Olde Buzzard TX Project

[N8UH]

Hello everyone!

Thanks to a couple of recent acquisitions of 40's era parts, I have found my big rig plans drifting towards an earlier era. That, and the size and power consumption of the super-maul I had planned were just not going to work out for my location.

So, I've got the idea running through my head of building a 40's era CW/Phone transmitter using as many period components as possible. I'm gravitating towards the designs in the 1940 Orr book and the 1941 ARRL handbook for 400W-1KW class rigs. Push-pull finals, plug-in coils, link-coupled - the real deal. I'd like to keep everything within the engineering practices and component capabilities of the time, making concessions only where I must. Having it look correct is important also! So, this leads me to a couple of questions:

1. PA and Mod tubes: I was thinking of going with 2-810s X 2-810s. They seem somewhat plentiful, affordable, reliable and are correct for the time. Are these a good choice, or are there better tubes for the job?

2. Speech amp section: The common practice of the time seemed to favor "communications grade" audio. What would be a good way of improving audio fidelity while staying within the engineering practices of the period?

3. Modulator: This is one of those areas where I'm going to have to make a concession or two, due to the fact that ancient mod iron of the size needed is incredibly scarce. Here, I'll be using my 5KW broadcast iron.  I have a transformer and reactor, so I could go with a modified Heising setup. Should I go modified Heising, or go with just a transformer, as the schematics show?

4. Power supplies: I have 2 plate transformers that are each large enough to power both the modulator and PA. Should I go with a common supply or split supplies? I would think split supplies would end up being better, but what are the opinions here?

5. Oscillator: There seems to be little mention of VFOs in those days, even though the technology was there. What gives? I'd rather not have to start buying up crystals...

So anyway, that should give everyone enough to chew on for now 

What are your thoughts? I imagine there will be some lively discussion on this one...

Also, I'll be posting some for sale/trade and wanted items soon in the appropriate sections. Thanks guys!

[Opcom]

I like separate variable power supplies because it lets you electronically 'change the ratio; of the mod xfmr a little..

If you are running 810's, you should be able to get away with DC in the secondary of a 5KW transformer.

If you make the speech amp oversize, the fidelity will be much better at the necessary level.
The same goes for the speech amp iron. If the whole thing is 4-5X more than needed, should be very good.

I am using a 200W public address amp to drive a pair of 3-500Z modulators. At the 30W or so the p.a. amp has to run, it is perfect in about every way.

[KM1H]

1. PA and Mod tubes: I was thinking of going with 2-810s X 2-810s. They seem somewhat plentiful, affordable, reliable and are correct for the time. Are these a good choice, or are there better tubes for the job?

They will work but 100TH/250TH glow better under a nice load.

3. Modulator: This is one of those areas where I'm going to have to make a concession or two, due to the fact that ancient mod iron of the size needed is incredibly scarce. Here, I'll be using my 5KW broadcast iron.   I have a transformer and reactor, so I could go with a modified Heising setup. Should I go modified Heising, or go with just a transformer, as the schematics show?

Can you back into that BC iron with similar specs? I had the same problem trying to find a decent mod xfmr and wound up with a giveaway 7500W rated one that holds down that part of the basement. I picked up iron from a Gates BC-1T a few months ago but havent retrofitted it yet.

5. Oscillator: There seems to be little mention of VFOs in those days, even though the technology was there. What gives? I'd rather not have to start buying up crystals...

Crystals are for the birds. I use a 1939 Meissner VFO loaded down a bit to drive the 1940 HT-9 which drives 250TH's with 304TL modulators.  Using modern circuitry and some TC caps you can build or retrofit a stable VFO using most any tube.

Carl

[KC2ZFA]

Here's a VFO/exciter from the 1942 "RCA Guide for Transmitting Tubes" page 52:

[KE6DF]

813's came out near the end of the 1930's -- so that's another option.

803's a bit earlier. Sometimes you hear comments that 803's are hard to find, but there are almost always a few for sale on ePay. Right now there are four.

I like the idea of targeting a particular time frame for a HB rig.

When I get around to building a high power rig, I was planning to base it on technology from 1949.

That was the year I was born, but also I think tube AM transmitters hit some kind of high point right about 1948-1950 for a few reasons:

1.) You had the innovations from WWII in tube technology to build on.

2.) The bean counters hadn't seized control yet so you see BC TX's with 845s or 2A3's driving 833 or 810 modulators through high quality driver transformers. Not as many corners cut as later in the 1950's and 1960's.

3.) The R&D dollars were still focused on tubes. By the late 1950's, the smartest and best designers and research dollars moved to semiconductors.

[k4kyv]

1. PA and Mod tubes: I was thinking of going with 2-810s X 2-810s. They seem somewhat plentiful, affordable, reliable and are correct for the time. Are these a good choice, or are there better tubes for the job?

Those would be fine, but if you can score a pair of 8000s for  the final, they take about 70% the rf driving power as 810s in class C plate modulated service.  It's the same tube, except for a slightly different grid structure giving it a lower mu. I use a pair of Amperex HF-300s in mine, which are electrically similar to 8000s, kind of like an 8000 on steroids, but they have become extremely hard to find in usable  condition.

