John Curl's Blowtorch preamplifier

[lineup]

Yes. If you take an ideal transmission line with a 50-ohm characteristic impedance and terminate it at one end with a 50-ohm resistor, you will see a resistive 50-ohm impedance looking into it at the other end, regardless of its length. This is the beauty of transmission lines.

If you mis-terminate it by some amount, as with, say, 40 or 60 ohms, then you will see non-idealities at the other end which are length-dependent.

Obviously, the best scenario is for a transmission line to be properly terminated at both ends. Under those conditions, a reflection will not occur at either end.

When unable to terminate properly at both ends, one can get good transmission line behavior by series terminating at the source or by shunt terminating at the receiver.

Cheers, Bob

hi john curl
hi bob cordell

Cordell,
looks like whatever you state, there will be some questions.
Some times probably because of slight misunderstandings.
I am not of the konwledge quality regarding the deeper technical side issues in audio, to be in a position to question your great knowledge bank.
But, do not be mistaken. I think I know VERY WELL what is good audio practice & what is not so good.
I can tell bad audio thinking / amplifiers, just by looking at a schmematic, by now.
Has taken me like 10-15 years of reading audio stuff to come to this
--------------------------------


We had a few posts in a preamplifier topic where we at the diy audio floor had some posts
regarding audio signal cables and the transfering of such signals to and from a preamplifier.

In this post
http://www.diyaudio.com/forums/showthread.php?postid=1522467#post1522467
I present the Audio Buffer / Cable Driver offered by member ACD from Denmark.
profile: http://www.diyaudio.com/forums/member.php?s=&action=getinfo&userid=6273

He has got a small audio hobby business, that offers good quality AUDIO KITS he has designed.

Here is the idea, in a capture I did from his PDF Product info:
Why use Impedance Line Level Buffers
http://www.diyaudio.com/forums/attachment.php?s=&postid=1522495&stamp=1211967404
And here is the PDF in question:
http://audio-innovation.eu/White Papers/Why use a Buffer.pdf

-----------------------------------------------

Most anything in audio chain has Pros & Cons.
* A 50 Ohm ( or some other impedance!!! ) cable driver can be the optimal thing.
Not only for HF signals, but there are benefits for BASS tones as well!

* On the downside, I would say:
Any active element in an audio chain creates non linear distortions.
Very high quality amplifier / active elements however
has got so low distortion, it does not matter anything.

With active I refer to anything that is not:
wire, resistor, capacitor

-----------------------------------------------



Would be good to hear what John or Robert
has got to add to this my summary of drive audio cables with a specific impedance,
using hifi buffers.
Today, there are plenty of Op-Amps that can be used for this.
Isn't this so, Mr Scott Wurcer?

 

[Bonsai]

Post #7159

Qoute
One thing that I like to do is right at the RCA connector where the signal comes in, run it through a series 49.9 ohm resistor and then a shunt 100 pF capacitor to the connector ground so that the interconnect is effectively terminated in 50 ohms at very high frequencies. This, of course, is a compromise, as not all interconnects have a characteristic impedance of 50 ohms, but I think it is better than nothing, and it forms a first line of defense at the very high frequencies where things get difficult to manage.

Cheers,
Bob

Bob, can you clarify this with respect to terminating correctly:-

Input side: If we terminate th e receiving end the signal passes through a 50 Ohm resistor and then onto the input device base or gate. on the base or gate side, a 100pf cap is connected to ground.

Output side of the driving source: it was not clear to me - a straight series 50 ohm resistor (obviously this assumes near zero output impedance)

Any practical ideas on how to deal with phono amp RFI susceptability? I guess on an MC amp, the 50 Ohm approach is not going to cut it. Maybe an RFI bead followed by a 50Ohm + 100pf series arrangement from input signal line to ground - so no resistive series connection in th e signal line.

 

[Steve Eddy]

I got a question regarding termination, say the source has 50ohm series resistor, the receiver has 50ohm+100pf termination.
Is this law still works if we put isolation transformer in the middle of the cable? 10k/10k isolation transfomer needs the receiver impedance >10k, 50ohm is way too low.

No, you wouldn't want to put a 10k:10k transformer ahead of a 50 ohm load. Though I suppose technically you could add some networks to the primary and the secondary that would present a 50 ohm load at RF frequencies.

se

 

[lumanauw]

Hi, Steve Eddy,

From your Belden link, I see this picture. For lower cable, the graph shows 3 area, below 1khz is low frequency area, between 1khz-10khz is transition area and above 10khz is high frequency area. The impedance is stable at high frequency area.
Is this mean the termination equations don't work in low frequency and transition frequency area?

 

[lumanauw]

Hi, Charles Hansen,

Thanks for the explenation

In this CD player's output schematic, after opamp, the signal goes to 22uf/35V capacitor, then to 220ohm, then to 100ohm before goes to output terminal.
Is this mean the output impedance of this CD player is 320ohm (220ohm+100ohm)? Is this source impedance (320ohm) is too high for 50 or 75ohm cable where the receiver end is not terminated (high impedance)?

