diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Tubes / Valves (http://www.diyaudio.com/forums/tubes-valves/)
-   -   Signal phase shift through a tube amp (http://www.diyaudio.com/forums/tubes-valves/90230-signal-phase-shift-through-tube-amp.html)

d1camero 13th November 2006 06:23 AM

Signal phase shift through a tube amp
 
Hi all, new to this forum, but an experienced builder in guitar amps.

I tried posting this question on one of the best guitar amp forums, but basically got shunned - they think I am being far to technical, but I really want to get an answer to a perplexing anomoly.

A few of my associates and I have played through many amplifiers, and from time to time, the amp did not seem to respond to the guitar - it was if there was a disconnect of some type. This sensation is reproduceable any time, at will and it is quite noticeable.

One hypothesis is that the was some kind of a delay in the amp, maybe incduced by the R-C tone stack. (??)

So... being an engineer/scientist, I dragged out the DSO and pumped both a sine wave and a guitar signal through the amp, scoping it at various points. It came up with some interesing and perplexing results. Here are the traces:


http://www.cameronsoftware.com/misc/...ine%20Wave.jpg


http://www.cameronsoftware.com/misc/...tar%20Note.jpg

The DSO is being triggered only on one channel. - the input signal.

Blue: input
Red: output after second gain stage
Green: input to power tubes
Orange: at speaker output

When I look at these, two quesions come to mind:

1) why does the sine wave have a phase shift (between input and later stages), and the guitar signal does not? Check out the zero crossing line on the sine wave, there is a shift of 127us. See the peaks of the guitar signal, the alignment is **very** close.

2) why is the phase shifted sine wave out of order?
Blue (input) ahead of Red (preamp) - good
Red (preamp) behind Green (power) - ??
Greem (power) begind Orange (speaker) - ??

I can only explain the last two is that between the preamp and power input a large shift (e.g. 340 degrees) occurs.


If this is not the right forum for this type of question, please let me know (also, if you know the right place please tell me, it will help me greatly).

I am just trying to understand what I see here.

thanks
Don

SY 13th November 2006 11:52 AM

This is indeed the right place, lots of techie tube types. In any case, the delay through a tube amp is unbelievably small compared with human perception- on the order of a few nanoseconds. A perception of "delay" is no doubt due to other psychoacoustic factors.

If someone else doesn't do it first, I'll pile through your waveforms a bit later when I've got a few more minutes.

I envy your DSO!:D

EC8010 13th November 2006 12:59 PM

Microseconds. Assume a -3dB point of 100kHz (a little generous) and a first order roll-off, then there's 45 degrees of phase at 100kHz, which corresponds to 5us. Make another sweeping assumption and say that phase against frequency is linear (it isn't, but it's near enough for this argument) and you have typical valve amplifiers contributing somewhere around 5-10us of delay. To put that into perspective, it's the time taken for sound to travel 2-4mm* in air. Only headphone listeners have their ears positioned that precisely...

* That's about an eight of an inch to those of you still using Imperial measurements and wooden school rulers.

SY 13th November 2006 01:24 PM

Quote:

Originally posted by EC8010
Microseconds.
What's a factor of 1000 between friends?

David Davenport 13th November 2006 01:28 PM

Re: Signal phase shift through a tube amp
 
Quote:

Originally posted by d1camero
This sensation is reproduceable any time, at will and it is quite noticeable.



Do you mean that there is something that you can do to make it happen or not happen? If so, what is it?

Dave

Robert McLean 13th November 2006 01:28 PM

Some observations :

first the signal waveforms
- frequency is about 1KHz

- a 340 degree phase shift is just a 20 degree phase shift with polarity reversal.

- Input and Speaker output are in phase, despite intermediate stages being out of phase, this appears to be nfb loop doing its job.

now the guitar waveforms
- fundamental frequency about 350 Hz, harmonics at about 4KHz ( I am just counting the peaks against the time base scale )

- at first glance it appears that all the signals are in phase, or in anti-phase, so it seems to contradict the signal waveforms. But looking more closely you see that it is the high frequency peaks that are lining up in phase. If you look at the zero crossing points you see that the different signals are not crossing zero at the same time. So the low frequency, ie fundamental, signals are being phase shifted.

