boards as we did not have any real SMPS experts on staff (myself included). I found it interesting that some SMPS manufacturers recommended using Y caps and others explicitly said to not use them. Power supplies are not really my thing, but I got sucked into helping design the decoupling networks and board layout for the analog sections, and of course that trickled into everything.

Nothing specific about the capacitors, and "lower leakage capacitor" is not the issue -- it is having capacitors there that causes the leakage. In the right place, Y-caps are great, as they provide a small (physically) wideband solution that works well when you have something like two signals that benefit from being tightly coupled through the ground path. That can cause offsetting currents that lower the overall noise. The problem with using Y capacitors "across" the SMPS terminals is that you couple input to output, a no-no when you are trying to isolate the two sides. It can make the input look better but couples more noise to the common ground and thus increases leakage into the signal path.

Bruno mentions some of this and how Y caps are "unavoidable". At least that was his opinion at the time. 

For me, it was another case of needing to watch very carefully the current paths, keep reminding everyone and myself that current is a loop and that the return path (ground) is part of that loop, and think (and analyze) carefully to ensure currents flow where desired without unintended consequences.

This sounds rather comparable to the referenced amp situation and ensuring signal grounds and earth grounds are kept isolated from each other. Or are there broader considerations?

 

[AnalogSteph]

Re: purely AC inlet filters:

Many designs of inlet filter will have one (X) cap each from live and neutral to ground. Which is all fine and dandy as long as we're talking an IEC Class I (earthed) device and secondary ground is firmly bonded to PE.

However, some power supply designs will turn a construction like that into the floating (IEC Class II) variety merely by snipping off the PE and turning X into Y caps between L and N and secondary-side ground. Now as long as you're in Norway or on a ship or elsewhere with "technical" mains with V(L) + V(N) = V(PE) = 0 (i.e. symmetrical / opposite L and N voltages), that is not a problem, however in many places we have single-phase AC that is grounded on the neutral leg. Suddenly you have several nF worth of coupling between the AC live and your secondary-side ground, and once you introduce a second earth connection somewhere else, the resulting leakage currents will duly make their way there over whatever connections there are in between, unbalanced audio cable shields included. Dumb.
Moreover, if there is no second earth connection, your secondary-side ground will be pulled towards Vmains/2. If you are powering an amplifier with unbalanced inputs like that and it sits there on its own, any open input will hum due to parasitic coupling to the (earthed) environment. Not great.

This stuff does in fact happen - the Yamaha A-x60 series of ca. 1980 integrated amps (760, 960, 1060 and II) sports exactly this kind of arrangement because they had a funky transformer voltage regulation thingy that basically controlled a dimmer on the primary side and generated corresponding amounts of noise. I guess the leakage current was supposed to be dumped via the external antenna on the FM tuner that went with them - very much of the times.

This sort of issue can be mitigated quite effectively just as long as you have grounded outlets, which I know is not a luxury that all countries have.
If you do, however, bringing PE into the chassis via a new mains lead opens up some exciting possibilities. The Y cap midpoint can be redirected to PE, and the direct connection to secondary-side ground replaced by another ca. 4.7 nF Y cap. At this point the mains leakage gets dumped straight into PE, and you only have a few nF of coupling to PE where your system's ground potential should roughly be anyway, and that is not enough to create any major ground loop issues. (Plus, even a miswired outlet only makes the chassis a bit tingly at best.) The 220/240V versions of said Yamahas make this mod very easy thanks to a convenient wire bridge on the PCB, for others it gets a bit more complex.

A similar approach can also be used on power or integrated amplifiers with large transformers that have substantial primary-secondary parasitic coupling capacitance (think mid-upper hundreds of pF) but are supposed to be floating IEC Class II affairs. One 4.7 nF Y cap from ground to newly-introduced PE should be quite effective at reducing the hum on open inputs without introducing any major ground loop issues otherwise. Some manufacturers have been smart (and/or rich) enough to use transformers with shield windings to cut down on coupling capacitance, a lot have not. (A shield winding connected to secondary-side ground basically forms a Faraday cage that makes E fields from the primary winding more or less invisible.)

It goes without saying that PE conductors are not actually infinitely low resistance (which is why PCs tend to be so troublesome), but a direct PE connection will go a long way.

Y capacitors inside SMPS are a slightly different story.
The crux of the matter is that primary-side ground is just one end of the rectifier. It's not live or neutral potential, it's a buzzy distorted something. I can fully understand if you don't want to couple that mess into your nice floating secondary-side ground. I could see knocking it down via a C or R||C combo from secondary-side ground to PE to effectively form a voltage divider (e.g. 100 nF || 1k à la old Thinkpad power supplies or whatever), but that's not totally ideal. You really want to be making things easier first by tackling the root cause with a shield winding in the transformer, which obviously is a matter of complexity and cost.

 

[amirm]

Y caps are essentially a standard practice in off-the-shelf switchmode power supplies. It was said that "medical grade" ones don't do it but that turned out false as well. I was consulting with a company on this and the so called medical one also had Y caps and bled mains noise into the DC ground output of the power supply. As Don mentioned, a work-around is to remove the Y cap but then the power supply is no longer compliant with FCC conducted emissions. Best to use balanced interconnects to avoid this problem.

 

[DonH56]

Thanks again!