 

[lumanauw]

Hi, PB2,

I'm not familiar with this termination/transmission line.
Usually the graphs shows digital data (square wave).
What happens in a not-properly terminated line cable (where there are reflections), when the signal is audio signal.
Will the original signal and reflection signals superimpose/subtracting each other?

 

[Charles Hansen]

What happens in a not-properly terminated line cable (where there are reflections), when the signal is audio signal.

Go back and read Bob Cordell's post 7196.

Characteristic impedance, reflections, terminations, and so forth are really only of interest to signals where the dimension of a wavelength are comparable to that of the cable. This is not the case with audio.

At least as long as we are talking about stereo systems. On the other hand, in the olden (analog) days, the telephone company did have to worry about this. Think about how long a telephone cable can be. Then if you are curious, try and figure out why the standard impedance for studio work came to be 600 ohms.

 

[1audio]

Reflections in audio are not really an issue since you need a significant part of a wavelength to get a reflection. In wire with an 85% velocity of propagation (typical of a solid insulation like PE) the wavelength of 20 KHz is approx 12 KM. This can be significant on phone communications in the old analog days but no one runs analog baseband signals over such long runs anymore.

Its the leading edges of fast rise events that have harmonics that may cause problems in an audio system. Or RFI/EMI that snuck into the system somewhere.

I have always tried to terminate cables to try to keep the RFI conducted on the cable from "spraying"into the system. The 50 Ohm/100 pF seems to be a fine solution for practical reality. Its also important to shunt the shields to the chassis with a small low inductance cap or several at each RCA.

I used to use Ferrite beads but they have enough of a bad rap I don't for audio any more.

 

[scott wurcer]

using hifi buffers.
Today, there are plenty of Op-Amps that can be used for this.
Isn't this so, Mr Scott Wurcer?

Yes, if you want to go that route there are amplifiers that have no problem with 50 Ohm single ended or ones that will drive terminated 100 Ohm twisted pair. Walt Jung is a big advocate for high power line stages. I used his AD823/AD815 composite amp for a long time. The AD815 would even drive 8 Ohms at low amplitude, 50 Ohms at 20V p-p no problem.

 

[Juergen Knoop]

lumanauw, you don't need to change your CDP.
320R is not to high for RG58 or others and there is no need for termination.
Pavel showed you some distracting simulations. The sub-ns risetime of the stimulus represents a frequency in the GHz range. This is not relevant for analog audio or the RF interference discussed by Glen/Edmond.
I'm using unterminated RG58 with my home stereo for years.
Of course this means nothing, because of my lousy equipment and ears.
regards

 

[chascode]

When unable to terminate properly at both ends, one can get good transmission line behavior by series terminating at the source or by shunt terminating at the receiver.

You are correct on this Bob. I prefer the shunt end termination. If you look at the Tek digital scopes they have an input impedance selector on one of the menus for 1 Meg or 50 ohms. In the "old" days Tek and HP sold an accessory 50 ohm input module that you connected to the input BNC connector for use with 50 ohm sources such as current probe amps (AM503), RF spectrum analyzers, etc.

My Sound Tech 1000A Stereo FM Alignment Generator has a 50 ohm output. For best performance with 75 ohm FM receivers I had to wind a custom 50:75 ohm balun.

Best, Chuck Hansen

 

[chascode]

Some amplifiers will overload and produce gross THD with very little injected RF carrier and some will do similar with significantly higher amounts of RF carrier.

One of the toughest tests in the EMI susceptibility test suite we use in aerospace, both MIL-STD-461/462 and RTCA DO-160, is the "buzzing relay" test, and it is one that is easily duplicated without expensive EMI test gear and a screen room.

You wire the normally-closed contact of a 28 vdc relay (we use a Mil-R-6061 relay) in series with its un-supressed coil and energize it. It will "door bell" and produce inductive discharge spikes in the conductors. The spikes are coupled to the input wires of the unit under test by securing the relay wiring to the input wires. This is done using both polarities of the 28 vdc source.

Of course, the abused relay contacts do not last very long in this test. We use a 4PDT relay so we have two sets of contacts to burn out and can extend the test setup test time.

You can also increase the distance between the relay wiring and input wiring to determine the audible and measurable effect on the input circuit.

Best, Chuck Hansen

 

[Sigurd Ruschkow]

50 or 75 Ohm coaxial cables are easy to find
but not only the cable is of importance but also the
connectors' char. impedance. BNC is 50 Ohm but what
is the char impedance of the RCA jacks and connectors we use?

I have only seen one specification of char imp for an RCA connector and that is WBT's NextGen which is 75 Ohms.

For balanced connectors, I think that Neutrik's are 110 Ohms.

Anyone seen specification for char imp for RCA jacks and connectors?




Sigurd

 

[jan.didden]

There is an AES Recommended Practice to use standard XLRs for 110 ohms digital interface. AFAIK there's no similar recommendation for RCAs.

Jan Didden

 

[PMA]

You guys should reconsider what you are trying to discuss. Cable termination is of no importance for audio signals, it is important only for EMI induced voltages or fast D/A residuals in audio. BNC or F-SMA are great, but again, we do not speak about 2GHz signals. The weakest point of RCA-CINCH is its construction, and the fact that it connects live earlier than ground wire, this results in SDI potential damage.

