Modifying the Inrad Mark V Roofing Filter Board

I have noticed over the years that I may have some IMD issues with moderately strong close in signals despite my installation of the new roofing filter.

In researching this issue, I found some interesting data on the W8JI website here: http://w8ji.com/receiver_tests.htm

Pretty interesting, notice the rather dramatic improvement in close spacing IMDR and BDR when the Inrad mod is removed. I suspect that this will be true for all spacing within the bandwidth of the filter. Now, I also suspect that the filter bandwidth is actually wider than the advertised spec of 4 KHz, more like something between 5 and 6 KHz.

I conjecture that there is too much gain in the 1st IF. I believe this because after installation of the roofing filter, it is necessary to reduce the IF gain in the transceiver menu to return the S meter readings to be equal to those before installation. When you reduce the gain in the menu it is in the third IF, not in the first. Therefore, since I believe that most of the IMD originates in the second IF mixer, reducing the gain in the menu does nothing to reduce IMD created by too much gain in the first.

In discussing this issue with Inrad, I was informed that the prototype board filters had greater loss by about 3.5 dB than actual production boards, and the amplifier design results in 3.5 dB net gain rather than exactly balancing the filter insertion losses. Inrad didn't think this was a big deal since it was only 3.5 dB, but I look at this issue somewhat differently.

As I wrote to Inrad:

"It does make sense to me that the higher gain in the 1st IF would lead to lower IMD numbers, if the IMD is generated mostly in the 2nd IF (mixer). This seems most reasonable since for 3rd order IMD the rule is 3 dB for every 1 dB of tone power increase. Thus, 3.5 dB more power per tone would increase the IM power by 10.5 dB. This exactly corresponds with W8JI's experience with the module and is most interesting."

So I conjecture that this gain would result in considerably lower IMDR and BDR numbers, like those corresponding with the W8JI table.

So, how do we get the narrow roofing filter benefits without reducing the dynamic range so greatly? We reduce the gain of the on board Inrad two stage amplifier.

Here's how to do it:

In order to keep the proper input impedance for loading the filter, two resistors must be changed. R5, the 12 ohm resistor should go to 15 ohms. R6, the 220 ohm resistor should go to 150 ohms. This will lower the gain by about 3.5 dB.

You can find the schematic of the roofing filter module on the Inrad website for more details.

You should be able to return the IF gain menu setting to the factory setting after making the changes on the filter board. Obviously, in the ideal situation, you would be able to adjust the gain for the particular insertion losses of your filter. No two filters will be exactly alike, and this is a close approximation of the average production filter.

Now, I do not understand why the ARRL and RSGB labs didn't find the same results, except, perhaps they received the early prototype production units for their testing.

Now increasing the IF gain setting in the menu will slightly increase the noise floor again back to the factory condition. You will loose the improvement there, but it only a dB or two. Personally, I think my new BHI DSP module, the ANEM, of which I wrote about in the March QST, will easily take care of that issue.

This is a simple change, and can be made easily. Your radio should be less susceptible to intermod generated from strong close in signals, particularly from garbage being thrown around the band from very strong close by signals.

Operators with large antennas in great locations should benefit from this small change. When the signals are of lower strength, this change will not do anything for the receiver. BUT, the new sun spot cycle probably has begun, and we will see more and more days when every signal will be S9 and the band will be wall to wall. It should really help then.

See what you think. Mike W2AJI