nfb issue

[nickb]

Yes all tests was done with signal aplied at the input of the amp. The remark about PI voltage was relative to the amount of nfb

the input of PI with 2.7V
[ATTACH=CONFIG]37404[/ATTACH]
and 2V
[ATTACH=CONFIG]37405[/ATTACH]

I don't think I can see it there - can you? It's pretty hard to tell as the pics are very very small.

What I don't understand is that in post #1 the blips are in three different places. Why? Did you change the controls between images?

[catalin gramada]

I think I have to clarify, please
All the test I did, except the last one was with signal aplied at input and OT output scoped. At 400 hz with all tone control and presence dimmed looks like that:

The pics from the last post are PI input scoped with 2.7v and 2 v measured in the same point also with 400 hz. I don't know what voltage was in the previous tests cause I just ilustrate the efect and anyhow my scope is not calibrate to be useful to pick values, but it was under limit of clipping
Now from the last test shows the blips is even present in the PI input at 2.7v and 2V (pi input- I picked up the voltages with my trusty Fluke)
IN PI input the blip is in the top of sine while in the same time the OT output shows the blip at the right side

[catalin gramada]

I don't think I can see it there - can you? It's pretty hard to tell as the pics are very very small.

What I don't understand is that in post #1 the blips are in three different places. Why? Did you change the controls between images?

Yes in first post the blips are in diferent places cause I had the intention just to show how it is dependent by controls. so The first image shows tone controls and presence dimed. second one nothing mod but trigered a bit for a better view. the third one is with tone controls dimed and presence turned down to "0". so If I move the presence the blip changed its position from the right to the left. The same is hapen if I play with tone controls turned down
All this s...is not happen if I disconect nfb from circuit

[catalin gramada]

I can repeat any test by you instruction if you like. Thank's for You time

[J M Fahey]

I don't see any problem at all.

The waveform extreme peaks when fed squarewave come from the heavy treble boost present in a Marshall, with many bright circuits in cascade, the Presence exacerbates this even more, and account for the "buzzy" sound (justly) attributed to Master Volume Marshalls when used way below maximum.

Once the power stage clips, all those peaks and the inclined or twisted waveform tops become crushed by the power tube saturation clipping which is much "rounder" or tlo be more precise, basically square without the sharp peaks.

So much confirmation for the mythical "tubes clip rounded" fairytale.

The blip is minuscule, absolutely non audible, and *may* be an artifact of some tube nonlinearity complicated by accumulated phase shifts, but no big deal anyway.

[Dave H]

It looks like there could be some instability in the power amp. What value are the screen grid and control grid stoppers? I notice the JTM45 doesn't have any control grid stoppers. If your clone is the same I'd try adding a couple (5k6?) to see if that affects the blips.

[catalin gramada]

Hi, now it have 5.6K grids as standard value but I just put in trying to cure this blip. It was a nonsense in my case this amp works well without those resistors. Even max dimed cannot do any diference with or without it. The screens are now individual 560/3w ohm also for mentioned reason for max 12w full disipation in screens when dimmed but was single 470/7w before, The resistor was mouted in one tube socket and I passed about 3" solid wire to the next screen . As I said this did not do any diference regards stability for my layout. All stable, no osscilation issues, just this .... bug I cannot get rid of. Without nfb conected the signal looks like from generator but it break-up too early,have not enough headroom for cleans.
Maybe like J M Fahey said,I bother too much with a problem which is not a real one in effect. Like Nickb said I also don.t think is a stability problem those blip is a constant no matter what changes I do in circuit: tubes change, caps change, voltage tweak,wiring ... the only thing I can do it now is to disconect nfb from circuit
Thank You all for support
If still have any ideas I'm ready to try,I'm off
Cheers

[Dave H]

Maybe like J M Fahey said,I bother too much with a problem which is not a real one in effect

You don't hear with your eyes but it would bug me too if I could see it on the scope even if it was inaudible.

[nickb]

You don't hear with your eyes but it would bug me too if I could see it on the scope even if it was inaudible.

Daveh -I't s certainly bugging the c^&p out of me

Catalin - Thanks for the better pics. I think what is going on is you have a small non-linearity which is being exaggerated by the all the treble boost. If a tube is causing this, I would be really surprised since they just don't work that way - especially as its so symmetrical.

