[HBR] Yet Another HBR Project -- Chapter 1B

[hbr@mailman.qth.net]

The excitement -- is that the right word? -- continues ...

The first big question is whether I can build a sufficiently linearly 
tuning oscillator to justify a dial calibrated directly to 10 cps 
increments.   That's especially a dicey issue with a tuning cap of 
unknown characteristics -- but this cap (which rotates 240 degrees to 
cover its range) is the only thing I've got that will allow tuning 180 
degrees without end effects.   And it is of outstanding quality so if it 
can be linearized, it'll be good.

I've got the oscillator mechanicals roughed out.   The posts on the 
right side of the dial mechanism formerly supported the stator of the 
built-on tuning cap; now they will hold the large face of a 3x4x5 box 
in which the new tuning cap and other oscillator parts will live.   To 
save space, I shortened the posts about 5/8".   That turned out to be 
more interesting than expected because the posts are surface 
hardened -- anyone know what process gives a silver color?   Seems 
like overkill to me ... what sort of wear would such posts experience? 

The trickiest part was having to drill and tap 5/8" deeper in the cut-off 
end.   Three of them were no problem; the fourth somehow was 
surface hardened in the bottom of the hole -- a crystal or two of 
whatever was used (cyanide?) must have gotten down there.   And I 
had to Dremel a notch at the right point on each post before I could 
cut them off with a hacksaw.   

Lengths had to be exactly the same: I cut the posts a hair long, then 
drilled through a block of wood (of proper thickness) and used a belt 
sander to grind the ends off flush.  

The assembly went together without problems -- as Jim has noted it 
takes some time to drill and tap a bunch of holes, but I had only 24 
of them (3-48 and 4-40) so it wasn't that bad.

Tapping holes in an aluminum chassis or box usually isn't too 
successful because the material is too thin and the usual chassis is 
dead soft.   The following trick makes it work for 3-48 machine 
screws (a la command sets):  Drill a tiny hole -- say #50 or a bit 
smaller. If you're attaching a cover, drill through both pieces.  By 
using a Dremmel while spinning the bit in an electric drill, taper the 
butt end of the correct drill to tap for the screw you want -- #39 for 3-
48 -- to a sharp point.   Chuck with tapered end out in drillpress (or 
electric hand drill), lube with silicone grease, and quickly force (while 
spinning) through your pilot hole.   You now have a hole of the correct 
tapping diameter but in metal 2x to 3x the thickness of the rest of the 
sheet -- of course you do not remove the upset material!   This area 
has also been work hardened, so the thread will be much stronger 
than in unworked metal of the same thickness.   Holes can be 
tapped in the usual way -- I usually chuck the tap in a variable speed 
reversing hand drill but one of the old hand crank drills also works 
well.   The hole in the cover plate should be enlarged to fit over the 
upset; a #31 drill works for 3-48 screws.   

Try this on scrap first -- practice makes perfect!

It's more work than sheetmetal screws, but it goes quickly and gives 
a strong and good looking result on the usual 0.050 chassis.   Also, 
you can remove and replace the machine screws many times, if 
necessary.   This sort of thing may be unnecessary for a simple box 
cover, but for an oscillator assembly you get close to the stiffness of 
a cast box with the ease of working of sheet metal and a fraction the 
weight.

I needed mainly the stiffness (with minimum weight) because the 
entire oscillator -- box, tuning cap, sockets, tubes, small parts, and 
some wiring -- is supported on the four posts of the LM worm gear 
drive.   Looking down, the assembly is basically a backwards 'L' 
supported only at the short end of the bottom -- the cast gear frame.  
 I may eventually add a single support at the top (rear) of the 'L', but 
since it will go to the center of an aluminum chassis, a high 
stiffness/weight ratio is still a good idea.

Pictures -- does anyone recognize that tuning cap?   The box doesn't 
actually touch the panel, tho it is a bit closer than I wanted.   Looking 
at the 'C' photo (oscillator box open) the coil mounts by one end on 
the larger upper hole of the two in the space between the cap and the 
tube sockets.   It's a plastic (plexglas) tube about 3/4" dia x 2-1/2" 
long that held craft 'glitter' -- available at Wal-Mart and doubtless 
many other places.

http://www.collie-rescuers.com/ebay/11803A.jpg
http://www.collie-rescuers.com/ebay/11803B.jpg
http://www.collie-rescuers.com/ebay/11803C.jpg

I spent several hours researching tubes.   Because they were never 
cheap (and transistors were starting to come on the market), most 
subminiature tubes were used only by the military.    There are at 
least four battery filament triodes but all have 100 ma (or more) 
filaments and some are rated only for push-to-talk service -- 100 hour 
filament life.   There probably was a triode designed for hearing aid 
service (remember the cigarette pack size hearing aids?) but I don't 
have a suitable index of tube types to numbers; I have to start with a 
type number and retrieve characteristics.  

