Hoffman Amplifiers Tube Amplifier Forum
Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: Yeatzee on October 25, 2024, 04:30:11 pm
-
I've got a friends AC30 C2 model in right now, first brought to me because it blew the HT fuse and had a JJ EL84 go lightning in a bottle
All good, replaced the HT fuse he bought a new quad and the amp continued to run fine for almost 2 months or so with regular weekly playing. Then, out of nowhere, he turns it on at rehearsal and bad noise fuse blown same deal. Same exact tube socket as well which had me really scratching my head.
He brought it back over, I tossed in a spare EL84 in that same tube socket replacing the bad tube and everything seems to test fine from what I can tell. Amp ran no problem, sounded fine, biased a hair over 100% dissipation as you'd expect with the stock 50R resistor. I started doing some research however because it just seemed so odd for this to happen, and I found a thread on here talking about how the spade connectors for the OT that go to the PCB with the tube sockets can sometimes arc and give voltage to the grids resulting in the tube dying like his did. In that thread there were photo's that clearly show arcing so I looked closer at the amp and there's nothing, no signs of any damage and all of the resistors off the tube sockets measure fine.
I measured all of the voltages on every pin of every socket, and they all matched eachother. The "potential" problem socket shows no difference between the others, so now I'm feeling like I'm in a weird place where I don't really trust the amp but can't find anything wrong. What are the odds two tubes ended up dying in the same socket, and it's purely the tubes. Just a total coincidence they happened to be in the same socket...?
With my spare EL84 in the amp, keeping the other 3 that it came with in it I let the amp run both idling for a couple hours as well as running a signal generator through it for an hour (just a constant tone) to see if I could get anything to show itself as faulty and neither test revealed anything. So I put the amp chassis back in the cabinet hooked it's speakers back up and played the amp, again sounded great no problems. I re-measured everything and then just let it sit for a while idling. I noticed the grid voltage had climbed somewhat. I don't remember off hand what it was, still mV range but it had definitely climbed a touch on ALL of the tube sockets. Idk if that's normal or not. But if there is a touch of DC getting to the grids would that degrade the tubes over time eventually resulting in this behavior of catastrophic tube failure randomly after working for a long time? Or is this a red herring?
Appreciate it! Don't want to give the amp back and have it kill another set of nice tubes if I'm missing something. Thanks!
Edit: I'm back home so I re-measured and I'm getting about 90mv dc on pin 2 of all of the sockets. And seems to be increasing slightly. I'll check back after an hour and report back where it's at.
-
Alright well now I'm more confused than before. I let the amp idle again for another couple of hours and the grid DC voltage stabilized to about 30mv dc. So I thought ok all is well, seems to all check out ok I guess but I looked at the tubes and realized the two middle tubes clearly got a lot hotter based on the TAD stickers.
(https://i.imgur.com/AMRmKUG.jpg)
So what am I missing here? All 4 tubes show the exact same cathode voltage, every pin reads basically identical voltages across all 4 tube sockets.'
-
One or more EL84 sockets may be defective and need replacing. Also re-solder the sockets.
Also check the 470 ohms.
The same thing happened to me; a contact in a socket was burnt and blew the fuse and sometimes even the tube.
-
One or more EL84 sockets may be defective and need replacing. Also re-solder the sockets.
Also check the 470 ohms.
The same thing happened to me; a contact in a socket was burnt and blew the fuse and sometimes even the tube.
Yeah the screen and grid stop resistors all measure fine, so odd. When you say defective, what would the defect be exactly? I don't really understand what it could be if the voltages are measuring fine, I feel like there's something super stupid I'm missing here. Not opposed to swapping sockets, I just want to know what's going on
-
When you say defective, what would the defect be exactly? I don't really understand what it could be if the voltages are measuring fine, .....
The same thing happened to me; a contact in a socket was burnt and blew the fuse and sometimes even the tube.
Those modern 9 pin ceramic sockets with the y fork pins that grab the tubes pins are often junk. I don't trust their pins. You can't see down in there to see whats going on like the old cup 9 pin sockets. We saw 1 here the pin was broken inside the ceramic, couldn't see it. I won't use them! :BangHead: :cussing:
I built amp, had tremolo, oscillator would not oscillate, it was the modern 9 pin ceramic socket. (My solder joints were fine.) I only found out it was the socket because I changed it.