What about your signal generator? Do you know for sure that it's got a pure output? Feed it into your sound card line-in and get the frequency spectrum - thee are plenty of tools on the Internet to let you do this. I'd be looking for the 3rd harmonic perhaps 40dB down as it's very small blip.

Scope use / calibration

You have knobs to set the vertical scale marked in V/div and horizontal scale ms/div. You just need to state what these are set to when providing scope shots. Your 'scope may not have been calibrated for some time but unless it has developed a fault it won't be too far off.

Many scopes have a calibration output which provides a known frequency and amplitude. If you have one, hook your probe to it and see what your scope reads. Failing that measure a 9V battery with your fluke and then again with the scope (on DC coupling) and see how it compares. You can use your signal generator to check the time base to see if it's in the ballpark.

[Mick Bailey]

Only yesterday I was scoping the output of a Bogner Shiva and it had those same little blips and showing a very similar scope trace to those pictured. Usually I ignore them and adjust the amp's controls to get the best waveform possible. In the case of the Shiva, turning the bass up full and treble almost right down with no presence totally eliminated them.

If the bumps are unstable or excessive, or can't be tuned out with the tone controls, then I'll take a closer look as they could be a result of oscillation or instability. Largely, though they're down to the characteristics of the particular amp.

I once has a determination to find out why a waveform had a little 'dent' in the side of it and couldn't establish the cause. Well, not until I later got an identical model amp (Laney VC30) and it did the same thing. I removed the amp from the chassis and had it on the bench. Problem gone. Returned it to the cabinet and the problem was back. It turned out that the proximity of the power tubes to the speaker magnet was the problem.

So, the issue is deciding whether something is a fault, or a characteristic.

[catalin gramada]

Only yesterday I was scoping the output of a Bogner Shiva and it had those same little blips and showing a very similar scope trace to those pictured. Usually I ignore them and adjust the amp's controls to get the best waveform possible. In the case of the Shiva, turning the bass up full and treble almost right down with no presence totally eliminated them.

If the bumps are unstable or excessive, or can't be tuned out with the tone controls, then I'll take a closer look as they could be a result of oscillation or instability. Largely, though they're down to the characteristics of the particular amp.

I once has a determination to find out why a waveform had a little 'dent' in the side of it and couldn't establish the cause. Well, not until I later got an identical model amp (Laney VC30) and it did the same thing. I removed the amp from the chassis and had it on the bench. Problem gone. Returned it to the cabinet and the problem was back. It turned out that the proximity of the power tubes to the speaker magnet was the problem.

So, the issue is deciding whether something is a fault, or a characteristic.

Hi
Yes , the same manifestation like mine
I read You post and thought how dumb I could be cause I diagnosed the amp turned up side down in top of cabinet but was not lukky. I put it on a table right now and is identicaly the same. More, I covered the chassis with a sheet of metal but nothing changed.
I have not another pair of power tube right now to do a test, but preamp tubes did not any diference

[nickb]

Hi
Yes , the same manifestation like mine
I read You post and thought how dumb I could be cause I diagnosed the amp turned up side down in top of cabinet but was not lukky. I put it on a table right now and is identicaly the same. More, I covered the chassis with a sheet of metal but nothing changed.
I have not another pair of power tube right now to do a test, but preamp tubes did not any diference

So, do you think you'll leave it there as an inaudible mystery? Or are you the kind that wants to try get to the bottom it?

[catalin gramada]

No,definely. Strange things need a explanation in the end. But I see is not a easy way to diagnose. I think Mick Bailey well pointed. The question is: is a fault(a defective bug) or a characteristic and the bug is manifest from design(a consequence of a bad design)? At this point I think the investigation need a plan.I have to take a time to think about and stopped now I will start with checking my test equipment generator,scope , probes...) as you suggest. Have to remove any doubt something in back of test going wrong. Thank You
Catalin

[The Dude]

I admire your curiosity. I would have called it "good enough". I'm curious how this will end.

[pdf64]

Have you tried inserting the test signal directly at the LTP phase splitter input yet?
I think that may be helpful, eg is it present with the pre-amp eliminated, nickb's 'ringing squarewave' test.