D C Blocking on mains supply

[Guest Muon]

Hmm...I have a bad buzz going on.............I better get with this.

 

Edit: typo..typing in the dark.

Edited August 6, 2014 by ortofun

[Guest guru]

That's what they were constructed for,the sonic benefits were a bonus.

[Guest Muon]

Mmmm..sonic benefits

 

Valued thread, and timely for myself

[Guest guru]

glad to be of assistance, you will be quite delighted with the effort. I just bought another 160 caps from mouser to make up modules as the caps I use are being deleted from production.

[HdB]

Just noticed this thread - it's really good to see some of these simple projects coming into hifi use instead of only being seen in 'Pro-audio' gear

 

A cautionary note about using some of those smaller 2,200uF caps in parallel to replace single 4700uF cap - shared/parallel caps aren't at all guaranteed to always evenly share the load current so each 2,200uF must be rated well above the load current and also must be quite capable to withstand much larger amplifier startup currents, etc - I would suggest you use very generously rated electros here and not be wanting to go just for small size -

 

See the heavy duty caps that Per Anders (Sjostrom) uses in his dc trap modules (see link post #49)

 

Also, it's good to keep in mind that when you 'happen apon' something that works well, simple or not, there's a tendency to keep it in your system for years, and years so some power electros do have short lives and other ones have quite long lives,  so some otherwise good quality NOS caps may/maynot be well on their way towards the end of their life - just something to keep in mind ,,,  (I often use the  Siemens/Epcos long life thick film caps here (or the Rifas) and incorporate a slow start circuit for power amps, etc)

 

Isolation transformers are not guaranteed to get rid of all dc components on the incoming ac supply - all it takes is for the positive going half wave to be 'slightly' larger/smaller than the 'other half and you have a dc component that the isolating transformer is quite happy to pass on -  this can also occur even when using a centre tapped secondary isolation transformer for balanced power supply.

 

 Unfortunately, things are generally never quite as simple as they appear when talking about power distribution systems and the quality of our mains power supply isn't actually improving either despite the dramatic price increases.
 

[ThirdDrawerDown]

 

I recently made up several 'Isotran' units for myself, using a transformer customised for me by SES, here in Melbourne.

 

I have A/B tested these and the improvement on turntable and phono stage was considerable and worth chasing (one Isotran for each component).  A tweak I am happy to recommend.

[andyr]

Just noticed this thread - it's really good to see some of these simple projects coming into hifi use instead of only being seen in 'Pro-audio' gear

 

A cautionary note about using some of those smaller 2,200uF caps in parallel to replace single 4700uF cap - shared/parallel caps aren't at all guaranteed to always evenly share the load current so each 2,200uF must be rated well above the load current and also must be quite capable to withstand much larger amplifier startup currents, etc - I would suggest you use very generously rated electros here and not be wanting to go just for small size -

 

See the heavy duty caps that Per Anders (Sjostrom) uses in his dc trap modules (see link post #49)

 

Also, it's good to keep in mind that when you 'happen upon' something that works well, simple or not, there's a tendency to keep it in your system for years, and years so some power electros do have short lives and other ones have quite long lives,  so some otherwise good quality NOS caps may/may not be well on their way towards the end of their life - just something to keep in mind ,,,  (I often use the  Siemens/Epcos long life thick film caps here (or the Rifas) and incorporate a slow start circuit for power amps, etc)

 

Isolation transformers are not guaranteed to get rid of all dc components on the incoming ac supply - all it takes is for the positive going half wave to be 'slightly' larger/smaller than the 'other half and you have a dc component that the isolating transformer is quite happy to pass on -  this can also occur even when using a centre tapped secondary isolation transformer for balanced power supply.

 

Thank you for your input, @@jrhill

 

I was unaware of your last point.  Are you saying that if there is a DC component on the mains ... this will cause asymmetry in the 'positive-going half wave' vs. the 'negative-going half wave' ... which the isotran will be unable to correct?

