New Electro Harmonix 12AT7 in reverb driver (Fender Deluxe), won't hold 410v plate?

[mikepukmel]

Has anyone had problems with new electro harmonix 12AT7's going bad, in the reverb driver position in Fender Deluxe AB763 or clones?

Nick did some great research about high plate voltage (and a lot of other stuff I don't understand) in the cathode follower spot in certain Marshall amps. I didn't understand the whole topic, but it sounded like the plate voltage or plate to heater difference was too large. in these amps, the reverb plate has 410 or higher on the plate.

The data sheet says not more than 300v plate, and not more than 100v heater to cathode.

Had my amp running great for a few days, the reverb started crackling and popping badly. Pull the reverb driver tube, goes away. Can't see well into the tube due to the socket rings that hold the aluminum guards on.

[Chuck H]

It may be a limitation of the EH tube. Probably not though. The standard 12ax7 data for brands from the golden age also lists a max Vp of 300V. Which is probably why the new data sheets say that as they all seem to be verbatim regurgitations, right down to things like example voltages and plate loads, of the old data sheets. But I digress.

It may have simply been a bad tube. It also may be something less obvious. There are a couple of popular primary impedances for reverb transformers. IIRC some are 18k and others are 23k, but reports range from as low as 15+k to as high as 28k. The reverb driver in a AB763 topology works very hard all the time the amp is on. If you happen to have one of the lower primary transformers in an amp with high-ish operating volts AND are using a schematic that specs a cathode resistor which creates a hot bias then it could just be a matter of mish mash parts and design parameters. That is, if the particular model that you built tends to be a little hot on the reverb driver with a 23k primary transformer and you happen to have purchased an 18k primary transformer then the tube will be dissipating just a bit more, won't it.?.

You might just need to cool the bias a little. Consider that the reverb driver tube IS operated as a power tube. Did you check bias and current before the failure?

[loudthud]

What is your plate dissipation?

[Enzo]

The problem in cathode followers is not plate to heater but cathode to heater. The cathode is wrapped around the heater. And so they are close together. Too much voltage and it breaks down. Plate to cathode won't be a problem.

[nickb]

Has anyone had problems with new electro harmonix 12AT7's going bad, in the reverb driver position in Fender Deluxe AB763 or clones?

Nick did some great research about high plate voltage (and a lot of other stuff I don't understand) in the cathode follower spot in certain Marshall amps. I didn't understand the whole topic, but it sounded like the plate voltage or plate to heater difference was too large. in these amps, the reverb plate has 410 or higher on the plate.

The data sheet says not more than 300v plate, and not more than 100v heater to cathode.

Had my amp running great for a few days, the reverb started crackling and popping badly. Pull the reverb driver tube, goes away. Can't see well into the tube due to the socket rings that hold the aluminum guards on.

The big question is does it go away if you fit a different 12AT7?

I don't want dis-information to spread so if I may sidetrack for a moment. On the other thread the conclusion was that excess current causes the failure as it damages the cathode. I proved that heater to cathode breakdown is not the cause, at least in all the cases I tested and so may be a myth. Plate voltage had nothing to do with it.

[mikepukmel]

Thanks everyone.
1) I haven't measured anything on the reverb transformer, but this amp does run a little high on all voltages. Is DC resistance of any use here?

2) Ive set the bias on the output tubes and checked it a few times, but haven't checked it recently.
3) I don't know how to check it on the reverb driver tube. For the reverb driver, do I just measure the voltage drop across the cathode resistor? Is there some calculation to account/adjust for the 25uf cap?

4) Unfortunately, I can't find the other new 12AT7. There is one in the phase inverter, but worried to swap a bad tube into that spot. I will dig around, thought I had one for a marshall clone build Im working on but can't locate it. If I do buy a couple of 12AT7's, should I get different brand? Understood re dis-information, I don't to get into that!

Anyone have better luck with a particular brand in that spot in old Fender circuits?

Looking for arcing inside the tube: I tried looking into the tube while the amp is running, but due to the little ring around the bottom of the socket, and the getter on top, I can't see down to the bottom of the tube where all the little connections are located.\

EDIT: Don't have an extra 12AT7, other amp has all 3 12AX7, ordering a couple now.

[mikepukmel]

Something interesting. I looked at the bottom of the tube with a magnifying glass, and under my mag light. On the EH tube, the pins go through the glass into the bottom of the tube, then there is very, very fine wire connecting some of the pins to the tube 'guts'. I just looked on line for some other brands, one has the pins go through the glass, up into the 'guts' of the tube, I don't see any long piece of very thin wire.

Curious why the construction is so different? I thought a 12AT7 is a 12AT7 is a ...

