Post Pics of tube amps here

[HiDef]

Well I wouldn't build an amplifier that turned on plate voltage with filament voltage anyway. Why? I like the ability to control stuff. Especially the old drain the life out of the filter before opening the lid trick. Would I put a delay in if it were a directly heated fialment? Nah

Broadcast transmitters are a very different type of service compared to hammy hambone or chickenband. The object there is to squeeze the last bit of life out of tubes and still transmit within regulations. Many of these transmitters ramp up filament voltages slowly to avoid exceeding a certain amount of current. Others use indirectly heated cathodes which DO need warmup time.

Ever use a G3SEK tetrode board?

 

[Shockwave]

I am familiar with both the triode board and the tetrode board. Problem is there aren't any FET's that can handle the screen voltage in large tetrodes. The design goal is 2 KV on the screen at about 225 ma plus or minus additional negative screen current with active shunt regulation. I have read the G3SEK article in the October 1997 edition of QEX and have been looking into ways of expanding it's screen operating voltage.

At one point I looked into using an insulated gate bipolar transistor as the active component. I even purchased this device based on the specifications published online by the manufacturer. Problem is once it arrived with paperwork, it became clear this part was not suitable for linear mode use. The prospect of building a high voltage switch mode shunt regulator for use in an RF environment was too complicated.

My next idea was to use the FET in the basic G3SEK circuit to control the voltage in the cathode circuit of an Amperex 8802 / 3-500Z tube. I figured if I replaced the typical zener diode bias used on a tube like a 3-500Z with the FET output device used in the G3SEK circuits it should be able to operate the tube as a voltage regulator.

Looking at the curve charts for the 3-500Z it seems possible to draw the right amount of shunt current at 2 KV on the anode and about 35 volts on the grid. If I just ground the grid on the 3-500Z then I can't get the 3-500Z to draw enough shunt current at only 2 KV on the plate without biasing the cathode positive. If I bias the cathode positive then it makes it complicated for the FET to swing the cathode negative when the tube needs to conduct less.

My questions are would it be better to try applying forward bias to the cathode to get it to draw current and then control the negative bias on the grid with an FET? It seems this would be somewhat complex too since the cathode current is trying to push the voltage negative with respect to ground here. Another thought is using a different tube then the 3-500Z that would draw more current at 2 KV on it's plate without applying positive grid voltage. Any ideas on this?

 

[HiDef]

I am familiar with both the triode board and the tetrode board. Problem is there aren't any FET's that can handle the screen voltage in large tetrodes. The design goal is 2 KV on the screen at about 225 ma plus or minus additional negative screen current with active shunt regulation. I have read the G3SEK article in the October 1997 edition of QEX and have been looking into ways of expanding it's screen operating voltage.

At one point I looked into using an insulated gate bipolar transistor as the active component. I even purchased this device based on the specifications published online by the manufacturer. Problem is once it arrived with paperwork, it became clear this part was not suitable for linear mode use. The prospect of building a high voltage switch mode shunt regulator for use in an RF environment was too complicated.

My next idea was to use the FET in the basic G3SEK circuit to control the voltage in the cathode circuit of an Amperex 8802 / 3-500Z tube. I figured if I replaced the typical zener diode bias used on a tube like a 3-500Z with the FET output device used in the G3SEK circuits it should be able to operate the tube as a voltage regulator.

Looking at the curve charts for the 3-500Z it seems possible to draw the right amount of shunt current at 2 KV on the anode and about 35 volts on the grid. If I just ground the grid on the 3-500Z then I can't get the 3-500Z to draw enough shunt current at only 2 KV on the plate without biasing the cathode positive. If I bias the cathode positive then it makes it complicated for the FET to swing the cathode negative when the tube needs to conduct less.

My questions are would it be better to try applying forward bias to the cathode to get it to draw current and then control the negative bias on the grid with an FET? It seems this would be somewhat complex too since the cathode current is trying to push the voltage negative with respect to ground here. Another thought is using a different tube then the 3-500Z that would draw more current at 2 KV on it's plate without applying positive grid voltage. Any ideas on this?

Why not look at what's used in an older FM broadcast tx with a similar tube?

Some of the older RCA TV txs with the 8807 might show an older pre microprocessor/optocoupler design.

 

[Shockwave]

The transmitter I'm working on runs a very similar tube to the 8807. It's the 4CX15,000. Just like in every other broadcast transmitter I've seen they do not use any active regulation on the screen. Typically they design the screen supply so that it has very low internal resistance. Basically they just load the heck out of the supply with a resistor so that any change in current flow to the tube represents a very small change in the overall current. Thereby providing a reasonable stable voltage. The resistor also sinks most of the negative screen current.

While this is acceptable, it's not ideal. As the tube approaches the end of it's life, secondary emissions begin to increase causing more negative screen current. By strictly controlling the screen voltage through active shunt regulation you can continue to use these tubes much longer. This also reduces IMD and increases linearity. These are the reasons I'm trying to adapt this equipment to make use of this type of regulation.

