Telling the difference between red-plating and red-gridding

[dtbradio]

Hi all!!

Still fairly new to tube amp design/building, and have a question with what is probably an obvious answer. How do I tell if its the plate or grid glowing at high volume? I assume the plate would basically be the entire plate surface glowing red, and grid would look like its coming from further inside the tube behind the plate (this is what's showing in my case). In either case, I already know what needs to be done to correct it, just need to be 100% sure which problem it is.

Always get great info from this forum!!

[Phil_S]

Red plating is obvious and unmistakable. You'll know it when you see it. The red glow inside the plates is going to be the heaters.
I don't know if there is such a thing as red gridding. The grid is wound fine gauge wire. I'm reasonably sure, if overloaded, it disintegrates very quickly.

[LOUDthud]

A Traynor YBA-1A can make the screens of EL34s glow, even with no signal. You can see it around the top mica spacer if the shinny silver getter around the top of the tube is small enough. On some tubes there is a round hole in the plate that also lets you see the screen grid. It's basically a thin wire glowing a dull orange/red.

[dtbradio]

Thanks for the reply! This is separate from the filament, which doesn't change in brightness, and based on your reply its not the plate. I've read elsewhere in this forum (several posts) about the screen grid glowing when the tube is pushed if the screen grid resistor is too low. Total idle plate current through the matched EL34's sits at around 75mA with about 480 plate volts, which is about right for 70% bias.

I just now did a controlled test into a 15-ohm 50-watt test load, and I can gradually bring the tube drive up to where the element in question just starts to glow. I'm about 99% sure at this point that its the screen grid, so I'll be upping the screen resistors at the shop tomorrow, and will post my results.

[Stevem]

Don't judge by looks!
Learn how to do it by the proper formula , the tubes design spec and the needed voltage readings .

The glow inside a power tube that gets brighter as signal is applied is the electron plume and is normal.

[martin manning]

I assume the plate would basically be the entire plate surface glowing red, and grid would look like its coming from further inside the tube behind the plate...

In the photo below the large black box-like structure is the plate, and "red-plating" will typically appear as patches of dull red glow centrally located on that part. The screen grid (on true pentodes like EL34) is a spiral-wound wire around the central cathode. If you look closely through the openings in the plate you can see the screen grid, and also the control grid, which is the smaller diameter coil closer to the cathode, which is in the center of he tube, behind the copper-colored support rod.

[dtbradio]

I swapped out my 430-ohm sand-block grids for 1000-ohm 20-watt 1% units. At idle, screen voltage is now 479, with 3.7 volts across the grid resistors. This comes out to just under 2 watts, and that value measures/calculates the same for both tubes. With a 1-KHz sine wave test signal applied to the amp input, and with gain set to max amp output before break-up (about 42 watts RMS into a 15-ohm test load with a visually-clean sine wave), I'm seeing a screen voltage of 435, with screen current calculated at 21.5 mA. This comes out to about 9.4 watts screen dissipation, which is just 1.4 watts over max grid rating. Since I will NEVER play at anything above about 60% volume (and usually a lot less) for more than a few seconds, I'm ok with the grid numbers now.


Edit:

For giggles, I added the original 430's in series with the 1K's, and lowered the grid dissipation to 8.9 watts at full volume. I'm still looking through my bits and bobs to see if I have anything in the 2K ohm range to try.

Edit 2:

Found another pair of 1K 1% 20-W and shoe-horned them in series with the first pair, for 2K total per screen. Now seeing a screen dissipation of 6.74 watts at full volume before breakup of a clean sine wave (42~ watts RMS into 15-ohm load). With an 800 mV sine wave amp input and all controls dimed in overdrive mode, screen dissipation is 13.44 watts

[dtbradio]

Thank you for your reply! I was 99% sure I was right about it NOT being the plate, your post confirms it. With the grid resistor changes, the tubes are now quite happy at all normal volumes.

I assume the plate would basically be the entire plate surface glowing red, and grid would look like its coming from further inside the tube behind the plate...

In the photo below the large black box-like structure is the plate, and "red-plating" will typically appear as patches of dull red glow centrally located on that part. The screen grid (on true pentodes like EL34) is a spiral-wound wire around the central cathode. If you look closely through the openings in the plate you can see the screen grid, and also the control grid, which is the smaller diameter coil closer to the cathode, which is in the center of he tube, behind the copper-colored support rod.