Difference between revisions of "UHF Preselector"

From W9CR
Jump to navigation Jump to search
Line 18: Line 18:
 
|}
 
|}
  
Note R3 and R4 use the same preselector.  The preselector hooks up to a
+
Note R3 and R4 use the same preselector.  The preselector hooks up to the RX input and can be removed for testing.
  
 
[[File:Preselector front.jpg|thumb|widths=300px|heights=300px|center|Preselector in receiver]]
 
[[File:Preselector front.jpg|thumb|widths=300px|heights=300px|center|Preselector in receiver]]
  
The internal design of the preselector is basically a tubular inductor and a top capacitance hat.  The center of this hat/inductor has a hole into which the tuning screws slide in to/out of.  The R0 preselector has a much longer screw, and the front panel has a protective metal part protecting the protruding parts.
+
The internal design of the preselector is basically a tubular inductor and a top capacitance hat.  Interesting to note the coupling takes place at the low side vs. the normal high side. The center of this hat/inductor has a hole into which the tuning screws slide in to/out of.  The R0 preselector has a much longer screw, and the front panel has a protective metal part protecting the protruding parts.
  
 
<gallery  widths=300px heights=300px class="center">
 
<gallery  widths=300px heights=300px class="center">
Line 40: Line 40:
 
| UHF R2 || .650
 
| UHF R2 || .650
 
|-
 
|-
| UHF R3 & R4 ||  
+
| UHF R3 & R4 || .437*
 
|-
 
|-
 
|}
 
|}
 +
Note the R3 and R4 is based on measurement from a friend, not measured with a caliper.
  
 
<gallery mode=packed widths=200px heights=200px class="center">  
 
<gallery mode=packed widths=200px heights=200px class="center">  
 
+
R4 and 3 Vs. R0.jpg|R0 in front vs a R3/4 preselector
R4 and 3 Vs. R0.jpg|thumb|R0 in front vs a R3/4 preselector
+
UHF Preselector diagram.png| Preselector diagram
 
</gallery>
 
</gallery>
  
File:UHF Preselector diagram.png|thumb|center|UHF Preselector diagram]]
+
Patterns of preselectors measured with a spectrum analyzer. Note the bit of ripple in the pass band.  Motorola sells a probe to be used to align the preselector via alternate peaking and diping it, but it's much easier to tune on a spectrum analyzer/tracking generator.  This has the advantage allowing you to tune for the exact pass band needed.  In many cases it's advantageous to tune the center off of the frequency.  I find it's easy to tune the entire US 440 ham band input range as a CF of 447.5.
 
 
Patterns of preselectors
 
  
 
<gallery mode=packed widths=200px heights=200px class="center">
 
<gallery mode=packed widths=200px heights=200px class="center">
 
QUANTAR R2 factory.png|UHF R2
 
QUANTAR R2 factory.png|UHF R2
 
+
QUANTAR R0 380 UN MOD.png|Quantar R0 at 380 MHz
QUANTAR R0 380 UN MOD.png|thumb|Quantar R0 at 380 MHz
+
QUANTAR R0 BEFORE MOD 1.png|Quantar R0 at center
QUANTAR R0 BEFORE MOD 1.png|thumb|Quantar R0 at center
+
QUANTAR R0 MAX FREQ.png|Quantar R0 at 437 MHz
QUANTAR R0 MAX FREQ.png|thumb|Quantar R0 at 437 MHz
+
R0 tuning.jpg|Tuning a R0 preselector
R0 tuning.jpg|thumb|Tuning a R0 preselector
 
 
</gallery>
 
</gallery>
  
Line 66: Line 64:
 
Two styles of modification have been attempted on these preselectors to raise the frequency.  Lowering the frequency has not been attempted (ie. R4 to R2), and all focus on moving R0/1 to R2, though the same techniques would work to move to R3/4.
 
Two styles of modification have been attempted on these preselectors to raise the frequency.  Lowering the frequency has not been attempted (ie. R4 to R2), and all focus on moving R0/1 to R2, though the same techniques would work to move to R3/4.
  
== Shorting the Rod ==
+
== Method 1: Shorting the Rod ==
  
Shortinign the rods was attempted on the first modification of the R0 to R2 preselector.  The cap hats were removed and the center rods were ground down .270" on each rod.   
+
Shortening the rods was attempted on the first modification of the R0 to R2 preselector.  The cap hats were removed and the center rods were ground down .270" on each rod.   
  
 
This became a problem as the rods are made of silver plated copper plated aluminum, which makes grinding a hard thing to do.  It removed the tinned part of the rods where they solder to the cap hats, and the alignment "key" notched into the top of the rods.  The final result was a filter which had a bit more ripple and wider than desired, but with good insertion loss.
 
This became a problem as the rods are made of silver plated copper plated aluminum, which makes grinding a hard thing to do.  It removed the tinned part of the rods where they solder to the cap hats, and the alignment "key" notched into the top of the rods.  The final result was a filter which had a bit more ripple and wider than desired, but with good insertion loss.
Line 74: Line 72:
 
I suspect the width was changed mostly due to moving the cap hats lower in the housing.
 
I suspect the width was changed mostly due to moving the cap hats lower in the housing.
  
