>...snip...
>Makes me wonder how many SB-220 series have lost the filament xfmr due to
>the infamous C-4??
>
>On a slightly different note Rich makes mention of filament xfmr
>protection from a tube short on his Web page but it is a bit confusing.
>Maybe you can elaborate a bit Rich.
In a SB-220, during intermittent push-push parasitic oscillations at
c.110MHz, lateral mechanical stress is applied to the 3-500Z's hot
tungsten filament (1800 degrees K). This stress is due to the apparently
large burst of grid current associated with such oscillations. (see
photo on page 15 in the September, 1994 *QST*). After several such
oscillations, the filament may be bent far enough to touch the grid,
resulting in a filament to grid short. The filament center-tap is the DC
cathode/filament return. The center-tap is connected to the +120V power
supply during receive in order to cut off the flow of anode current.
When the filament/cathode is bent far enough laterally to touch the
grounded-grid, a short to ground is placed on the +120V power supply
during receive. However, there is no fuse for the filament transformer
from which this supply receives its AC source. Thus, a short on the
+120V supply causes the unfused filament transformer to summarily melt
down. One solutions is to install a 10k ohm, 2w resistor between the
+120v supply and the relay's bias switching contacts. In the event that
a 3-500Z develops a grid/filament short, the short-circuit current will
be limited to about 0.12A. Another solution is to convert the SB-220
from voltage-cutoff bias to resistor-cutoff bias (a.k.a. 'self-bias').
The existing 100k, 0.5w resistor on the relay contacts can be used for
this purpose.
>Summertime suggestions for SB-220 series owners:
>
>1. Check those HV filter caps, replace if showing leakage. If no
>leakage, measure the voltage across each one. In the SSB position they
>should be roughly 370VDC and a +/- variant of 10-15V is OK with the 30K
>resistors. A wide range of readings and it is time to consider
>replacements. The Harbach assembly is an excellent choice since he uses
>1% matched CGS caps and 100K resistors. They still fit into the original
>coffin with no air circulation but it is better than original.
>
Matsushita/Panasonic also makes suitable electrolytics---which are
available from Digikey. The part # is P6443 for 330uF at 450 V. The
first price break occurs at 10 units, so if you buy 2 sets (16) and share
one set with a friend you can save some money (the price for 16 is $9.95
each). Digi-Key's telephone number is 1-800-344-4539. Heath was not
known for generously sized filter caps. IMO, a little extra C is better.
>2. Replace the 200pf red rectangular grid bypass caps with 500V dipped
>silver micas. Later amps used 115pf supposedly to better compensate for
>grid XL. I use 120-130pf since they are on hand...work fine. Those old
>block style caps are known to crack and go bang.
I would be inclined to replace the grid to ground 1A RFCs (whose fusing
current is around 15a) with frangible resistors or fast acting 0.25A
fuses. However, when this is done, an additional 1000pF or so is needed
to protect the grid fusing element from 10m RF.
- OTOH, The grids may also be directly grounded. If you choose to do
this, and you have a scientific bent, measure the 3-500Z's grid
self-resonant frequency before and after with a dipmeter. (not
recommended for members of the Rauchian camp)
>3. Check the output bandswitch for burnt contacts.
Some photographs of what seem to be parasitic-arced bandswitches are
available on my Web site.
>4. Measure the value of the parasitic suppressor resistor. It should be
>47 Ohms. If over 60 it is time to replace with a new one or alternative
>offerings.
>
If the original suppressor resistor burned up, IMO, it wasn't because the
resistor had a manufacturing defect.
>5. Check the 500pf 80M loading mica, C-56; they like to split open.
>
The 2000v rating is difficult to find. Try Spectrum Electronics, first.
>6. Check the plate blocking cap, C-29 for any signs of cracking. I
>sometimes wonder if this is really a RF rated cap since I have found many
>bad ones.
It is not an RF-rated cap.
>7. Install a pair of back to back diodes across the meter leads. Any
>1N400x series is fine. Original meters are virtually extinct and the
>diodes will protect from a tube flash over or short.
>
Original meters are still available: SPECTRUM ELECTRONICS (616)
742-0613 They reportedly carry the SB-220's current-limiting filament
transformer and well as other parts.
>8. Burnish the relay contacts. I have yet to replace one due to hot
>switching but a little PM goes a long way. Make sure it engages with a
>good clunk. Remove and compress the spring a bit if necessary for good
>action.
Hotswitching is assured when a stock SB-220 is used with a
faster-switching modern radio. The fix is to install high speed
switching. (see Fig 7 on my Web site)
http://www.vcnet.com/measures
Also see 1/94 *QST*, 'The Nearly Perfect Amplifier'. To test for hot
switching, press the push-to-talk button on a mobile microphone. If the
SWR indication on the radio temporarily surges at the instant the button
is pressed, hot switching is indicated.
>
>9. Most SB-220's and early 221's will have the small Johnson 154-9 Tune
>cap or the OEP 26-131. This cap is prone to arcing since it is only
>.075"/3KV spacing. Later amps and Heath replacements were the OEP 26-164
>which is about .090/3500V spacing. It will still arc on occasion when the
>amp is mistuned but not melt down as did the original. There is no
>Johnson/Cardwell equivalent. If there is enough interest I will have the
>26-164 reproduced.
>
- IMO, the voltage withstanding ability of the original tune-C is lower
than the voltage withstanding ability of the output bandswitch.
- IMO, the voltage withstanding ability of the wider-spaced tune-C is
higher than the voltage withstanding ability of the output bandswitch.
When something goes wrong, either the bandswitch OR the tune-C is likely
to arc. The choice is yours. Those who would rather file down arc marks
on the tune-C plates should probably opt for the closer spaced tune-C.
However, those who would rather replace crispy-crittered bandswitches
should probably go with the wider spaced tune-C.
>10. Check the fan. After many years and no lubrication they slow down a
>lot. Replace with the Harbach unit...spend the extra $1 for the high
>speed version, particularly if you live in a 50Hz world.
>
...and if you live in the 60Hz world, the high speed fan will reportedly
suck chrome plating off a trailer hitch.
Rich---
R. L. Measures, 805-386-3734, AG6K
--
FAQ on WWW: http://www.contesting.com/ampfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|