I did discover a whole line of subminiature tubes (roughly 5904... 
5908) designed for 26.5 volts on filament *and plate.*    I know of no 
military equipment that used such tubes -- does anyone?   'Far as I 
knew, the 26.5 volt plate era began and ended with the R-392 -- 
miniature tubes in that one.

I considered acorn tubes but the sockets are large and the only tube 
I could find is the 958 which has a 100 mA filament.

I was unable to locate even one miniature battery operated single 
triode.   The 3A5 is a dual triode, but has a 100 mA filament.   It 
looks like the best bet is triode connected 1U4's (minature, 1.4 volt 
0.05 amp filament, sharp cutoff pentode) and I'll try those first.  There 
are various stability questions with these tubes but for the more the 
more than 10-fold reduction in total heat production, it's worth a shot. 
Worst comes to worst I can replace with 6AF4's.

An invaluable resource for researching tubes is:

http://home.wxs.nl/~frank.philipse/frank/frank.html  

*19,000* tube types are indexed there, complete with data sheets. I 
found a few oddballs that were missing but I'll bet they'll be there in a 
few months -- the site owner does regular updates and posts 
requests for help on unknown types.

The oscillator box will be warmed by the other tubes in the set but 
that'll be only via the air, so it will happen very slowly.   Since the box 
is closed, all the parts affecting the frequency will warm up at the 
same rate, meaning that temperature compensation should be 
simple.   In fact I'll use a plastic form for the oscillator coil: If properly 
designed (a bit of space between the turns) such coils are self-
compensating and can even have a negative temp coefficient so they 
compensate for the positive coefficient of the other parts.

(The miniductor local osc. coils you see in some HBR projects are a 
really bad idea -- the wire is often not firmly gripped in the support 
bars and as the coil gradually warms, it will from time to time jump to 
a new position ... I only tried that once!

One must of course watch out for self-heating of oscillator tank parts 
due to circulating RF currents.   That means heavier wire than you'd 
expect for the coil and using multiple smaller NP0 caps in parallel for 
the necessary shunt cap.   Also the oscillator power level must be 
kept low.

Since I'm not using the FT-101 capacitor, there's no clear advantage 
to using the FT-101 tuning range.   I'll try tuning the oscillator 5.2-5.7 
Mcs instead of 8.7-9.2 Mcs.   That should be easier to linearize and 
definitely will help stability.   I haven't yet done the spur checks for 
that, (I will before winding any coils!) but it ought to be okay.  

('Ought to be okay' ... Hehehe ... ) 

This oscillator and dial mechanism is *big* -- several times that of the 
FT-101 and maybe double that of the 1MHBR -- but if the 
performance the individual parts are capable of can be realized in the 
assembly, it'll be well worth the extra space.   And it's all above 
chassis, so the space below is available for other circuits provided 
they do not generate much heat. 

I've made up the plate transformer -- two Radio Schlock 120 to 6.3-0-
6.3 x 3A filament transformers donated their primaries to make a 
single 120 to 120 transformer. They are bobbin-wound so you 
disassemble, unwind each secondary, cut off that end of the bobbin, 
use a sander to grind the center partitions to half thickness, restack 
the laminations inside the paired primaries, and reassemble.   Now I 
just have to wire up the power supply in the junk chassis, wire the 
oscillator and go from there.

Not being able to find suitable scrap, I ordered the front panel plate 
today.   About $29 for two 8" x 14" x 0.125" plates, half of that is 
shipping from WA via UPS.   Expensive I suppose, but their service 
is outstanding and the site works beautifully:

www.onlinemetals.com

Naw, I don't get anything by recommending them -- I could get a 5% 
discount, if everything worked right and I ordered again within 30 
days but I'm afraid I can't afford to save that much money.   

If I can get linear tuning then the next step is to wire up the Hahnel 
oscillator (multiples of 1.75 Mcs) and premixer and get that stuff 
working.   Then start over with the real receiver ... probably a couple 
months down the road, at least.

Walt
KJ4KV