Vox AC4TV - Adding a Screen Resistor?

[Bobwith2Bs]

Just got one of these little guys on the bench. It came in with the description of "having issues" and found that the volume would cut out intermittently. It ended up being a failed EL84 which then blew the heater fuse (160mA Slow Blow) as i was testing it.

New EL84 tube is in but the thing absolutely GLOWS when playing through the amp. I noticed there is no screen resistor in these amps, so i am now worried it is just going to blow through another power tube if i send it out the door in its stock state.

Do people typically mod these to have a screen resistor?

I saw a few threads here and elsewhere about mods for different tones or for the attenuator, but not much on just reliability updates.

[Masco]

I would suspect the 160mA fuse was not the heater fuse. How's the bias?
Coupling cap from previous stage leaking?

[Bobwith2Bs]

You are correct!

I misspoke and this is actually the fuse on the center tap of the windings for the B+ supply. it is listed as 120mA in this schematic; however, the combo version of the amp that i have had a 160mA fuse in place which matched the associated screen printing on the PCB.

All other voltages are within the ballpark of what i would expect so I would be surprised if there is a leaky coupling cap or something. At idle all seems good, but the screen REALLLY lights up when signal is going through it.

I know these small EL84 based amps tend to run the power tubes at, or beyond, their data sheet capacity but am wondering if this is worth modifying before handing back to the customer or just chalking it up to being how the amp was designed.

[cdemike]

I don't think a 470R screens resistor would hurt. It may just be worth explaining the situation to the owner and determine if they're willing to keep spare EL84's on-hand if they want to push the amp's output section. I'm guessing the answer will to go ahead with installing the resistor since it's such a small amp.

To your point about over-dissipating the output tube, I noticed the factory manual specifies 10.2V at the cathode side of the 220R bias resistor, which translates to 14.7W total idle dissipation. Parsing that further by subtracting the current drawn through the preamp taking the drop across the 22K resistor following the screens node leaves 3.3mA flowing through the 470R resistor separating the reservoir and screens node, so the screen is dissipating a little over 1W at idle. Subtracting 3.3mA from the 46.36mA cathode current leaves 43.06mA of plate current, making for a total plate dissipation of 13.65W. That's toastier than I'd like for the screen grid at idle and obviously above the datasheet maximum for the plate dissipation, so I agree with you and would also be inclined to bias it a little cooler if it were my amp.

[Bobwith2Bs]

Thanks for providing the numbers to back up my suspicion!

And you are totally correct about putting the decision on the owner of the amp. If this were my amp i would not be opposed to doing the repair and dumping some time into it.

The fun part is this is a shop owned amp, so they are going to have to decide if it is worth paying me for more labor to do the fix or just putting this thing out and warning customers about how hard these amps push their power tubes.

[nuke]

I might suggest increasing the value of R16 instead of adding a screen resistor. Maybe double R16 to 1k or 1.2k even and see how your screens are doing. Most likely, reduce the screen draw and still supply a nice stiff (de-coupled by C14) screen for good sound.

That may also reduce your plate current as well, since lower screen will reduce the current through the tube. I would probably limit a screen resistor to 100 ohms though. Use a metal oxide flame-proof.

Check C8 as well, it is under biased and subject to relatively high AC current in that spot. Cheap enough to replace with a low-ESR, 105c long life capacitor. Probably cost all of a dollar.

[Olomert Brever]

I'm not sure its going to be very easy to add a screen grid resistor. I just modified a couple of these for a player. Does your PCB look like this?

20250227_091027.jpg

[nuke]

Yeah, that's gonna be hard, but not impossible. You only need a 1/2 watt 100-ohm if you really, really want to add one.

Otherwise, just increasing the power supply resistor to the screen stage a bit will do the trick.

[Bobwith2Bs]

I'm not sure its going to be very easy to add a screen grid resistor. I just modified a couple of these for a player. Does your PCB look like this?20250227_091027.jpg

Yep! That is the same PCB. Definitely would take drilling a hole or two.

This is a customer amp who basically just wanted it running so that they could sell it.

I have warned them about the current situation (it has, and will continue to, eat EL84 tubes until the design is fixed) and they have said they will pass that info along to buyers.

Sounds like I am off the hook for doing any mods on this one, but it is an interesting problem. I know the Fender Pro Jr amps run EL84s very hot as well. Do these power tubes just not sound good at datasheet specs? Why do companies keep doing this?

[nuke]

The Blues Junior and Pro Junior run push-pull, pretty hot, and get about 15 watts RMS.

There are some data sheets that suggest running 6BQ5A tubes pretty hot, and you can. The king of the type was the 7189A, which was a very stout tube, high ratings, allows 440V on the plates and 400v on the screens, and 14 watts plate.

The basic 6BQ5/EL84 datasheet conforms to the original spec, and in theory, you shouldn't exceed it. But, it was used as a vertical sweep output tube in television, and as TV screen sizes got bigger, the TV makers pushed for higher ratings. That's where the 10BQ5 came from (same as 6BQ5/EL84, but 10v filament for series wiring) and the 6GK6 (different pinout, higher dissipation). In TV it's more of a switch to make the sawtooth at 59.9hz for vertical sweep.

The practical 6BQ5 as time evolved, got better than the data sheet says, but you can't count on it being true. Unless, you and your purchasing department have a secret handshake with the tube makers.

So you can usually get away with brutal use of the EL84/6BQ5 and they were historically cheap, so, burning them up, not concerning. The guitar amp world was obsessed with POWER and watts sold amps, back in the day.

In the typical amp you can buy or build, they tend to run well over the 300v max rating and as a result, to get the right class AB1 operating point, you have to run them right up to the plate dissipation max.

