ad: M2Ant-1

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-MFJ
ad: Left-2
ad: Subscribe
ad: MessiPaoloni-1
  1. KM1H

    KM1H Ham Member QRZ Page

    With zero on his QRZ page to give credence to his qualifications I wouldnt put him even close to G3TXQ (SK) or VK1OD who hasnt posted on here in years after getting into a couple of engineering related discussions and was questioned about his understandings.

    HZU seems to have a problem with others also when it comes to parasitic suppressors and appears to enjoy confrontations....do a simple search.
     
  2. G0HZU

    G0HZU QRZ Member

    Hey, I'm just a bit baffled by the lack of understanding (wrt parasitic suppressors) by the established experts who have been playing with amplifiers like this for maybe 50 years or so. I don't understand why you can't all agree on the subject.

    For example, If I single you out, I'm baffled why you think a typical 85nH suppressor coil is designed to be 'self resonant' at the oscillation frequency. Can you demonstrate why you think this is? My only guess is that you have probed around the complete system with/without the suppressor with a GDO and made this conclusion. I'd expect the coil (on its own) to be self resonant with various modes but all of these are at a much higher frequency. Adding a 47R carbon resistor inside the coil will introduce some extra capacitance but not that much...

    Probably the crudest way to describe the stability issue is to look at the single tube amplifier case first and model the tube as a 1 port device where the 1 port is the anode. This reduces the tube to a small signal model of a black box with just one port connection. Because of the stray capacitances shown on the datasheet and because of the inductance in the grid assembly it should be possible to use this data along with the tube transconductance to demonstrate that the tube will generate negative resistance at this port at the anode. I think this negative resistance will be generated across a wide frequency range extending up into VHF. I think this will in the order of about -10 to -20 ohms up at VHF and this will be in series with the net tube capacitance of maybe 5pF in the case of the 3-500Z tube. But that is just based on an initial crude datasheet based model (this could be refined after some real world tests on some real hardware).

    To prevent instability, the net series resistance of the parallel LR network in the suppressor (and the tank) needs to be greater than the net negative resistance in the tube at frequencies where instability happens. This will usually be at a frequency where the net inductance of the entire tank circuit resonates with the capacitance inside the tube. This entire (equivalent) series LCR network that includes the series -R and C in the tube itself needs to have a net positive R in order to be stable.

    Having 47R in parallel with 85nH will typically deliver a net series resistance of maybe 25 ohms up at about 100MHz for example. This is greater than the negative resistance in the tube that the model predicts.

    Adding the grid caps to float the grids may be done for various reasons but it will also serve to prevent negative resistance at lower frequencies. This capacitive reactance can also help to offset the inductance in the grid across quite a wide frequency range up at VHF and this can (in theory) help to minimise the negative resistance across the frequency range where the instability can happen. But it won't be enough to make the negative resistance go away in this range so the suppressor is still needed.
     
    Last edited: Jun 17, 2019
    G0GSR likes this.
  3. K2XT

    K2XT Premium Subscriber QRZ Page

    What I've come to understand is if an amp used to be stable and as it ages it isn't any more you don't need to redesign the parasitic suppressors. You might need to replace old carbon comp resistors that have drifted high in value with new ones like Carl has suggested and has experience with. That should get it going again. Simple. Thanks, Carl, that one is put to bed.

    (I was typing this as the previous post was being sent. Hopefully we don't disagree)
     
    KM1H likes this.
  4. G0HZU

    G0HZU QRZ Member

    Yes, I'd expect that a lot of the stability problems will be caused by ageing of components. I think this will include the components that are used to float the grids as well as the ageing of the resistors in the suppressors. So unless the amp is occasionally serviced with this in mind I'd expect it to start becoming unstable and unreliable with time...
     
    K2XT likes this.
  5. G0HZU

    G0HZU QRZ Member

    When it comes to 'directly grounding the grid' (which is the main thread topic) I can only offer my initial thoughts on why this may introduce some vulnerability to the amplifier in terms of instability. I haven't looked into this in any great depth but a basic datasheet model of the tube will show that hard grounding the grid will extend the range of frequencies where negative resistance occurs and this will bring it down through the HF bands. This means that if you actually wanted to 'explore' the theoretical range of frequencies where instability can happen then it could become quite alarming how low in frequency you could go if you had hard grounded the grid with copper straps. In reality, you might not be able to present the anode with enough net inductance to see instability this low in frequency but the suppressors won't offer much help once you get down towards HF.

    Note: When I first modelled this tube a few years ago I included all the external components in the tank circuit including the anode choke and its DC blocking cap. This model threw up the scary possibility of squegging at VHF. If this happens the model predicts very high RF voltages can be generated as the VHF oscillation can be gated on and off at a lower frequency. I'd expect this mode of instability to be highly destructive as the summed voltage of the squegged LF and VHF waveform at the tube anode can be very high indeed. I'd expect something to break down inside the tube if this ever happens but I have no idea how much the tube can tolerate here.
     
  6. WB2WIK

    WB2WIK Platinum Subscriber Platinum Subscriber QRZ Page

    https://www.google.com/search?safe=...iAhUNsZ4KHaM6BsUQ7Al6BAgAEA0&biw=1536&bih=871
    https://community.keysight.com/people/G0HZU
     
  7. KM1H

    KM1H Ham Member QRZ Page

  8. KM1H

    KM1H Ham Member QRZ Page

  9. G0HZU

    G0HZU QRZ Member

    This is getting silly. You can't judge (dismiss?) someone because they haven't described their background on various public forum pages. Besides, this topic isn't about me, it's a discussion about tube stability. I'm trying to contribute to it.

    It's getting late here in the UK but maybe tomorrow I'll try and post up some initial modelling techniques that can be used to predict negative resistance.
    The easiest place to start is with something simple like a JFET in grounded gate where the gate isn't grounded directly because of stray inductance. These devices have similar capacitances and transconductance to an RF tube so this makes a reasonable starting point. It should be possible to use the datasheet to predict the negative resistance and also where it will occur in the VHF band. A suitable suppressor can then be designed and evaluated on an RF simulator.
     
    Last edited: Jun 18, 2019
    N2EY likes this.
  10. HAMHOCK75

    HAMHOCK75 QRZ Member

    This is true. I designed solid state amplifiers for years but it is all changing. The type of broadband RF power amplifier exemplified in many ham transceivers which remind me of audio amplifiers with their transformer coupling is giving way to totally solid state, DC coupled designs. Once that is achieved, the power amplifier can be assembled like the Sanken DC coupled audio amplifiers of old. A ceramic package with big flat leads to carry the heat away. I can see that that one day the garden variety 100 watt output amplifier will just be single component.

    Not sure if I feel any sadness about it though. Fewer parts, fewer interconnects tends to improve reliability and skills lost in one area result in new skills in other areas.
     

Share This Page

ad: chuckmartin