[[File:R0 cut down tubes.jpg|thumb|R0 preseletor during modification.]]
+
<gallery mode=packed widths=200px heights=200px class="center">
 +
R0 cut down tubes.jpg|R0 preselector during modification
 +
R0 post mod drills.jpg|Quantar R0 converted to R2 by cutting tubes showing post modification drilling to lower capacitance
 +
Quantar preselector R0 to R2 tube cutting mod.png|Quantar R0 converted to R2 via tub cutting sweep
 +
QUANTAR R0 CONVERT OVERLAY ON R2.png|R0 converted to R2 by tube cutting over lay on real R2
 +
</gallery>
 +
 
 +
For the R0 filter at normal frequency the tuning slugs should be roughly as follows:
 +
{| class="wikitable sortable"
 +
|-
 +
! Center !! length sticking out (inches) || insertion loss
 +
|-
 +
|  380 MHz || .275 || .65 dB
 +
|-
 +
|  433 MHz || 1.140 || .60 dB
 +
|-
 +
| 437 MHz || 1.200 || .40
 +
|-
 +
|}
 +
 
 +
After grinding the stubs down it was necessary to use an aluminum flux and soldering iron to get the solder to adhere to the metal of the studs.  This flux destroyed my Metcal tip and it had to be replaced, at a cost of $30.  After this it was easy to use a heat gun and reflow the solder on the caps.  The issue here is to make sure they line up as the notches are removed from the stubs during grinding.
  
[[File:R0 post mod drills.jpg|thumb|Quantar R0 converted to R2 by cutting tubes showing post modification drilling to lower capacitance]]
+
Over all this method works, but it's much easier to cut the hats down in length.
  
[[File:Quantar preselector R0 to R2 tube cutting mod.png|thumb|Quantar R0 converted to R2 via tub cutting sweep]]
+
== Method 2: shortening the cap length ==
  
[[File:QUANTAR R0 CONVERT OVERLAY ON R2.png|thumb|R0 converted to R2 by tube cutting over lay on real R2]]
+
This is the preferred method to raise the frequency as it's more in line with the construction Motorola uses.

Revision as of 00:06, 4 December 2015

Basics

The UHF preselector is a 3 pole, low insertion loss (1.2 dB) 4 MHz wide filter. In practice this is closer to 5MHz wide.

The preselector varies based on the band of the RX as follows

P/N Range Description
CRX4001 UHF R0 UHF Preselector 380-433 MHz
TLE5991 UHF R1 UHF Preselector 403-433 MHz
TLE5992 UHF R2 UHF Preselector 438-470 MHz
TLE5993 UHF R3 & R4 UHF Preselector 470-520 MHz

Note R3 and R4 use the same preselector. The preselector hooks up to the RX input and can be removed for testing.

Preselector in receiver

The internal design of the preselector is basically a tubular inductor and a top capacitance hat. Interesting to note the coupling takes place at the low side vs. the normal high side. The center of this hat/inductor has a hole into which the tuning screws slide in to/out of. The R0 preselector has a much longer screw, and the front panel has a protective metal part protecting the protruding parts.

The bottom of the preselector has T10 size screws on it and they need to be tightened and loosened in order of the numbers on the bottom plate. Once apart the variance in the length of the cap hat should be noticed showing the range of the preselector.

Range Cap Hat length (inches)
UHF R0
UHF R1 .880
UHF R2 .650
UHF R3 & R4 .437*

Note the R3 and R4 is based on measurement from a friend, not measured with a caliper.

Patterns of preselectors measured with a spectrum analyzer. Note the bit of ripple in the pass band. Motorola sells a probe to be used to align the preselector via alternate peaking and diping it, but it's much easier to tune on a spectrum analyzer/tracking generator. This has the advantage allowing you to tune for the exact pass band needed. In many cases it's advantageous to tune the center off of the frequency. I find it's easy to tune the entire US 440 ham band input range as a CF of 447.5.

Modification

Two styles of modification have been attempted on these preselectors to raise the frequency. Lowering the frequency has not been attempted (ie. R4 to R2), and all focus on moving R0/1 to R2, though the same techniques would work to move to R3/4.

Method 1: Shorting the Rod

Shortening the rods was attempted on the first modification of the R0 to R2 preselector. The cap hats were removed and the center rods were ground down .270" on each rod.

This became a problem as the rods are made of silver plated copper plated aluminum, which makes grinding a hard thing to do. It removed the tinned part of the rods where they solder to the cap hats, and the alignment "key" notched into the top of the rods. The final result was a filter which had a bit more ripple and wider than desired, but with good insertion loss.

I suspect the width was changed mostly due to moving the cap hats lower in the housing.

For the R0 filter at normal frequency the tuning slugs should be roughly as follows:

Center length sticking out (inches) insertion loss
380 MHz .275 .65 dB
433 MHz 1.140 .60 dB
437 MHz 1.200 .40

After grinding the stubs down it was necessary to use an aluminum flux and soldering iron to get the solder to adhere to the metal of the studs. This flux destroyed my Metcal tip and it had to be replaced, at a cost of $30. After this it was easy to use a heat gun and reflow the solder on the caps. The issue here is to make sure they line up as the notches are removed from the stubs during grinding.

Over all this method works, but it's much easier to cut the hats down in length.

Method 2: shortening the cap length

This is the preferred method to raise the frequency as it's more in line with the construction Motorola uses.