Fuse protection for tubes?

[Shockwave]

Wouldn't it be nice if you could gain a reasonable level of protection for your expensive output tubes by adding a simple fuse inline with the tube? I've done this with every tube from a 6LF6 to a 3CX1200A7 and it works. You have to install the fuse on the cathode circuit of the tube in the DC bias line. Since plate current and cathode current are nearly the same, you can prevent excessive plate dissipation by fusing the cathode side of the tube. This offers protection from over driving, incorrect tuning, or a shorted antenna.

For tubes like the 3-500Z you would install the fuse between the filament transformer center tap and the bias zener diode. This has only been tested on grounded grid, cathode driven amps. You must place a resistor across the fuse to prevent excessive cathode voltage when the fuse blows under load. Otherwise all of the plate voltage will attempt to reach the cathode when the fuse blows and this will short the tube. Think of the tube, fuse, and plate supply being in one series circuit. In a series circuit the majority of the voltage drops across the point of highest resistance. If the fuse opens under load without the resistor to limit the voltage drop, the open fuse is now the point of highest resistance.

The fuse value should be determined by the tube specification sheet for maximum plate current. Using a lower value fuse offers more protection. Not a bad idea since this fuse does not directly protect the grid of the tube. The value of the resistor across the fuse is very important so that it places the tube into cutoff but not with an extreme voltage spike. Get this wrong and you may have been better off without the fuse altogether. Get it right and the tube as well as the amp should survive most common mistakes at the expense of a fuse.

One value that works well with the range of tubes mentioned is 10,000 ohms wire wound at 10 watts. Use 25 watts if you have several tubes in one stage. Just increase the fuse size according to the number of matched tubes in the stage. If the current is over 2 amps, dropping the resistor down to 5 or 6 thousand ohms may be appropriate. I suspect this will work on larger tubes but have avoided trying it on the 3CX1500A7 because that tube is very delicate and should have electronic protection on both the grid and plate.

Electronic protection that uses opto couplers driven off the plate and grid current meter shunts is a superior method since it's reaction time is faster and it protects the grid. The fact is there are countless homebrew and amps similar to the SB-220 that can benefit from this simple fuse / resistor modification while the electronic protection may be might be more then the person is willing to install. It was nice to hear customers say they made the same mistake that blew up their amp the first time but this time it only burnt up a 25 cent fuse. Make sure you use fast acting fuses rated at least 250 volts.

 

[Crusher]

This works good as you stated above shockwave. I used this inline with bias circuit for my 2 tube 35B box. I think i used a 15K-25k but then those tubes like 40Volts of bias on them to really calm them down.

 

[paws264]

This works good as you stated above shockwave. I used this inline with bias circuit for my 2 tube 35B box. I think i used a 15K-25k but then those tubes like 40Volts of bias on them to really calm them down.

I have separate Filament and bias supplies for each tube, each circuit has an in-line fuse in the cathode circuit, I used 1.0 amp fuses to test but now have 2.0 Fas-Blow fuses in there now. Each fuse block is parallel with a 30 Kilo-ohm resistor.

Tell me about your tuned input circuit.

.

 

[Shockwave]

You can use larger value resistors. I used 33 K on a pair of 3-500Z's before and it worked. I'm probably overly cautious here but I had one situation where a higher ohm resistor dropped too much voltage when the tube was conducting heavily. When the fuse blew it caused the resistor to open and ruined the tubes with a short to the grid. Since then I tend to use higher wattage and lower resistance to prevent this. As long as you reach cutoff without resistor or tube failure when the fuse blows, all is good.

Most input circuits use one coil and two capacitors. Either the coil or capacitor can be adjustable to provide the tuning range required. Most slug tuned coils are not appropriate for panel mounting. The alternative is an more expensive roller inductor or mounting a pair of variable caps. I prefer to use two coils and one variable cap on the input. Feed the drive trough both coils in series to the cathode. Install the variable cap from the point were both coils are connected together and ground.

The coils were approximately 11 turns of between # 14 and # 18 wire depending on power and Q desired and it was wrapped on a 1/2 inch form for 10 meters. The cap should be able to tune up to around 200 pf. This allows a wide range of adjustment with just the variable cap. The circuit is very efficient and provides a perfect match to almost any cathode driven amp. It may require initial adjusting of the coils (expanding or contracting) to get the match in the ballpark.

This circuit uses more inductance in the coils and has more RF voltage across the cap then most. Don't be surprised if the coils look bigger then you may be familiar with on this band. If your drive is over 100 watts compression trimmers like the 464 will fail. You might be able to get away with the 303M series. Anything more then 250 watts and you'll need at least an air variable. At 500 watts drive or more expect the RF voltage on the cap to reach 2 KV on peaks.