If you designed an EL84 amp with a maximum 290v plate, 8k plate-to-plate load, 250v screens, you'd get about 10-12 watts of clean power. But 15 watts RMS power sells better than 10-watts RMS power. But the tubes would last long time.

[Olomert Brever]

It really is tough to know for sure you are shipping an amp with reliable power tubes, especially since 2022. I have started burning in new tubes on a device I made for myself where I can monitor screen, plate, cathode, and grid current before installing them in customers' amps (when time permits). If the amp is healthy all you can really do is test it a bit and send it.

[Bobwith2Bs]

It really is tough to know for sure you are shipping an amp with reliable power tubes, especially since 2022. I have started burning in new tubes on a device I made for myself where I can monitor screen, plate, cathode, and grid current before installing them in customers' amps (when time permits). If the amp is healthy all you can really do is test it a bit and send it.

I have been thinking of building something similar!

Do you send a test signal through the tubes or just apply plate voltage, or something else entirely? Any fault protection built in for the situations where tube DOES fail on the burn-in device? I may need to start a new topic just to bug y'all about the best path forward for something like this.

[Bobwith2Bs]

The Blues Junior and Pro Junior run push-pull, pretty hot, and get about 15 watts RMS.

There are some data sheets that suggest running 6BQ5A tubes pretty hot, and you can. The king of the type was the 7189A, which was a very stout tube, high ratings, allows 440V on the plates and 400v on the screens, and 14 watts plate.

The basic 6BQ5/EL84 datasheet conforms to the original spec, and in theory, you shouldn't exceed it. But, it was used as a vertical sweep output tube in television, and as TV screen sizes got bigger, the TV makers pushed for higher ratings. That's where the 10BQ5 came from (same as 6BQ5/EL84, but 10v filament for series wiring) and the 6GK6 (different pinout, higher dissipation). In TV it's more of a switch to make the sawtooth at 59.9hz for vertical sweep.

The practical 6BQ5 as time evolved, got better than the data sheet says, but you can't count on it being true. Unless, you and your purchasing department have a secret handshake with the tube makers.

So you can usually get away with brutal use of the EL84/6BQ5 and they were historically cheap, so, burning them up, not concerning. The guitar amp world was obsessed with POWER and watts sold amps, back in the day.

In the typical amp you can buy or build, they tend to run well over the 300v max rating and as a result, to get the right class AB1 operating point, you have to run them right up to the plate dissipation max.

If you designed an EL84 amp with a maximum 290v plate, 8k plate-to-plate load, 250v screens, you'd get about 10-12 watts of clean power. But 15 watts RMS power sells better than 10-watts RMS power. But the tubes would last long time.

This is some great information! Really appreciate you giving some perspective.

I have somehow never owned an EL84 based amp so i only ever really interact with them when they come in for repairs and they ALWAYS seem to be running the power tubes so hot - which then leads to a failure that i am addressing.

[wpaulvogel]

It has a 470 ohm screen resistor. This is a Class A amp and you would need to monitor screen voltage drop while in a dynamic operation with your meter and a dummy load to confirm operating points before doing any type of modification. Idle operation has nothing to do with dynamic operation and plate and screen dissipation is designed into the amp for when it’s being played. All this time spent assessing amplifiers during idle operation is useless. It must be tested with a signal passing at all situations between idle to full power.

[Olomert Brever]

It really is tough to know for sure you are shipping an amp with reliable power tubes, especially since 2022. I have started burning in new tubes on a device I made for myself where I can monitor screen, plate, cathode, and grid current before installing them in customers' amps (when time permits). If the amp is healthy all you can really do is test it a bit and send it.

I have been thinking of building something similar!

Do you send a test signal through the tubes or just apply plate voltage, or something else entirely? Any fault protection built in for the situations where tube DOES fail on the burn-in device? I may need to start a new topic just to bug y'all about the best path forward for something like this.

I have found that regular tube testers are great but do have some deficiencies. For my test device, I started with an old Hammond chassis that I found in my shop. (I have no idea where this came from. Maybe an ex-employee or intern left it?) I used a Hot Rod Deville OT from my used parts bin, a generic used Fender 60 watt replacement PT from the same bin, and a small PT that can supply about 4 amps of 6.3 volts for filaments and a secondary that can power two 12AX7s with about 310 volts. This third transformer is supplied with its own IEC and power switch. I use two variacs. First I set the filaments to 6.3 volts with the small PT from one variac and leave this be. Next I power up the used Fender PT with the second variac and adjust it to get the B+ I want. Usually this is close to what the B+ is in the amp I am working on and will be installing the tubes I am testing in. First I bias the power tubes and let the thing sit for 10-90 minutes. I periodically check for any drifted voltages and tap the tubes with a chopstick to listen for noise. Then at some point I test play the amp with a guitar or bass. The play testing gives me an idea of how the tubes sound. Sometimes tubes will test ok on a standard tester but not sound good or crackle or something when in a working audio circuit. I have 1 ohm resistors at each anode, screen, cathode pin on the power tube sockets with test probe jacks on either end of the resistors so I can measure current through them. On the control girds I had test jacks on either end of the 100K grid leak resistors to measure if a tube is pulling an unusual amount of grid current. I currently have these grid jacks unhooked but plan to reconnect them later this year.

Right now I am not able to test EL84s as I had stolen some parts for a rush repair but plan on rectifying this before the end of the year.

As an example of utility, in 2023 I was building a JTM45 clone for a man. He sent me some NOS KT66s. I was skeptical of these so put them on my tester. They behaved well for about 90 minutes but after that, 2 of the 4 of them would slowly begin pulling more and more screen current to the point of failure. I am thankful I took the time to test these tubes and save any problems they may have caused in my customer's new amp.