Non top boost AC-30 in for repair.

[Colossal]

I tried a different set of output valves and much the same.

I thought the bright channel sounds very shrill through my 18 watt Marshall cab (which in fact was made with the same dimensions as an AC-30 cab, the cab is loaded with a Celestion Blue and a Celestion H30. I have seen this combination in one of Brian May’s AC-30’s).

Hello Mark,

I've wondered about this combination of speakers. Are you using the Celestion H30 Anniversary edition?

You have probably seen this from Greg Fryer's site:
https://fryerguitars.files.wordpress.co ... an-may.png

I have removed the three zeners and put a 75 ohm resistor in its place. The rail dropped to 315VDC, I have 307VDC on the EL-84 plates and 11.3VDC on the cathode. The valve 29mA of current going through it per a 1 ohm resistor mounted between the plate and the output transformer primary. I estimate that the valve is dissipating 9 watts which is a good spot for a class A/B amp. I’m using a 100uF bypass cap at the moment as it is what I had on hand. I can’t say I’ve noticed a massive tonal difference between the zeners and the resistor and cap.

My experience with my Trainwreck Rocket was that I’d need much larger cathode resistor to stem the current flow. So this amp has me a little perplexed.

What are your thoughts on current flow through the output stage of an AC-30?

I'm sorry for not replying about the zener bias condition. I have been moving the past couple of days and getting my place sorted. That is peculiar. I used the triode grid curve method by Kuehnel, shown here in his analysis of the AC30, to estimate the tube currents:

https://www.ampbooks.com/mobile/classic ... /vox-ac30/

He says it provides a reasonable approximation where the screens and plates are very close together in voltage. I was going to ask if you were planning to replace the zeners with a resistor, and you have, to a good outcome. Your new rail voltages seem spot on and the PT is sagging nicely. What is your Mains voltage at these settings? According to my calculator, with your new conditions of 11.3VDC cathode, 307V plates, and bias at 75 ohms, it is estimating 151mA for a quad (37.7mA single tube), 11.1W (per tube) and 93% of max dissipation, so your results are fairly close. Can you put in a 1R across the rectifier to B+1 to measure total current drawn by the amp to see how this compares to what you are getting with a single tube measurement? Thanks for doing all this work. It will help me greatly with my own AC30/4 build. I have estimated the power amp conditions. I will be using a Mercury Haddon clone PT, choke, and OT and took secondary resistance and inductance measurements on both to see how the predicted values for B+ will compare to the as-built. The original PTs were wound for 115-160-225-245V primaries. The Mercury runs 100-117-220-230-240V with a 285-0-285V (unloaded) secondary. My mains run hotter at 123VAC, so using the 117VAC primary, I am planning to add some additional series resistance to each leg of the secondary and a small value, post rectifier, if needed to get the rail where I want it.

[Mark]

Thanks for taking time to reply to me. It turns out the sag was a weary GZ-34. I put another GZ-34 in and my rail is back to 354VDC, the cathode voltage is 12.43VDC, the plate voltage is 347VDC, and plate current is 37.8mA. The valve is dissipating 12.6 watts with a 75 ohm cathode resistor.

I will try an 82 ohm cathode resistor that is good for at least 25 watts. The 75 ohm one I used got warm/hot, I was a bit surprised actually.

I’m planning on an AC-30 build too. Unfortunately, my Mercury transformers have a secondary of 303-0-303, so these are the voltages I’ll be dealing with. I’m thinking of using the Zener from the centre tap to the chassis trick to dump some voltage. I’ve spoken to Patrick at Mercury and Don Butler about it. He likes the higher voltages for clean tones but advises against driving it hard.

I would have bought the lower voltage Mercury power transformer if I had known more. One of those live and learn situations.

EFF63564-880F-4922-ABE3-BC7780118BB0.jpeg

[Colossal]

Thanks for taking time to reply to me. It turns out the sag was a weary GZ-34. I put another GZ-34 in and my rail is back to 354VDC, the cathode voltage is 12.43VDC, the plate voltage is 347VDC, and plate current is 37.8mA. The valve is dissipating 12.6 watts with a 75 ohm cathode resistor.

