Fuzz/Phase shifting effect (muskito flying by) in the mids

Froumy · Post by **Froumy** » Thu Jul 17, 2008 8:04 am

Damn. I heard it on those old ACDC clips, But I've yet to notice it on my build, a 2203 or a 2205. Maybe I'm just not tuned in, or it's just more obvious in different builds. Bahh. If it sounded even slightly like your clip Dan, it would drive me absolutely batty(I'd think everyone else, too). If it is more apparent in some builds than others, you'd think you could at least tame it.

5150, any luck with the scope? Did you try those shielded caps you mentioned? You guys try Larry's grounding suggestions/thicker wire?

Billy Batz · Post by **Billy Batz** » Thu Jul 17, 2008 2:33 pm

Well, as I said, those clips are an anomoly. This noise never sounds this easy to discern. You can clearly hear the noise separate from the note int he clip. It really sounds like its buried back there a ways very much like ghosting.

5150loveeddie · Post by **5150loveeddie** » Thu Jul 17, 2008 8:58 pm

Froumy wrote: 5150, any luck with the scope? Did you try those shielded caps you mentioned? You guys try Larry's grounding suggestions/thicker wire?

Not yet for the scope test, caps are no where to be find for now, it is not a grounding issue, tried all of that with no improuvement.

Froumy · Post by **Froumy** » Sun Jul 20, 2008 10:10 pm

Apparently, crossover distortion occurs in all class AB amps. But you can tame it.

Do all of your amps have .1uf PI caps? Lowering them is supposed to help. Upping the grid resistors to 50k -100k can help. You can increase the value of the tail resistor, or a add a series resistance between the coupling caps and the grid bias resistors. You can add a resistor across the phase inverter outputs after the coupling caps. You can limit some clipping on the grids with a zener diode. Too much negative feedback can make the problem worse.. Changing the bias will affect it, too. It can happen in the preamp section as well, which explains why removing a cathode bypass cap will reduce the trouble.

Seems like there are lots of things you can do, The question is how they'll sound. Of course, this assumes you have blocking or crossover distortion.

I wouldn't take my word for it, But Aiken had a pretty reasonable explanation on his site. Check out the scope views 5150lovedie. You should be able to prove that you indeed have this problem.

http://www.aikenamps.com/CrossoverDistortion.htm
http://www.aikenamps.com/BlockingDistortion.html

Try 'em out. It'll be interesting to see if any work, and if it comes with a sacrifice in tone. I think the solution is in there somewhere. Good luck.

5150loveeddie · Post by **5150loveeddie** » Mon Jul 21, 2008 1:26 am

THx for those ideas, I will look upon them for sure thx man!!..

I think I don't bias the amp correctly using the 1ohm resistors, this will under read reality and end up with cross over distortion most probably.. I'm working on a probe to be able to bias direct from the plates and not the cathode, I know this is more dangerous or less safe to accomplish but it is much more presice, if this works I'll let you know. Those 1ohm resistors never are at 1ohm they all show around 1.4-1.6ohm, right there it won't work properly..

novosibir · Post by **novosibir** » Mon Jul 21, 2008 10:01 am

5150loveeddie wrote:... I'm working on a probe to be able to bias direct from the plates and not the cathode, I know this is more dangerous or less safe to accomplish but it is much more presice, if this works I'll let you know. Those 1ohm resistors never are at 1ohm they all show around 1.4-1.6ohm, right there it won't work properly..

Right! This is much more precise!

You can try the 'transformer shunt method', but most people, who do it this way, don't take the internal resistance of the meter in account. Depending on the internal resistance of the meter (to what also the resistance of the meter leads & alligator leads has to be added!), a more or less amount of plate current still is flowing through the primary - and hence isn't measured!

The 'most accurate method' w/o desoldering anything is, to measure the resistance of both halfs of the OT's primary, then measure the voltage drop across the primary's halfes:

voltage drop divided by winding resistance is the real plate current!

Shure, in a 100W amp you only can measure the sum of the plate current of both output tubes of one side, so the individual plate current, each tube is drawing still may differ a bit from the 50% average - depending on, how accurate your quad of tubes is matched.

To measure each tube's plate current of a 100W amp accurate, you'd have to unsolder the plate pins and to put the meter in between.