 

[HiDef]

The transmitter I'm working on runs a very similar tube to the 8807. It's the 4CX15,000. Just like in every other broadcast transmitter I've seen they do not use any active regulation on the screen. Typically they design the screen supply so that it has very low internal resistance. Basically they just load the heck out of the supply with a resistor so that any change in current flow to the tube represents a very small change in the overall current. Thereby providing a reasonable stable voltage. The resistor also sinks most of the negative screen current.

While this is acceptable, it's not ideal. As the tube approaches the end of it's life, secondary emissions begin to increase causing more negative screen current. By strictly controlling the screen voltage through active shunt regulation you can continue to use these tubes much longer. This also reduces IMD and increases linearity. These are the reasons I'm trying to adapt this equipment to make use of this type of regulation.

The older brute force low source impedance screen supplies seem to be pretty forgiving when it comes to tOOb arcs. Maybe you can take a few steps to build one with even lower source impedance for the negative screen current and improved regulation. It will of course waste a lot of power in heat by design as you say. The only special thing needed would be an overload relay.

Some old timers have a procedure for scrubbing a grid. The idea is to dissipate the grid very carefully near it's limit without plate voltage over a period a few hours to lose the effects of cathode material buildup which can cause secondary emission. I've not seen it work myself.

I'd be interested to hear more about end of life power tubes and what happens to different types. I've seen older hiMu triodes with high grid current and lots of thoriated tungston types with low emission.

 

[Shockwave]

Look at that, we opened up a whole new interesting topic of discussion Hi Def. Tube rejuvenation. I have had limited luck with tubes that developed internal shorts. So long as the short was there with no voltages applied I found that a car battery can be useful in "popping" the short out of the tube. I've had about a 50% chance of this working or of course destroying the tube. However, you have nothing to loose if the grid is already shorted to the cathode.

I also remember my grandfather who was a service tech had a device he could rejuvenate cathode ray tubes with. I'm not sure of this but I think it manipulated the filament voltage so that it was possible to bring a new layer of thorium up to the surface as long as it wasn't completely depleted. I know this method does work on thoriated tungsten filaments depending on what caused the tube to have reduced output. For example if the tube was overloaded this can damage the top layer of thorium and still have a good layer underneath that can be brought to the surface with more filament voltage and no plate for a short time.

I understand how secondary emissions are increased because of thorium that has migrated off the filament to other areas of the tube. With this in mind it seems possible to "scrub" the grid or screen of this thorium since heat will boil thorium off the surface of an electrode. Next time I get a tube with unstable emissions I may try your suggestion. One thing I can say is that true shunt regulation of the screen definitely makes a huge difference in the ability to continue to use an otherwise unstable tube. If keying a tetrode transmitter for a period of time causes the output power to rise, screen shunt regulation will stop this.

As you point out with a triode, secondary emissions can also cause the grid current to go up. Since the grids are usually grounded this can't cause the voltage to rise but it can cause current to go up. I suspect this is happening with the screen in a tetrode too once it's shunt regulated and experiencing secondary emissions. My first experience with secondary emissions was in a Yeasu FT-101E. When those 6JS6C tubes got hard to get I started using my old pull outs. Key the mic for a few seconds and plate current along with output power would rise to unsafe levels. Here a simple zener diode on the screen solved the problem.

I also agree that the brute force method is more forgiving and in a simple circuit design. To really simulate voltage regulation here one should be sinking about ten times the current the tube draws in resistor current. They usually don't go this far. In the transmitter I'm working on this would equate to 4.5 KW heat dissipation in the resistors.
That's why active shunt regulation is looking so attractive to me now. I figure I'll either take my time and get it right, or have an early 4th of July fireworks show.

With the use of a set of MOV's on the screen terminal of the socket and optocouplers on the meter shunt resistor to drive the protection circuit I think I can make it reasonably safe. My only hang-up now is figuring out the best way to bias an appropriate tube for use as the active component in the shunt regulator. If I wait long enough maybe a new 2 KV FET will be on the market

 

[bigred222]

well obviously wouldnt you notice a difference in output power when things start to fail inside the tube. dont you guys have meters on these TETETETETETETRODES
RED LISTENING ON 27.1350 shockwave give me a shout

 

[HiDef]

Look at that, we opened up a whole new interesting topic of discussion Hi Def. Tube rejuvenation. I have had limited luck with tubes that developed internal shorts. So long as the short was there with no voltages applied I found that a car battery can be useful in "popping" the short out of the tube. I've had about a 50% chance of this working or of course destroying the tube. However, you have nothing to loose if the grid is already shorted to the cathode.

I also remember my grandfather who was a service tech had a device he could rejuvenate cathode ray tubes with. I'm not sure of this but I think it manipulated the filament voltage so that it was possible to bring a new layer of thorium up to the surface as long as it wasn't completely depleted. I know this method does work on thoriated tungsten filaments depending on what caused the tube to have reduced output. For example if the tube was overloaded this can damage the top layer of thorium and still have a good layer underneath that can be brought to the surface with more filament voltage and no plate for a short time.

