New RF Deck

[W8ACR]

Just in time for top band season, here is what I have been working on for the past week or so. This is a classic push pull RF deck using HK-54's and plug in coils. It is designed for 160 meters only. I will run the HK-54's at 1800VDC and 200mA. Maiden voyage tomorrow.

Ron W8ACR

[W8ACR]

More pics. Also, I noticed that I labelled the filament voltage as 5VDC. That is wrong, it is 5VAC as usual. All knobs are on verniers, and I managed to put this together without having to drill a single extra hole! Pretty good for me! 

[W8ACR]

more pics.

[KL7OF]

nice...look for you on 1945....Steve

[W7TFO]

Kudos, OM!

Braid on plate connections is not so hot, I'll send you some plated copper ribbon;  keeps little parasitics at bay.

73DG

[w0lpg]

Ron, I can't tell from the photo for sure but you may have the neutralizing capacitors wire to the wrong grids.  They should be cross connected.

Val  W0LPG.  Watertown, SD  on 1980 kHz early evenings with HB 813  Mod. P-P 811

[W8ACR]

Hi Dennis,

No need to send copper strap, I have some and have replaced the braid with the strap. (I could use a few extra HK-54's though) 

Val, Thanks for making me double check, but the grid leads are crossed. I'm pretty sure it is correct. Smoke test coming up later today.

Steve, sure would be good to hear you on the air again. Been a long time. If I get it working, maybe we can try to set up a SKED.

73, Ron

[W4RFM]

I love the vintage look and especially the craftsmanship in your construction.  Well done.
BOB

[WBear2GCR]

Wants and needs a window/porthole... imho.

                     _-_-bear

[W8ACR]

Initial tune up tests done, no major problems. The Rf deck neutralized easily, no evidence of parasitics. However, I can't get as much plate current as I should be able to. Max plate current at full loading is about 130mA. I should be able to get up to about 250mA. I have tried multiple plate voltages from 1500 to 2000. I have tried grid bias from -150 to -250. SWR is 1:1 via a homebrew antenna tuner.

All other parameters are right on. Filament voltage 5.0 VAC. Grid current 50mA. Bird wattmeter shows about 180 watts of carrier output with about 250 watts input.

Two thoughts that I have had.

1. soft tubes - these HK54's are probably 70 years old or more. I have four of them. I have tried all different combinations of the four tubes. All different combinations have about the same amount of output.

2. Could the L/C ratio of the final amp be affecting this? I am using a B&W TVH 160 coil. This is designed for 500W input, so I am thinking that it would typically be used with plate voltages of 2000 or more, and it may have too much inductance for a lower plate voltage/plate current ratio.

Any suggestions?

Thanks, Ron

[W7TFO]

With the bias you have, I'd say more drive within the limits of grid current.

73DG

[W8ACR]

Hi Dennis,

Max permissible grid current is 30mA per tube, 60 mA total. I have driven them up to 60mA with minimal effect on plate current compared to 50mA. 25mA per tube is the recommended amount for plate modulated class C service.

Ron

[W7TFO]

Maybe raise the bias, makes it further into class C.  Will take more drive to get the grids a bit positive over that.

Also might check the L/C ratio in the output net.   A Q of 12 is supposedly the target.

73DG

[KC2ZFA]

Ron,

I suspect the number of turns on your output link are insufficient when feeding them to 50 ohms.

[WD5JKO]

Ron,

I suspect the number of turns on your output link are insufficient when feeding them to 50 ohms.

I agree.

Saying that, a few other things will also help. Resonating the link inductance with a series capacitor will act somewhat as a loading control along with the position of the swinging link. This could be a fixed capacitor (Micamold, or RF door knob rated for RF current) once the value is known.

Another is to up the circulating Q when loaded. I ran into this issue once with a 1KW P-P circuit, and I was using the B&W TVL (?) 500 watt plate coil when on 75m. I needed to shed some turns off each side, and then needed to add fixed capacitors across each coil. As I recall, I used two Centralab mini doorknob 100pf in series (50pf 10KV) on each side. That brought up the Q, and I was able to get 400ma at 2500v with the link half way in, and the link series capacitor optimized. I used a medium size air variable dual section cap from one side of the link to ground. The cap was something like 400pf per section.

Jim
Wd5JKO

[KC2ZFA]

indeed !