2. Speech amp section: The common practice of the time seemed to favor "communications grade" audio. What would be a good way of improving audio fidelity while staying within the engineering practices of the period?

Use the best quality audio transformers you can find, such as transformers salvaged from junked broadcast equipment, e.g. UTC LS series or Thordarson Tru-Fidelity series. Look in an ARRL handbook of that period; they have a chart for R-C coupled audio amplifiers, which I believe is normalised for something like 100-5000~ frequency response, but I am quoting from memory; you'll have to check for the book for the exact figures. In any case, adjust the values of coupling capacitors proportionally; if the chart is for the above frequency response, doubling the values of the coupling capacitors will  halve the low end of the response, so you would multiply it by 3 to achieve roughly  flat response down to 30~.

3. Modulator: This is one of those areas where I'm going to have to make a concession or two, due to the fact that ancient mod iron of the size needed is incredibly scarce. Here, I'll be using my 5KW broadcast iron.  I have a transformer and reactor, so I could go with a modified Heising setup. Should I go modified Heising, or go with just a transformer, as the schematics show?

That depends on the transformer. Most Hammy Hambone mod transformers are designed to  run the PA plate current through the  secondary. But most broadcast quality transformers are designed to use a reactor.  You  can  get some improvement in a Hammy Hambone  transformer by using a reactor with it, but OTOH, running the PA plate current through BC iron will most likely degrade its performance.  Of course, 5 KW broadcast iron might work OK at <1 KW power levels without the reactor.

4. Power supplies: I have 2 plate transformers that are each large enough to power both the modulator and PA. Should I go with a common supply or split supplies? I would think split supplies would end up being better, but what are the opinions here?

The common power supply gives you better voltage regulation for the modulator stage, since the PA serves as a very low resistance bleeder to hold the voltage steady. But, the split power supply allows you to adjust PA plate voltage independently of the modulator plate voltage and vice-versa. If the mod transformer has too much step-down to allow the audio head-room you want, you can increase the modulator plate voltage while leaving the PA plate voltage the same, and vice-versa.

5. Oscillator: There seems to be little mention of VFOs in those days, even though the technology was there. What gives? I'd rather not have to start buying up crystals...

To really stay period, you need the old style xtals in a round holder about the size of a half-dollar that plug into pins 2 & 4 of a 5-pin tube socket. They were manufactured in quantity by the Bliley and Valpey crystal companies. These xtals still show up at hamfests. But don't be surprised if the actual crystal frequency is off several hundred cycles, up to a  kc/s or so from what is marked on the holder.  They tend to drift off frequency with age. To really get into the "period" experience, try grinding a lower frequency xtal up to the exact frequency you want to operate.  Instructions can be found in the older handbooks. VFOs were used in that era, but they were rare. Receivers like the National HRO had stable HF oscillators, so there is no reason why astable transmitter VFO couldn't be built to that period with similar oscillator circuitry.

You don't have to come up with your ultimate design at your first attempt. Build the rig with whatever parts are available, get it working, and then, as you attend hamfests or peruse on-line sales, look for higher quality and more authentically period components.  Whenever you score a promising component, stick it in the transmitter to replace the lower quality or non-period component.  If it works OK, wire it in permanently.

My HF-300 transmitter was initially built with a lot of Hammy Hambone components, which have been gradually replaced one at a time with top-of-the-line commercial grade stuff of the late 30s era. I originally built it in the early 70s largely with pre-WW2 parts, but it is still a work-in-progress. In its present form, very few hams could have afforded to build an exact copy of what I have in the era when those components were in their heyday, especially considering that this was during the Great Depression. But I built mine at a tiny fraction of the cost, mostly using stuff from estate sales and flea markets, stuff that somebody else didn't want and had discarded as junk.

[N8UH]

Thanks for the excellent replies!

'ZFA:

That VFO looks great. I might give that a go. I think I'd probably need to build a PA stage to give the 810's enough drive. Quite doable though. Nice, elegant circuit!

Carl:
Yeah, 100THs or 250THs would be my first choice, but they seem to be getting a king's ransom for those things. Must be the audiophools buying them up because they look cool. I've also heard that it's pretty difficult to find ones that haven't gone gassy?

Don:
Excellent info, as usual! I looked into 8000s, but they are quite scarce. Not as bad as finding HF300s as you mentioned. One of the things drawing me to 810s is the fact that they are still being manufactured, albeit in China. But at least that means there is a large number of them out there.

Good advice on the transformers. Any idea why it is that the earlier equipment used transformer coupling for each audio stage? Seems to me like simple R-C coupling would be an overall better engineering decision.

I may just give the split power supply a try. I think it would be great to build both supplies with variacs to vary the output. Would that be a big issue line regulation-wise?

I looked up those crystal holders you mentioned, and I have to say... they are quite appealing! They seem quite scarce though. If I had a good supply of those, I'd think twice about building a VFO!

[k4kyv]

Any idea why it is that the earlier equipment used transformer coupling for each audio stage? Seems to me like simple R-C coupling would be an overall better engineering decision.

