Simulating tube amplifiers

Discussion in 'Amateur Radio Amplifiers' started by W8JI, Dec 29, 2011.

Thread Status:
Not open for further replies.
ad: L-HROutlet
ad: l-rl
ad: Left-2
ad: L-MFJ
ad: MessiPaoloni-1
ad: Subscribe
ad: K5AB-Elect-1
  1. G0HZU

    G0HZU QRZ Member

    Hi Ian
    The big choke is a potential issue but I would expect that Heathkit will have carefully chosen this inductor design such that it behaves as a very small and very lossy capacitor above 500kHz.

    For sure, if you swapped it for something that could regularly lap the outside of the smith chart and become a highish Q inductor in places then there would be a stability issue at HF and VHF.

    However, I would suggest you try putting one on a VNA and seeing what it shows from 1MHz to 200MHz. I would hope (and expect) it to reside over at the right hand side of the chart showing lots of tiny (insignificant) loops where it changes its ESR and capacitance according to frequency.
    eg it will probably look like 3pF in series with several hundred ohms resistance. The resistance will vary a LOT across HF to VHF but I would expect the response on the smith chart to show a lossy capacitor with a few pF.

    So no real threat to HF or VHF stability :)

    If it doesn't behave like this then it could be swapped for an alternative choke but I would expect it to look pretty good on a VNA.

    Regards
    Jeremy
     
  2. G0HZU

    G0HZU QRZ Member

    I just downloaded the SB-220 manual and these chokes are RFC4 and RFC5.

    There are images of these chokes and they each appear to be 4 stacked chokes in series. I would
    expect they did this to reduce the net series capacitance up at RF once beyond the parallel resonant
    frequency of each section.

    So this choke could look like 1 or 2pF in series with a resistance that varies with frequency across HF
    to VHF. There will obviously be various resonance modes within the choke but I would expect these to
    appear as tiny loops on the smith chart that effectively vary the ESR of the 2pF capacitor.

    So I would suggest Heathkit chose this 1mH inductor design with some care. I can't see how this
    design of choke could cross over to the other side of the smith chart and present an inductive
    reactance with a lowish ESR that could cause instability at HF or up to 200MHz by resonating with leaded caps of 200pF x3.

    Has anyone put one on a VNA?

    Note that I should have said it will behave as a small cap above several MHz rather than 500kHz as the parallel resonance of each section will probably be a few MHz.
     
    Last edited: Jan 13, 2012
  3. W1QJ

    W1QJ Ham Member QRZ Page

    To further complicate the issue Rich AG6K removes the 1MH choke that is 26 ohms resistive and puts a 22 ohm 1/2 watt resistor in its place. Now what happens?
     
  4. G0HZU

    G0HZU QRZ Member

    Funnily enough I was going to suggest the possiblity of replacing the inductor with a big 4.7R resistor that has low inductance and keeping the 3x 220pF caps. Go much lower in resistance than this and instability returns (on the model) with a vengeance at low VHF. Eg 1R would not be a wise choice.

    However, I was a bit nervous of suggesting swapping out for a resistor in case there's a mode of instability that could thermally damage the resistor and then go on to do even more damage elsewhere.

    22R and 1/2 watt rating sounds a bit vulnerable to me if it hooted at the wrong frequency. However, it does have the caps across it to protect it but I'd have to think about that one for a while :)
     
    Last edited: Jan 13, 2012
  5. G0HZU

    G0HZU QRZ Member

    I think a lot will depend on the stray inductance associated with the 22R install. My first thought would be that it would be difficult to get the strays low enough to keep the 22R 1/2W resistor undamaged if there was a hoot at a frequency where the 22R + stray lead inductance set up a resonance (at the hoot frequency) with the 200pF cap(s) and assoc lead inductance.

    It must also get quite warm under normal operation on the lower bands? Does it ever show signs of overheating?
     
    Last edited: Jan 13, 2012
  6. W1QJ

    W1QJ Ham Member QRZ Page

    Yes, many times they get cooked because the wattage is low. It has to pass the grid current. I someone overdrives the amp it will have to pass > 150ma They act as a fuse should the tube arc or grid to filament short.
     
  7. GM3SEK

    GM3SEK Ham Member QRZ Page

    That is a standard technique for winding RF chokes, almost 100 years old, but it can only be partially successful in removing additional parallel and series resonances at higher frequencies.

    Yes - that is how I know that LF-style RF chokes are complex and unpredictable at VHF. Their properties can only be measured, which wasn't possible in the 1960s when the SB-220 was designed. Worse still, in those days they didn't even have the tools to think about the problem in the ways we do today.

    Back then, the best available design technique was to apply basic principles, and then project them forward into the unknown. They had no way to measure or predict the effects of that 1mH RF choke on the VHF stability of the SB-220, so that part of the design was strictly cut-and-try. Certainly it was done "with some care", but not in the same way that we'd do that today.



    73 from Ian GM3SEK
     
  8. G0HZU

    G0HZU QRZ Member

    I'm trying to take you seriously Ian but I can't...

    In 1944 there were soldiers in WW2 in Normandy using wireless sets operating above 4GHz. This was about 30 years before stuff like the SB-220 was introduced. To imply people couldn't easily analyse the impedance of a choke back in around 1970 is silly.

    Also, modern stability analysis techniques date back to the late 1950s. Try googling John Rollett and K Factor.

    If you like I'll show you how I would 'turn back time' to the 1950s (way before the SB-220 existed and I was born) and how I would initially attempt to assess that choke using homemade apparatus i.e. a homemade directional coupler(s) and an old Tek 585 dual beam scope and a signal generator.
     
    Last edited: Jan 14, 2012
  9. AG6K

    AG6K QRZ Member QRZ Page

     It's a 30Ω, 0.25w carbon resistor and its easier to blow than a grid RFC made with #28awg Cu wire. I've heard of no one having a 3-500Z short who used a frangible resistor to fuse the grid.
    • Rich, ag6k
     
  10. W1BR

    W1BR Ham Member QRZ Page

    I vaguely recall an April article in HR where the author suggested using three 2-watt 100-ohm resistors for each grid. Now we are
    down to 0.25 watts instead of six watts total?
     
Thread Status:
Not open for further replies.

Share This Page

ad: elecraft