SUMMARY There are two relatively easy ways to improve RFX-250 performance. I recommend using method 1 on EVERY RFX-250 installation as it greatly improved perceived RF interference and costs less than $1. Method 2 is useful to reliably extend RFX-250 operation to over 500 feet (which I have tested) (and probably to about 1,000 feet though I have not tested that length) and costs about $20. Method 2 should always be used with method 1.
DETAILS OF WHY These methods are proven to greatly reduce perceived incidences of RF interference, which are really caused by noisy and/or insufficient power to the RFX-250. The RFX-250 requires a power supply of between 4.5 and 5.5 volts. On average it draws about 8ma when idle and up to 13ma when receiving data. The “on average” part is the kicker. Digital devices like the RFX-250 pull lots of current for very short bursts of time—typically in the microsecond or even nanosecond range. These draws can be many time the ‘average” current draw and, if the power supply cannot keep up, the device will malfunction.
The RFX-250 takes its power from whatever device it is attached to, such as the MRF-350 I used for my testing. Power is supplied over a long, thin pair of wires, which have resistance and capacitance attributes that limit how well changes in power can be accommodated: the longer the wire, the worse the problem. That’s why UR recommends doubling or quadrupling wire pairs in CAT-5 cables, so as to reduce wire resistance.
I have great respect for UR and their products. However, they are simply off the mark on this advice. Good digital electronic design requires the use of one, or several, bulk decoupling capacitors at the power supply input. This is especially critical if even several inches of wire are used. The UR design supplies a 10 foot cable for attaching the RFX-250—that’s 120 inches of wire. And there is ZERO decoupling capacitance at the input.
The problem is this: when the 250 is just sitting idle, it pulls a fairly constant current. But when it receives an RF input, it goes into action and drives a signal to the MRF-350 and also turns on an LED indicator. These actions nearly double the power consumption of the 250 and also cause short term current consumption of even greater amounts.
Without a decoupling capacitor, the 250 will often malfunction and simply light up the RF indicator even when RF isn’t present. You think there’s radio interference when there is not—or not nearly as much as the 250 seems to indicate.
The ideal fix is for UR to build bulk decoupling capacitance into every RFX-250 they sell. I have not done a full analysis, but were I the UR engineer, I’d put in a 100uF, a 10uF, a 1uF and a .1uF capacitor right at the power entry point. Different size capacitors respond to different duration transients. The biggest capacitor (100uf) provides a long term reserve storage of power that could last for milliseconds. The smaller ones respond to shorter requirements and are needed because larger capacitors take a bit of time to start supplying power. These changes will greatly enhance the performance of the RFX-250 system and increase production costs by < $1. I vastly improved my own system’s performance by installing a 47uf tantalum capacitor (Digikey part# 478-4180-ND, cost = $0.91).across the +5 and GND inputs to the 250. I started by simply inserting the capacitor into the screw terminals outside the box. I ultimately opened the 25 box and soldered the capacitor in place on its PC board for neater and more secure mounting. You can mount the capacitor using screw terminals even if your connection to the MRF-350 (or other device) uses the plug. Both connections are in parallel and both can be used simultaneously.
I used 47uF simply because I happened to have that value of capacitor lying around last night (yes, that’s what I was doing on Christmas night) when I was experimenting. I tried using two of them in parallel (that’s about like using a 100uF cap) but found no real improvement. Power supply issues are very difficult to measure though. What works in my installation might not completely resolve the problem in yours. That’s why I’d recommend using the 100, 10, 1 and .1 combo described earlier. That range of values provides the broadest protection. In production quantities, these parts are only a few cents each and UR should install them on EVERY product it makes in this line!
MY INSTALLATION My house consists of two separate buildings. I recently built a home theater and consolidated my two DirecTV HR10 HDTV satellite receivers into a single location. One satellite is for my wife, the other is mine. We can watch either satellite from any television (7 televisions total). One of the televisions is in the second building, away from where the satellite receivers are. I had two problems: getting hi-def audio/video to that place and conveying the remote control signals. I already had several CAT-5 cables running between the buildings and two were available. To run more would have been very difficult and I’m lazy.
