MikeB Posted October 9, 2009 Posted October 9, 2009 Wondering if any INSTEON/X10 filter/noise/signal experts might be able to offer some advice. I have a workbench in my office that often has several PCs and/or servers on it. Occasionally I will have a PC hooked up that will start disrupting my INSTEON signals and dropping reliability througout my house. The most troublesome power supplies seem to be certain model server-class power supplies from Intel. I installed a FilterLinc on the outlet, but it does not seem to be helping. Any other way to isolate these outlets? Anything better than a FilterLinc?
Illusion Posted October 9, 2009 Posted October 9, 2009 I am certainly no expert Mike, but I have personally found the 6288 by Leviton to be superior to Filterlincs. I would think the ultimate solution may be to put an isolation transformer on that circuit. I have one of those on my entire entertainment system. Of course, IndyMike is probably the answer guy on this one.
IndyMike Posted October 12, 2009 Posted October 12, 2009 Hello Mike, I've managed to recover from Illusion's accolade (still blushing a bit) and would like to second his recommendation of the Leviton 6288. The Filterlinc is a notch filter that is tuned to the X10 120K carrier. Being a notch, it does a very good job of attenuation noise in a narrow band around it's resonant frequency. The Leviton 6288 is a low pass filter that kicks in around 62Khz. By the numbers, it's not as good as the Filterlinc at 120Khz. However, being a low-pass, it provides increasing attenuation from 62Khz on up in frequency. As a result, it is far more effective at reducing out of band noise that could affect your Insteon signals. In general, I've found that "high end" power supplies are "signal suckers" rather than noise generators. The 6288 should provide better isolation here are well since it uses rather huge, hulking inductors. I am a bit curious about your description - you mentioned that plugging in the server disrupted communication "throughout the house". Is your PLM in the vicinity (subject to signal absorption)? If not, it sounds like your server is actually generating noise. Not what I would expect from a high end power supply. IM
Algorithm Posted October 12, 2009 Posted October 12, 2009 Hello IndyMike, Thanks for all the above information. Do you know whether the ACT AF120 (15A plug-in) filter is a low-pass filer? I have an AF120 in my computer room (location of the PLM), feeding virtually all the equipment (not the PLM, of course), including two APC UPS (an 800VA and a 900VA). The filter appears to make no difference. I would like to try another type, but I'm sure the 6288's 5A rating is insufficient. I'm even doubtful whether a 10A FilterLinc would be sufficient, but am not sure about that.
IndyMike Posted October 13, 2009 Posted October 13, 2009 Hi Darrell, Had a discussion on this filter with Jeff Volp some time ago. The AF120 is apparently a L-C-L style bandstop filter. ACT claims a filtering range of 60KHz to 200KHz but does not include a graph of the filter effectiveness (they mention a peak of 30:1). The wire in XPF (20A) filter is about the only unit that I know of that can handle this type of load. This unit uses a L-C-L-C-L configuration with parallel caps on the inductors. Not a true low pass (like the 6288) since it's tough to get an large inductor that can handle 20A. Could you possibly separate your loads on different filters? This might allow you to determine which device is the real offender. IM
Algorithm Posted October 13, 2009 Posted October 13, 2009 Hi IndyMike, You are a wealth of information; thanks! For the most part, the devices are run off the two UPS units, other than obvious exceptions such as laser printer and so forth; but all pass through the AF120, which is rated 15A, the same as the breaker. Yes, I could separate the loads into three: UPS1, UPS2, and non-UPS devices; and use three filters. But that still means using a 10A FilterLinc (not a low-pass filter) for the loads which exceed the rating of the 5A 6288 low-pass filter. I suspect that 800VA, 900VA, and even the laser printer might all fall into this category, but hope that none would overload the 10A FilterLinc. Thoughts?