[mikepukmel]

It may be a limitation of the EH tube. Probably not though. The standard 12ax7 data for brands from the golden age also lists a max Vp of 300V. Which is probably why the new data sheets say that as they all seem to be verbatim regurgitations, right down to things like example voltages and plate loads, of the old data sheets. But I digress.

It may have simply been a bad tube. It also may be something less obvious. There are a couple of popular primary impedances for reverb transformers. IIRC some are 18k and others are 23k, but reports range from as low as 15+k to as high as 28k. The reverb driver in a AB763 topology works very hard all the time the amp is on. If you happen to have one of the lower primary transformers in an amp with high-ish operating volts AND are using a schematic that specs a cathode resistor which creates a hot bias then it could just be a matter of mish mash parts and design parameters. That is, if the particular model that you built tends to be a little hot on the reverb driver with a 23k primary transformer and you happen to have purchased an 18k primary transformer then the tube will be dissipating just a bit more, won't it.?.

You might just need to cool the bias a little. Consider that the reverb driver tube IS operated as a power tube. Did you check bias and current before the failure?

Hi Chuck, re bad tubes, if a tube is working OK, do they usually go bad right away, or could they go bad any time? This tube has been running for a few months now, although I don't play the amp all that much. Most, a few hours a month or so. I ordered new tube + extra, will check voltages. As mentioned I think V is high in this spot.

[nosaj]

Hi Chuck, re bad tubes, if a tube is working OK, do they usually go bad right away, or could they go bad any time? This tube has been running for a few months now, although I don't play the amp all that much. Most, a few hours a month or so. I ordered new tube + extra, will check voltages. As mentioned I think V is high in this spot.

Think of it like a tire that may have issues. it could blow out today or it could be 3 years from now. How you drive it can influence how much faster or slower it goes. A tube is just like any other mechanical device.

nosaj

[Leo_Gnardo]

You might just need to cool the bias a little. Consider that the reverb driver tube IS operated as a power tube. Did you check bias and current before the failure?

Can't hurt. I've seen Fenders with 2K2 resistors in the Rk position on the reverb drive, and converted many to this value.

Bypass cap optional. And in that statement, there are options. You may choose a cap that's smaller than spec in order to "voice" the reverb drive, choose to accentuate higher frequencies & roll off lower ones. As in the preamp stages, the typical 25 or 22 uF bypass is a wideband choice.

[mikepukmel]

Can't hurt. I've seen Fenders with 2K2 resistors in the Rk position on the reverb drive, and converted many to this value.

Bypass cap optional. And in that statement, there are options. You may choose a cap that's smaller than spec in order to "voice" the reverb drive, choose to accentuate higher frequencies & roll off lower ones. As in the preamp stages, the typical 25 or 22 uF bypass is a wideband choice.

OK thanks, work to do. I will double check that it is a 2.2K in that spot. Pretty sure it is (I usually check with ohm meter before installing and soldering in). For power tubes, check bias by measuring the voltage drop across the cathode cap, and measuring the plate voltage. OK to do that here? It still plays, so I can get quick measurements, hopefully it doesn't go completely out and blow something else up.

Dang thing is that it played very well for a while after fiddling with the grounding.

[mikepukmel]

One other question for you before I put my safety goggles on: is it safe to switch the 12AT7 in the reverb driver with the phase inverter and see what happens? If the tube is arcing inside, will that risk taking out other components?

[mikepukmel]

'Fired' up the amp, crackling worse now.

Edit OOPS:

Plate voltage on the driver = 429.2v
Measured resistance across cathode res = 2.194k ohm
Measured voltage drop across 2.194k resistor = 8.96V (was a little jittery but close to 8.9 and change)

I = 8.96v/2194 = 0.004A

Plate voltage 429.2v half the current across each half of tube.

P = (0.002)^2 * 2194 = 1.75/2 =
about 0.876 watts across each half.

[Leo_Gnardo]

So, half that across each side, 0.096a Can this be right 96ma per side? This can't be right.

Gotta be way wrong. 96 mA on an output tube would have its plates glowing orange. In a pre tube, white hot with the glass nearby melting into slag. That's not happening is it?

9V div by 2200R = 4 milliamps (lets forget multiple decimal point accuracy for now, ballpark figure is close enough) for both triodes, or 2 mA per plate / cathode.

[Enzo]

12AX7s and 12AT7s don't often burn up other parts when they fail. But even if they do, it would just be a resistor. A couple pennies.


Asked earlier, a 12AT7 is a 12AT7, right? "12AT7" is a set of performance specs, not a set of mechanical plans. A 12AT7 specifies the standard pattern of 9 pins in a circle, it does not specify how the pins are shaped inside the glass. If you read writings about comparing 12AX7s, you will hear about long plate versus short plate, different getter arrangements, different color plates, and so on. In fact we often identify the brand of a tube by its appearance.