Good to know about the GZ34! The voltages are similar to an 18W Marshall. I'm presuming your screen voltage is a little lower with the drop across the choke and your 270R screen resistors. You could add a 25-40R current limiting resistor, one to each leg of the secondary, to add to the total series resistance the GZ34 sees. This would drop the rail more as well as limit any hot switching current (if the amp as a standby switch). A GZ34 generally has an ra of 50R.

I will try an 82 ohm cathode resistor that is good for at least 25 watts. The 75 ohm one I used got warm/hot, I was a bit surprised actually.

The power the 75R was dissipating is (0.0378A+ ~0.005A) * 12.43V = 0.53W x 4 tubes = 2.12W

I’m planning on an AC-30 build too. Unfortunately, my Mercury transformers have a secondary of 303-0-303, so these are the voltages I’ll be dealing with. I’m thinking of using the Zener from the centre tap to the chassis trick to dump some voltage. I’ve spoken to Patrick at Mercury and Don Butler about it. He likes the higher voltages for clean tones but advises against driving it hard.

I would have bought the lower voltage Mercury power transformer if I had known more. One of those live and learn situations.

Mercury's secondary voltages are given as unloaded. I would guess you would get 303 * ~1.17 (sag factor) = 354VDC rectified. Again, additional series resistance can be added to the secondaries, and a small value post rectifier to massage the B+0 rail voltage. Yeah, Mercury seems to select for high rail voltages as their sonic preference.

[Mark]

I had considered the resistors too, but the size due to the wattage would be a bit cumbersome. What wattage do you estimate will be needed?

It should be a rule of thumb to always order the lower voltage Mercury power transformers. I have their kit for the Fender Champ 600, I never used the power transformer as it is 115/120, and the output from it is insanely high for a class A amp. It measured 445VDC on a 6V6!!!

[Colossal]

I had considered the resistors too, but the size due to the wattage would be a bit cumbersome. What wattage do you estimate will be needed?

So from Blencowe, for a tube rectified amp, the total RMS ripple current is about 1.5 times the DC current. We'll say your amp is pulling 170mA of current. I measured my Vox PT to have a source resistance of about 35R per leg of the secondary, so will add another 45R. Using 45R of added resistance to each secondary, this brings Rs to 80R (per the GZ34 datasheet, Rs = Rsec + (Vpri/Vsec)^2*Rpri). I won't be including a standby and hot switching, but my mains run 123V on a 117V primary so I will want to dump a little voltage off the secondary. Using 45R of added series resistance:

P = I^2*R
P = (1.5 x 0.170A)^2 * 45R
P = (0.255)^2 * 45
P = 2.92W

A 10-25W wirewound chassis mount resistor would be fine and run cool. I used 50R 25W in a Rocket type build once running 300-260-0-260-300; worked great.

It should be a rule of thumb to always order the lower voltage Mercury power transformers. I have their kit for the Fender Champ 600, I never used the power transformer as it is 115/120, and the output from it is insanely high for a class A amp. It measured 445VDC on a 6V6!!!

Yeah, I always read the fine print. I get the specification sheet for the transformer beforehand to make sure what I am getting, as well as present them with my scenario and expectations for a build. I ran a 2x6V6 Plexi (Mark Huss design) at that plate voltage (used a Vibroverb PT), but the screens were supplied by a 10H 280R choke with 5k in series to knock down the B+2 voltage. Sounded great!

[Mark]

I haven’t replied yet as I haven’t had anything meaningful to say. I will have to give it some thought.

One thing I have wondered about having the brilliant channel and the normal channel working together is do they interact?

I found if I plugged into one channel and turned up the other channel. Nothing happens till I get to 12:00 on the pot, then there was an increase in volume, the brilliant channel sounded more like the normal channel. Plugging into the normal channel and turned up the brilliant channel, the volume did increase but not to the same degree as the other way around.

The reason for this could be the phenolic board being conductive, the filter caps may not be smoothing well enough, there might be cross talk within the valve itself.

I will look into this further.

The reason I wanted to try this is, I wish to build an AC-30, and I’d like to add a EF86 channel, but I wondered how this would effect the rest of the amp.

What are your thoughts?

[Mark]

I experimented a little more today and I’m getting amplified signal when all the volume controls are turned down. The normal channel and the brilliant channel are effected. If I plug into the brilliant channel I can turn up the normal channel and hear the guitar but at a much quieter volume.

