PDA

View Full Version : Biamp phase match tube and solid state?



speakerdave
12-31-2005, 12:12 PM
Memo from the TechNOT's furrowed brow:

On another thread a poster said that there can be phase problems when biamping with tube and solid state amps. Is this true? It occurs to me there might be two issues.

One of the amps could be inverting and the other non-inverting. This could be true of any two different amps no matter what type of amplification device is used. Right?

The other issue might be how long it takes for the signal to get from the input to the output of any particular amp? Is there any way of knowing? Are all possible differences infinitesimal? Is this like rise time or is that something else?

Is this a non-issue dreamed up by hypochondriacal tweeks in the middle of the night?

David

Ken Pachkowsky
12-31-2005, 12:30 PM
Is this a non-issue dreamed up by hypochondriacal tweeks in the middle of the night?

David

Now, there is something that deserves to be a cliche`

Well said!

Now, for the answer! I have no %^%$#@ idea...

Zilch
12-31-2005, 02:31 PM
Hmmmm. Seems like it could be an issue.

Easy to measure, tho. Same signal into the two amps and look at both outputs on a dual-trace scope.

Rise time is something else.

Don C
12-31-2005, 02:32 PM
I have a collection of old Audio magazine annual equipment directories. These list the specifications for many components, and one thing listed is whether an amplifier inverts phase or not. From looking at this, it seems rare to find an amp that inverts phase, but there are a few. Some manufacturers don't specify though, and there are some that have "var" or "sel" in this space, indicating, (I think) that there is a switch or control for phase. If you tell me which amplifiers you are looking at, and narrow down the apprximate year that they were made, I will see if I can find that spec for them.

speakerdave
12-31-2005, 04:16 PM
I will be using Audio Researce VT60 and either McIntosh MC2200 or JBL 62xx series. None of those are listed, probably.

David

Zilch
12-31-2005, 05:17 PM
62xx is inverted from current pro standard.

[Always messes me up here. When I rewire stuff, I gotta THINK what I'm doin'! :p ]

Ian Mackenzie
12-31-2005, 05:24 PM
Provided you are not starting a design from a clean sheet of paper it will either sound right or wrong.

What are you bi amping and biamping with? Some crossovers have a 180 degree phase invert switch for your scenario.

Perhaps the more pertinent issue is will the tube amp and SS amp sound blend well enough? I prefer to use identical amps for both hi and low end..but that is an ideal situation.

The phase shifts you are referring to are usually at the audio extremes either leading or lagging a few degrees. Of course depending on stabiliy margin and open and closed loop gain phase change will occur but not in the audible audio band.

Ian


Memo from the TechNOT's furrowed brow:

On another thread a poster said that there can be phase problems when biamping with tube and solid state amps. Is this true? It occurs to me there might be two issues.

One of the amps could be inverting and the other non-inverting. This could be true of any two different amps no matter what type of amplification device is used. Right?

The other issue might be how long it takes for the signal to get from the input to the output of any particular amp? Is there any way of knowing? Are all possible differences infinitesimal? Is this like rise time or is that something else?

Is this a non-issue dreamed up by hypochondriacal tweeks in the middle of the night?

David

speakerdave
12-31-2005, 07:12 PM
62xx is inverted from current pro standard.
Only as regards their use of the XLR connector. JBL says it this way in the manual for the 6290: “In normal stereo operation (or dual mono), a positive-going signal applied to an input’s pin #3, the phono jack tip, or the + terminal of the barrier strip will cause a positive-going signal to appear at the corresponding channel’s red output binding post.” P.8

I think that means non-inverting.

David

ralphs99
12-31-2005, 09:36 PM
Hi Dave,

The inverting/non-inverting issue is something that can affect any amplifier regardless of the circuit technology employed.

I would hope that every amplifier with unbalanced inputs is non-inverting, but it's a good question as to whether this is always the case. I guess the only way to be certain is to inspect the instruction manual or measure it.

With an amplifier with balanced inputs, it's a bit of a lottery.
The XLR connector has historically used both pin3 and pin2 as non-inverting in US and European made equipment respectively.
An IEC standard followed the German DIN standard of the 60's that declared pin2 to be non-inverting.
In the US there has never been any offical standard as far as I know. But the industry there self organised themselves into using pin3 hot.
The difference wasn't a big issue until the 1980's as the European and US markets were largely separate and a lot of European gear used DIN connectors instead. In Australia however, we were getting equipment from both continents, and this caused much confusion here!
In the mid 80's the US finally agreed to do what the rest of the world had been doing for a long time, and comitted to using pin2 as non-inverting.
So older US made equipment has pin3 hot, while everthing else should be pin 2.
As far as the length of time it takes the signal to get from the input to output (propagation delay), this extremely small in analogue equipment. Electricity flows at close to the speed of light (about 0.5x) and so propagation delay doesn't become a consideration until the amplifier has a bandwith extending into the GHz range.

The rise time (and fall time) of an amplifier is unrelated to the propagation delay. The rise time is another way of specifying frequency response and phase margin. The higher the upper cut-off frequency, the faster the rise time. The greater the phase margin, the less ringing will be evident. As long as the amplifier in question has enough bandwidth and phase margin for the application, the rise time can be ignored. Again this is a parameter that is independant of circuit technology.