 

The 2,200uF electros I was thinking of using are 100v Nichicon KG Gold Tune - it seems they have a ripple rating of 3.2a.  One of the posters here said that caps in parallel 'share' the actual current draw - so even if this is not done equally, I would've thought that these are sufficiently "generously rated"?

 

 

Regards,

 

Andy

[HdB]

Not exactly Andy,

 

The mains supply to your house may have a number of different non-ideal qualities on the power waveform that's known as a dc condition, or dc component -  they're already part of the supply and this can be asymmetry between the +ve and -ve parts of the waveform, can be a 'flat-top' on the waveform (voltage limiting, overloaded distribution system, etc) on one half, both halves, etc, - it can be a 'mis-shapen or twisted form of sine wave supply, and again, mayn't be the same for the +ve half and the -ve, and so on - not sure if they've continued to add some ac low freq signals for testing and communication purposes too.

 

I think Rod has covered a lot of this on his site - it's irritated the hell out of him for years so he's written a fair bit about it - it's interesting just how a simple pair of electros and a few other bits can make such a big difference and is so easy to build into all our gears, not just power amps, distribution boxes, etc

 

Those Gold Tunes of yours will actually take quite a heavy beating over long periods without much complaint, so shouldn't have any problems there - I think there's a big jump in the current rating about 50 volts, but not sure - something about the thickness of the film inside - my memory's not like it used to be - approaching it's 'used-by' date!

 

Some thing to keep in mind is some of the bigger amps that can draw a lot of power particularly in pro-audio use, or amps that're sometimes used as emergency backup and suddenly they're running 'flat stick' and everything starts to be approaching it's limits, including these dc trap capacitors.

 ... 

[Conster]

Hi,

 

I've been looking into DC blockers and spotted this active device which claims to monitor and correct the AC waveform for DC, supposedly without the limitations of blockers. I'm not electronically minded, so would be very interested in the thoughts any of you may have on this product.

 

http://www.hifi12a.de/HORCH-LINE-SILENCER/en

 

Thanks,

Con.

[Guest guru]

Funnily enough, I haven't heard any limitations of blockers in any systems I have used them in, only good things.

[Conster]

Guru, does that apply to external inline units as well. I'm wondering if they could be current limiting in any way? Thanks.

[andyr]

Not exactly Andy,

 

The mains supply to your house may have a number of different non-ideal qualities on the power waveform that's known as a dc condition, or dc component -  they're already part of the supply and this can be asymmetry between the +ve and -ve parts of the waveform, can be a 'flat-top' on the waveform (voltage limiting, overloaded distribution system, etc) on one half, both halves, etc, - it can be a 'mis-shapen or twisted form of sine wave supply, and again, mayn't be the same for the +ve half and the -ve, and so on - not sure if they've continued to add some ac low freq signals for testing and communication purposes too.

 

I think Rod has covered a lot of this on his site - it's irritated the hell out of him for years so he's written a fair bit about it - it's interesting just how a simple pair of electros and a few other bits can make such a big difference and is so easy to build into all our gears, not just power amps, distribution boxes, etc

 

Those Gold Tunes of yours will actually take quite a heavy beating over long periods without much complaint, so shouldn't have any problems there - I think there's a big jump in the current rating about 50 volts, but not sure - something about the thickness of the film inside - my memory's not like it used to be - approaching it's 'used-by' date!

 

Some thing to keep in mind is some of the bigger amps that can draw a lot of power particularly in pro-audio use, or amps that're sometimes used as emergency backup and suddenly they're running 'flat stick' and everything starts to be approaching it's limits, including these dc trap capacitors.

 ... 

 

 

Thanks again for your input.

 

Andy

[TMM]

Isolation transformers are not guaranteed to get rid of all dc components on the incoming ac supply - all it takes is for the positive going half wave to be 'slightly' larger/smaller than the 'other half and you have a dc component that the isolating transformer is quite happy to pass on -  this can also occur even when using a centre tapped secondary isolation transformer for balanced power supply.