NP. As I said, this is not really my areas of expertise, take with block of salt.

Bruno mentions some of this and how Y caps are "unavoidable". At least that was his opinion at the time. 

Power factor correction and RFI/EMI suppression may make them necessary in some applications. See comment above. That is what makes them commonplace in COTS (commercial off-the-shelf) supplies, and makes designing them out difficult.

This sounds rather comparable to the referenced amp situation and ensuring signal grounds and earth grounds are kept isolated from each other. Or are there brooder considerations?

Way too broad a subject to go into here. Mixed-signal circuits with very wide bandwidth yet needing precision DC references; multi-GHz high-dynamic range, low-jitter signal paths; and gobs of digital processing on-chip can be a nightmare from the IC design through the system level implementation. Noise coupling paths are legion and difficult to isolate. We ran into places where a very short ground trace between pads was a huge problem, and another place where a similar trace coupled noise single-endedly into a differential pair and trashed the performance of one receiver (of >100).

For audio (and other) amps, leakage paths can induce unexpected hum and power line spurs, e.g. when you have a class D amp and SMPS that exhibits relatively high 50/60 Hz powerline spurs at the output, though there is nothing overtly power-line related anywhere in the circuit. That is one of the consequences of inappropriate or improperly placed Y capacitors (among other things). I first ran into this many years ago working with high-impedance sensors targeting medical applications; tiny signal buried by an unanticipated 60 Hz leakage spur. Another fine example is the ~60 kHz frequency of most modern fluorescent lights that gets into everything. Our screen rooms used DC lighting, and in the open lab sometimes we'd test with the lights off using flashlights to see.

FWIWFM - Don

 

[RickS]

After finding the Y caps used in the Delta inlet filter, made me wonder about common household surge protectors. Could not confirm but I did find one site claiming older Tripplite Isobar used them. Are Y caps common in surge protectors?

 

[JSmith]

Are Y caps common in surge protectors?

Yes, quite common in decent ones.


JSmith

 

[Sokel]

Had a look at a very popular IEC filter and it's four version are like this:

https://docs.rs-online.com/d9f1/0900766b8144830d.pdf

Medical versions don't appear to have any.

 

[BR52]

Had a look at a very popular IEC filter and it's four version are like this:

View attachment 393028

https://docs.rs-online.com/d9f1/0900766b8144830d.pdf

Medical versions don't appear to have any.

In most European countries, the mains consumer plug is symmetrical, so the L and the N are not defined depending on how you turn the plug. The leakage current to the protective earth is very limited, depending on the classification (portable or stationary). For this reason, designers try to avoid Y caps. I know it has nothing to do with audio; it's only for security reasons. The bleed resistor is for discharging if unplugging to avoid possible el. shock.

 

[Sokel]

In most European countries, the mains consumer plug is symmetrical, so the L and the N are not defined depending on how you turn the plug. The leakage current to the protective earth is very limited, depending on the classification (portable or stationary). For this reason, designers try to avoid Y caps. I know it has nothing to do with audio; it's only for security reasons. The bleed resistor is for discharging if unplugging to avoid possible el. shock.

Yep,I have measured stuff where at the right direction had 130dBr mains noise and at the wrong direction mains noise jumped at 90dBr.
There's also a difference in case potential,some can measure 70-120V (harmless though) at the wrong direction and zero at the right one.

Goes without question that everything is marked at my gear.

 

[BR52]

Yep,I have measured stuff where at the right direction had 130dBr mains noise and at the wrong direction mains noise jumped at 90dBr.
There's also a difference in case potential,some can measure 70-120V (harmless though) at the wrong direction and zero at the right one.

Goes without question that everything is marked at my gear.

If you measure the 80 to 120 volts, your protective earth is broken! I remember we taped up the PE on some oscilloscopes, and then you touched the scope and another class 1 appliance, had a wake-up call.

 

[Sokel]

If you measure the 80 to 120 volts, your protective earth is broken! I remember we taped up the PE on some oscilloscopes, and then you touched the scope and another class 1 appliance, had a wake-up call.

At the device or in general?
Cause I was there when then putted this rods deep tn the ground with that braided copper huge ribbons when I built my house and measured them (I *think* the regulation states that this must be a few Ohm,6?,8? I don't remember)

Also outlets are also checked and I have a dedicated phase for my gear,I have seen what led,motors,etc do when they share the same phase when I do my silly measurements.
The nicest gear about it had an indicator which showed if it was at the wrong direction.
I have also seen wrongly internal wiring.

 

[Speedskater]

The connection to Planet Earth has little to do with day-to-day AC power quality.
That 80 to 120 Volts may be what electricians refer to as 'phantom voltage'. Add the smallest load or resistance and the voltage disappears.

 

[RickS]

In most European countries, the mains consumer plug is symmetrical, so the L and the N are not defined depending on how you turn the plug. The leakage current to the protective earth is very limited, depending on the classification (portable or stationary). For this reason, designers try to avoid Y caps. I know it has nothing to do with audio; it's only for security reasons. The bleed resistor is for discharging if unplugging to avoid possible el. shock.

Older houses here only have hot and neutral, symmetrical outlets and no ground! Newer (circa 1940) ones have polarized outlets with grounds. My house is newer. My son’s has hybrid wiring. Kitchen and baths got newer wiring and GFI outlets before he bought.