On 02-Apr-10 12:53 AM, Jim Thomson wrote:
>
> ## Eimac sez a 3x3 is supposed to be 24.0 pf from anode to grid
> on a wooden bench. I measure exactly 24.0 pf on the 875.
> Eimac sez the YU-148 is supposed to be 24.5 pf, from anode to grid.
> I measure exactly 24.5 pf on the 875. That's good enough for me.
*** Agreed, no reason not to use the 875 or equivalent for this
measurement. I have used my old homebrew C meter that tests at audio
frequencies with success. Also used my MFJ successfully.
throw away those knobs and use fixed capacitors!
>
> ## which is why I don't use fixed caps, nor 9 x bandswitched
> tuned inputs anymore. Those days of fixed SM caps and a
> tunable Millen 60946 slug or XR-50 are over. Plan B was
> to use variable arco's and fixed T-50 torroids. Plan C is to use
> 1 x tapped coil, and 2 x broadcast variables.
*** The comment I made about throwing away the knobs was made tongue in
cheek. I actually have done that. Interesting experiment but impractical
unless you ONLY run the amp into a 50 ohm dummy load.
*** I disagree, but only on a philosophical level. I absolutely refuse
to consider any solution which requires more knobs. Even with logging
scales and charts, it increases the band change effort well past my
comfort zone. My eyes are not what they used to be. I can't see the damn
charts or numbers on the knobs easily any more. I would (and do) spend
as much time as necessary to implement fixed tuned input networks. Those
stupid slug tuned coils are about the worst thing I can think of to do
for an input network. T50-2 and T50-6 toroid inductor are the only thing
that should be considered (for a fixed tuned network). Arco adjustable
caps are good but still not as good as fixed silver mica caps. All my
input networks are constructed using toroid inductors and fixed silver
mica caps. They are tuned on the bench (usually one time only but
unfortunately I don't get it right every time) and then forgotten until
I rebuild the amp for some reason. (I'm never satisfied!) I usually get
them close with the MFJ and then "fine tweak" for at least 30 dB return
loss in the middle of each band using the VNA. This will give better
than 1.1:1 input SWR over the band, including 80/75. If it is an 8 band
amp, then there are 8 tuned networks.
>
>
> ## I only have the trial version of KM5KG's RF network designer.
> The full blown version is next weeks project, which should have been
> done a long time ago.
*** The new trial version is crippled excessively. I have an older trial
version that has the solenoid coil design feature enabled. I keep
meaning to get the paid version because it is such a great tool but
never seem to get around to it.
>
> ## when the tank coil cooks, and BW is narrow, that's usually
> a subtle hint, that loaded Q is too high. A loaded Q of 8-10
> using GM3SEK's sheet works good. If drive power is then reduced,
> plate load Z will rise, and so will loaded Q. A 8-10 loaded
> Q will rise to 10-12. Eff is still good, no coils get cooked...
> [ and no bandswitch's getting cooked either]
*** I design for a Q=12 but I also have a separate band switch position
for each and every band, no doubling up. I have trouble imagining a tank
so marginal that it melts down when used at a lower power level.
*** Checking SWR bandwidth is a great "feel good" test that can easily
be done with instruments everybody *should* own. I prefer to actually
measure the loaded Q (since I have the instruments). For instance the
last tank I did had design goals of Q=12 for 80-12 and Q=14 for 10. I
tuned up the finished amp on each band with full drive and recorded the
logging scale numbers. Returned the amp to the bench, set and measured
the loaded Q on each band for the actual tuning used for full power. No
resistors, no guessing. Measured Q's came out to be in the range of 12.1
to 12.6 on 80-12 and 14.6 on 10.
>In all
> three cases, the MFJ produced measurement results that were at least an
> order of magnitude better than what VE7RF reported.
>
> ## Maybe mine is blown up ?
*** "Blown up" is probably too harsh. The few detector diode failures
I've seen resulted in reading that were incorrect but close enough that
you could talk yourself into believing them. The only way to tell is to
check cal with resistors and capacitors (and inductors if possible) of
known accuracy at a number of frequencies. Check for both R and X. 50
ohm microwave terminations work good for R. X should be checked with
good caps that have a reactance approximately equal to 50 ohms at the
test frequency.
>
>
> ## u could hook interconnecting leads to mfj.. and measure it... then
> subtract those readings from the completed coil + interconnects.
*** Sometimes that works OK, sometimes not. The actual inductance of the
"test leads" varies quite a bit as you move them around. On large coils
such as used for 80 meters, it's usually insignificant. But on smaller
coils such as used for 10/12 or 6, it can result in unusable results. A
guy needs to take each measurement on a case by case basis and PAY
ATTENTION! Take nothing for granted.
>
>
>
> ## The b+k has spring loaded 3/4" wide slots on front, so bannana
> jacks/test leads don't have to be used at all. Any strap up to 3/4"
> wide, can be used as a test lead. Also, things like .01 uf caps can
> have their leads stuffed deep into the same spring loaded slots,
> eliminating the stray L from their 20 ga long leads.
*** Seems like no matter how many "fixtures" I dream up for the various
instruments I have, none are 100% effective. Also seems like simpler is
usually better. For leaded parts using instruments with UHF connectors,
I mostly just bend the leads around so they jam into the UHF connector.
This works well for caps and toroids, don't work at all for air core
solenoid coils. (Proximity to instrument and fingers change the
inductance and Q) Those BNC to binding post adapters work pretty good
for most leaded parts when you need to "screw" them down.
73, Larry
Larry - W7IUV
DN07dg - central WA
http://w7iuv.com
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|