In the pic I use link tuning as loading on my "millen" exciter knock-off...the cap on the right.

Also, bear in mind that the ancient scrolls say the 160TVL is to be used with 90pF total effective capacitance across the coil for resonance at the _lowest_ end of the band.

Peter

[K3YA]

I suspect that KC2FZA may be right about needing more turns on the output link.  A quick check for the tubes is to briefly tune the plate off resonance.  If the plate current goes way up, like 300mA, the tubes have plenty of emission.  In that case increase the coupling (loading) with more turns on the link or a higher Q in the tank circuit.

Nice looking RF deck, Ron

[W8ACR]

OK, I'm making progress. I put in a four turn link coil instead of three turn, and I also tapped the TVH 160 coil four turns from the end. This of course would lower the inductance of the plate tank circuit which I believe would also raise the Q of the tank circuit.

Here are the new operating conditions and results.

Eb=1700V
Grid voltage= -250
Grid current= 50mA
Plate current= 250mA
Power output= 280W

Eb=1500V
Grid V= -250
Grid current= 50mA
Plate current= 180mA
Power output= 190W

My circuit is still a bit inefficient, but these numbers are much better than before, and I am able to get much more plate current and power output. Perhaps the efficiency of the tank circuit is likely to improve if I raise the Q even higher. I'm not sure. It would be very easy to tap the coil further in.

I now have a new problem though. I get an occasional arc from the 160TVH to the link. The arc occurs at the moment that I remove RF drive by turning off the exciter output. The four turn link is a bit wider than the three turn link, and comes very close to the TVH coil as it sits in the coil gap. I assume that this is a DC arc, since the DC is still present, but perhaps it could be a parasitic??

To deal with the arc, I was thinking that I could either try the three turn coil again, or maybe wrap the four turn coil with glass tape.

Thanks to everyone for their input
Ron

[KC2ZFA]

is this arcing happening on the side of the link coil that’s grounded ?

what if you keep the old link and just tap inwards from the ends of the
tank coil ?

[WD5JKO]

Must read info on this topic from 2007:

http://amfone.net/Amforum/index.php?topic=9518.0

Also here is a PDF with a novel way to HB a new coil:
http://ad5x.com/images/Articles/CoilRevB.pdf

Another via QST written by W3JIP:
http://www.arrl.org/files/file/Technology/tis/info/pdf/9708033.pdf

[W8ACR]

OK, looks like things are finally working like they should. The arc is gone. All I had to do was gently adjust the entry angle of the link coil into the gap of the TVH coil. I also moved the tap to the sixth turn from the end. I also had to add a 75pf fixed vacuum padding cap across the coil. I now get 200W output with 300W input, so 66% efficiency seems to be acceptable. No AM activity on 160 here tonight, called CQ on 1945 a few times but no reply. The modulation envelope looks very good on the scope also. Anyway, it looks like it's working well. Thanks again for all the suggestions, it looks like my main problem at first was due to suboptimal L/C ratio. Too much L.

73, Ron W8ACR

[W8ACR]

160 RF deck is working well. I have made one contact and got a good report. I can get 200W output with 1500V on the plates, or 250W output with 1750V. Audio looks good on the scope.

I did have one additional problem that I had to work on for a while. I found that if I drove the tubes hard, my grid current would initially be good, but would then go from 60mA down to zero over the span of about 3-4 seconds. Just a slow steady drop to zero. If I kept the grid drive to 50mA, and the plate voltage to 1300 or so, it didn't seem to happen. When I turned up the plate voltage to 1500 or more, and then applied more drive to get the grid current up to 60mA, the grid current would momentarily go to 60, then just fade away. I had not seen this behavior before. I have four HK-54's, so I tried different combinations, and found one tube that seemed to be the culprit. Once I removed that tube, the problem stopped. Any ideas of what could be going on with that tube?

I'm surprised that there doesn't seem to be much 160 AM activity at night. Anyway, we'll keep at it.

73, Ron

[KL7OF]

I was on 1945 on Friday nite and condx were not good....

[W2PFY]

quote] Once I removed that tube, the problem stopped. Any ideas of what could be going on with that tube?[/quote]

Hi Ron, It would be interesting to see if that tube will operate as a stand alone tube at full output. Will it show the same effects operating as a single tube? I think it may be easier to understand what may be happening with that tube by testing it this way?