That is not quite correct. The type of line distortion you describe is when the positive going part of the AC wave is loaded differently to the negative going part. It is the reason why we get DC offset in the first place. The resulting distortion can be described by a DC offset and strong even order harmonics.

When you pass this through a transformer or DC blocking capacitor the DC component will be removed but the harmonic distortion will remain. The waveform will look the same but it will have been 're-centered' so that the area under the waveform for both the positive going and negative going parts of the wave are equal.

This is an exaggerated version of the type of waveform you'd see on the output of your isolation transformer or DC blocking cap. Notice how although one half cycle reaches over 200V and the other only gets to ~175V, the shaded area under each half cycle is the same because the half cycle that peaks at a lower voltage is wider. Therefore, no DC component exists.

See the following simulation:

R1 is the impedance of the line. R2,R3,D1,D2 serve to load the line asymmetrically and create some distortion and DC offset on it (V2). C1 is the DC blocking capacitor and R4 is the load (device you don't want DC sent to).

V1 is the ideal supply voltage, V2 would be your distorted line voltage and V3 would be after your isolation transformer or DC blocking cap.

The tiny DC offsets on V1 and V3 are artifacts due to the resolution of the simulation (i didn't want it to take forever to simulate )

 

Hi,

I've been looking into DC blockers and spotted this active device which claims to monitor and correct the AC waveform for DC, supposedly without the limitations of blockers. I'm not electronically minded, so would be very interested in the thoughts any of you may have on this product.

http://www.hifi12a.de/HORCH-LINE-SILENCER/en

Thanks,

Con.

Inside (link) I spy a bunch of diodes and power resistors. I think the white 6-pin devices are solid state relays. The blue transformer just provides a low voltage output to power all the 'sense' circuitry i'd imagine.

If I had to guess how it works i'd say they are monitoring the level of DC on the line and actively switching in enough diodes so that the forward voltage drop of the diodes is greater than the DC component present on the line, so the DC cannot pass. With the number of diodes they have, it could probably only suppress a few volts of DC.

If there is no DC on the line, no diodes are switched in and the input goes directly to the output.

The problem I see with this method is that adding some diodes between the input and output there will be a 'flat spot' created on the output waveform - like crossover distortion in a Class-B amplifier - which will be described by high order harmonic distortion. This could be even more detrimental than just having the DC offset present... It's possible to filter out such a 'flat spot' slightly but I don't see anywhere near enough capacitance or inductance present for that to happen.

It is also very likely that it would produce a small 'pop' noise on the line as it transitions between states.

Of course, i'm just speculating how it works and I could be completely wrong. Either way it doesn't look like it would handle a lot of power, and the price is ridiculous Even at unit quantities I only see maybe $50 worth of stuff there and the circuit board looks like something i'd produce in my garage.

If you are going to spend that much on a device, get a Online (double conversion) Uninterruptable Power Supply with a True Sine output inverter. The only problem is that it will probably be noisy (fans etc), and you have to replace the batteries eventually. The best no-maintenance and completely silent solution is an isolation transformer and the second best is a capacitor blocker like in the OP.

Edited August 11, 2014 by TMM

[andyr]

That is not quite correct. The type of line distortion you describe is when the positive going part of the AC wave is loaded differently to the negative going part. It is the reason why we get DC offset in the first place. The resulting distortion can be described by a DC offset and strong even order harmonics.

When you pass this through a transformer or DC blocking capacitor the DC component will be removed but the harmonic distortion will remain. The waveform will look the same but it will have been 're-centered' so that the area under the waveform for both the positive going and negative going parts of the wave are equal.

 

This is an exaggerated version of the type of waveform you'd see on the output of your isolation transformer or DC blocking cap. Notice how although one half cycle reaches over 200V and the other only gets to ~175V, the shaded area under each half cycle is the same because the half cycle that peaks at a lower voltage is wider. Therefore, no DC component exists.

 

Thanks for your input, TMM.

 

What issues for the downstream component being powered are there, when the mains is not a "theoretical" sine wave but is "lop-sided" ... because:

• the +ve side goes higher than the -ve side goes lower, and
• the -ve half-wave has a longer duration than the +ve half-wave?