Larry

raiken · Post by **raiken** » Mon Jul 21, 2008 12:26 pm

5150loveeddie wrote:Those 1ohm resistors never are at 1ohm they all show around 1.4-1.6ohm, right there it won't work properly..

If they are 1% resistors, they are very likely to be actually extremely close to 1 ohm - within 1%, actually.

The problem is that your meter cannot properly read a 1 ohm resistance because the meter lead resistance is too large in comparison to the resistance of the 1 ohm resistor. Some high-end meters, like the venerable Fluke 8012A, have a "lead zero" capability on their low-ohms measurement scale that can be used to zero out the meter for an accurate reading.

This is not significant, however, because when checking bias current, you are measuring the voltage drop across the 1 ohm resistor, not the resistance across it, and the meter lead resistance is insignificant compared to the internal impedance of the meter when measuring voltage (the internal resistance is typically many megohms). In essence, your bias voltage readings will be correct, even though your resistance readings will be way off.

Randall Aiken

Froumy · Post by **Froumy** » Mon Jul 21, 2008 2:32 pm

Mr. Aiken,
Do you think these guys are experiencing crossover or blocking distortion? Thanks for all the great tech info that you've shared on your site! Very informative. Yours was the the only web reference I found with both explanations, and possible solutions. Thanks again.

raiken · Post by **raiken** » Mon Jul 21, 2008 3:20 pm

Froumy wrote:Mr. Aiken,
Do you think these guys are experiencing crossover or blocking distortion? Thanks for all the great tech info that you've shared on your site! Very informative.

At first glance, er, listen, it sounds like crossover distortion to me, from too much drive to the output stage. This is very closely related to blocking distortion, except that blocking distortion usually refers to a type of distortion with a long time constant, where the entire signal cuts off, or is "blocked" for a period of time following a transient overload, resulting in a "farty" or "splatty" sound.

Crossover distortion in an AC-coupled output stage is greatly exaggerated when that stage is being overdriven, because the AC-coupling of the PI into the output tube grids creates a "clamp", where the peak of the driving signal is clamped to ground, and the signal continues to increase in the negative direction. This has the same effect as turning down the bias voltage to bias the stage colder, with the resulting increase in crossover distortion.

If your output transformer has an extended high frequency response, or a high leakage inductance, it can make this crossover distortion more audible, as can a speaker with extended high frequency response. Different transformers, as well as different output tube types, can show different "looking" crossover distortion because of their interaction with the circuit.

A quick test to see if this is causing the problem is to put a unidirectional zener clipper circuit on the bias feed resistors to clip the negative-going transitions to a level equal to the positive-going transitions, as referenced to the idle bias voltage. This keeps the average signal level from going negative, which stops the resulting bias shift and "extra" crossover distortion.

You can make a unidirectional clipper with a zener diode in series with a regular diode, with the zener's cathode and regular diode's cathode connected together, and this assembly connected across the bias feed resistors, with the zener's anode connected to the coupling cap/output tube grid stopper connection and the regular diode's anode connected to the other end of the bias feed resistor that goes to the bias supply, so you only clip negative-going signals.

Select a zener diode whose reverse breakdown voltage is slightly higher by 5 or 10 volts than your negative grid bias voltage when the tubes are biased to the current you like. For example, if you have your bias voltage set to -35V, try a 39V zener. The closer you get to symmetrical clipping, the less extra bias shift you will have, so you can experiment with different zener voltages.

Since the clipper is only turned on when the tube hits cutoff, it doesn't affect the "normal" tone, other than to eliminate the additional crossover distortion caused by bias shift.

Randall Aiken

Billy Batz · Post by **Billy Batz** » Mon Jul 21, 2008 3:47 pm

Great info. Thanks a million and I will try that.

What exactly would you do about it if this proves to be the cause? All my amps do this to some degree. Fenders and marshalls. It could be that I get it more then most people here because my cab is an H30-55 type while 90% of people here seem to use 25Wers. Also I assume attenuating may extend the speakers high end response since your running it much cleaner so that could factor in tho you can hear it unattenuated if you try real hard. If it depends on each OT that may explain why some amps do it more then others. I can hear it in vintage amp clips but its a lot more mellow and smothered then clips of peoples new builds. Possibly the OT isnt as clear int he highs?