I haven’t looked into it a lot yet. It could be valves, or power supply, or the board itself.

Any thoughts or experiences to share?

[pdf64]

Depending on the particular version, those channels share cathodes (later treble models don’t). If so with yours, that another opportunity for crosstalk; full bypassing there is necessary to prevent it, so perhaps there’s a tired cap with excessive esr?
What fixed the vibrato?

[Mark]

Thanks for your reply, I had a open circuit 180K resistor, and the footswitch was replaced with a female jack and it was very poorly wired up.

I will have to check the cap in the preamp but I suspect that I replaced it.

[Mark]

The plot thickens with the problem of bleed through coming through the amp. I decided to remove the 220K resistors from the wipers of the volume pots to the P.I. stage.

To my surprise, the volume of the leaking signal increased when I removed the two resistors (any input to the P.I. stage.) God only knows how the signal is getting into the P.I. stage?

I’ll keep plodding along with this one.

6352777B-68F2-43BA-BA24-8D258480DE79.jpeg

[Mark]

I replaced the stryo 470pF cap with a SoZo 500pF cap, it sounded a lot better than the stryo cap that was in the amp.

The bleed through is still occurring and I found out that the last tech who worked on it couldn’t fix it. The earth buss bar had been cut near the normal volume pot.

I’ll have to start again with the bleed through.

[Mark]

While I didn’t think to add the DC voltages for this amp. It might be a good idea to mention them.

Power supply nodes

B+ 346vdc, screen node 340vdc, PI stage node 308vdc, brilliant/normal channel node 203vdc.

Output valve measurements.

plate 341-342vdc, screen grid 335vdc, control grid 80mV to 100mV, cathode 12.3vdc.

PI stage

Plate 242vdc, cathode 61vdc, 40/42vdc

Brilliant channel

Plate 203vdc, grid 0.0vdc, cathode 1.76vdc

Normal channel

Plate 202vdc, grid 0.2mVdc, cathode 1.76vdc.

I don’t see anything out of the ordinary except the tiny voltage on the normal channel grid.

[Mark]

I will add to this as I keep fault finding.

If I input 301mVpp into the brilliant channel I get an output of roughly 19vpp, and I get a distorted 64mVpp from the normal channel.

EF9E09E2-0598-48DB-874A-3B68B22429FD.jpeg

As I have a small DC voltage on the normal grid, I thought I would remove the input from the grid and earth the grid. As shown the waveform is larger and cleaner.

6283CDC5-7930-47F3-93BF-238B3DA3DDAF.jpeg

Have reattached the input to the grid and removed the bias resistor from the normal cathode so it is unable to amplify. I have 48mVpp on the plate. The voltage isn’t present on the power supply node.

BE2E5377-C017-42BB-AB28-C0AF8C6ABDB0.jpeg

I decided to lift the plate resistor to see if that killed the signal. It did, but it’s strange, I have 5mVpp on the plate which what I saw on the input of the PI stage, and I mentioned that I disconnected the wipers from the brilliant and normal channels and the voltage was still present on the PI cathode and grid leak resistors.

9C6E0251-2A30-4DC2-B7C3-F1469629C569.jpeg

The output on the Plates is about 60mVpp, I will try putting a short across the 1M grid leak resistor to see if the output can be stopped.

B7054777-7C70-448E-A2D3-1909A1B3F21A.jpeg

At this point it is hard to know if I’m seeing symptoms of one fault of two different faults.

Any thoughts and observations are welcome.

[pdf64]

I suggest to try a different tube; dunno what the failure mode is, but some ECC83 seem have an excessive degree of coupling between their sections.
Also check for ESR in the HT node and cathode decoupling caps - the second scope photo may indicate that the latter is higher than ideal.

[pompeiisneaks]

I haven't read the entire thread... do you get fully 0 ohms resistance to ground on the individual channel volume pots when you rotate them all the way down but bleed is still happening? I know I've heard of pots that don't actually get to a true dead short at lowest setting, and this can mean due to 'some resistance to ground' they also allow some signal to pass through.

Probably an obvious thing you've already tested, but just test with resistance or continuity from wiper to ground at the lowest setting of both pots.

~Phil