Belton's very good 9 pin sockets. Belton probably also makes those sockets for using EL84 tubes with base pins for PCB.
-
When you say defective, what would the defect be exactly? I don't really understand what it could be if the voltages are measuring fine, .....
The same thing happened to me; a contact in a socket was burnt and blew the fuse and sometimes even the tube.
Those modern 9 pin ceramic sockets with the y fork pins that grab the tubes pins are often junk. I don't trust their pins. You can't see down in there to see whats going on like the old cup 9 pin sockets. We saw 1 here the pin was broken inside the ceramic, couldn't see it. I won't use them! :BangHead: :cussing:
I built amp, had tremolo, oscillator would not oscillate, it was the modern 9 pin ceramic socket. (My solder joints were fine.) I only found out it was the socket because I changed it.
Belton's very good 9 pin sockets. Belton probably also makes those sockets for using EL84 tubes with base pins for PCB.
Like this?
https://www.tubesandmore.com/products/socket-9-pin-34-mounting-hole-pc-mount
This is the only one I could find, I'm not near the amp right now but I don't believe this would fit based on photo's.
-
Look inside the contacts in the socket, if necessary with a magnifying glass.
Replacing a socket is not easy. To do so unnecessarily would be disappointing.
-
Most likely the power tubes are drawing control grid current or coupling caps from phase inverter are leaking dc. Take out the power tubes and measure if there's any DC voltage on pin 2 of the power tube sockets. If there's not then the coupling caps are fine and you'll need a new set of power tubes.
-
... All 4 tubes show the exact same cathode voltage ...
It's a shared cathode resistor isn't it, so that has to be the case, as all cathodes are wired together? It certainly doesn't mean they're all passing the same cathode current though.
If the EL84 sockets aren't on a pcb, I suggest to wire a 1R current sensing resistor between each cathode socket lug and the 50R bias resistor // bypass cap.
Then you can check that the valves are a reasonable match.
Otherwise use a bias probe to check each valve.
Upside down EL84 in an AC30 seems a really bad design.
Control grid current is related to valve dissipation and wear.
The hotter the cathode, the hotter the control grid, and it'll tend to emit a few electrons itself.
The more worn a valve, the more gas gets released from within the metal electrode structures.
Beware, a valve with excessive grid current can spiral into self destruction, EL84 / 34 seem especially liable to this failure mode.
And with parallel push pull amps such as AC30, a bad valve can kill its parallel wired neighbour.
-
Look inside the contacts in the socket, if necessary with a magnifying glass.
If those are ceramic 9 pin tube sockets, you can't see anything looking into the tiny hole in the socket.
-
Like this?
https://www.tubesandmore.com/products/socket-9-pin-34-mounting-hole-pc-mount (https://www.tubesandmore.com/products/socket-9-pin-34-mounting-hole-pc-mount)
This is the only one I could find, I'm not near the amp right now but I don't believe this would fit based on photo's.
Yes, like those. But it might not fit, have to measure the chassis hole and/or PCB hole spacing.
And you'll need some type of tube retainer/holder for the socket.
But try what PDF64 and Lauri said 1st.
-
Most likely the power tubes are drawing control grid current or coupling caps from phase inverter are leaking dc. Take out the power tubes and measure if there's any DC voltage on pin 2 of the power tube sockets. If there's not then the coupling caps are fine and you'll need a new set of power tubes.
No dc on the grids pin 2 of the sockets with the power tubes pulled. This is a new quad of tad el84's I just put in it.
-
... All 4 tubes show the exact same cathode voltage ...
It's a shared cathode resistor isn't it, so that has to be the case, as all cathodes are wired together? It certainly doesn't mean they're all passing the same cathode current though.
If the EL84 sockets aren't on a pcb, I suggest to wire a 1R current sensing resistor between each cathode socket lug and the 50R bias resistor // bypass cap.
Then you can check that the valves are a reasonable match.
Otherwise use a bias probe to check each valve.
Upside down EL84 in an AC30 seems a really bad design.
Control grid current is related to valve dissipation and wear.
The hotter the cathode, the hotter the control grid, and it'll tend to emit a few electrons itself.
The more worn a valve, the more gas gets released from within the metal electrode structures.
Beware, a valve with excessive grid current can spiral into self destruction, EL84 / 34 seem especially liable to this failure mode.
