Problem with this fizzy sound

[kleuck]

Sorry, i should have read the whole thread

[Silvio55]

It's ok man! at least it's good to see people giving ideas!

[Ross Grey]

I got up early this AM, cut the trace and added a 1Meg grid stopper to V4B. No detectable change in volume or tone as promised by Merlin. In OD, the sound is really smooth and there are plenty of even-order harmonics that want to sing out.

But with clean drive, after adding the grid stopper and re-balancing the PI, I still get the frizzy sounds when the 'gain' control gets past 10-11 o'clock.

On the other hand, it is amazing on how good this amp sounds when the PI gets unbalanced. My favorite configuration is (original) 56k at the plate and 65k at the cathode. With these values, I get huge clean, no frizz sound until about 2-o'clock on the gain control. This is plenty of control from high-headroom clean, to good clean compression before the amp starts to distort.

When distortion arrives around 2'oclock, it is very much like frizzy decay, but not too harsh now. In a performance setting it will never be noticed. At practice levels, you can tune this out with mock-triode mode (low power) or the tone control. The 'thick' setting is still way cool for OD.

So this is a better configuration, but the PI is still not balanced. Here are some thoughts on where to go next.

1) cut the input drive with a more Fender-like divider into the first preamp stage. Better yet, put a volume control here to adjust for different pickups and pedals. In almost any setting, I like cutting the volume on my guitar to improve the sound of this amp. This mod would make it official.

2) perhaps cutting the PI drive would keep it away from frizzy territory. There is plenty of gain in this amp already. If trimming the gain in the right spot would help cut frizz, I would be willing to try. Smaller plate resistor or more grid leak for V4A perhaps?

3) i wonder if diode biasing for the V4B cathode might eliminate voltage doubling -- if this is the issue.


For the time being, the grid stopper stays in place and I'll drop in a more permanent fix for R31 - probably 68k and work with this for awhile.

I'm always open to ideas.

Ross

[kleuck]

Apart from the cathodyne issue, this looks to me like a fizziness issue often seen with 1974X clones, there's a cure : use a big cathode cap on the power tubes, like 500-1000µf (but gives a sorta "fixed-bias stiffness to the amp).
That's not always enough, and there's a second trick : use of zener diodes to limit the power tube bias drift.
I can't remember exactly, but there's a lot of information on 18watt.com.
Search "Paul Ruby fizz fix" or : http://www.18watt.com/storage/18-wat...z_info_311.pdf (must be registered i suppose)

Sorry if it has been said before, the forum too slow for me at the time, can't read all the thread)

[Silvio55]

No kleuck, this fizzyness is another thing, the paul ruby mod takes care for the fizzyness caused by crossover distortion, when the power tubes are into cutoff, so you hear it when you crank the volume, and sounds like "mosquitos", is very different. This fizzyness is more like a raspy sound and is more evident with lower volume, I could say even like it's fixed volume, no matter the volume knob position, you hear it always at the same level, but disappears at higher setting cause the louder sound masks it.

[Ross Grey]

Thanks Kleuck,

I also looked again at the Paul Ruby fix, and it is pretty clear that Paul's fix addresses artifacts that are apparent well into overdrive. The frizzy decay that some experience with the NT is like a rude onset of distortion that appears and disappears abruptly as the signal envelope increases and decreases, even with fairly low settings of the gain control. This can really trash what is good clean tone otherwise.

There is pretty strong consensus that this is a product of the cathodyne PI, which has been the target of recent approaches to a solution. In my experience, unbalancing the PI helps to the extent in that the onset of frizzy decay can be pushed out to higher gain control settings where the onset sounds more like distortion onset in a poor germanium transistor fuzz box -- in my amp anyway.

My next experiments may be with mods to the PI drive or output load to see how these affect this issue. Still thinking about this.

Ross

[kleuck]

Sounds like blocking distortion, apart from the PI, did someone just tried to add a grid-stopper to each (second and third) triode ?

[Ross Grey]

Kleuck,

This idea came up fairly early in the thread and a quick review of the chain shows that Silvio added a 10k stopper on V3B with no effect as I recall. There was some confusion on the need to cut a trace when adding a stopper. This was cleared up, but not before someone had added stoppers that couldn't possible be effective because a trace was not cut on the board.

Since the level of the gain control seems to have so much influence on the arrival of frizzy decay, I want to try a stopper on V4A too.

My quick re-read of the thread reminded me that SickMan recommended stoppers on both V3B and V4A, but I didn't see a comment on the effect of a stopper on V4A.

This will probably me my next change before making modes to shunt more of the V4A input to ground if the stopper does not help.

Thanks again.

Ross

[kleuck]

I would try 33K on both V3B and V4A.
10K is enough for RF issues, but we are not in this territory.

[Ross Grey]

Kleuck,

Thank you. Good advice it turns out.

My current stopper configuration: 33k stopper at V3A (post tone stack), 33k stopper at V4B (PI driver), 1Meg stopper at V4A (PI).

