Economy Half Kilowatt HB Project

[W9ZSL]

Yo, is it spring yet?  

I'm still working on the AM rig...modulator first.  I was thinking about the 4-65A for a final when I ran across a GE Ham News publication from 1948 on eBag.  It describes a two-tube rig designed for CW, 2 band 80/40 using a 4-125A.  The first tube is actually a VFO that drives the 4-125A.  It is so simple it hurts!

So I bought the pamphlet.  First thing I noticed was the VFO tube is a 1614.  I'm totally unfamiliar with that tube.  The final still interested me but I would substitute a Pi net for the plate tank in the diagram.  Another thing that put me off was the odd-ball Millen 74001 slug-tuned plug in assembly in the oscillator plate circuit.  Where would I find one of THOSE?

Well, I found one on eBag this aft...brand new!  What are the chances?  I'm being enticed.  Since I'm only interested in working 75 either AM or CW, seems to me this rig just might do the job.  The construction instructions speak of eliminating some plates on the primary variable of the VFO and considering this design is for CW, low end of the band, I don't think I have to worry about the variable.

So my question: is there a good substitute for a 1614 and what should I watch out for if I substitute a Pi net tank?

I will build my dream AM rig before I die.  Of course, my eye sight is going bad so it may have a lot of cold solder joints.  I shouldn't say joint.  I may relapse! LOL!

Mike - W9ZSL

[WA7BBI]

The 1614 is the industrial version of the 6L6. Exact same pin out too. Looks like a fun project.

73,  Jim, WA7BBI

[K9DXL]

N9TEW hamtubes has this listed for $18.00

I'm not connected with this company in any way, but have ordered from him and got very good service.  Probably cheaper and more reliable than Ebay.

[K9DXL]

P.S.  If you decide to order from Bob, the URL is http://www.hamtubes.com/tubes/list5.htm

Or you can build THIS lil' beauty: http://n4trb.com/AmateurRadio/GE_HamNews/issues/GE%20Ham%20News%20Vol%2002%20No%201.pdf

[w3jn]

mmmmmm I dunno, this sounds like a recipe for disappointment.

A high power, self-excited oscillator is generally a really bad idea.   There's way too much power circulating in the tank for good stability; as the tank circuit heats this thing is gonna drift something awful.  Without a buffer between the oscillator and the final, any load change on the final is going to be reflected somewhat back to the oscillator and THAT will cause drift, FMing, and yooping.

I'd use a 12AU7 or a 6J6 as the oscillator, using one section as the osc and the second section as a buffer, then couple that to the 1614 via a resonant circuit.

I'd build it as is, see where its weaknesses are, but leave plenty of room to punch extra holes in the chassis cause I think you're gonna need it.

[WQ9E]

I second JN's thoughts   As designed that would be an interesting demonstration rig for a vintage contest but it is far from a daily driver.  Under modulation it is going to exhibit an interesting mix of modulation types.  Under modulation the classic MOPA rigs with no buffering/isolation stages tend to not do so well.

As JN suggested couple a good VFO to the 1614/6L6 and you could use any of the decent external VFO units or use a Meissner Signal Shifter (they really don't drift that badly, I have several versions of it from early to the turret EX model) to directly drive the final so you won't need a 1614.

[W9ZSL]

I was pretty much looking for opinions.  I'd rather go with a 6AG7 that would work with either a crystal or my Heathkit VFO followed by a 6L6 driver.  Considering the GE bulletin states specifically how to build the thing in order to assure stability, adding a Pi Net would muck it up no doubt.

[K1JJ]

I like the article's comment that it's a great rig for passing traffic...  

Imagine that in a current SDR ad.  How times have changed.


Yes, JN is right. Leave a lot of room for adding on. Or better yet, go with a conventional tube rig design and you will save time, money  and effort in the end.

T

[Chuck...K1KW]

That actually is a decent and simple circuit that will work well.  It's an ECO or electron coupled oscillator.  The screen of the 1614 is the "plate" of the triode based variable oscillator and the 1614 plate is the output tank, isolated by the RF grounded screen.  This provides isolation from the oscillator section and is almost as good as two separate tubes.  The final is straight through and you can use a pi net output tank with no difficulties whatsoever.

Build it as is, a 5881 or any variety of 6L6 will work in place of the 1614.

BTW, you could use any parallel resonant tank components on the output side of the 1614, there is no magic in using the exact specified components. These types of circuits are extremely tolerant of component changes so long as the values are close.  Just keep the resonant circuits shielded from each other since everything is at the same frequency.  