 

[1audio]

For what its worth an RCA type connector cannot be higher that about 35 Ohms impedance (OD/ID and air dielectric, it in the math). There are some tricks to making it less disruptive but it will never be 50 Ohms or 75 Ohms. I have looked all all of the alternatives with a TDR and confirmed that its not happening. But at 30 MHz (HD video analog) its not a gross issue.

For audio this is not a big issue. And a cable constructed with two signal conductors and a shield (current premium practice for audio interconnects) is far from impedance controlled anyway.

I think the more relevant issues here are controlling EMI ingress and sensitivity without adding big hum loop and grounding problems. This is where technique and lots of painful experience trumps simulations.

 

[lumanauw]

PCB tracks as short as possible?

 

[scott wurcer]

We used to keep an old Craftsman drill around the lab to generate RFI. You just run it right up against your circuit.

 

[Sigurd Ruschkow]

I agree.

The SDI issue is why I always have a small filter at the input even though strictly audio-wise it is not necessary nor wanted.

Mentioning RCA jacks, my favorite ones are Furutech's FP-106R.
Haven't found any better.



Sigurd

You guys should reconsider what you are trying to discuss. Cable termination is of no importance for audio signals, it is important only for EMI induced voltages or fast D/A residuals in audio. BNC or F-SMA are great, but again, we do not speak about 2GHz signals. The weakest point of RCA-CINCH is its construction, and the fact that it connects live earlier than ground wire, this results in SDI potential damage.

 

[Bob Cordell]

Post #7159

Qoute
One thing that I like to do is right at the RCA connector where the signal comes in, run it through a series 49.9 ohm resistor and then a shunt 100 pF capacitor to the connector ground so that the interconnect is effectively terminated in 50 ohms at very high frequencies. This, of course, is a compromise, as not all interconnects have a characteristic impedance of 50 ohms, but I think it is better than nothing, and it forms a first line of defense at the very high frequencies where things get difficult to manage.

Cheers,
Bob

Bob, can you clarify this with respect to terminating correctly:-

Input side: If we terminate th e receiving end the signal passes through a 50 Ohm resistor and then onto the input device base or gate. on the base or gate side, a 100pf cap is connected to ground.

Output side of the driving source: it was not clear to me - a straight series 50 ohm resistor (obviously this assumes near zero output impedance)

Any practical ideas on how to deal with phono amp RFI susceptability? I guess on an MC amp, the 50 Ohm approach is not going to cut it. Maybe an RFI bead followed by a 50Ohm + 100pf series arrangement from input signal line to ground - so no resistive series connection in th e signal line.

Hi Bonsai,

I'm sorry if my post created so much confusion. I'll try and clarify it and complete the thought.

When I mentioned the 49.9 ohm series resistor at the RCA followed by a shunt 100 pF, I was thinking as the amplifier designer who wants to build his amplifier to be as best-performing with whatever unknown source and interconnect the user employs.

The 49.9 ohms and 100 pF is just the first line of input EMI ingress defense, right at the connector. It properly terminates the interconnect with the best compromise at very high frequencies without placing an undue capacitive load on the preamp that is driving it. A 100 pF directly across the input would not be as desirable because it would mis-match the interconnect at very high frequencies and possibly cause resonances. The 49.9 ohm resistor in that regard provides good damping.

When the input lead gets to the circuit board, there is additional lowpass filtering of the kind ordinarily found at the input of an amplifier, usually one or two poles, and perhaps staggered to produce a linear phase rolloff approximation. These LPF sections have higher resistive impedances associated with them. Together, they should have less than perhaps 0.1 dB of attenuation at 20 kHz.

Now let me put my preamp manufacturer hat on. In an earlier post, Edmond stated that he would not feed the output of a preamp to the interconnect without at least some series resistance. I agree with him 100% and always follow this practice. This is an easy way to achieve a series source termination of 50 ohms (the considerations may be a little different for preamps with naturally higher output impedances than 50 ohms, like tube preamps).

So if you follow these practices at both ends, you will have a transmission line that is properly terminated at very high frequencies at both ends if it is 50 ohms. It will always be properly terminated at the source end. Even if the interconnect is not 50 ohms, perhaps 75 ohms, these practices will be quite effective. Transmission lines behave best when they see a resistive termination of some sort at both ends, so there are few oppunities for resonances.

Finally, it is important to think of the loudspeaker cable as a transmission line with a characteristic impedance as well. Ideally, the loudspeaker should present a resistive load at its input terminals out to VHF frequencies to give that transmission line at least some kind of resistive termination, even if the impedance will not be right. There are those who advocate putting a Zobel network at the loudspeaker terminals for this reason. 100 ohms and 0.05 uF in series is not a bad choice. Some loudspeakers that use resistive pads for the tweeters have an input impedance that looks pretty resistive up to quite high frequencies, so they might not benefit as much, but this depends somewhat on their internal wiring.

Cheers,
Bob