restoring Millen GDO

[w4bfs]

I noticed I've been away for a while from this forum.  I also have been off-the-air for more than 3 years.  I finally have a modest antenna up and slowly getting some gear to work.  Part of this needed a GDO hence the restoration.

After recapping the very old 20 uF  160 V electros with 10uF  350V  Nichicons, I noticed that the 9002 tube filament seemed to be very bright.  A dmm check found 7.5 Vac on the fil.  looking a little farther along the brown wire I saw where it came to a terminal strip with a jumper.  I removed the jumper and put in a 7.5 Ohm  1/2 W. resistor.  now have 6.5V with 120V line voltage.  since 9002 tubes will get scarce I recommend this mod

[K9MB]

I wonder if you have considered putting a rectifier in the 6.3vac line and adding a 7806 regulator?
Old signal generators that use the 955 benefit from this, making them more stable and the note cleaner.
Would need a bridge, so one lead will need to be lifted off the ground and bypassed with a low inductance cap.

[w4bfs]

I would want to know the current capability of the 6.3 V winding before raising the demand by your suggestion. The power xfmr is tiny and I don't know how much power it can reliably deliver as it is already stressed some by higher line voltage
(120V vs 115V)

[Detroit47]

The wheel is already round.

[WQ9E]

I would keep AC on the filament and use series resistance to reduce the voltage as needed.  Filaments generally last longer on AC, running on DC they will potentially develop hot spots leading to more rapid failure.

GDOs are still very useful pieces of gear.  Several years ago I repaired a Johnson 500 for a friend, the problem was it wouldn't work on 40M phone and output was low on the CW segment.  I assumed a failed switch contact in the output network but everything looked fine and using an antenna analyzer connected to the RF output with a simulated plate load resistor to ground showed reasonable results when tuning the output network on 40.  But the grid dip meter told the tale and the tank circuit was resonating just above 6.9 Mhz with the tuning cap at minimum capacitance.  It turned out this Viking 500 had never worked properly on 40 because the original owner/builder had soldered the 40 meter tank connection one turn off.  An easy repair once the issue was identified.

I have a Millen GDO with all of the coils including the LF set and a Measurements Model 59 with the LF, HF/VHF, and UHF heads.  Both work well but I am partial to the Measurements 59.

Rodger WQ9E

[K9MB]

I would want to know the current capability of the 6.3 V winding before raising the demand by your suggestion. The power xfmr is tiny and I don't know how much power it can reliably deliver as it is already stressed some by higher line voltage
(120V vs 115V)

Very good point.
This problem with stuff we bought in the 60s or before is real. The resistor is likely the best option. I had not considered it.
In bigger equipment, a second filament transformer can be added to tue primary 73, assuming it is rated at high enough voltage as a bucking winding, so that the original equipment sees 110- 117vac, but that makes no sense here. 73, Mike

[Carl WA1KPD]

I run my whole bench off a strip with a variac set to about 112V.
It's the same thing with the strip for BA's in the shack.

[KD1SH]

  Curious about this. Not doubting what you say—just wondering if there's been any sort of study out there regarding tube heater life when run on DC vs. AC. I've seen mention of it on audio-phool forums, but it's all anecdotal. Think "tube rings" and "oxygen-free wire," to keep things in perspective when reading those forums.
  My Clegg Zeus 6 meter AM transmitter uses a ballast tube to stabilize the heater voltage to the oscillator tube in an attempt—perhaps a bit overzealous—to reduce VFO drift. That particular ballast tube is hard to find these days, so I've considered running regulated DC to the tube instead.

Filaments generally last longer on AC, running on DC they will potentially develop hot spots leading to more rapid failure.
Rodger WQ9E

[WQ9E]

  Curious about this. Not doubting what you say—just wondering if there's been any sort of study out there regarding tube heater life when run on DC vs. AC. I've seen mention of it on audio-phool forums, but it's all anecdotal. Think "tube rings" and "oxygen-free wire," to keep things in perspective when reading those forums.
  My Clegg Zeus 6 meter AM transmitter uses a ballast tube to stabilize the heater voltage to the oscillator tube in an attempt—perhaps a bit overzealous—to reduce VFO drift. That particular ballast tube is hard to find these days, so I've considered running regulated DC to the tube instead.