I chose an Audio Authority 9870 product to transfer all signals. This system includes a model 9871 driver that converts component video and analog audio into CAT-5 compatible signals. At the other end, a 9878 wallplate converts CAT-5 back to component video and analog audio. Each signal is carried over a separate wire pair in the CAT-5 cables. That’s 3 for component video, two for audio, and 1 for power. Two pairs remain unused and Audio Authority attaches them to a jack on each end for infrared remote control: Two wires are used for ground, 1 for +V and 1 for data.
If I could get the RFX-250 to work over the Audio Authority infrared circuit, I’d be in great shape. I was discouraged in reading this forum a few weeks ago when the idea first occurred. I read about the unreliable RF operation and widely varying experiences of people. I also read of people trying the jack connection and the wired connection, thinking one was better than the other. Other people talked about the need to use multiple wires for power and ground.
In a prior life I was an electronic design engineer in the casino business. I learned a long time ago that when a device suffered mysterious problems, you should first check the power supply. Because of the long wires involved, and because the RFX-250 got its power from long wires, I was doubly suspicious. I ended up speaking with Eric Johnson. He told me they had systems in their lab working over long wires but “out in the field, things are different”. I remembered how many times I thought that as a young engineer and how many times I heard young engineers tell me that when I became a manager. This reinforced my power supply suspicion.
In the end though, it was Eric’s enthusiasm, commitment of support and commitment to this board, that sold me to buy some UR equipment and give it a try. I was already a happy MX-950 user from my smaller system that I am now upgrading. I logged onto surfremotecontrol.com and ordered several RFX-250s, 3 additional MX-950s a MSC-400 and an MRF-350 to supplement the 2 MX-950s I already had.
I was struck by the frustration that several users described in working with this system and decided to take the engineering approach: proceed in small steps and measure everything. First I connected the MRF-350 and single RFX-250 together in the same room and made sure everything worked. Next I added in the Audio Authority system. Once everything worked in the same room, it was time to expand.
Before that though, I wanted to get some idea of what each device needed to be happy. I used an ammeter to measure current to the RFX-250. At idle, it was about 8ma. Next I measured the load at the MRF-350 by connecting the ammeter between the GND and DATA lines, which showed how much current is needed to reliably drive the signal line (1.8ma). I knew that since multiple RFX-250 devices could be connected, the system had to be “open collector”. In other words, the RFX-250 signals only to ground, not to +V. A pullup device inside the MRF-350 does that chore.
Next I opened the RFX-250 to see a single integrated circuit, a number of supporting surface mount capacitors and resistors, a couple of LEDs, a crystal and another device I did not recognize. I also saw a nicely laid-out PC board with a great ground plane. What I didn’t see though confirmed my power supply suspicions: no bulk decoupling capacitors of any kind! I knew my chances of getting a system to work were nearly 100% now.
The chip is a maxim MAX1473. A quick look at their website got me a PDF of the data sheet. It is an 315Mhz-433Mhz ASK Super heterodyne receiver chip. The part I did not recognize was an IF transformer. Essentially the MAX1473 is a glorified and miniaturized FM radio receiver adapted for narrowband data reception.
It is easy to be misled by the RFX-250 nomenclature. While the connector says RF out, it really isn’t Radio Frequency that is output. It is a digital signal. That’s important because the digital signal, at 1.8ma or thereabouts, is far less susceptible to noise than the RF antenna input. I was no longer worried about shielding on the wire.
I also measured the voltage output of the MRF-350 at 4.99 volts and knew that was a problem for long wire-runs. The Maxim data sheet shows the MAX1473 needs a 4.5 to 5.5 volt power supply. Since the MRF-350 starts with only 5 volts, that leaves on a 0.5 volt (500 millivolt) margin for voltage loss along the entire cable length. That’s not a lot of margin. Moreover, there could be variations from one MRF-350 to another. I wouldn’t be surprised to find that some put out 5.25 volts (giving more room for loss) or even 4.75 volts, giving far less margin for loss. Such tolerances are necessary for economical manufacturing but can partly account for why different users experience different results.