So a tube amp and a transistor amp both with a frequency response that extends to say 100kHz or more can be used together at audio frequencies with no phase issues arising.
Cheers,

Ralph.

grumpy
12-31-2005, 10:07 PM
One of the amps could be inverting and the other non-inverting. This could be true of any two different amps no matter what type of amplification device is used. Right?

Inverting or not is easily resolved once you determine if the amps don't match.
(swap speaker leads + for -, if necessary)

If you need to empirically determine if your amps match phase, you can set them up
with a mono signal to a pair of like speakers (FM inter-station noise works well)... if
the sound is obviously -not- coming from between the speakers, try swapping the
leads on one speaker (heck, try it anyway so you'll know what it sounds like).
If the amps are both inverting or both non-inverting, then you should hear a stable
image between the speakers, if one amp inverts and the other not, you should hear
a diffuse, indistinct sound that goes away if the leads of one speaker are reversed.

I don't know if the version of VT60 you have has balanced inputs or not, but make sure
you're testing with the preamp you intend for your system if you'll be mixing balanced
with single ended... some preamps invert some output types and not others :p

Amp in-out delay matching should not be an issue for any quality
audio-bandwidth amplifier, especially at typical bi-amp crossover frequencies.
Crossover matching (or lack thereof) is a significantly larger issue.

Amp gain, input matching, grounding, and "sound" will likely be areas to watch.
Using similarly designed amps generally simplifies this. That said, I've been running
an AR VT100mkII above 45Hz and a 6260 below... haven't ever noticed the 6260 clipping to warrant hauling in the noiser (fan) 6290. Quite happy with this arrangement.
You might need to consider gain structure if noise becomes an issue.

Cheers,

-grumpy

speakerdave
12-31-2005, 10:27 PM
This is really so cool. Thank you, gentlemen,

Zilch
12-31-2005, 10:47 PM
Just finished my first reading of D'Appolito. On P. 116, in discussing CLIO calibration:

"CLIO can also apply an additional correction file to any measurement through its Reference command. For example, closing S2 with S1 and S3 open will allow CLIO to measure the power amplifier frequency and phase response, which can be saved as a reference file and to correct measured loudspeaker response for an frequency-response variations or phase shift in the power amp." [Emphasis added.]

There's filters in them amps' signal paths, no?

CLIO can also be used as a dual-channel digital 'scope. I would now not be surprised to find, in detail, some phase differences between the two channels of the same amplifier, even.

Give me a couple of weeks here.... :)

grumpy
12-31-2005, 11:04 PM
I would now not be surprised to find, in detail, some phase differences between the two channels of the same amplifier, even.

No doubt you will. Low tens in degrees of phase isn't abnormal. I just make the claim
that a -few- degrees between channels, or difference at crossover between amps, will
be fairly moot. I'd guess the more audio-educated here could produce documented
evidence of this that predates my birth by a generation or two.

Even an old MAC 2100 isn't too bad (poked around re the "noisy L250 post") see some
measurements posted here that seem pertinent:

http://home.insightbb.com/~stephenwmoore/Speakers/Measurements/McIntosh_Amp.htm (http://home.insightbb.com/%7Estephenwmoore/Speakers/Measurements/McIntosh_Amp.htm)

Let us know how it goes :D.

-grumpy

(ah. re filters, yes... including output inductors to help clamp ultrasonic oscillation... my
understanding is that delay is inversely related to bandwidth and that most modern
amps have somewhat similar characteristics in that regard) good point.

LE15-Thumper
01-01-2006, 04:04 AM
The phase problem might not always be exactly 180 degrees. Propogation delay, the time it takes for the signal to get through an amp from input to output, can cause phase shifts that may be some odd value from amp to amp. Especially tube amps that may have capacitor coupled stages and transformer outputs. These devices have their own delaying affect on the signal flowing through it. Especially at different frequencies as well.

So I would say that using tube and solid state could pose some tough phase issues to correct.

ralphs99
01-01-2006, 06:30 AM
Hi Thumper,

I'll have to take issue with your assertion that the propagation delay can cause phase shifts. It tried to explain that the two effects are independant in an amplifier that doesn't bandwidth limit the signal of interest.

The delay from input to output for amplifier that doesn't bandwidth limit the signal of interest is defined as the propagation delay. It is a time domain effect. It not a term that is mentioned in the context of audio frequency amplifiers very often as it has no relevance.

Phase shift and the delay it produces (group delay) is a measure of bandwidth of the amplifier. It is more of a frequency domain effect. It is often spoken of in audio circles as many amplifiers and processors do not pass audio frequency signals without introducing a rolled-off amplitude response and attendant phase shift (and hence group delay).

Of course the bandwith of an amplifier is largely independant of the technology employed ie. tubes or transistors. A wide bandwidth (for audio frequency signals) amplifier can be designed with either technology.

I hope that makes things a bit clearer.

Cheers,
Ralph.

ralphs99
01-02-2006, 07:37 PM
Hi Dave,

Looking back at my posts in this thread I think I took things a bit off topic and never really answered your question.
Apart from the possible issue with each of your amps having a different polarity, there are no other phase issues to worry about.

Cheers,
Ralph.

moldyoldy
01-02-2006, 08:39 PM
Even if there are delay issues with your selected amps, most active XOs have independant band-delay settings. You should be able to compensate up to 180 degrees just like you would for driver (mis-)alignment during XO setup.