Chuck K1KW

[AB2EZ]

Mike
et al.

I recently built a 2-tube, plate modulated AM transmitter that is a scaled down version of the transmitter that is the subject of this thread. It works very well. The frequency stability is excellent (no apparent phase modulation, no significant frequency drift after key down), and the modulation (-95% to over 125%)  is very linear.

Along the way, I learned a few things that were important.

I'm using a 6AG7 (octal version of a 6CL6) as the oscillator, in a standard electron coupled configuration. I'm using 40m crystals. You will probably be using a VFO configuration. In any event, as a crystal oscillator, the 6AG7 works much better than a 6L6. (This has been documented long ago). Using a 6L6 or a 6V6 (as in the Ameco AC-1) to make the crystal oscillator results in too much crystal current... and causes the crystal to overheat and drift.

The amplifier tube in my transmitter is a 6550, running in class C, with a modulated B+ that varies between around +15V (limited by a negative peak limiter) and +700V. The carrier B+ is 300V. The modulated screen voltage is obtained by a 7500 ohm + 7500 ohm resistive voltage divider connected between the modulated B+ and ground (with the screen connected to the center point). This wastes some power compared to using just a dropping resistor, but it improves the tracking between the modulated B+ and the modulated screen voltage.

I am using a tuned circuit between the oscillator output and the amplifier input. This was necessary because... on 40 meters... the parallel combination of the oscillator's output capacitance and the amplifier's input capacitance produced too low a load impedance on the oscillator to obtain the desired swing in the amplifier's grid voltage.

Important note:

At first, I saw a lot of instability in the amplifier stage... apparently caused by capacitive coupling from the plate to the grid of the 6550. This was greatly exacerbated by the resonant tuned circuit at the input of the amplifier tube. I added a 33pF capacitor, in series, between the tuned circuit and the grid of the 6550. Using a 25-50 pF capacitor, to couple from each stage to the next, is a pretty standard thing to do in a vintage-style tube transmitter. Doing so made the problem go away. Since the 6550 is being plate modulated, I checked to see if there was any vestige of the modulation on the RF envelope of the voltage on the grid of the 6550. There was essentially none.

The physical mechanism that causes the extra series capacitor to cure the plate-to-grid feedback problem is not entirely clear to me. I analyzed a few simple, candidate explanations... but they don't explain the effect.

I think the effect is somehow associated with the non-linear impedance looking into the grid of the 6550 (grid current flowing when the grid voltage exceeds the cathode voltage, and also the voltage-dependent capacitance between the grid and the cathode).

The presence of the 33pF series capacitor means that some of the RF voltage across the input tuned circuit is dropped across the impedance of the capacitor... but there is still enough voltage swing on the grid of the 6550 to drive the tube into class C.

This scaled down version of the transmitter under discussion in this thread produces 8 watts of carrier. The B+ at carrier is +300V; and the plate current, at carrier, is 50mA. The voltage divider for the screen of the 6550 draws an additional 20mA from the modulated B+ supply at carrier... so, at this power level, using a resistive voltage divider to obtain the modulated screen voltage is not very energy efficient.

Stu


 

[W9ZSL]

As far as an exciter goes, I have a Kenwood TS-440 with AT, a DX-20 and a DX-60.  Someone mentioned that using the DX-60 would work in the AM mode because unmodulated the output is quite low.  On the other hand, the Kenwood's power is adjustable.  Of course, I wouldn't use this circuit.  My first choice has always been the "all purpose" amp described in the 1962 Orr Radio Handbook.  It uses a 4-250A or 400A but since I'm going to settle for 300 to 400 watts, a 4-125A would be just fine.

[K1JJ]

For RF simplicity, take a look at this hi-fi, one-tube (RF) series modulated self excited VXO rig.

One xtal can be VXO pulled about +- 3 KHz before modulation causes it to warble. It is very stable when within 1 KHz.  Big audio peaks, being a series modulated rig with no transformers.

I still run it today, occasionally.   The  final runs class C, but the modulator is class A.  Could be scaled up and class B plate modulated, I suppose.

http://www.amwindow.org/tech/htm/series.htm

T

[W9ZSL]

I'm also seriously considering using a single 813.  The '57 ARRL Handbook has one shown on page 207.  All I would need in addition to the tube and socket would be the filament iron.

[WU2D]

I agree that less tank current is a good idea for stability. With a ceramic form, you can do fine at higher power. The ARC-5 proved that. The PTO ECO in .the ART 13 is an 837, a beefy pentode with 500V on the plate. It is a better VFO than many of the time period and it must be running 5 Watts.