|Oct 24, 2012, 09:49 PM|
Build a cheap 5 minute Notch Filter
Ever try flying with a friend and have their video system mess up yours when flying close? Ever wish you could use the 1.2GHz band with your 2.4GHz radio? Many times in FPV, the small things make more of a difference than the big ones. As the World gets cluttered with more RF traffic, filtration becomes increasingly beneficial. One way to knock out interference is with a simple notch filter. And luckily, you can make one in under 5 minutes! Typically, you will be using a band reject filter.
Band reject filter:
The band reject filter is used to suppress a specific band of interference from entering your video system or control system. This is for times when the RF environment is known (which most of the time it will be).
A few places to use a band reject filter:
1. Flying 1.2GHz video with 2.4GHz control - Place a 2.4GHz filter on the 1.2GHz video
2. Flying with multiple FPV pilots at once - Put a filter on your VRX tuned to their frequency
3. Eliminate interference from your UHF radio on your video - Place a filter tuned to your video frequency on your UHF radio
What we are doing is using a small section of coaxial cable to short out a particular band. I will go over the physics behind this in a later post. First the details on the build.
I give you this tutorial in the name of my Lord and Savior, Jesus Christ. I hope that you will find this both educational helpful in your endeavors.
|Oct 24, 2012, 10:07 PM|
How it is made
The notch filter is really nothing more than using a section of coaxial cable as an impedance transformer. This might seem complicated, but it is really quite easy. There are two ways to make these: cable splice, and SMA Tee. All you are doing is connecting a coaxial cable stub to your antenna line. In this tutorial, I am using RG316 cable.
For RG316 or RG142:
Stub total length (in inches) = (2N+1) * 2053/f in MHz
Where N is either 0, 1, or 2.
In most cases, the formula is simply:
2053/f in MHz = Length in inches
Or in metric:
5214/f in MHz = Length in Centimeters
SMA Tee method
This is the easiest method, but is a bit more expensive and bulkier than a splice. You also might need to go to 3/4 wavelength (N= 1). You need the following:
- SMA female 3 sided Tee with one SMA male end
- A short SMA male to SMA male jumper
First, disconnect the antenna on the transmitter or receiver and hook up the jumper and then connect the tee. Connect your antenna to the other side. Screw your second SMA jumper into the center pin. Measuring from the center of the SMA Tee measure off ¼ wavelength (adjusting for velocity factor) and mark your cable. Cut it off right there. If ¼ wavelength is too short (if you are trying to filter out 5.8GHz for example) ¾ or 5/4 wavelength may be used.
Length in inches = 2053.1/f
Either formula works. That’s it! You are done!
So for some common frequencies:
910 MHz = 2.256"/5.73 cm
1280MHz = 1.604"/4.07cm
2.4GHz = 2.514"/6.39cm (this is 3/4 wavelength)
Cable splice method
For this method, you only need a short section of coaxial cable and a soldering iron. It is a bit harder to use this method, but it works just the same and is lighter. Splicing a cable isn’t easy by any means. You must connect both the center conductor and the shield. You want to have at least 60% of the shield intact and wrapped around your splice. This makes certain the no signal is emitted from the splice, but sill maintained inside the coaxial cable.
There are many ways to do this. The easiest way I found is to first strip back the outer jacket about ½” or so. I then take a knife and press it down so it cuts through the shield and the inner insulator. From here I peel back the shielding and then cut a V-notch in the coaxial cable on either side of my previous cut and dig out the Teflon. I then tin only the center conductor and prepare my feeder cable.
The easiest way I have found to make the feeder is to expose a very short section of the center element and then tin the shield right behind it to avoid it fraying while stripping the cable. From here I solder my center conductors together so they just barely mate. I use a lot of heat here when I have a Teflon insulted cable (such as RG316) to be sure I have a good bond because Teflon can take a lot of heat. If your antenna has a black jacketed cable (such as RG58), then it is a low heat insulation and you cannot super heat it and instead must be conservative with heat. From there, I join the shield as close to the center joint as possible without shorting it out. You can place a small section of metal overtop of the exposed joint if it makes you feel better.
Length in inches = 2053/f
910 MHz = 2.256"/5.73 cm
1208MHz = 1.604"/4.07cm
2.4GHz = .838"/2.12cm
Band Pas filter
This is the same as the band reject filter except you make it so only the frequency of choice you are using passes through the filter. Inevitably, the 3rd, 5th, and 7th harmonics will also pass through this filter as well.
How is it made? Simple, build a band reject and short it. Yep, that’s it. Sound strange? It kind of is. See the next post for an explanation of how it works.
|Oct 24, 2012, 10:07 PM|
Retuning your system from the band reject filter
Special thank you to HugeOne for pointing this out to me! - Thanks, buddy!
Once you have a tuning stub in place, it can throw off your SWR by fair margin and reducing your desired signal strength to as little as 50%. This results in a 3db reduction in performance. However, if your stub is an EVEN multiple of 1/2 wavelength of your desired frequency, the loss is neglegible and therefore not necessary. Thus if you are trying to filter the 2.4GHz harmonic from 1.2GHz video, there is no need for retuning.
