John Curl's Blowtorch preamplifier part III

Status
Not open for further replies.
Jump to Latest
#20,465
RNMarsh said:
In other words, do these 2 Z curves sound different on any particular PA?
Click to expand...
Seems easy to figure out.
You need two amps (or an amp and a headphone amp).
The first one (in an another room acousticaly isolated enough) charged alternatively with R and Speaker, the second one, or the headphone) taking its signal from the two sides of the speaker cable (why not trying this ?) to listen.
( Sighted, flawed, subjective. As it should be ;-)


To measure, it is easy, only one amp and have some ear plugs, because it is painful.

I bet you will not find great difference, if you take the signal right at the speaker's output of the first amp. More if it is the other side of the Speaker cord (serial impedance of the cable)

What do you think ?
 
#20,466
Markw4 said:
Are there publications or formal theories that say a whole wave needs to fit inside a line before anything starts traveling, or whatever it is you claim?
Click to expand...

Never spoke about whole wave. Our audio waves are several thousand times longer than the cable. Yes there is a transition zone that you could see in the simulation I posted - provided that you understand the plot.
 
#20,467
simon7000 said:
It requires only one adjustable resistor to work. You can buy the transformers off the shelf. Pretty standard circuit and usage. Used by conference call telephone folks, call in talk shows, even old fashioned computer modems.

https://catalog.triadmagnetics.com/Asset/TY-146P.pdf
Click to expand...

RNMarsh said:
a directional coupler for 10Khz is available.

Pulsar directional coupler C3-P2-411 covers the range 0.01-35 MHz,
Click to expand...

Way under what would be required to discriminate the level of reflections discussed here. 26dB directivity is low, would not allow measuring SWR values of under 1.3 - 1.4 with a reasonable incertitude. Anyway, it would be interesting attempting to repeat CBs experiment using such transformers/directional bridges. Using a directional bridge also requires calibration at the characteristic impedance of the cable under investigation, around the frequency range of interest, to eliminate as much as possible systematic/correlated errors. Something I don't see CB doing.
 
#20,468
jneutron said:
in his part two article, he demonstrates reflections of a 10 Khz signal in a 4.9 meter long cable. So, 5 times longer wavelength, 80% of the length..

Even in his part 1 article, figure 8 on page 7 (which I previously pointed out and you ignored), a 10Khz wave showing reflection in a 4.9 meter cable.

What is of interest in figure 8, is that he shows the delay. That delay is the multiple reflection pass composite, and it will exactly match the lumped LCR element analysis.

Or, will you ignore that as well?

jn
Click to expand...


jn - you have got to be kidding, right?

Its not the cable doing the 'reflecting', its the reactive speaker + cross over you are seeing. It doesn't matter what approach you use to analyse it, the cable LC is swamped by the load LC.

Are we talking cross purposes here or are you just having fun(again)?
 
#20,471
PMA said:
Never spoke about whole wave. Our audio waves are several thousand times longer than the cable. Yes there is a transition zone that you could see in the simulation I posted - provided that you understand the plot.
Click to expand...

Didn't read every message in this discussion (too much garbage challenged my patience) but I think some are talking about a 10KHz square wave, with fast rise/fall edges (something that is never to be found in audio, discussions about "attack" is junk). Such a signal has high order harmonics along with the 10KHz fundamental, which could reflect on a mismatched long enough cable and as a result affect the waveform along the cable. That's a fact, but its audio relevance is IMO less than the effect of moonshine on your next amplifier stability.
 
#20,474
PMA said:
I am permanently repeating that any reflection of such wave in some 4 - 6 m cable is a nonsense, if we avoid sudden turn-on/off. Web discussions are really exhausting.
Click to expand...

Let me pose a simpler example a match terminated source and a 10m 100 Ohm line terminated by an 8 Ohm resistor. The solution here in t-line terms is simply the superposition of the forward and reflected wave (at all frequencies). Are you telling me this is not a solution to this system?

One of the problems with simulators BTW is that all components have zero physical extent.
 
Last edited:
#20,475
PMA said:
any reflection of such wave in some 4 - 6 m cable is a nonsense, if we avoid sudden turn-on/off.
Click to expand...

Why would reflections just cease, when a very long cable (with reflections) is shortened?
What is the actual mechanism for that? Only the propagation delay of the cable changes.
The lumped approximation is called such for a reason.


Here is a decent demo for those new to transmission lines (poor production values, though).
YouTube
 
Last edited:
#20,476
syn08 said:
Way under what would be required to discriminate the level of reflections discussed here. 26dB directivity is low, would not allow measuring SWR values of under 1.3 - 1.4 with a reasonable incertitude. Anyway, it would be interesting attempting to repeat CBs experiment using such transformers/directional bridges. Using a directional bridge also requires calibration at the characteristic impedance of the cable under investigation, around the frequency range of interest, to eliminate as much as possible systematic/correlated errors. Something I don't see CB doing.
Click to expand...

Been using similar for years now as I do actually specify custom cables for my installations. Those particular transformers are way better than their minimum specs.

I did have a problem with Belden delivering one order. The sales folks kept slipping the shipment date. So I mentioned to one of the product engineers the penalty was $2,500,000.00 for the first day late. They had no problem after that getting it to me on time.
 
#20,478
simon7000 said:
It is easy to distinguish between audio frequency signals coming and going. The original circuit used what is called a hybrid transformer. Today it is done with opamps and a balanced bridge. Yes there is some signal reflection whenever there is mismatch. Perhaps you have used a telephone and noticed what is called the sidetone or the amount of your voice that is in your handset varies. That is due to reflections from the old fashioned two wire telephone cable.

As to loudspeaker parameter shifts the DC resistance can go from 4 ohms to 16 ohms in a fully driven professional loudspeaker. Consumer grade midranges might go from 6 to 9 ohms. The inductance on a consumer midrange might go from .2 mH all the way to .4 mH. These parameters used to be monitored as the first large scale sound systems first had the ability to do so. Once the parameters were known it was much easier to set power ratings and limiters. Some gear will still shown the total impedance change.

Attached is a transformer hybrid circuit.
Click to expand...

Having built and purposed these for radio call-in use, I will say they are very sensitive to input and output port impedances. Not shown in this schematic are the RC networks required to balance the port impedances over the 300-3000 Hz voice range of the phone application. Just the small imbalances between incoming lines in a multi-line system can cause a 10+ dB increase in bleedthrough between ports....which in the case of a radio talk show can cause feedback.

In the application proposed here, the variable impedance of the speaker would cause the port imbalance to mirror the impedance curve of the speaker itself.

Or so I suspect...
Howie
 
Status
Not open for further replies.
Top Bottom