This worked so well, I rebalanced my PI voltage drop resistors back to 56k/56k from my unbalanced configuration of 56k/68k -- and it still works well. I have no frizzy decay when I want clean tone.

With that said, as gain is increased past 12 o'clock with my stock strat, things start to distort as expected, but the arrival of distortion is more typical. In my limited testing so far, the onset of distortion is even smoother when the master volume is turn up past practice levels.

While I think the stoppers are critical in addressing this issue, I need to mention that my NT now has many mods that may have contributed to more headroom and smoother tone. In particular, Sozo caps, 100k plate resistors for the early preamp stages, I changed R22 from 27k to 47k and I removed R38 to remove the extra goofy input to the PI driver. Clipping C27 was one of the first mods I made.

I am continuing to evaluate this configuration with a patch for balancing the PI before I make this change permanent but won't be able to get to this until later tonight if something else doesn't come up.

Look for a follow-on post. Thanks.

Ross

[kleuck]

Cool so far
So not (or not mostly) a Concertina issue at the end.

[Ross Grey]

Very cool. Perhaps the PI design could be better, but this is heading in a good direction.

The nature of frizzy decay in this amp is that it sneaks into clean settings in an abrupt and annoying way that is not controlled by pick attack or the tone stack. This is especially true as the gain control setting gets closer to distortion, but frizzy decay is more than the typical onset of tube distortion.

Adding the 33k grid stoppers on the second and third preamp stages took care of the frizzy decay in my assessment.

I rebalanced the PI, and returned R28 to the factory setting of 2k7 in my amp. This made the sound of the amp big and bold again. The distortion sounds a bit brighter with these factory settings, but the onset of distortion can now be controlled in a more predictable way with the gain control, pick attack, the tone stack or the guitar volume setting. There is no frizzy decay plaguing the clean sounds with the current configuration.

The onset of distortion occurs a bit sooner (on the gain control setting) now that the PI is balanced and R28 is reinstated, but I am really liking the clean tone below the 12 o'clock setting and some great sounding compression as the gain approaches 12 o'clock.

Even with the plate resistor and input divider changes I put in place in attempt to improve head room, the gain in this amp gets way over the top very quickly as the gain control is increased. For this reason, my next experiment will be to shunt more of the signal to ground at the input of the third gain stage (R27) to see if the gain level can be managed over a wider range before the sound goes into bleak saturation.

I'm starting to like this amp.

Thanks. You guys (an assumption) are the greatest (a fact).

Ross

[kleuck]

my next experiment will be to shunt more of the signal to ground at the input of the third gain stage (R27) to see if the gain level can be managed over a wider range before the sound goes into bleak saturation.

Won't be useful, as the grid resistor is bootstrapped, the impedance would still be high.
But there's ways to tame the gain, the simplest one is to use a 12AY7 in place of the first 12AX7, and i would go this way.
Or you can if you want o stick to 12AX7 use a 47K plate load, along with a 1K5 cathode resistor.
The volume pot too can be tuned, you can use a 330 to 470K resistor between the wiper and ground to get a more useful range in the low part of the travel, at the expense of less max gain and (perhaps) a somewhat less bright tone.
You can too add a 330 to 470K series resistor before the volume pot, added to the aforementioned resistor, to avoid loss of brightness, but the max gain would be lower too.

[Ross Grey]

Ok, Tried the easy mods first - some tube changes. Here is the short version for the setup that worked the best.

I took the Mullard ECC83 out of V3, moved it to V4 and put in a Sovtek 12AX7 in V3. Huge benefit. The Mullard is apparently a very different design -- the guts are twice the size of the Sovtek internals. This change moved a lot of gain from the early preamp stages to the third stage and PI so the gain control is much more useful. The Sovtek in V3 also seemed to clean things up nicely.

I tried lower gain tubes in V3 and V4, but found that the preamps did not have enough drive with these changes (in my opinion). A nice jazz amp perhaps.

The Mullard sounds really nice in V4 with a small bright speaker cab because it warms things up a bit and the overdrive is noticeably smoother. Another Sovtek in V4 is great with a bigger warmer cab because the sound is more defined. Other brands of tubes in various locations were not interesting. I still use JJs for the finals and prefer these.

This is still sounding better, but I still have occasional overdrive sprites that may still be some blocking distortion. Additionally, my notes indicate that the overdrive got 'seductively smooth' after adding the (1meg) stopper for the PI stage. This means I thought the change too far and made the sound a little squashed.

Today, I am going to update all three stoppers. If I am using the calculators correctly, there should be no noticeable high frequency loss when taking the V3A and V4B stoppers to 47k. I'll also take the stopper on the PI lower than 1meg. I'll need to see what resistors I have on hand that are between 500k and 1 meg.

[kleuck]

The squashiness you report is an interesting issue.Did you try 500K on the Pi grid ? May be enough now.