Pentode theory, am I crazy?

[Bob-I]

I built this amp using an ECL86 for a reverb tube, pentode side to drive and triode side for recovery. In this thread I had some issues that folks helped me with but I'm still having issues.

http://ampgarage.com/forum/viewtopic.php?t=1709

When I push the amp I get arcing between the plate, pin 6 and cathode, pin 7. I'm thinking this through while waiting on an airplane and I think I understand the issue.

With a pentode the screen acts as an electron excelerator, and therefore should be at a lower potential than the plate. On this amp the screen is 20V higher than the plate but someone on another forum said that this is common with small pentodes so I moved on. In order to stop the arcing I needed to drastically increase the cathode resistor and lower the bias.

My thought is that the electrons are looking for a path to ground and since the plate is negetive with respect to the plate, the easiest path is from pin 6 to 7 by arcing. If I lower the screen voltage the arc should stop and the tube should work as designed.

I really want to solve this one. Not only does the reverb sound GREAT at low volumes with the proper cathode resistor, but I just want to beat it.

Anyone?

[skyboltone]

Bob:
Your thinking seems plausable to me. On the same topic too, I think it's worth mentioning that once an arc track as occured, pretty much regardless of the substrate material, an arc will occur there more easily thereafter.

I think the screen of a pentode should be either at a lower potential or tied to the plate in triode mode. On the other hand, the suppressor grid is closest to the plate and is internally tied to the cathode. Would that be an easier path to ground than the tube socket pins? I dunno.

The dunce.

[Bob-I]

Bob:
Your thinking seems plausable to me. On the same topic too, I think it's worth mentioning that once an arc track as occured, pretty much regardless of the substrate material, an arc will occur there more easily thereafter.

Yes, I plan on replacing the tube socket once I get this solved.

I think the screen of a pentode should be either at a lower potential or tied to the plate in triode mode. On the other hand, the suppressor grid is closest to the plate and is internally tied to the cathode. Would that be an easier path to ground than the tube socket pins? I dunno.

The dunce.

I forgot about the supressor grid. Damnit, I'm a dunce too. But I'll still try the larger screen resistor.

[Bob-I]

I'm finally back home so I had the chance to try a larger screen resistor. I took out the 2.2K and installed a 330K (the largest I had). The arcing is gone. So next I rebiased the pentode side to closer to spec, about 8 watts and still no arcing. Now the screen is at 280 and the plate is 345.

Conclusion is that these small pentodes need to keep the screen voltage low.

Oh... and the reverb sound great!

[UR12]

With a pentode the screen acts as an electron excelerator, and therefore should be at a lower potential than the plate.

My thought is that the electrons are looking for a path to ground and since the plate is negative with respect to the plate, the easiest path is from pin 6 to 7 by arcing. If I lower the screen voltage the arc should stop and the tube should work as designed.

Anyone?

Bob

Your first statement is true. Your second statement is not true. Electrons do not look for a ground unless you are running a positive ground.

Maybe this will help……..electrons are boiled off of the cathode of a tube because it is heated to a high temperature. Once the electrons are boiled off they are attracted like little magnets to anything with a positive potential. The first grid, the control grid has a varying signal applied and if the signal makes it positive in respect to the cathode then the electrons are attracted towards the control grid. If you think of the control grid as a screen mesh, the electrons can pass through the control grid. If the signal applied to the control grid is more negatively charged then the electrons are “repelled back toward the cathode and very little or no current flows through the tube. When the control grid becomes positive enough the electrons are pulled towards the control grid. As the approach the control grid they see a higher positive charged grid, the screen grid, and because the screen grid is more positively charged than the control grid they forget about the control grid, pass on through and head towards the screen grid. As they approach the screen grid they pick up speed (speed of light) and they are attracted even more by the higher voltage of the plate. Because they are traveling so fast and the mesh of the Suppressor grid is bigger they fly past/through the suppressor grid and hit the plate. Some of these electrons are traveling so fast they bounce off the plate and the suppressor being a like charge (negative or ground) repels it back to the plate. The surface area of the plate can only hold so many electrons at any one time. When you have so many electrons striking the plate that there is no more room then the plate is saturated with electrons and you have reached the maximum amount of current the tube can pass at the voltages you have on it. Granted this is a much simpler explanation than what it could be.

Having your screens running higher voltages than you plate can be disastrous as you are seeing with all of the arcing going on. Glad you got it running!

[Bob-I]

Thx UR12. That really clears things up. One of my problems is that my electronics teacher in HS was an old Military guy, they learn electrons flow TO the - not to the + so in some ways my thinking is backwards.

I like to think of every problem as an education, so this was a huge help, not to mention I now have an excellent reverb ckt.