I set up a small spreadsheet to figure out my projected voltage loss across 600 feet of cable. (I know that my run is >500 feet and <600feet, and it is best to use the worst case scenario). CAT-5 cable uses 24 AWG wire, which has a resistance of .03 ohms/foot. At 600 feet x 2 (two wires, +5 and GND), total resistance is 36.24 ohms. I rounded up total current requirements to 15ma, to provide additional room for error. The voltage drop is R x I or 36.24 x .015 = 0.54 volts. That means I could expect to lose about 0.54 volts over the wire run. My MRF-350 supplies 4.99V – 0.54 = 4.45 volts—which is already below the lower limit of 4.5V. I decide to try it anyway because the number is close, I’ve overestimated the current (maybe), my run is < 600 feet and there are two wires in the Audio Authority ground link. I also know I am right on the edge though.
My CAT-5 run is in three stages. The first is about 100 feet from the satellite receivers. The second is 250 feet past there and the last stop is at the full 500+ feet. At each point there are female cat-5 connectors which are jumpered to the next segment. I decide to test at each stage and using a stock, unmodified MFX-250 to see what happens.
When everything was in the same room, I almost never saw a flicker of the RF light except when I pressed the remote. About 95 of 100 remote presses were acted upon properly. The remainder were simply lost. Even with line of sight, I occasionally experience a slow or lost command on the DirecTV units and didn’t think much about it.
At the 100 foot mark, everything still worked but only about 90 of every 100 commands get through. At the 350 foot point, the RF indicator was lighting and only 2 of 3 commands came through.
At the end point, no remote control commands would work at all. The RF light would flicker regularly. When I pressed a remote control button the RF light would illuminate and stay on for several seconds. Was there really that much interference in the room?
I installed my 47uF capacitor by simply sliding in the screw connector. The RF light activity dropped considerably but it still required careful placement in the room. If I was within five feet of the television (a brand-new Toshiba 56MX196), the RF light flickered continuously. Only about 1 in 2 remote commands were executed.
I removed the whip antenna from the MFX-250. The RF light turned off and the remote worked well over 90% of the time. Perhaps this proved that there was RF interference in the area but I still had doubts. I found a wall-wart 5V 200ma power adapter from an old piece of equipment and cut the connector off. I removed the capacitor from the screw plug and attached the +5V and GND from the adapter to the screw connector.
It is important to not connect two power supplies together, so I disconnected the +5V wire feeding from the MRF-350. Now I measured 5.05 volts into the MFX-250. I reattached the antenna and, WOW, everything seemed to work well. After that initial success I started moving the 250 closer to the TV. Surprisingly, the RF light came on again—not because I got closer to the TV but because the AC line from the power pack got tangled with the +5V power pack output. If such a coupling caused malfunction, I certainly didn’t have a reliable system. Then I remembered the decoupling capacitor I removed.
The power pack had six feet of wire on it—it still needed a capacitor. So I opened the MFX-250 and soldered the capacitor across the power inputs. (I could not get the capacitor leads and the power pack leads to reliably co-exist in the screw terminals and so I soldered the capacitor in place.
When the 250 was reconnected, everything worked beautifully. I got 100 of 100 remote presses detected. The only time the RF light flickered even a bit was when it was placed within 6 inches of the TV on the lower left side. I’ve only used this system for a few hours but it seems rock solid.
I’ve attached a second MFX-250 to the system for use in the other building but I installed a capacitor in it too. I also disconnected the +5V wire going to the 2nd building through that long run from the MRF-350 so it doesn’t pick up noise and cause unexpected problems.
I’ll let you know after another week of operation but right now things are great. The bottom line is: the UR system is a fine one that suffers from a lack of attention to the power supplies for the remote RF antennas. That can be readily fixed by UR and I certainly hope they do so. I’d also like to see UR offer a power pack for remote installs of MFX-250s across hundreds of feet of wire. UR could probably retail these devices for < $20.
In the meantime, you can greatly improve your MFX-250 performance by adding a capacitor to your power inputs on EACH MFX-250 regardless of wire length. You can also create your own long-range systems by adding an external 5v supply and decoupling capacitor. Just don’t forget to disconnect the +5V line at both the source (MRF-350, etc.) and the MFX-250.
Finally, I know this forum is for professional installers. And I know that UR is working hard to insure professional installers get paid a margin for their expertise. I ask both the members of this forum and UR to remember us end-users too though. While I am not a professional, I do try to be diligent in how I use the equipment. Please don’t cut guys like me out of the support loop. We need information and sometimes we can help you too.