 

 

Regards,

 

Andy

Edited August 11, 2014 by andyr

[Conster]

Inside (link) I spy a bunch of diodes and power resistors. I think the white 6-pin devices are solid state relays. The blue transformer just provides a low voltage output to power all the 'sense' circuitry i'd imagine.

If I had to guess how it works i'd say they are monitoring the level of DC on the line and actively switching in enough diodes so that the forward voltage drop of the diodes is greater than the DC component present on the line, so the DC cannot pass. With the number of diodes they have, it could probably only suppress a few volts of DC.

If there is no DC on the line, no diodes are switched in and the input goes directly to the output.

The problem I see with this method is that adding some diodes between the input and output there will be a 'flat spot' created on the output waveform - like crossover distortion in a Class-B amplifier - which will be described by high order harmonic distortion. This could be even more detrimental than just having the DC offset present... It's possible to filter out such a 'flat spot' slightly but I don't see anywhere near enough capacitance or inductance present for that to happen.

It is also very likely that it would produce a small 'pop' noise on the line as it transitions between states.

Of course, i'm just speculating how it works and I could be completely wrong. Either way it doesn't look like it would handle a lot of power, and the price is ridiculous Even at unit quantities I only see maybe $50 worth of stuff there and the circuit board looks like something i'd produce in my garage.

If you are going to spend that much on a device, get a Online (double conversion) Uninterruptable Power Supply with a True Sine output inverter. The only problem is that it will probably be noisy (fans etc), and you have to replace the batteries eventually. The best no-maintenance and completely silent solution is an isolation transformer and the second best is a capacitor blocker like in the OP.

 

Thank you very much TMM. The thing that had me intrigued about the device was the piggy-back power plug. Got me thinking that there was some injecting instead of blocking going on.

 

I'm off to research UPS's now!

 

Cheers, Con.

Edited August 12, 2014 by Conster

[TMM]

Thanks for your input, TMM.

 

What issues for the downstream component being powered are there, when the mains is not a "theoretical" sine wave but is "lop-sided" ... because:

• the +ve side goes higher than the -ve side goes lower, and
• the -ve half-wave has a longer duration than the +ve half-wave?

 

 

Regards,

 

Andy

The only implication i can think of is that it could increase the ripple on a full wave rectified powersupply by causing it to operate as half wave as the lower voltage peak isn't enough to charge the smoothing capacitors. This would only happen at very low power draw where the ripple on the powersupply is almost non-existent anyway.

In reality that type of asymmetric distortion on the mains will be very small, 200V vs 175V is an extreme scenario. I'm not an expert on the topic but i'm lead to believe the bulk of distortion on mains is odd-order (flattening of both peaks) due to devices that load the +ve and -ve half cycles evenly but have poor power factor.

[andyr]

The only implication i can think of is that it could increase the ripple on a full wave rectified power supply by causing it to operate as half wave as the lower voltage peak isn't enough to charge the smoothing capacitors. This would only happen at very low power draw where the ripple on the power supply is almost non-existent anyway.

In reality that type of asymmetric distortion on the mains will be very small, 200V vs 175V is an extreme scenario. I'm not an expert on the topic but i'm lead to believe the bulk of distortion on mains is odd-order (flattening of both peaks) due to devices that load the +ve and -ve half cycles evenly but have poor power factor.

 

Aah, OK - thanks, TMM.

 

With your first point, I would think that this (the non-charging on the lower voltage peak) would only happen on an extreme case?  Which could very well occur in one's house if your neighbours' power activities are extreme, but would normally not ... as there would be only, say, 20v max difference between +ve voltage peak and -ve voltage peak?

 

Re. your second point ... it seems to me that 'flattening of the peaks' might be "fixed" by going through an isolation transformer?

 

 

Regards,

 

Andy

[Addicted to music]

If you are really fixated pursuing distorted non perfect sine wave on the supply and think that the  -/+ peaks are uneven, one way to check is to ensure the +/- DC rails on the supply is equal using a CRO or a MM.