Conclusion :
The low frequency ( approx 350 Hz) fundamentals are being phase shifted. High frequency components ( 4KHz ) are not being phase shifted, at least not as noticeably.

The most likely cause of this in my view is simply RC coupling between stages, which affects low frequencies, and lets the high frequencies through unaffected.

Brian Beck 13th November 2006 01:28 PM

You’re both right (great harmonizer that I am). Nanoseconds and microseconds. There will be a group delay in the passband of microseconds, as EC says. However, when the input experiences a step in voltage, the output will START responding almost instantaneously, within nanoseconds. However the response will be sluggish (relative to nanoseconds) and will take microseconds to build. Either way, these delay figures are not perceptible to humans.

I have heard of the problem that “d1camero” experiences. I would have thought it was related to blocking of coupling and/or bypass caps following a severe overload. Even plugging in the guitar jack can induce a jolting transient. I recall Tubelab (George) commenting on this phenomena in a prior thread. Perhaps he will respond.

Brian Beck 13th November 2006 01:43 PM

PS: I had a guitar amp that would go quiet for a second or two after if I bumped into or knocked the cabinet. I think it was due to the blocking overload problem I mentioned above. In my case there was an internal intermittent connection that would drive the output stage deep into overload for a bit of time.

I don’t think your problem is a phase-shift related issue. The DSO pictures are interesting, but perhaps are not pointing you in the right direction.

tubelab.com 13th November 2006 03:18 PM

All amplifiers (tube or solid state) introduce some phase shift. This phase shift usually varies with the frequency of the applied signal, especially if there are any tone controls involved. It is doubtful that any microsecond long delay effects can be heard by the human ear. It is certain that they are not responsible for what you are hearing.

I have heard a delayed sound effect on some guitar amps, especially modified old Fenders. In some cases the delay can be a handful of milliseconds. In the cases I am familiar with it is usually on a quiet note that follows a loud blast. It can also be the first note after some silence if an overdrive pedal is used. In these cases the delay is related to "blocking distortion" or "farting out" as it is refered to by some guitar players.

For a detailed explanation of this effect go to:

http://www.aikenamps.com/

Click on Tech Info, then on Advanced, then go down to what is "blocking distortion". There is a lot of good info for technical guitar amp people on that site.

I have been able to capture the effect on my old Tek 2232 scope. If you can trigger your scope on a delay event good, but I couldn't so:

I put one channel at the guitar input and walked the other channel down the signal chain taking several scope traces at each point. It may take several tries to capture a delay event. You need to set the scopes sweep speed slow enough to see the envelope of the guitar signal (the attack and decay of each note, not the note itself). Depending on playing style this may be 10 to 100 mS per division. Concentrate on the notes attack since this is where the problem is. I let the capture buffer fill up and then scroll through it examining each note (compare the signal from the guitar input to the other trace taken further down the chain looking for the delay). Sometimes it helps to set the scope on DC coupling and look for a DC shift at the attack of a delayed note. I bet that you will find a DC shift at the grid of the output tube (tubes) even though the signal is not affected. In some cases the timing of the notes is such that the next note arrives during the recovery time from the last note causing the bias of the output tubes to shift and in some cases cut the output tube off causing a momentary silence and a delay on the beginning of the next note.

The chopped off notes may not be visible on the grid of the output tube. It may be visible on the plate, but be carefull of the extreme voltage that may be present if the output tube gets cutoff.

In extreme cases the whole amp gets upset. This can come from two sources or a combination of both. I have seen (old Fenders again) the power supply voltage drop 50 to 100 volts at the attack of a heavilly overdriven note causing a DC shift through out the amp. New rectifier tube and power supply caps helps. Sometimes a transient can be intensified through the negative feedback loop causing an upset in the driver stages. Often this effect can be speaker dependent. Try disconnecting the feedback.


All times are GMT. The time now is 08:11 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2