I understand how secondary emissions are increased because of thorium that has migrated off the filament to other areas of the tube. With this in mind it seems possible to "scrub" the grid or screen of this thorium since heat will boil thorium off the surface of an electrode. Next time I get a tube with unstable emissions I may try your suggestion. One thing I can say is that true shunt regulation of the screen definitely makes a huge difference in the ability to continue to use an otherwise unstable tube. If keying a tetrode transmitter for a period of time causes the output power to rise, screen shunt regulation will stop this.

As you point out with a triode, secondary emissions can also cause the grid current to go up. Since the grids are usually grounded this can't cause the voltage to rise but it can cause current to go up. I suspect this is happening with the screen in a tetrode too once it's shunt regulated and experiencing secondary emissions. My first experience with secondary emissions was in a Yeasu FT-101E. When those 6JS6C tubes got hard to get I started using my old pull outs. Key the mic for a few seconds and plate current along with output power would rise to unsafe levels. Here a simple zener diode on the screen solved the problem.

I also agree that the brute force method is more forgiving and in a simple circuit design. To really simulate voltage regulation here one should be sinking about ten times the current the tube draws in resistor current. They usually don't go this far. In the transmitter I'm working on this would equate to 4.5 KW heat dissipation in the resistors.
That's why active shunt regulation is looking so attractive to me now. I figure I'll either take my time and get it right, or have an early 4th of July fireworks show.

With the use of a set of MOV's on the screen terminal of the socket and optocouplers on the meter shunt resistor to drive the protection circuit I think I can make it reasonably safe. My only hang-up now is figuring out the best way to bias an appropriate tube for use as the active component in the shunt regulator. If I wait long enough maybe a new 2 KV FET will be on the market

Been there with the car battery and cheap 3-500Zs

The thing about grid scrubbing is one needs to know the right level to dissipate the grid during the process. I know suggested numbers for tubes that will never be seen on H.F. Dunno any for 4CXxx000s. You usually use a 0-100VDC bench supply and set it so the product of the v and a is a certain wattage. Then let it cook for an hour or 2 with filament and no plate. The control grid would obviously need to be biased somewhere where the screen would conduct and would also need to be monitored for overdissipation in some tube types.

Of interest on this subject, I have a 4CX1500B which is an oldie but makes more output than all my other similar tubes. The tube draws increasing grid current after time and will run away. The bias supply has only a very tiny current capability because it was designed for a 4CX1000. I'd like to use this tOOb. It seems to behave slightly better after sitting with the filament on for a few hours.

My ultimate cure for screen grid designs is to avoid them altogether. The power gain found in tetrodes comes at the expense of additional support circuitry and additional IMD. I personally don't think it's worth the trouble anymore. There are surplus triodes out there which are very easy to construct a large power amplifier with. I'm done with tetrode designs.

 

[bigred222]

Whats a IMD?

 

[gc505]

A few from my collection:

Here is my Maco 750, runs eight M-2057's (similar to 8950's):

This one is a fairly rare Thunderbolt 750, runs three 6LF6's into four 8908's:

And here is my favorite boat anchor, the workhorse, a vintage Drake L-4B with two 3-500z's. This unit has a modified power supply that provides it with 3300 plate volts. I use this one solely on the 15, 20, 40, and 80 meter SSB HF bands.

 

[bigred222]

what kind of shape is that Maco in? new tubes weak tubes. LETS SEE THE INSIDES take some more pics. and i have never seen a thunderbolt like that. SWEET MAN NUMBERS TO YA
-Red

 

[TonyV225]

 

[HiDef]

Here is my Maco 750, runs eight M-2057's (similar to 8950's):

This one is a fairly rare Thunderbolt 750, runs three 6LF6's into four 8908's:

And here is my favorite boat anchor, the workhorse, a vintage Drake L-4B with two 3-500z's. This unit has a modified power supply that provides it with 3300 plate volts. I use this one solely on the 15, 20, 40, and 80 meter SSB HF bands.

Nice firebottle pics!

 

[gc505]

Nice firebottle pics!

Yep, few things are more beautiful than a glowing set of red/white-hot 3-500z's.

what kind of shape is that Maco in? new tubes weak tubes. LETS SEE THE INSIDES take some more pics. and i have never seen a thunderbolt like that. SWEET MAN NUMBERS TO YA
-Red

The Maco is in excellent physical condition... one of the best I've seen. The tubes are in fair/good condition. The six final tubes test above 90% while the two in the driver stage test at 80% and 85%.

The Thunderbolt was bought from a local a couple years back, the only history of them that I have heard is that they were manufactured by a guy in Tennessee back in the late 60's and early 70's. I've only seen one other one like it (although I am sure there are more out there somewhere).

The Maco and Thunderbird go unused, covered and sitting on the closet shelf, I'll try to pull them out and get some more pics of the guts soon.

Here's a pic of the interior of the RF Deck of the Drake L-4B:

 

[scottdarling1221]

Yeah i hope that the 3-500 z ends up being an adequate tube for making our active shunt regulator work would save me a bunch of money.