 

Just remember that you will never get a perfect 230 sine wave out of the mains.  its impossible as all the issues mentioned in previous posts on this thread.  If its that stuffed it will never be corrected with an isolation transformer.  Your next best thing is to regenerate using UPS.   Its why so many industrial machinery like the manufacturer I support will never run motors directly of the AC mains.  Most have gone to DC conversion drive DC motors to protect and isolate  from AC supply quality and variation.  

[HdB]

Rod Elliot has a rather good intro article about all this in his "General Information"  articles  called "DC Offset and Transformers" (see www.sound.westhost.com) and there's a followup set of articles in the Power Supply section called "Transformers I, II & III"

 

His summary/design of a useful dc trap is quite effective, simple and cheap - just don't underrate the caps  - spend a few more $s - and make sure some adequate protection from accidentally touching the live mains, not just insulation tape.

[Addicted to music]

Rod Elliot has a rather good intro article about all this in his "General Information"  articles  called "DC Offset and Transformers" (see www.sound.westhost.com) and there's a followup set of articles in the Power Supply section called "Transformers I, II & III"

 

His summary/design of a useful dc trap is quite effective, simple and cheap - just don't underrate the caps  - spend a few more $s - and make sure some adequate protection from accidentally touching the live mains, not just insulation tape.

 

I have read this article and many like it and Im not disputing the claims.  However I like someone to shows me measured results on before and after implementation would be good to see.  I have worked in the reprographics area supporting industrial printing all my life, I have never seen an issue with mains subjected with DC offset.  The only DC offset I have ever encountered was self induced by the equipment I support, i.e. a faulty SSR injecting DC to a transformer it was driving, intern heating the transformer up to the point of it smoking the room out: thats it in 30yrs + in the electronics industry thats all  the DC injection I have encountered, and it wasn't from the mains supply.

[andyr]

If you are really fixated pursuing distorted non perfect sine wave on the supply and think that the  -/+ peaks are uneven, one way to check is to ensure the +/- DC rails on the supply is equal using a CRO or a MM.

 

Just remember that you will never get a perfect 230 sine wave out of the mains.  its impossible as all the issues mentioned in previous posts on this thread.  If its that stuffed it will never be corrected with an isolation transformer.  Your next best thing is to regenerate using UPS.   It's why so many industrial machinery like the manufacturer I support will never run motors directly of the AC mains.  Most have gone to DC conversion drive DC motors to protect and isolate from AC supply quality and variation.  

 

Aah, Peter ... I suggest you have mixed up 2 things re. the AC mains feed:

1. Driving AC TT motors, and
2. Supplying, say, a PS in an amplifier, which generates +/- DC rails for the amplifier PCB.

Yes, primitive AC motor drives (like the original Linn one) use the AC mains straight off the wall.  However, Linn's Lingo ... and my 'Number9' ... and, I'm sure, many other AC motor controllers generate a 50Hz signal from a 'crystal' (I have no idea what the device really is ... but I've heard it called a 'crystal' ).  So the question is ... what does a non-perfect sine wave (and, hence, potentially differential DC rails) do to the 50Hz signal produced by the crystal?  This is in fact the same issue as #2, above.

 

So ... what does having differential DC rails do to the performance of an amplifier?  I suggest nothing - until you get near clipping?  (The DC rail with the lower max voltage will clip first.)

 

But I'm very interested to hear from an expert who can tell me how my thinking is wrong.

 

 

Regards,

 

Andy

[Guest guru]

regarding the printer influence, why don't you build one and have a listen to it or run your plasma or lcd of it and see what it does to the picture.

Edited August 12, 2014 by guru

[Addicted to music]

@@andyr,

 

Im not sure of AC motors on TT, but most low voltage/wattage AC motors that I deal with normally comes with its driver pcb, these run of a DC supply that is converted to support the AC motor.  However even these are now rare,  as non contactless high torque DC stepper motors are now cheap and common and are so much more accurate  , las the life of the machine.

,