And with parallel push pull amps such as AC30, a bad valve can kill its parallel wired neighbour.
Unfortunately all of the sockets are pcb mounted. The issue is like a ghost issue to me, I don't understand how I can actually "see" it apart from the clear affects on the tube sticker being darker than the others showing it's running hotter. What can I do to actually detect the issue without messing with the pcb?
-
Unfortunately all of the sockets are pcb mounted. The issue is like a ghost issue to me, I don't understand how I can actually "see" it apart from the clear affects on the tube sticker being darker than the others showing it's running hotter. What can I do to actually detect the issue without messing with the pcb?
You could measure plate dissipation across the OT. There's lots of resources for that on the interwebs. Uncle Doug, Rob Robinette have detailed explanations. That would at least separate both sides of the transformer, which in your case if it was 1 bad socket, you could figure it out. That would also expose a bad trace in one side of the PCB.
You need some basic understanding of how the circuit works in order to utilize it.
-
Look inside the contacts in the socket, if necessary with a magnifying glass.
If those are ceramic 9 pin tube sockets, you can't see anything looking into the tiny hole in the socket.
Yes it is very difficult to see, ceramic or not.
One thing to do, is to clean them with good contacts cleaner
-
If the EL84 sockets aren't on a pcb, I suggest to wire a 1R current sensing resistor between each cathode socket lug and the 50R bias resistor // bypass cap.
Then you can check that the valves are a reasonable match.
Otherwise use a bias probe to check each valve.
Yes 1 ohm resistor is the way to go .
Or bias probe
Picture of EL84's pcb may help
-
As sockets are pcb mounted, it's usually feasible with at least one of them to cut the pcb track that carries only the cathode current of a single valve, and bridge the cut with a 1R.
Then the cathode current of each can be checked by swapping the valves around.
Obviously it's more long winded than using a 2 or 4 way bias probe.
... You could measure plate dissipation across the OT...
With shared cathode bias, using the OT half primary resistance to assess current can't check each valve of the quad.
-
As sockets are pcb mounted, it's usually feasible with at least one of them to cut the pcb track that carries only the cathode current of a single valve, and bridge the cut with a 1R.
Then the cathode current of each can be checked by swapping the valves around.
Obviously it's more long winded than using a 2 or 4 way bias probe.
... You could measure plate dissipation across the OT...
With shared cathode bias, using the OT half primary resistance to assess current can't check each valve of the quad.
Unless I'm misunderstanding, I'm not sure how that'd help without doing it to at least 2 of the sockets / PCB traces. The valves are a brand new quad, I'm not concerned with them being bad. If socket #3 is the problem socket, and I do that mod to measure cathode current off socket #3 swapping the tubes isn't really going to tell me anything right? I'd have to also do the mod to say socket #1 as a baseline to compare.
Unfortunately all of the sockets are pcb mounted. The issue is like a ghost issue to me, I don't understand how I can actually "see" it apart from the clear affects on the tube sticker being darker than the others showing it's running hotter. What can I do to actually detect the issue without messing with the pcb?
You could measure plate dissipation across the OT. There's lots of resources for that on the interwebs. Uncle Doug, Rob Robinette have detailed explanations. That would at least separate both sides of the transformer, which in your case if it was 1 bad socket, you could figure it out. That would also expose a bad trace in one side of the PCB.
You need some basic understanding of how the circuit works in order to utilize it.
Are you talking about the output transformer resistance method? Where you measure the voltage at the center tap and subtract the plate voltage from that and then get the resistance from CT to the plates and divide the voltage drop by the ohms? I've done that in the past, I can give that a try to see if it reveals anything
-
AC30 is notorious for eating EL84s. That single shared cathode resistor exaggerates differences between tubes; the strongest one will pull a little more current, raising the bias voltage, which forces the weakest one to run abnormally cool.
Anyway, it wouldn't surprise me if there was intermittent arcing or leakage between a spade /socket pin/ solder pad somewhere. I don't know if that model has clamping diodes at the anodes, but if not, during overdrive the anodes can swing to a kilovolt no problem, so any leakage across the PCB or socket could raise the grid voltage at the same moment, and poof!
-
A lot of AC30's also run the tubes too hot and have too high a dcv on the plates.
