Grounding the Grids vs. Cap/RFC Grounding

Discussion in 'Amateur Radio Amplifiers' started by W4LAC, Jun 13, 2019.

ad: L-HROutlet
ad: l-rl
ad: l-BCInc
ad: abrind-2
ad: Left-3
ad: Left-2
ad: L-MFJ
ad: l-gcopper
ad: QSOToday-1
  1. G0HZU

    G0HZU QRZ Member

    When it comes to describing the role and design of parasitic suppressors, I think that applies to pretty much every post I've ever seen (by anyone) across various forums over the last 15-20 years. I've not seen anyone produce even a basic model or calculation to predict 'why' the suppressor R and L values are chosen for each amplifier type. However, there is no shortage of people (experts?) willing to give vague answers about VHF Q and resonances but nothing of any substance that would withstand scrutiny in a formal engineering design review.
    Last edited: Jun 16, 2019
  2. SM0AOM

    SM0AOM Ham Member QRZ Page

    What I have seen (and used) are methods, both analytical and empirical, of finding the spurious poles in the transfer functions for amplifier stages. By adding loss (or "damping") they may be rendered harmless by moving them away from the right half plane.

    Warren Bruene addressed this question in some articles.

    A form of parasitic suppressor which is quite difficult to understand is the one where the cavity formed by the tube anode and the surrounding metal is "de-Q:ed" by mounting resistors that are only connected at one end close to the anode.

    When these were encountered in a MF/HF transmitter in the 80s, it certainly puzzled me and my colleagues. Someone suggested testing what should happen if they were removed, and the result was a power oscillator at around 200 MHz.

    There was a test jack connected to a very small coil in the anode compartment, and a scalar network analyser showed marked resonances in this frequency range, which vanished by re-installing the resistors.

    I believe that the art of AC circuit analysis in tube circuit is becoming lost. Very few engineers and even fewer radio amateurs possess this knowledge today.

    Earlier experts were able to "tame" extreme designs such as four 4cx5000A in parallel.

    N2EY likes this.
  3. G0HZU

    G0HZU QRZ Member

    Sadly, I think this is now slowly happening with discrete semiconductors. Certainly where I work, modern RF design is becoming more and more like datasheet based LEGO using complex integrated circuits or chipsets. The skill/art is in keeping up to date with the latest ICs and chipsets and making the best use of the 'LEGO' to do what the customer wants.
  4. KM1H

    KM1H Ham Member QRZ Page

    The basics are simple and a lot of repetitive angst over the subject is not needed to build a good one, just a bit of common sense and seeing what commercial amps have used for many decades and not trying to reinvent the wheel. The simple resistor and coil has been around since the 50's and before that various versions of a tuned circuit was used

    Start first by determining the parasitic frequency by eliminating the suppressor. I started with a GDO in the 60's and went with a ~ 1971 design surplus HP 141T system in the early 90's that I got for cheap from work. Besides heating the basement work area it allowed all sorts of signal analysis.

    You will find that for most tubes in any well built amp stage from receiving tubes to tubes with handles that the frequency is pretty repetitive with that type of tube. Ive found a 10-15% variance to be the norm across many amps and tube brands.

    The inductor should be broadly self resonant around that frequency.

    The R does not act as a load and should consume little to no power. Its sole purpose is as a suppressor that stops the parasitic from even starting. Too low a value deQ's the circuit excessively so that it starts to accept power from the desired RF. This becomes a problem with ancient high gain tube designs such as the 811 when used on 10M. Too high a value and it starts becoming ineffective. It is also the reason that AG6K's nichrome voodoo get red hot on 10M with any amp approaching serious power.
    Carbon composition resistors all heat from the ambient tube heat and some small conductive path and change value, normally up in a normal circuit not experiencing a hard failure.

    The Ohmite OY resistor is the best type to come along since carbon and is available in the common values used which are typically 33 to 56 Ohms with some at the bottom of the bell curve at both ends. I use 47 for just about everything or a pair of 100 in parallel when pushing ancient tube designs such as running a pair of 572B's on 6M at 600-700W out with no complaints.

    Why waste time on something that is long proven to work and any engineer/good tech worth the name can figure it out without hand holding?

  5. G0HZU

    G0HZU QRZ Member

    Your vague answer reinforces my belief that a lot of the forum experts here are 'experts in fiddling' with these amplifiers and they each have their own vague explanations for how these suppressors function. I think this is why there has been so much disagreement over the years.
    N2EY likes this.
  6. K2XT

    K2XT Premium Subscriber QRZ Page

    Good advice from the master!
    I might add from experience with my first homebrew pair of 4-400s using a discarded L4B tank circuit back in the 70s. It used to start arcing across the loading cap. Driver was a TR-4. I used to think the problem was the loading cap had too close spacing. In reality it was probably hot switching because I had a big old vacuum relay on the output which I now know had a 30 ms switching time (meant for some other purpose than switching rf at the output of an amp)? And of course the TR-4 had no software and menus to set the transmit delay time!
    KM1H likes this.
  7. KM1H

    KM1H Ham Member QRZ Page

    My explanation was far from vague and I suspect those who cant accept a simple explanation based upon decades of factual experience should take up another pastime in order to annoy others.

    I also find it odd that your call does not show in the database.
  8. WB2WIK

    WB2WIK Platinum Subscriber Platinum Subscriber QRZ Page

    Common problem, and what I do with old rigs (even my TR-7 has no sequencing for keying an amp) is just use an external sequencer; key the sequencer, and it keys the amp first, waits 30mS and then keys the transceiver...and performs the operations in reverse (unkey transceiver, then unkey amp) when going back to RX.

    That does not allow for QSK (full break-in) operation but I don't use full QSK on CW anyway, except when operating barefoot with rigs designed for that like my Ten Tec stuff. A sequencer can be built for $30 or so like this, and are also available commercially built and packaged for quite a bit more, like this.

    Chip Angle N6CA used to sell one for $99 years ago, intended for sequencing masthead preamps and stuff for weak-signal VHF/UHF work and I still have one of his -- it works fine after 35 years.:)
  9. SM0AOM

    SM0AOM Ham Member QRZ Page

    The presence of G0HZU on this forum is very refreshing, as he is one of the few posters presenting real engineering knowledge posting here.

    If he ceased posting, the loss for the community could be compared to G3TXQ or VK1OD.

    N2EY likes this.
  10. G0GSR

    G0GSR Ham Member QRZ Page

    Yes, we had those on the Marconi 250kW transmitters (BD272)

    The story goes that the Marconi factory were having problems with parasitics which they couldn't solve.
    A Marconi engineer put his head in the final RF cubicle whilst on power and the parasitic cleared.
    They then made an" artificial head" using metal plates and resistors to earth each side of the final RF circuit.

    The resistors were shorted on most bands but had to be in circuit on 13MHz IIRC.
    I don't know how true that is but that's how it was told to me and its a good story.

    Those TX's also had resistors from the top to the bottom of the copper, final valve boilers. DC shorted but were apparently necessary.

    Last edited: Jun 17, 2019

Share This Page