 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
206.228.220.37
I decided to build a parallel AC line filter along the lines of the ones outlined here, with a couple twists. First step was to determine the values of the components needed. Since I wanted to terminate the RF transmission line formed by my dedicated AC line, I wanted to match the impedance of the filter to the characteristic impedance (Zo) of 12/2 Romex at RF frequencies, which I managed to find in a reference as 143ohms. Since I had some 1 watt 68ohm PRP resistors on hand, I decided to use two of those (68*2=136, close enough to 143) in an R-C-R configuration. I decided against using multiple R-C-R in parallel, opting for an R-(multiple C)-R configuration to keep the impedance close to Zo for as much of the frequency range as possible.
I opted for Wima caps for the filter, namely 0.47uF MP3X2, .047uF MP3X2 and .0047uF Y2, sourced from TAW Electronics. I used Mortite to damp between and around the caps.
I found a nice empty Wall Wart shell from Polycase.com which is perfect for this job, complete with Hot, Neutral and ground prongs installed. I made an inner shell of TI-Shield for the inside of the wall wart, with fully soldered joints to minimize RF leakage from the back of the unit.
Finally, in a bit of a departure from the norm, I decided to pot the guts of the unit inside the TI-Shield inner box using potting epoxy filled with carbon black. Carbon black is a strong RF absorber, so it may actually help in preventing shunted RF escaping from the unit, and may increase the RF-absorbing ability of the filter. The Mortite constrained in the rock-hard epoxy should provide good damping for the cap vibrations - the caps are already encased in epoxy for fire-proofing, so I don't think it can hurt. Even though the epoxy contains carbon black, it is not conductive (8x10^14 ohm-cm resistivity), so no shorting worries. It is certainly fireproof now, as well! The only concern might be dielectric absorption effects from the carbon-filled epoxy (dielectric constant = 3.5), but time will tell on that one.
Here are some pictures at various stages of completion:
So how does it sound, you ask? Well, I plugged it into one side of the duplex outlet into which my PS Audio PPP is plugged (everything audio runs off the PP) and took a listen. First impressions were not very good; everything sounded a bit flat, and lacking in dynamics - a bit worrying. I decided to let it break in for a few days to see if anything changed - day two was about the same. Day three was this morning, and I have to say it is starting to come around; the system has much improved dynamics and sounds very clear and clean now. I'd say it is slightly better than what it was before the filter was installed. I'm hoping the improvements will continue over the next few days - I'll report back in a couple.
Follow Ups:
For whatever reason, it seems like there is a bit of breakin to these things. Day six and the system sounds excellent. Well worth the effort.
My DIY AC LIne filter, per a magazine (Popular Electronics?) article I once read. It's just a capacitor in a plug. Was it a .01uf and 600 volts?
out into the room, probably better off keeping it in the wall! I have found that caps subjected to AC are huge noise polluters. In another project I built a shield for the inside of the EQ unit that came with my Infinity RSIIb speakers, around the rectifier section of the board, using TI-Shield (picture #1). Two 1000uF caps in that circuit were spewing noise out into the rest of the circuit, and this did a great job of reducing that. The reduction in background noise from the unit was substantial, as shown in the pictures #2 (before) and #3 (after) below.
There is always some parasitic inductance associated with capacitor leads and internal wiring. This forms a tank circuit with the capacitance and very large currents can flow if the circuit environment permits the cap to ring.
Your example shows how easily paid circuit designers can cause trouble by overlooking this behavior. I'm sure in their design model, the filter caps were represented as very low impedances at all elevated frequencies.
An alternative approach would have been to provide snubbers (R-C series networks) around each of the rectifier diodes, or even to replace the rectifiers with HexFRED or similar low storage devices. It is charge storage in the rectifiers that stimulates the ringing.
