Increasing filtering in 5F4

[RobBozic]

I've built a 5F4 with the PaulC mod, Ragin Cajun speakers, JJ6L6's. At low volume it sounds great, but when you crank it sounds like there are some ghost notes trailing each note. The layout is nice & tidy, therefore I don't think it's a bad solder joint etc..

Would increasing the first filter cap value to say 40uf help decrease any possible ghost notes etc..?

Thanks
Rob

[loudthud]

The 5F4 schematic shows a full Pi B+ filter. Is that what you have? Ghost noting is usually attributed to poor power supply filtering so 40uF will probably fix it right up. How is the filtering on the bias supply?

[RobBozic]

'Loudthud', I copied the Ceriatone layout because it was easy to read. The filtering on the bias is 100uf.

Here's a link:http://www.ceriatone.com/images/layo...one_100508.jpg

[loudthud]

The Ceriatone layout is true to the original as far as schematic is concerned. Many people use the more conventional approach of connecting the OT CT to the first filter cap. Another change that is frequently made is the addition of 'Screen stoppers', a 470 ohm 1W resistor in series with each 6L6 pin 4.

I guess the next question to ask is what did you use for a choke? Some chokes designed for the typical Fender application might not be able to handle the current demand of a 5F4. Current rating might be too low and resistance might be too high. If the choke saturates, lots of ripple gets to the screen grids and that will cause ghost noting. So you might try just moving the OT CT to the first filter cap to side-step that issue.

Another issue is the Paul C mod. In some amps that can really hammer the headroom. This is usually not described as ghost noting, it's a warm tone with low volume. Try increasing the 2.2 meg resistor to 4.7 meg or for trial purposes lift one end of the 2.2 meg and add two or three 1 meg resistors in series with it.

[RobBozic]

'Loudthud',

Mine's wired as per the original with regarsds to the choke & OT centre tap.

I've also added the 470R resistors to the screens.

The choke is a Ceriatone unit, it's very big for a choke so I assume it's the correct unit.

Regarding the PaulC mod, I wired the amp with it from the start so I never did a comparison against the original PI schematic. I've read it smooths out the tone at high volumes & increases headroom, what's your impression of it?

Thanks again
Rob

[loudthud]

Back to the original problem, speakers can cause ghost noting. Have you tried another speaker?

The choke should be OK but try moving the OT CT to see if that fixes the ghost noting.

The 470 ohm resistors will improve the life of the output tubes, but they are usually not associated with ghost noting. However, power supply ripple on the screens can cause ghost noting so that's why you should try separating the plates and screens.

The PaulC mod doesn't work well with 12AX7s. It sounds 'warm' because it clips the top part of the waveform on the cathode side. It seems to increase headroom because you have to drive the PI harder to get hard clipping of the output tubes. If you have an oscilloscope, you'ld see that you actually get less clean power, especially when driving 6L6's. It's not bad conceptually, try the little change I suggested and I'll bet you hear the difference immediately.

Another headroom bottleneck in the 5F4 could be the cathode follower stage (the middle preamp tube in the layout). The resistor connected to pin 3 should be reduced to 1K or 820 like in the 5F6. I don't know if it will make an audible difference.

[RobBozic]

Thanks again 'Loudthud',

I'll move the OT CT to the first filter cap as per your suggestion, then I'll mess with the PaulC mod to see if I like it better stock.

Regarding the 1.5K on V2a (3), reducing this to 820R will increase gain, as I've already running a 12AX7 in V1 this might be overkill, ..then again I might like it better.

The speakers are brand new Ragin Cajun's, so I don't think I'll need to replace those.

Thanks for your help, I'll let you know how it turns out.

Rob

[loudthud]

Although reducing the cathode resistor will increase the gain slightly, the capacitor bypassing it does the bulk of the work and the feedback around the stage stablizes the gain. The increase in headroom is the main benefit but that's probably not a huge problem in your amp.

Speakers can get damaged by rough shipping. Installing speakers on a warped baffle will sometimes cause problems. I got a pair of 'like new' Eminence speakers on ebay that the owner didn't like for some reason. I put them in a cabinet with them wired in series and... nothing. One had an open voice coil. The previous owner had wired them in parallel and didn't notice one of them was dead. When something unusual is going on, it's best to verify each component in the signal chain by direct substitution, try it (them) with another amp. You are right that the chances they are bad is probably less than 1%. I just thought I mention the possibility.

[RobBozic]

Loudthud, I did some tweaks on the weekend and ended up removing the PaulC mod, I also beefed up the filtering to the screens as per the 5E8A and I left the OT centre tap as per the stock arrangement. I also installed a GZ34.

To me the amp sounds better now. I tried increasing the first filter cap to 40uf but I felt it stiffened the amp too much, even with the 5U4GB.

Thanks for your help.
Rob