I’ll bet that most of us have never actually known how to adjust or “tune” the IF filters in our radios, regardless if HF/MF or VHF/UHF. All I was exposed to was to tune for best match to my EAR – if it sounds good, then it’s good enough..!!
.
Although true – and certainly everything has to (eventually) pass to our ears, there may be a better way to be able to understand how filters respond – and to take advantage of their characteristics in ways that we’ve never experienced in the past. Also take into consideration that the new digital modes will be able to “hear” even below the noise-floor - so the ear may no longer be a factor in the near future (or current). I used the following technique when I needed to adjust the bandpass characteristics of the Elecraft K2′s IF filters. It was quick, easy, highly visual and very accurate.
.
The technique starts with injection of broadband, but stable (consistent) noise into the front-end of a receiver. For this, I used the Elecraft N-gen – an inexpensive, yet very useful tool that provides a noise signature that is within 3db from 100khz to 500mhz, it’s useful for all our radios to lower-UHF. The unit comes as a kit (typical Elecraft) that is very easy to assemble and has a single ON-OFF switch and a BNC-F connector for it’s output. It’s stand-alone, powered by a single 9V battery. You simply connect the N-gen using a short BNC M-to BNC-M cable, or a simple BNC-MM adapter works nicely as well. Turn the N-gen ON and you have a noise source being injected.
.
The next part requires a laptop with a microphone port to the sound card, as well as some software, as we’ll now be looking at the AUDIO that results to the external speaker port of the radio (an Elecraft K2, in my case). Fortunately, most radios have standardized connectors – most HF radios use 1/4″ phono plugs, while more compact radios use the 3.5mm stereo plug. Coincidentally, most laptops also have a 3.5mm stereo receptacle for mic-INPUT (as well as output), so it was a simple case of having a short 3.5mm male-to-male stereo cable to make this test between my MacBookPro and the Elecraft K2 (about $4 at Fry’s or other similar stores).
.
Software for the Audio Spectrum Analyzer comes in all flavors, as well as prices. Although we could have used many VERY expensive tools destined for the professional sound industry, there are a number of Freeware packages available.
- For the Windows environment, most folks are using Spectrogram Version 5.17 (last freeware version). http://www.w5big.com/spectrogram.htm
. - For Mac users, I have found iSpectrum to be a great tool for filters as well as V/UHF repeater audio alignment.
http://www.dogparksoftware.com/iSpectrum.html
.
The technique is as follows :
- Connect the audio OUTPUT from the radio, to the sound card INPUT (Mic-input in most cases) of the laptop.
. - Start Spectrogram or iSpectrum, depending on your hardware platform.
. - You should see a representation of the radio audio-output on the screen – if not, switch to the other channel (left, or right).
. - Once you’re able to see the signal displayed as a moving graph on the screen, narrow the bandwidth of the audio sweep from a low of ZERO, to a high of about 2.5 or 3 khz. This will focus our analysis on that portion of the usable radio audio spectrum.
. - Now start adjusting a single parameter of the current IF filter and watch the response on the Audio Spectrum Analyzer – you’ll see the noise move from left to right. If the filter is operating, you’ll see cut-off either on the left-side (low cut) or on the right-side (high-cut). In some cases, you might see both sides cut-off, signifying that you’ve engaged an IF bandpass filter.
. - Okay… now we understand cause-versus-effect. We’ve made a change thru one of the controls, then SEEN the effect of that change to the audio spectrum. Remember the N-gen generates broadband noise that is quite uniform across the spectrum of 100khz to 500mhz – to within 3db (about 1/2 of an S-Unit), so what you see, is pretty-much what you get (remember that the human ear can only distinguish about a 4db change in sound volume – best case).
. - Time for some real alignment – I’ll use the 8-pole CW filter that has a bandwidth of 400hz. In this case, we want the (passed) noise to be dead-center on the our SideTone frequency (600hz in my case) so I’ll set a marker at 600hz. We want the noise to be centered here – with equal fall-off on both sides of the 600hz marker.
.
Let’s start with a wider filter, so that we “zoom” into the target bandwidth – here’s the progression : 
Above is actually of a fairly wide-band CW filter (1.0khz)..
Above is from a 400hz wide filter – notice that the skirts only go about 200hz on either side of the 600hz marker line…- SSB filters (or anything non-CW) have a broader bandpass, so instead of aligning to a center frequency like in CW (at the side-tone frequency), we normally try to align the filter to be centered BETWEEN two frequencies (low-and-high limits). The standard objective is to get as much of the received noise to be within the low and high cut-off points as possible – and not have some of the signal ‘roll-off” either the high, or the low ends.
.
Above : an example of an SSB filter that is biased too far to the low-side – while this might be good for low frequency fidelity, it will compromise high frequency response – an area where most of the consonants are in the English language.
Above : a better aligned SSB IF filter that keeps most of the signal within the limits of the filter bandwidth.
