On May 4, 2006, at 12:00 AM, Steve Thompson wrote:
>
>
> Bill Turner wrote:
>> For years VHF parasitic suppression has been done in the anode
>> circuit of an amplifier. While it does work, I wonder if it might be
>> better to apply suppression measures in the grid circuit, especially
>> with a grounded-grid triode.
>>
>> One of the chronic problems with anode suppression is getting the VHF
>> impedance high enough, but not so high that the suppression resistors
>> overheat when operated on 28MHz. This is a delicate balance and one
>> that not many amplifiers - even expensive commercial ones - get
>> right. Many of us have found commercial amps with the suppressor
>> resistors seriously overheated, having their values changed and in
>> some cases, burned open, making them totally ineffective.
>
> Part of the problem has been carbon comp resistors which were (still
> are?) widely used. They are unsuited to the application and there have
> been better alternatives for at least 30 years.
>>
>> Since it takes two resonant circuits for a g-g triode to oscillate,
>> one in the plate and one in the grid-cathode, I'm wondering why there
>> hasn't been more focus on the grid side? If the resonance in the grid
>> side could be dampened out sufficiently, the anode side could be left
>> alone with no suppressor at all. In fact, some amps such as the
>> Ameritron AL-1500 already take this approach. The power level in the
>> grid circuit is much less than in the anode circuit and components
>> will be stressed much less.
>
> Being pedantic, it doesn't require any tuned circuits at all. The
> condition is gain>1 and phase=0. How you avoid this doesn't matter.
> How
> it arises can vary wildly, with very different optimum cures. What
> goes
> on with circulating currents and stray fields outside the tube can be
> more significant than internal feedback. A resonance gives you rapid
> phase changes over a narrow frequency, which increases the
> likelyhood of
> getting the right phase for oscillation, but it's not a prerequisite.
>>
>> I recently built an 8877 amp using one of AG6K's parasitic suppressor
>> kits. While it worked ok on 28MHz SSB and CW, when using constant
>> carrier RTTY, the resistors burned up within about 4-5 minutes of
>> steady operation. One of the resistors got so hot it unsoldered
>> itself, cracked in two and fell down on top of the tube! Since I
>> operate almost exclusively RTTY, something obviously needed to be
>> done.
>>
>> Thinking about the AL-1500, which uses the same tube, I decided to
>> try my amp with no anode suppressor at all. Gingerly, I fired it up,
>> expecting massive fireworks, arcing and strange noises. Guess what?
>> Nothing. It worked perfectly. I ran the plate tune cap through it's
>> range, ran the input tune circuit through it's range, tried
>> everything I could to make it oscillate but it would not. Neat!
>>
>> The only parasitic suppressor the amp has is the 33 pF in series with
>> a ten ohm resistor, connected from grid to cathode, which Rich
>> recommends. Since the un-suppressed anode circuit grid-dips at about
>> 145 MHz, the 33 pF has a reactance at that frequency of about 33
>> ohms. That value in series with a ten ohm resistor seems to be enough
>> to dampen any resonance in the grid circuit. At least, I can't find
>> any VHF resonances in the grid area with a grid-dip meter.
>
> It might be gain reduction rather than damping resonance that's
> working
> here. The capacitor from cathode to grid acts as a divider with the
> plate to cathode feedback C. For tubes with VHF grid structures, I'd
> guess that a low inductance 100pf grid to cathode would go a long
> way to
> killing all vhf parasitics in a hf amp.
To me, a 100pF C without a series-R is likely to create a worrisome
high-Q resonance in conjunction with its Cu leads.
>>
>> There are probably other methods which could be used in the grid
>> circuit. For example, a 145 MHz series resonant circuit with a Q of
>> 3-4 in series with a ten or twenty ohm resistor from grid to cathode
>> might work too.
>
> Your dip meter tells you there's a resonance. It doesn't tell you if
> it's series or parallel, which matters, it doesn't tell you it's the
> frequency of oscillation, and it doesn't tell you the plate
> impedance at
> that frequency.
Since it is an indicator of VHF gain, an important measurement is the
parallel-equivalent R presented to the anode/plate at that frequency.
> Don't get me wrong - plate suppressors do help stop
> oscillation, but copying something used in a another design using the
> same tube without considering differences in all the external
> circuitry
> and/or targetting something picked up by a dip meter means that
> success
> might be pure luck. How many amplifiers, like yours, don't need
> suppressors but have never been run without one?
All AL-1500s have no parasite suppressor whatsoever, and they have a
reputation for needing replacement 8877s from time to time. Since
tube failure in AL-1500s often occurs during the warranty period, and
Eimac's warranty policy requires turning in the dud, finding a dud
8877 from a AL-1500 that I could autopsy took several years.
The result:
http://www.somis.org/8877-gs2.JPEG
Curiously, to be able to purchase this tube, I had to agree not to
disclose the name of the seller because he said he did not want to
incur the wrath of the AL-1500's designer - the "recognized expert"
who decided that the feedback C in an 8877 was insignificant, and
that a parasitic suppressor was unnecessary.
>
> Unfortunately few have access to equipment which allows them to
> investigate why an oscillating amplifier does so.
IMO, there are too many variables to go much beyond making educated
guesses based on the available measurements.
> Common factors in some
> tubes lead to the 'I did this to fix it' folklore, which is fine - but
> then it can start being seen as a universal cure, when each situation
> and/or tube type needs individual consideration.
In my experiences, the fastest way to cause a parasite is to announce
absolute victory over them on a HF Ham band.
cheers, Steve
>
> Steve
> _______________________________________________
> Amps mailing list
> Amps@contesting.com
> http://lists.contesting.com/mailman/listinfo/amps
>
R L MEASURES, AG6K. 805-386-3734
r@somis.org
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|