The reson it came out so clear and 'seperate' in that clip is I was using a MASS150. It does it horribly when I use that attenuator and set it low for some reason. What Im using now is a much superior attenuator and it sounds closer to how it does when unattenuated. Like ghosting. Back there somewhere but not so in your face.

novosibir · Post by **novosibir** » Mon Jul 21, 2008 3:57 pm

Billy Batz wrote:The reson it came out so clear and 'seperate' in that clip is I was using a MASS150. It does it horribly when I use that attenuator and set it low for some reason. What Im using now is a much superior attenuator and it sounds closer to how it does when unattenuated.

Which one?

Larry

raiken · Post by **raiken** » Mon Jul 21, 2008 4:00 pm

Billy Batz wrote:Great info. Thanks a million and I will try that.

What exactly would you do about it if this proves to be the cause? All my amps do this to some degree. Fenders and marshalls. It could be that I get it more then most people here because my cab is an H30-55 type while 90% of people here seem to use 25Wers. Also I assume attenuating may extend the speakers high end response since your running it much cleaner so that could factor in tho you can hear it unattenuated if you try real hard. If it depends on each OT that may explain why some amps do it more then others.

The reason it came out so clear and 'separate' in that clip is I was using a MASS150. It does it horribly when I use that attenuator and set it low for some reason. What I'm using now is a much superior attenuator and it sounds closer to how it does when unattenuated. Like ghosting. Back there somewhere but not so in your face.

The MASS is a fully reactive load, which will greatly exaggerate the high-frequency content of the crossover distortion. You can see this on a scope if you compare the signal into a resistive load, like a Z airbrake, and the reactive load, like the MASS or Marshall PowerBrake.

If it does prove to be the case, and the zener clipper fixes it, I'd just leave it in there.

Alternately, you can redesign the phase inverter for less drive (try increasing the tail resistor to an abnormally large value) or redesign it to make it clip naturally on the low side before the high side, or design for DC-coupled output stage driven from cathode followers, or other such things. However, first try the zener clipper to make sure this is the problem.

FWIW, I am very sensitive to the sound of crossover distortion, it drives me nuts. I went to great lengths to design it out of my Tomcat amp, which started out life based on the Marshall 18W output stage, but I couldn't stand the crossover distortion at full drive. I'm now to the point where I think I prefer a good, properly "organically" designed preamp distortion and an unclipped, high-power output stage just so I don't have to hear that little buzz riding on the note decay!

Randall Aiken

raiken · Post by **raiken** » Mon Jul 21, 2008 5:13 pm

The other thing that can cause funny noises riding on your notes is parasitic oscillations. These usually only show up in reactive loads, or when your lead dress is not optimal.

Proper location of grid stoppers helps, and if the amp has global negative feedback, it can usually help to put a 33K or 47K stopper right on the input side grid of the phase inverter, as it is usually the PI/output tube/OT/speaker loop that oscillates.

The oscillation sometimes won't show up on a scope until you agitate it with a signal, then you can see a little fuzzy "blip" riding on your sine wave, usually at the peak, but sometimes at the crossover point. The oscillation will usually also respond to various settings of the controls, particularly the presence and treble controls. It may completely disappear at some settings.

This is why you should always test your amps into a good resistive and reactive load, in addition to a real speaker. Keep a sharp eye on the scope for any anomalies, and exercise all knobs in all combinations end-to-end.

Randall Aiken

Billy Batz · Post by **Billy Batz** » Mon Jul 21, 2008 5:27 pm

There is a possibility thats it because one person has a PPIVMV and it didnt cure the problem when turned low. Im not sure if I tried to separate the grid wires from the plates more on V3. I may try that but I think I have. I think the grid resistor there is worth a try.

raiken · Post by **raiken** » Mon Jul 21, 2008 5:36 pm

Billy Batz wrote:There is a possibility thats it because one person has a PPIVMV and it didnt cure the problem when turned low. Im not sure if I tried to separate the grid wires from the plates more on V3. I may try that but I think I have. I think the grid resistor there is worth a try.

If the PPIMV is implemented properly, there should be no increased crossover distortion when playing at levels below the point where the signal peaks hit zero volts at the output tube grids. The danger with most PPIMV implementations is that people run the wires too near other things which causes them to oscillate.

Randall Aiken

Metropoulos Forum

Fuzz/Phase shifting effect (muskito flying by) in the mids

Who has Aluminium OR Steal chassis OR both with the SNAG???