Transformer coupling allows a voltage step-up between stages, giving more gain for the number of tubes in the circuit. It also allows the transition from single ended to push-pull without a phase inverter. Nothing wrong with it if good quality audio transformers are available. With crappy transformers, you are better off with R-C coupling.

I may just give the split power supply a try. I think it would be great to build both supplies with variacs to vary the output. Would that be a big issue line regulation-wise?

I have found  that variacs noticeably degrade the voltage regulation, probably due to voltage drop in the carbon brushes. I once used a variable tapped transformer that came out of an X-ray machine. It had much better regulation, but I foolishly let another ham have it, who never used it. In my present set-up, the modulator always runs at full voltage, but I run the PA power supply through a variac so I can adjust carrier power level. Regulation is less critical with the PA, since it pulls a steady plate current. At reduced plate voltage I turn the audio gain down to stay within 100% modulation.  That simplifies things, because varying the modulator plate voltage would require fiddling with the bias supply as well. Varying the bias is no problem, but each tube has its own separate variable bias adjustment, so that would require a lot of "fiddling" to change plate voltage.

I looked up those crystal holders you mentioned, and I have to say... they are quite appealing! They seem quite scarce though. If I had a good supply of those, I'd think twice about building a VFO!

As I said before, you don't have to complete the rig to its final version on the first try. It might take years to find and accumulate all the appropriate components for  the targeted period in history. Use what you have, with the idea of morphing to the final version as stuff turns up. Maybe leave enough room on the chassis for a xtal later on, or use an FT-243 at first and keep an eye out for Bliley/Valpey crystals. The FT-243 plugs directly into two pins of an octal socket, so use an octal socket at first, and if you later on find some of the round holders, replace the octal socket with a 5-pin one.

I can go either way, xtal or VFO. I use a type 59 broadcast receiver audio tube as the xtal oscillator, in the conventional pentode circuit from the 30s handbook. The 59 is special in that it has a separate base pin for the suppressor grid, allowing it to be grounded directly, not internally connected to the cathode. So for VFO control, I use a switching arrangement that grounds the control grid and un-shorts an RF choke from cathode to ground. I inject the VFO voltage to the cathode, making the stage work in grounded-grid when serving as an amplifier for the external VFO. The VFO has a low-Z output via a coupling link, which directly feeds the cathode pin through a coupling capacitor. This arrangement is very stable, with no tendency to self-oscillate, as might have been the case if I had driven the control grid with the VFO, since pentode audio tubes are not shielded well enough to be 100% effective at RF. That imperfect shielding is what allows the audio tube work as a xtal oscillator, by providing the necessary feedback through the grid-plate capacitance. A well shielded pentode like the 802 or 6AG7 would be stable as a grid-driven amplifier, but would likely not work very well as a pentode xtal oscillator.

I decided to build that transmitter to "period", since the cabinet originally held a "composite" (homebrew) mid-30s broadcast transmitter. Unfortunately, it was totally gutted when I acquired it, but it still had the original meters, name plates, tuning knobs and only a few of the original transmitter parts. After I built the initial version of the rig with parts on hand, I thought it would be cool to retro-fit as much as possible to the period of the original BC transmitter. I initially used some of the few original components that came with the cabinet (including the UTC LS-103 modulation reactor). The rest were either WW2 surplus or cheap-ass Hammy Hambone parts. Over the years, as I have run across them, I have replaced as many as possible of the "ham radio quality" and/or non-period components with top-of-the-line commercial-grade pre-WW2 parts. I still have a few to go, and I originally built the transmitter in 1970.

[KM1H]

Lots of vintage tubes go gassy and they can be saved if its not too bad.

Keep the HV around 1000V as it wont sustain an arc and then apply just enough bias to get some good color in the anodes. Do not apply RF but do use an air flow.
Ive run tubes that way from about 15 minutes to almost a day and have salvaged most of them. Residual gas is one thing, going to air of most any degree thru leaky seals is unrecoverable unless it is very minor and even then the peak useable voltage will be lower.


Carl

[KE6DF]

Speaking of period pieces (transformers) does anyone know when UTC switched from the black/charcoal color for the LS series transformers to the blue/grey color used in the later transformers?

Thordarson made some nice transformers in the late '30's in the CHT series. I don't think they were quite broadcast quality, but were considerably better than the base level Stancor or UTC S series pieces. They also look very buzzardly, IMO.

[k4kyv]

I think the UTC change was sometime about 1949. I have a photocopy of the catalogue that includes the new style, along with a special close-out section for the old style, but I would have to dig it out.  Not only did they change shades of grey of the paint, they rotated the core 90° and used new style castings, so the larger open frame transformers changed from being wide and low, to tall and narrow. The LS series reflected this change in style, and so did the PA series which was renamed "CG".

[W7TFO]

Lots of vintage tubes go gassy and they can be saved if its not too bad.

I've brought this up before.  I have an article on how to deagss transmitting tubes (and MV rectifiers--yes, they can get gassy and arc-back) without lighting the filament.

It was published back in the 50's by Machlett Labs.

E-mail me for a copy...it works!

73DG