Print

[VE7RF]

A buddy is trying to get a 1:9 un-un to work on his grid driven tetrode. Terminated in a  450 ohm globar,  SWR is  1:1   but only up to aprx  20 mhz.
Above  20 mhz,  it rapidly all goes to hell....reaching  2.5:1   at 29.00  mhz.   ( this on the test bench)

He needs the  1:9  un-un to work from  80-10m.   What's the trick here ?   What  is the ideal material to wind the trifilar onto ?

Once eventually installed into the amp, a small  coil is wired in parallel  with the 450 ohm globar to  cancel out the XC from the tube,  but that is a normal procedure,
and will not alleviate the high swr issue on the upper bands.

Jim   VE7RF

[K1KP]

Even a few inches of wire has significant inductance at 29 MHz. The 'globar' sounds like a large resistor. Try it with a 450 ohm 1/8W resistor at low levels to see if its any better... Look at the impedance with a VNA and see if it's inductive or capacitive; if it's inductive, try padding with a small value capacitor.

Off the top of my head, without knowing more details about the unun, I'd guess mix 31 or 43.

-Tony, K1KP

[KM1H]

That should be right at the socket with wide copper strap at both ends and then as Tony said if needed.

[W9IQ]

Use mix 43 ferrite. Use a compensating capacitor on the primary for the upper frequencies.

- Glenn W9IQ

[VE7RF]

it is wide cu strap to the 450 ohm globar.  The normal deal with grid driven tetrodes is the tube C is  XC..and esp on the upper HF bands..which ends up in parallel with the globar.  Normal deal is to use a small coil in parallel with the globar, to cancel out the tubes  XC.   I see on the alpha amps that are grid driven, they relay switch in the coil, but only on the upper HF bands.  The grid driven alpha amps use a  1:4 un-un and a 200 ohm non inductive resistor.

In this case, we have lousy swr... before the tube  is even in the picture.   The  450 ohm globar has to handle a fair whack of power for this application, so 1/8 watt resistor is a moot point.

I suspect the wrong type of core was used, so it would be better to optimize that....1st.   He has some  type  67 coming next  week.  He will try that, and also some type  61....and also type 43.   The trifilar winding consist of teflon coated wire.

[VE7RF]

Use mix 43 ferrite. Use a compensating capacitor on the primary for the upper frequencies.

- Glenn W9IQ

Tnx, we will try that. The compensation cap might have to be switched out on the lower freqs, and maybe not. No big deal, as one end could be opened off with a relay.

Ideally we want the swr flat up to 10m band.  Then install into lower RF deck, connected to the tetrode.   Then compensate for the  tubes XC  on upper HF bands.... with the usual small coil in parallel with the 450 ohm  globar. 

I have not been involved with tetrodes in passive grid in a while.  Have also seen where a small 0- 4.4 uh roller is wired across the 450 ohm globar... to cancel out the Xc of the tube.

We also have to consider the C from the neutralizing cap.....which is a special gas filled type, designed to handle the B+  involved.

[LZ1HD]

Making 1: 9 un-un is not easy. The three wires must be spaced apart. The dg0sa.de website describes in detail how such un-un can be done. In short, if you use a wire with a diameter of 1 mm. the three wires must be wound on the toroid by 5 mm. distance from each other.
Regarding the ferrite material that should be used, a choice should be made between Fair-Rite Mix 61 and 52, but the first is preferred.
Impedance raising is usually required for old tetrodes, which have low gain and 70 V Eff from the transceiver are not enough for normal driving.  With this 1:9 un-un maximum 210 V will be available.

Veselin LZ1HD

[WB6BYU]

One of the issues as you go up in frequency is that windings
become a more significant portion of a wavelength, so the
characteristic impedance of the transmission lines used
is more important.

That's one advantage of the higher permeability types of
ferrite - they require fewer turns to cover the lowest bands,
hence shorter wires for the same core size.

Ideally the transformer would be trifiliar wound, with the
wire spacing adjusted for the optimum characteristic
impedance.  A simple scramble-wound transformer doesn't
always work as expected, as the coupling between turns
isn't as good as many analyses assume.

[N8CBX]

...a small  coil is wired in parallel  with the 450 ohm globar to  cancel out the XC from the tube, 

That is a confusing way of saying it. You're describing LC compensation, shunted by a 450 ohm resistor. That's the easiest method to compensate the tube's input capacitance is with a parallel inductance. This, of course, cannot result in broadband matching, which is what your buddy is seeing.
A 1:9 transformer ("un-un") is increasing the RF voltage by 9 times to the grid....Wow, typically a 1:4 transformer is all that's needed for the larger tetrodes.
Jan N8CBX

[W1VT]

I believe Jerry Sevick found that low impedances worked best with the typical trifilar 1:9 tranformer made out of #14 wire placed side by side.  Going from 5.55 to 50 ohms instead of 50 to 450 ohms.  It may be useful to see what impedances work best with the transformer you have created.

[N8CBX]

My 1:4 step-up transformer (50ohm to 200 ohm) is made from 20ga. mag wire; Evenly spaced, 14Turns winded parallel (not twisted) on a FT125-43 core. The return loss was about -30 at about 36mhz as tested on the bench.
Jan N8CBX

[N8CBX]

If one wants to use a tuned grid circuit using parallel inductance, one can make a switchable input to switch to two or three different parallel inductances. And could result in a good VSWR from 1.8 to 30 MHz.
I made a switchable input circuit in my big tetrode amplifier.
Jan N8CBX

[LZ1HD]

In this case, another approach can be used - wideband matching. For example LPF Chebyshev 5th order. The capacitor in the middle will be grid capacity. However, for 450:450 ohm matching, it should be approximately 22pF. Depending on the tube, when the grid capacity is greater, another lower impedance can be selected for which the transformer is feasible.

Veselin LZ1HD

[G3RZP]

Using a low pass filter as LZ1HD suggests is very practical at the 200 ohm level, and I first did that some 35 years ago with a pair of 4-250As. It is just  variation on the wideband distributed amplifier...

[VE7RF]

Making 1: 9 un-un is not easy. The three wires must be spaced apart. The dg0sa.de website describes in detail how such un-un can be done. In short, if you use a wire with a diameter of 1 mm. the three wires must be wound on the toroid by 5 mm. distance from each other.
Regarding the ferrite material that should be used, a choice should be made between Fair-Rite Mix 61 and 52, but the first is preferred.
Impedance raising is usually required for old tetrodes, which have low gain and 70 V Eff from the transceiver are not enough for normal driving.  With this 1:9 un-un maximum 210 V will be available.

Veselin LZ1HD

I'm having one helluva time trying to translate the DG0SA   website... into english. ( google sez it will translate to english..then doesn't).    By looking at his info in German,  I believe  he has nailed it.   His worse case  return loss on 29 mhz is  16 db. (1.35:1  swr).     His design works from  160-10m.

The tube is an  8281 tetrode for a commercial application.  There is no way that a  50 ohm  globar would work...nor a  200 ohm globar with a 1:4 un-un.   The only thing feasible is a 1:9   un-un.

The compensating inductance is typ used on the higher bands....and one end of a coil can easily be opened off..... or a 2-3 relays used, to tap a coil.

I need an english version of his 9:1  info.   As is,  all that I can figure out is he used 1.5mm  wire.   I can't figure out what type of core was used, or core OD,  ID, and height.