Alex's fix - Method 1: Inverse reactive stub at the tuner
The simplest fix brings us right back to the physics of a coaxial cable. Luckily the computation is very simple. I will get into the physics of it later but here's the quick and dirty. For this example, I will use a 1280MHz Band reject filter on my 433 control:
First - Calculate 1 wavelength of cable for the frequency of your system (the frequency you want to keep)
1 wavelength in inches = 8212.4/f in MHZ
For my 1280 reject filter on 433 control the computation would be:
8218.4/433 = 18.966"
Second - Subtract your filter length from this. Then cut that length of cable and add it to your splice.
So for a stub of 1.603" (1280 Band reject)
18.966-1.603 = 17.36"
Simply add that to the opposite side of your splice (or coupler). Thus if you measured your filter and your correction tip to tip, you would have 1 wavelength of cable.
Alex's Method #2: Reactive stub 1/4 wavelength away
This might seem really strange, but it works the same way as the above stub, but you simply make a second splice in another location 1/4 wavelength away.
First, measure down the cable from your splice 1/4 wavelength in the frequency you intend to keep (again, going with 1280 band reject on 433 control).
Distance in inches = 2053.1/f in MHz
Thus we get 2053.1/433 = 4.742"
Put another filter of the same type in that spot. That's all there is to it!
Below, you will see my updated 1280 band reject filter for use on 2305MHz. The stubs are 1/4 wave on 1280, the spacing is 1/4 wave on 2305.
|Oct 24, 2012, 10:08 PM|
How it works
This works using a simple coaxial cable transformation HAMs have used for years. Every 1/2 wavelength, the impedance of a system goes through one complete cycle. For odd multiples of 1/4 Wavelength, the impedance is transformed by the following formula:
Zin = Z0^2/ZL
Zin = input impedance
Z0 = Nominal impedance of the cable
ZL = Load impedance of the system at the end of the cable
So for a band reject filter, you get:
Zin = 50^2/Infinity = 0
That's a dead short!
For a band pass filter you get:
Zin = 50^2/0 = Infinite
That's an open circuit!
Incredible isn't it? A short looks like an open and an open appears to be a short for certain frequencies.
|Oct 24, 2012, 10:18 PM|
Just a little question and answer for you guys so you don't need to read the whole thread past here:
Q1: Can I put a 433MHz filter on my 1280MHz video?
A1: No! You will choke out your own video.
Q2: Can I choke out 1258MHz signals on my 1280MHz system
A2: No. The channels are too close together. You will choke out both. Generally you want the signal you are rejecting to be more than 110% of your frequency or less than 90%.
Q3: What kind of rejection should I see?
A3: I see between - 30 and -40db. This is a reduction of over 1000!
|Oct 24, 2012, 11:03 PM|
Hmm, if I already have a lowpass filter on my 1280vtx and I'm flying with FrSky, would I benefit from making a filter for the FrSky module?
Thanks for the great tutorial by the way. Really excited to see your future videos on youtube as well
|Oct 24, 2012, 11:42 PM|
|Oct 25, 2012, 02:07 AM|
Mike at DragonLink swears that his Tx has the best filtering available on its transmission to prevent interference on 1.3GHz Rx's. I had terrible interference on my Racewood Rx, so I bought the RMRC one which is a huge improvement. However, when pushing the outer edges of my signal strength I still get interference in the form of rhythmic "winking" and horizontal lines. Turning my DragonLink Tx output switch to low power helps clear up the picture, so it seems I am still suffering some interference from my DL, despite it's claimed excellent filtering. Now that I am making new personal distance records, this is becoming a limitation.
Do you think this would help me despite Mike's filtering claims? Or should I just buy a Lawmate which seems to work for others?
Thank you for all your selfless contributions to the FPV community.
|Oct 25, 2012, 07:07 AM|
When it comes to pushing the limits, fitration is the key to keeping it all clean and working well.
|Oct 25, 2012, 08:26 AM|
^Hopefully they get it right. I don't trust BEVRC to use the proper coaxial cable. Using a different cable changes the velocity factor and thus throws off the calculations. The reason I said to use RG316 or RG142 is because they have the same velocity factor as the SMA Tee. The cables they use have a different velocity factor than the ones I do. I don't know about their other components. I assume the Tee would be the same, so that wouldn't change.
This is so incredibly simple that any hobbyist can do it. If you can solder up your own video system, this is a walk in the park.
|Thread||Thread Starter||Forum||Replies||Last Post|
|Discussion How to build a wooden hull in 1 minute||hiljoball||Sailboats||0||May 29, 2012 11:44 AM|
|Wanted Low Pass or notch filter for 1.2-1.3ghz||Adidas4275||FPV/RPV - Equipment and Vehicles (FS/W)||3||Apr 13, 2012 03:28 PM|
|Video Yardbird RC YB-22 5-minute build||TimAZ||Electric Plane Talk||15||Apr 03, 2008 04:13 PM|
|Sold Ending in 5 minutes on E.B.A.Y look 5014 Receiver cheap!!||rcgalore||Aircraft - General - Radio Equipment (FS/W)||0||Feb 19, 2008 04:37 PM|