|
An amazing filter ,
by CT3FQ (1)
|
Roofing Filters
When ICOM IC-7800 appeared in the market, I was really curious and also suspicious with the IMD and IP3 numbers divulged for a double IF receptor. Years before, the manufacturers used the "quadruple conversion" as a marketing weapon for the top range transceivers.
Reading the IC-7800 catalog (it's truth is not elucidating) I discovered a new word for me: "Roofing Filter."
|
fig1.The FTDX-9000
"roofing filters" |
After reading some papers about the divulged Roofing Filter innovation of the 9000 euros transceivers I have noticed that was not a miracle of the other world, but just a consequence of the technological evolution.
When the Yaesu 1000mp was projected the narrow filter (<6KHz) technology didn't exist for hight 1st IF frequency (70MHz).
Anyway, what is the big problem of the 1st IF filters bandwidth? The 1000mp's 1st IF filter is a simple 14KHz LC circuit and we know that in the Amateur Bands is normal to be dozens of stations inside of a 14KHz bandwidth that will cause intermodultation and selectivity problems in the next IF stages of our receiver.
What happens is, although out of the bandwidth of the 2nd FI, 3rd and even 4th FI, the 1st IF signals stations will cause false signals in the others stages disabling the true signal. If we have a narrow filter in 1st stage, we will deny that the "garbage" saturates the nexts IF stages circuits. |
|
fig.2- The INRAD's roofing filter |
The acquisition of the Filter
I google's search the internet for "roofing filter" words I have discovered that INRAD have produced a roofing filter for the 1000mp. After having exchange some impressions with G3VXJ and ON4ADZ - that had already tested the filter and they were satisfied - I made the order to INRAD. Unhappily, I had to wait more than 2 months for the delivery because the 200 monthly units produced was not enougth for the market demand. |
fig3. The Assembly
|
The assembly
The most difficult of the Filter assembly was to open 1000mp MK5, without losing any screw. The INRAD Kit includes a special coax cable for the 70MHz and has the size for the perfect fitting in the "motherboard" of the 1000mp. The riskiest part was to weld two threads to the board for feeding of 5 Volts, because Yaesu "forgot" to put the feeding pins in Mark V as they existed in the old 1000mp. |
Fig.4 Yaesu 1000 MP MKV BLock Diagram
|
| Model |
Noise
Floor(dBm) |
AGC
Thrshld (uV) |
dB |
100kHz
Blocking(dB) |
Sensitivity(uV) |
LO Noise
Spacing(dBc) |
kHz
|
Front End
Selectivity |
Filter
Ultimate(dB) |
Dynamic Range
Wide Spaced(dB) |
kHz |
Dynamic Range
Narrow Spaced(dB)
|
KHz |
| Ten-Tec ORION |
-127
-135 |
0.8 |
3 |
137 |
0.6
0.25 |
130 |
10 |
B Band Pass |
100f |
96 |
20 |
93 |
2 |
Icom
IC-7800 |
-126
-136(1)
-139(2) |
4.5
1.2(1)
0.6(2) |
3 |
>135 |
0.60
0.15(1)
0.10(2) |
130 |
10 |
A Trk Presel |
100f |
102 |
20 |
80 |
2 |
Icom
756 Pro |
-127
-136(1)
-139(2) |
3.5
1.0(1)
0.5(2) |
3 |
132 |
0.55
0.21(1)
0.14(2) |
127 |
10 |
B 0.5 Octave |
90 |
86 |
20 |
71 |
2 |
Yaesu
FT-1000 D |
-128(2) |
6.0(2) |
3 |
>131 |
0.3(2) |
121 |
10 |
B 0.5 Octave |
90 |
90 |
20 |
69 |
2 |
Yaesu
FT-1000 MP |
-125
-134(1) |
3.8
1.1 |
3 |
>135 |
0.48
0.18(1) |
128 |
10 |
B 0.5 Octave |
90 |
97 |
20 |
68 |
2 |
Yaesu FT-1000
MP MKV Field |
-133(1) |
3.0(1) |
3 |
135 |
0.2(1) |
128 |
10 |
B 0.5 Octave |
90 |
88 |
20 |
69 |
2 |
Yaesu FT-1000
MP MKV Field
Inrad roofing filter mod |
-133(1) |
3.0(1) |
3 |
135 |
0.2(1) |
128 |
10 |
B 0.5 Octave |
90 |
91 |
20 |
77 |
2 |
fig.5 receivers tests table (source:
Sherwood Engineering Inc.)
(1) preamplifier 1 on
(2) preamplifier 2 on
|
fig.6 In red the new 1st IF bandwidth |
The first tests
When I power on the 1000mp was surprised because there was no difference between now and before! If we look for the table of the fig5. we verified that most of the values are equal before and after the "roofing filter", but in the two last columns of the table (Dynamic Range) we notice that exists some dBs difference between the original and the modified. The amazing number is the 2Khz +8 dB that, anyway, it doesn't even near of the ICOM IC-7800's 80 dB and far away of the TenTec ORION's 93 dB
But after all, what is the importance of a better Dynamic Range?
All receivers deteriorate when the intermodulation caused by strongs signals are superior to the own noise floor. Dynamic Range is measured putting two signals some Kiloherts distanced of the receiver signal . As we can see in the Table of fig.5 they were made two measures of Dynamic Ranges, one at 20KHz and other at 2KHz. Bigger is Dynamic Range better is the dynamics of the receiver when confronted with other strong adjacent signals.
A miracle called roofing filter.
In Contests is normal to have QRO stations spaced some Kilohertz. This fact disturbs the reception creating false audio images. I have tested the the 1000mp in the 40 meters - where each conquered kilohertz is a hard battle - and I was astonished when I was receiving the weak signals just some KHz down the QRO stations. I have listen the yaesu 2nd receiver in the same frequency and I didn't listen anything of nothing.
I also noticed a noise floor reduction. However, this modification has a small bug: The quality of the audio in AM and FM is worse, but as it is obvious, always we can use the 2nd receiver of the 1000mp to listen "broadcasting."
Will it worth the € 180 that I paid for the filter? |
73
CT3FQ
Carlos Neves, December 2004
(1) Translated from original Portuguese review by an Automatic Software Tool
|
|