So run the modern tubes too hot/too high dcv that have 3 actual grids spaced close together, use crummy new ceramic tube sockets and use a PC board with the sockets mounted to the PCB instead of the chassis, add them all up, and what could go wrong? :dontknow:
-
Yep! AC30 is almost a masterclass in how to push tubes just close enough to the edge of destruction for them to outlast the warranty :wink:
-
Yep! AC30 is almost a masterclass in how to push tubes just close enough to the edge of destruction for them to outlast the warranty :wink:
:laugh:
-
Look, I get get that guys love the sound of an AC30, I love Rory Gallaher's early sound, an AC30 in Taste, then AC30 and an old tweed Bassman or tweed Twin together for Irish Tour '74. Just love that sound! :blob8:
But a lot of Vox amps have problems, they just do.
I should say; Rory always used a Rangemaster Treble Boost with the AC30 and he later used a Hawk II Tonal Expander/Booster with the tweed amps.
-
... The valves are a brand new quad, I'm not concerned with them being bad. ...
Sorry, I misunderstood, I thought you were concerned about 2 of the EL84 running hot; checking for that was all my suggestion was intended to address.
-
This is where having a pro-grade bias meter would be very handy, as it can tell you exactly how much current is flowing through each tube without going through all kinds of trouble with mods/work arounds, etc...It will allow you to have some really useful information quickly.
If you find that one tube is drawing more current, then switch the tubes around and see if it follows the tube. Also having an entire extra set of high quality matched tubes to sub in may confirm what you found with the first set...you will be eliminating one bit of uncertainty.
The only EL84s I trust at all these days are the Apex Matched, BURNED-IN sets from Amplified Parts. They have the best system for matching and burning in of all the tube suppliers. According to them, the burning in process weeds out just a few more of the extra sensitive tubes that will die quicker than the others...important in an amp like an AC30.
-
No dc on the grids pin 2 of the sockets with the power tubes pulled. This is a new quad of tad el84's I just put in it.
Try it with new quad of el84's and preferably some other brand than tad. Recently I've come across plenty of new tad power tubes that are drawing control grid current and have a very short lifespan.
-
I literally cannot find replacement tube sockets for this amplifier, I don't see anything that look right on any of the website I know to look at. Anyone know of a place that carries these PCB mounted 9 pin sockets that screw into the outside of the chassis?
Regarding the tubes, I'm doubtful there's anything wrong with the set. I've been running the same redplate EL84 quad in my vintage AC30 with no problems through tons of long gigs. Two other tubes died in the same place this one TAD is showing excess heat signs, it's the amp I'm confident in that I just still don't quite understand how.. short of spending another $120 on a bias probe.. I can measure the current at this socket and figure out where the issue is.
-
You could disconnect the connectors coming from the output transformer to the power tube pcb and jury rig 1 ohm resistors from the connectors to plate pins on the tube sockets. That way you can measure tube currents individually.
edit:Nevermind I just realized this won't work because plates on each side are connected together on the pcb.
-
I literally cannot find replacement tube sockets for this amplifier, I don't see anything that look right on any of the website I know to look at.
Where are you looking? They're easy to find here in the UK (Langrex, Rapid etc) but I'm not sure who the go-to suppliers are in your country.
-
Yeatzee - the thing the screws in from the front on the EL84s are the tube and socket retainers. Is the socket top mount? Have you unscrewed one retainer to see the socket underneath?
Do they look like this https://www.tubesandmore.com/products/socket-belton-9-pin-pc-mount-standoff (https://www.tubesandmore.com/products/socket-belton-9-pin-pc-mount-standoff) OR this https://www.tubesandmore.com/products/socket-9-pin-34-mounting-hole-pc-mount
Here's all the pc mount sockets they have https://www.tubesandmore.com/search/node/pc?categories%5B0%5D=Vacuum%20Tube%20Accessories (https://www.tubesandmore.com/search/node/pc?categories%5B0%5D=Vacuum%20Tube%20Accessories)
-
"...it's the amp I'm confident in that I just still don't quite understand how.. short of spending another $120 on a bias probe.. I can measure the current at this socket and figure out where the issue is."
There's no more "understanding" to it. You need a bias probe.
-
Super stupid of me, it's got a mount on the outside of the chassis that makes it look like the socket is different from the typical belton options. I've got a quad on the way to swap in and see.