crosstalk and fizzyness in plexi se

[gtrboy]

anybod built a plexi se? im getting some strange extra fuzziness which is oscillation, crosstalk (feedback), or a poorly biased tube or something. anybody found faults in the design of this thing?

thanks

http://rh-tech.org/public/z_stingray...ematic_v11.png

[kleuck]

Try a Zobel on the output, and maybe some filtering (caps // with anode of the last 1/2 12AX7, say 1nf)

[gtrboy]

Try a Zobel on the output, and maybe some filtering (caps // with anode of the last 1/2 12AX7, say 1nf)

i tried a 470pf on the preamp plates. the first 1/2 then the second. i also put a low pass after the eq. most high gain designs put a treble cut in between the phase inverter and power. ive experimented but had little luck. its on a pcb so im not able to add grid stoppers.

[Enzo]

No offense meant, but when you say the problem is oscillation, feedback, bias trouble, or crosstalk, it sounds to me like you do not really know what is wrong and are making guesses. You need to find out just what the amp is doing, then the cure might be easier to effect.

DO you have access to a scope? That is the way to detect oscillation. And oscillation takes various forms, parasitics are different from regenerative feedback.

The power tube and the input tube already have grid stoppers. There is no phase inverter in a SE amp.

Make up a signal tracer with another amp and a probe with a series cap, then you can listen to the signal stage by stage.

It is built on a board? Are the tube sockets on the board? Or are they off-board with wires to them? In the latter case, wire dress is VERY important.

[gtrboy]

No offense meant, but when you say the problem is oscillation, feedback, bias trouble, or crosstalk, it sounds to me like you do not really know what is wrong and are making guesses. You need to find out just what the amp is doing, then the cure might be easier to effect.

DO you have access to a scope? That is the way to detect oscillation. And oscillation takes various forms, parasitics are different from regenerative feedback.

The power tube and the input tube already have grid stoppers. There is no phase inverter in a SE amp.

Make up a signal tracer with another amp and a probe with a series cap, then you can listen to the signal stage by stage.

It is built on a board? Are the tube sockets on the board? Or are they off-board with wires to them? In the latter case, wire dress is VERY important.

thanks enzo,

no offense. its hard sometimes for me to communicate via text.

im totally guessing but these are common design or build issues. i removed a component and cut the signal in the middle and still got signal at the end (although faint). thats some kind of crosstalk. when i say feedback i mean it in terms of "do you think i need feedback?" perhaps across a plate resistor or another spot. as for a phase inverter i meant many larger high gain amps place a high cut cap between the legs of the phase inverter but since i don have a PI is a low pass filter my only option near the power section? the hi octane has a similar topology but i dont see a small cap bleeding off the high end or caps across the plate resistors. this is built on a board and the sockets are on the board. i do not have a scope. its is very quiet otherwise. the grid resistors are very close to the grids.

[kleuck]

I made a little amp with a VJ OT, and it's not a good one
My design -wich is way lower gain- began to sound good when i fitted a Zobel on the output.I would replace the 470 ohms screen resistor with a 1K one too.
And the cathode cap of the EL84 with a 22µf, 470µf is way too much if you don't cut the lows before the power stage, and you don't (well, only on the first triode, but that's not a huge attenuation).

[gtrboy]

I made a little amp with a VJ OT, and it's not a good one
My design -wich is way lower gain- began to sound good when i fitted a Zobel on the output.I would replace the 470 ohms screen resistor with a 1K one too.
And the cathode cap of the EL84 with a 22µf, 470µf is way too much if you don't cut the lows before the power stage, and you don't (well, only on the first triode, but that's not a huge attenuation).

thanks kleuck,

do you have a specific zobel application that goes at the end? i mean is it a small cap across the speaker outs or across that last 220k to ground in the signal path? do you think that value is 470uf to assist with hum balance? theres a heater connection there.

[kleuck]

A big cap on the cathode resistor of the power tube usually does nothing for the hum, and the overall filtering seems very good already, what it does is "stiffening" the sound, making a cathode bias amp sound more like a fixed-bias one.
Can be useful, and i often use 100µf here, but the sound can become mushy if you do not strictly limit the lows before the power stage.
For the Zobel : try a 8.2 to 10 Ohms/10 Watts resistor in series with a non-polarized cap of 2,2µf across the 8 Ohms output (or similar across another output : 16 to 20 Ohms/1 µf across the 16 Ohms output etc)
What a Zobel does is taming the impedance rise of the speaker, which translates in an rising load on the primary of the OT, causing harshness and fizziness, depends on Both the OT and speaker, you don't always encounter this issue, but here we have : poor OT + no nfb + single-ended = worst scenario

[loudthud]

If there is an ultrasonic oscillation, the zobel resistor will burn up or get hot with no guitar signal. Many times an ultrasonic oscillation can be detected with a DMM set to AC volts. Noise at the speaker terminals should measure in the tens of millivolts. An oscillation will measure volts.

[Chuck H]

The "fuzziness" is probably from too much bottom end in the circuit combined with the comression due to the large G2 feed resistor. The OP could try changing the G2 rail resistor to a much smaller value and changing the tone stack slope resistor to a higher value. That could tame the fuzzies (although "fuzzy" in my term book isn't usually a bad thing. "fizzy" is bad). The crosstalk may be normal. It's nearly impossible to simply lift an AC component and not get a little bleed coming through from the power supply or grounding scheme. Idealizing ground scheme and power supply circuits and using low ESR (impedance?) caps can help. And should be a major design consideration from the start of a build IMHO. If there IS oscillation, it's probably due to lead dress, layout or grounding scheme.

[Enzo]

OK. Crosstalk is not the same as feedback. You really need to know what the amp is doing. We can hang every anti-oscillation measure in the book onto this amap, but if the problem is not really oscillation, it will do not good. Just as increasing the filtar caps multiple times will do nothing to abate hum coming from tube heaters. (Not suggesting you have that, just an example of wrong cure does not help) SO far all your approaches are assuming oscillation.

Google "RF probe." A VERY simple thing you can make for your meter with a diode and a cap. High frequency oscillation - above audio - may or may not show up on your meter. The RF probe allows any meter to read the amount of oscillation even though the freq is too high for the meter. (It is what we call a detector circuit, and is exactly what allows AM radios to work) You can then use it to look for oscillation through the circuit.

And even if you do have oscillation, it matters where it comes from. A zobel on the output may keep it out of the output stage, but would do nothing to stop it if V2a was feeding into V1a.

Various meters and/or meter adjuncts may be able to read steady oscillation, but none of those will read parasitics, which you could also have.


What does it sound like if you pull V1 and inject a signal at R15?