Since many cathode driven amps have an input impedance close to 50 ohms, you can test this circuit with a small 50 ohm carbon resistor from cathode to ground with an MFJ antenna analyzer while the amp is off and the relay is manually keyed. Sometimes this is the only way when the input circuit is located in the pressurized compartment under the tube. When the coils are right the cap should easily be adjusted for a flat VSWR. Usually well under the 200 pf full range. When the resistor is replaced by the hot tube, only a slight adjustment of the front panel cap should be required.

 

[Crusher]

Never did a T match, just used a PI network for the input. 1 inductor and 2 variable capacitors. I used the air variables used for plate circuit on the heathkit sb220. the ones rf parts sells. 3400 or 3600V 20-240pf. work great and will handle drive.

 

[Crusher]

I have separate Filament and bias supplies for each tube, each circuit has an in-line fuse in the cathode circuit, I used 1.0 amp fuses to test but now have 2.0 Fas-Blow fuses in there now. Each fuse block is parallel with a 30 Kilo-ohm resistor.

Tell me about your tuned input circuit.

.

I should have done seperate filaments and bias, but i was lazy. was my first tube box. I will say this, the 35B is not a good tube to start off with. Took me like 8 months start to finish. Did alot of reading, still reading. lots to learn.

 

[Shockwave]

The PI input is more forgiving on the capacitors since it places them both at low impedance points. The T input places the cap in the center of two coils where the impedance is high and so is the RF voltage. Both designs are very capable of matching the input. The PI input Crusher describes using the Heathkit plate caps would be nearly indestructible as long as the room is available to mount them.

 

[Crusher]

Yes the caps I put in were overkill, but thinking long term usage. For the bias I connected 59 of the 6A10 diodes in series. Then used I think 25K ohm 25W in series. I use a relay to bypass the resistor when amp is keyed. Went by the GS25B website by W4ZT's plans. No point in having to reinvent the wheel. I have never used a T match on an amp. Parts are certainly cheaper, But I guess I was lazy and it was easier drilling a few extra holes in the case.

 

[paws264]

Yes the caps I put in were overkill, but thinking long term usage. For the bias I connected 59 of the 6A10 diodes in series. Then used I think 25K ohm 25W in series. I use a relay to bypass the resistor when amp is keyed. Went by the GS25B website by W4ZT's plans. No point in having to reinvent the wheel. I have never used a T match on an amp. Parts are certainly cheaper, But I guess I was lazy and it was easier drilling a few extra holes in the case.

Update?

.

 

[Crusher]

The amp did very well. With 3950 no load and about 3700vdc under load. Used about 34V of bias. In cutoff, amp draws no current at all. Gives them a better chance to cool down. With a transistor driver. 350W pep gave me 5kW pep out. I'm sure if I used a tiny tube as a driver would have taken less. Also I wonder how it would have done if I had used a class A or class AB driver instead of class C. I wonder if it will require less drive? Wonder if higher conduction angle somehow affects drive ? Any thoughts or input anyone?

 

[Shockwave]

Unless there was something wrong with the transistor driver (oscillating, impedance mismatch or very rich with harmonics), it should require the same input power to the GS35B to make the same output as it would with a tube driver.

Recently I did some testing on the GS-35B. The input impedance is close to 50 ohms but a pair drives at 25 ohms. Many will arc internally when first fired up (confirmed with Hi-Pot tester) because they have been sitting on the shelf so long the vacuum may have leaked a bit.

To my surprise the conditioning recommended for these tubes prior to use is very beneficial in terms of breakdown voltage within the tube. The tube can hold off more than double it's 3000 volt military CW rating once this is done. The benefit here is greatly reduced grid current on a tube that already draws a lot of grid current.

 

[Crusher]

Yup, they need to sit and burn in just the filament for about 24 hours before plate V is applied. Done that and never had an issue so far. I keep then under rated plate V. Just seems like a good idea with these tubes.

 

[Shockwave]

One problem you're likely to find running the plate voltage towards the low end is it's difficult to approach the useable plate dissipation of 2500 watts without grossly exceeding the grid dissipation of 26 watts. Not an issue if you run the tube conservatively. On the other hand if you intend to use the full dissipation of the anode, you're going to want to apply more B+ voltage to keep the grid from over dissipating.

As most people know, this tube was not designed for HF linear applications. It was designed for pulsed military radar use. That's actually a good thing if we could provide some accurate specifications for our application. The fact the tube can be used to microwave frequencies means it's highly stable at HF and usually does not require a parasitic suppressor due to extremely low internal inductances.

 

[Crusher]

I have built several with very good results. Very nice tube for the price. Only thing I can say is, not the best tube for a beginner. I used this tube for my very first amp, no prior knowledge. It took me awhile since I had to figure out how to float the B- and since the tube takes a minimum of 30V of bias. Not like the more common US tubes that are zero bias. But a great way to really dig into it and learn from the get go.

 

[Mr Fix it]

t match

T match with 19 turns of #14 enamel wire on a 5/16 form for each coil with a 110 pf variable cap will give you a 1.1 swr with a single 3cx3000a7. This tube has a cathode impedance close to 50 ohms.