Google "filament notching" and you will find some information on it, operation on DC was identified by GE as a cause of this shortening of life.  Is it worth worrying about for small vacuum tube filaments?  My view is probably not but I wouldn't run them on DC unless there was a good reason to do so (like the need to keep hum at an absolute minimum, for example some of the old Tektronix 1 and letter series plug-ins for their lab grade scopes had some of the tube filaments in the low level stages powered by DC).

In the case of the OP, the use of a dropping resistor is clearly the most practical solution.

And someone is missing the boat on scamming the audiophools further, how about some nitrogen free wire???  I don't think anyone is selling it for $1,000 an ounce yet 🤣

Rodger WQ9E

[KD6VXI]

Lots of people have stated that reversing the DC bias on the filament reverses this issue.

Think.

Power up this time, pin 1 has +12v pin 9 or 12 has ground.

Next 'boot up' is I'll have pin 1 be ground and pin 9 or 12 as positive.

Don't know if this fixes it.


--Shane
WP2ASS / ex KD6VXI

[KD1SH]

  Thanks—interesting stuff. I just took a quick look on Google, and interestingly, some of what I found was related not to vacuum tube filaments but to incandescent lamp filaments. Much higher temperatures, but the same principles would apply, I'd imagine. The article I saw stated that, with AC in a lamp, the notching effect was greater toward the ends of the filament where, due to the heat-sinking effect of the filament supports, the temperature gradient was greater. Makes sense. With DC, however, the consistent direction of electron flow causes tungsten ions to flow also in the same direction, encouraging the notching effect to a more consistent degree along the entire filament.
  I agree, small tubes, okay if required, but with larger transmitting tubes, not so much. Aside from the risk of "notching" an expensive tube, the notion of rectifying and regulating with the currents involved...well...a pair of 3-500Z's will draw roughly 30 amps, so that's that.
  My Zeus uses a separate winding on the power supply transformer to supply 20 VAC, with the ballast—basically just a PTC resistor—in series with the oscillator tube's heater. I suppose that if I couldn't find a suitable ballast I could rewire the oscillator tube's socket to pull 6.3 VAC from the same filament circuit that heats the rest of the tubes, rather than go through the trouble to build a rectifier and regulator to run the tube on DC.
  Looking at some of those audio-phool sites, it seems the preferred wire isn't the only thing lacking oxygen: hypoxia will do strange things to you.

  Curious about this. Not doubting what you say—just wondering if there's been any sort of study out there regarding tube heater life when run on DC vs. AC. I've seen mention of it on audio-phool forums, but it's all anecdotal. Think "tube rings" and "oxygen-free wire," to keep things in perspective when reading those forums.
  My Clegg Zeus 6 meter AM transmitter uses a ballast tube to stabilize the heater voltage to the oscillator tube in an attempt—perhaps a bit overzealous—to reduce VFO drift. That particular ballast tube is hard to find these days, so I've considered running regulated DC to the tube instead.

Google "filament notching" and you will find some information on it, operation on DC was identified by GE as a cause of this shortening of life.  Is it worth worrying about for small vacuum tube filaments?  My view is probably not but I wouldn't run them on DC unless there was a good reason to do so (like the need to keep hum at an absolute minimum, for example some of the old Tektronix 1 and letter series plug-ins for their lab grade scopes had some of the tube filaments in the low level stages powered by DC).

In the case of the OP, the use of a dropping resistor is clearly the most practical solution.

And someone is missing the boat on scamming the audiophools further, how about some nitrogen free wire???  I don't think anyone is selling it for $1,000 an ounce yet 🤣

Rodger WQ9E

[WQ9E]

The 6BK7 tube used as the VFO for the Zeus (I the Clegg "twins" also) is very common so if your ballast tubes go out, I wouldn't hesitate to use a simple rectifier/regulator circuit to provide 6 volts to this tube filament.  There have been several articles on doing that to replace the ballast tube used in the R-390 series.

Using ballast tubes in parallel is an interesting approach, I guess they had to do that in order to hit the current/voltage value needed.

The Hallicrafters SX-73 uses a pair of C7 "night light" bulbs in parallel as the current limiter for the VR tube, the only place I have seen that done but it works well and looks cute warming up

Rodger WQ9E