MikeB Posted October 13, 2009 Author Posted October 13, 2009 I am a bit curious about your description - you mentioned that plugging in the server disrupted communication "throughout the house". Is your PLM in the vicinity (subject to signal absorption)? If not, it sounds like your server is actually generating noise. Not what I would expect from a high end power supply. No, the PLM is on a dedicated circuit on my main breaker while the troublesome server power supplies are located on a circuit off the sub-panel. The circuits I notice the most issues on are bedroom circuits which again are on the main panel. I see a lot of PCs and servers come and go in my office, and I've had issues with 2 types of power supplies: - High-end PC Power & Cooling power supplies caused issues even when behind an AF120 filter. These power my in-house servers. For the heck of it I replaced it with a 10 amp FilterLinc and the issues I was seeing stopped, so I assumed the AF120 had failed. Now I'm not so sure. - Intel server power supplies. Not necessarily all, but some. Most recently the power supplies in 1U 34xx series barebones servers. The servers do not need to be turned on. When simply plugged in the server fans run in a very low RPM "sleep mode", and I see intermittent comm issues around my home. These servers are plugged into 10 amp FilterLincs, but are still causing grief. Thanks, I will add a couple of Leviton filters to my next Smarthome order!
IndyMike Posted October 13, 2009 Posted October 13, 2009 Hello Darrell, The 5A FilterLincs are similar to the AF120 in configuration. These are true L-C-L notch filters with a very deep resonant notch at the X10 carrier frequency (120KHz). I haven't opened a 10A FilterLinc, but believe they are similar. If you are dealing with noise at the Insteon carrier frequency, the 10A Filterlinc will be an improvement. If you are dealing with out of band noise, the AF120 is a better choice (with the 6288 being better still). I'm not sure how to guide you on this one. You could try a 10A Filterlinc on the UPS(s) to see if it can catch the noise. If it doesn't work, you'll likely need the 20A XPF wire in filter. The problem is, these babies are not cheap. There may be a more affordable isolation device out there. I haven't searched in some time. IM Hi IndyMike,For the most part, the devices are run off the two UPS units, other than obvious exceptions such as laser printer and so forth; but all pass through the AF120, which is rated 15A, the same as the breaker. Yes, I could separate the loads into three: UPS1, UPS2, and non-UPS devices; and use three filters. But that still means using a 10A FilterLinc (not a low-pass filter) for the loads which exceed the rating of the 5A 6288 low-pass filter. I suspect that 800VA, 900VA, and even the laser printer might all fall into this category, but hope that none would overload the 10A FilterLinc. Thoughts?
IndyMike Posted October 13, 2009 Posted October 13, 2009 No, the PLM is on a dedicated circuit on my main breaker while the troublesome server power supplies are located on a circuit off the sub-panel. The circuits I notice the most issues on are bedroom circuits which again are on the main panel. I see a lot of PCs and servers come and go in my office, and I've had issues with 2 types of power supplies: - High-end PC Power & Cooling power supplies caused issues even when behind an AF120 filter. These power my in-house servers. For the heck of it I replaced it with a 10 amp FilterLinc and the issues I was seeing stopped, so I assumed the AF120 had failed. Now I'm not so sure. - Intel server power supplies. Not necessarily all, but some. Most recently the power supplies in 1U 34xx series barebones servers. The servers do not need to be turned on. When simply plugged in the server fans run in a very low RPM "sleep mode", and I see intermittent comm issues around my home. These servers are plugged into 10 amp FilterLincs, but are still causing grief. Thanks, I will add a couple of Leviton filters to my next Smarthome order! Funny you should mention that you were suspicious of the AF120 having failed. As it turns out, the AF120 uses 2 fuses. The First is a user accessible overload fuse rated at 15A. The second is a 1A fuse between the filter cap and ground. If this were to blow, you would loose much of the filter effectiveness. http://idobartana.com/hakb/AF120_Schematic_vB.pdf I'm a bit surprised at your 1U servers. When in low power sleep mode, you may be getting impulse noise that is making it past the power supply filtering. The manufacturer may be rationalizing this by saying that the units are typically operated off a UPS. Even so, having a rack of 1U servers banging away at each other through the powerline isn't an ideal environment. You may want to contact the manufacturer to avoid possible interaction problems in the field. Let me know if the 6288 helps with this, IM
MikeB Posted October 13, 2009 Author Posted October 13, 2009 I'm not sure I will still have these 1U servers in my office when I get the Levitons in, but I'll be sure to post back once I get the Levitons installed and am able to test their effectiveness. Thanks so much for all the info!
Algorithm Posted October 13, 2009 Posted October 13, 2009 IndyMike, thanks for the good information. I'll include some 10A FilterLincs with my next order, and see if they help.
Recommended Posts
Archived
This topic is now archived and is closed to further replies.