You are correct that noise suppression caps connected across the AC line also result in resonant behavior, unless appropriate series resistors are provided. Such caps only suppress noise in the frequency range between their effective corner frequency and their primary resonance.
are Cree SiC Schottky's, which have zero reverse recovery charge. I did a lot of hunting around the circuit looking for noise using a RTA spectrum analyzer and a grounded piece of TI-Shield, and by far the biggest source of noise was the two 1000uF caps associated with the rectifier diodes. You can see in the picture that there are small REL RTE PS caps right next to the shield (also right next to the offending 1000uF caps), and putting a piece of TI-Shield between the 1000uF electrolytics and the RELs made a big difference in noise. Also, orienting the RELs with the outer foil toward the lower impedance side of the circuit made a noticeable difference as well.
with similar complaints. In theory, there should be no charge stored, and no reverse pulses to stimulate resonance. However, something else seems to be going on here.
The capacitors will still resonate with stray inductance in the rectifier circuit if given suitable stimulation. The diodes may not be the source of noise in the conventional way, but I wonder if they have other noise issues.
TI-Shield is interesting stuff. It provides magnetic as well as electrostatic shielding, and its structure will allow magnetic fields to generate eddy currents in the copper layers that will dissipate some energy. It supports slow waves and can resonate itself electrically.
Being forced to admit that my DIY AC filter just wasn't good enough, I decided to bring it up to "Tweakers" standards and performed some major modifications to it.
Boy were you guys ever so right. The reds are redder, the greens are greener and the blues are bluer than ever before and I don't even like the Blues!
Nice try, though.
That's weird, because they use caps to suppress NOISE in all sorts of places. They use caps across rectifiers, they call them snubbers. They use caps across switching contacts, they call them sparking capacitors. They use caps across the contacts in car distributors, they call them "condensers." In fact if you DON'T use a capacitor you can hear the alternator/distributor whine through your car radio! Yes, I said hear. They use caps across the starter contacts in simple fluorescent light fixtures. They install caps across the brushes in many motors. My scale electric helicopter uses a simple DC motor to propel it and it has caps across the brushes/commutator to suppress radiated noise from interfering with the operation of the receiver, were talking about Radio Control and Radio Frequency Interference.
So, if what you are saying is true, then we should avoid caps in their entirety. But obviously we can't because caps serve a useful purpose and suppress all kinds of radiated noise.I have an entire book on Radio Frequency Interference both receivers and transmitters and it recommends various bypass capacitors depending on the particular problem. It also recommends the use of chokes too.
Edits: 02/09/11
Thanks for trolling my thread.Of course I understand that caps are used as snubbers, I use them all the time, and that's what the filter I discussed is based on. I was just pointing out that they can also radiate the noise back out into the environment, and that shielding to capture and drain that radiation might be a good idea.
Have a nice day.
Edits: 02/09/11
Oh no, it's a great day. The sky is blue and the birds are chirping here in sunny California.
You have a nice day too.
Glad to see this elabortately improved job! Looks like a lot of nice work and thinking.
What surprises me is that you didn't get good results right out of the box. Mine always do improve it immediately and then work in a bit better, so I never have had regrets on this.
I've never potted mine, though I did experiment in earlier iterations of R-Cs with sand fill. With the R-C-Rs, as long as the values of R are right and each C is surrounded by it's own Rs of matched values, it works best for me.
I do surround with TI and ground that as well. But I do not do total surround as you do, just more of a clamshell design (spiral?) padded by cotton, as Al Sekela showed me to do. With grounding TI, the AWG matters, and I found 12 AWG to be good for that purpose.
For the R-C-R wires, however, the smallest guage solid core that is safety rated for the task at hand is the best.
Keep listening, keep trying. These things work well and they work cumulatively and they're synergistic with other AC tweaks.
Yeah, I certainly wasn't expecting it to sound worse, either, but it did. Maybe it was the fact the I switched the socket for the PPP to the never-used top one of the duplex pair - I needed the bottom one for the filter to fit. That also means I had everything unplugged for a short while to make the switch. Maybe the carbon black/wire system needs time to set up; each particle of carbon insulated with epoxy can be thought of as a tiny capacitor that might have a breakin period too, perhaps.
Could also be that the PPP and the internal filters in my system do a pretty good job on RF as is, so additional filtering may not make as much of an impact. It is sounding very good today, though.
As I said in the OP, time will tell...
Post a Followup:
| FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
||||||||||||||
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2024, Audio Asylum, all rights reserved ]
