CMRR

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Gus

Well-known member
Joined
Jun 3, 2004
Messages
5,249
Location
n
I often wonder how many new "microphone designers" understand CMRR or even how an emitter follower works

Discuss
 
I guess that depends on what you mean by "new"... New microphones, or new designers? I expect new microphones from established microphone companies will understand the written and unwritten performance criteria. New designers in any area can have gaps in their understanding. The benefit for new designers inside existing mic companies is that they can draw upon experienced engineers to mentor them.

Last century while working at Peavey I took advantage of having an in house transducer design group to pick the brains of the senior microphone engineers. I was curious about optimal mic terminations and did not get a concise single answer, but these guys knew their business.

JR
 
I am not so sure, but I do not think anybody who does not know how an emitter follower works would be allowed anywhere near designing in any electronics manufacturer.

However, you certainly have a point when it comes to CMRR. I am not going to give names but I am currently reading a book on VLSI, the authors of which are university lecturers. One presented as a project scientist and the other one is proudly credited with having over five years of industrial experiences, and it certainly shows.

On differential signalling they state that , given the source and load impedances are equal, both conductors are induced with identical electromagnetic interference, and since the receiving circuitry only detects the difference between the two conductors, the technique rejects the noise.

Now, this is not only a very lousy description, as it implies that somehow keeping both the source and load impedances equal is the thing to do for digital signal transmission, but also clearly shows the authors' lack of knowledge on CMRR.

As we all know keeping the source and load impedances equal gives you the perfect power transfer but with 6dB signal amplitude attenuation, and this is certainly not the way to transmit digital signals let alone analogue. And for CMRR you certainly do not want high source impedance either.

After practicing electronics for long years as a self taught person, I bit the bullet and have recently completed my honours degree in EEE as a (very) mature student. I found that the younger generation lecturers are really like dynamite in digital, in particular with embedded, and naturally in programming. But they lack good analogue experience. The lecturer who taught analogue in year three was/is a very good all round engineer and excellent tutor. But even he made a fundamental mistake when explaining instrumentation amplifier application. He drew a schematic and showed the remote transducer driver circuit tied to case chassis ground. I told him that he would have to float that as through the mains earth one leg of the differential line would be grounded and his CMRR would go out of the window.

Things like that would not go past the eagle eyes of the masters here.
 
Last edited:
I am currently reading a book on VLSI, the authors of which are university lecturers. One presented as a project scientist and the other one is proudly credited with having over five years of industrial experiences, and it certainly shows.
On differential signalling they state that , given the source and load impedances are equal, both conductors are induced with identical electromagnetic interference, and since the receiving circuitry only detects the difference between the two conductors, the technique rejects the noise.
Now, this is not only a very lousy description, as it implies that somehow keeping both the source and load impedances equal is the thing to do for digital signal transmission, but also clearly shows that the authors desperately lack knowledge on CMRR.
As we all know keeping the source and load impedances equal gives you the perfect power transfer but with 6dB signal amplitude attenuation, and this is certainly not the way to transmit digital signals let alone analogue. And for CMRR you certainly do not want high source impedance either.

I apologize in advance if I misunderstood something, but I don’t see what that quoted author was wrong about. Hi speed (high frequency) systems are allways connected using equal source, load and characteristic impedance interconnection to avoid reflection in the cable.

IIRC, MisterCMRR wrote somewhere online about 600 ohms lines, where telephone connections in the beginning of telephony were very long and reflection could happen in the audio spectrum as well. This is generally not the case now so modern devices are no longer produced with 600 ohms inputs and outputs.

An example (AES3 interconection) is shown in the figure:
 

Attachments

  • aes.png
    aes.png
    89.9 KB · Views: 16
Last edited:
Indeed transmission lines are somewhat different considerations than relatively low frequency, short, audio interfaces. Long enough audio lines (like telephone) also exhibit transmission line characteristics. The old 600 ohm I/O standards for audio are related to this.

HF video kind of has one foot in both worlds demanding proper impedance terminations (more commonly 75 ohm).

JR
 
I completely understand that. But as I mentioned, the description is lousy within the context of CMRR. What if medium or low speed transmission? Low source impedance is still the desirable design parameter. But you are correct, say, if the authors limited the discussion to high speed.
 
Whether or not a signal cable is a "transmission line" can be a rather fuzzy definition. As a "rule of thumb," if a cable is physically longer than 1/10th of an electrical wavelength at the highest frequency of interest in the signal, the signal distortion due to reflections in an unterminated line will be on the order of 10% or about 1 dB. For analog audio (taking 20 kHz as the highest frequency of interest), one wavelength is on the order of 35,000 feet. So cables under about 3,500 feet will show only slight ripples in frequency response in the upper octave. These are very rough numbers but hopefully will make clear that, at GHz frequencies (or digital data rates, termination can become necessary for cables only a few inches long.

I know from my experience teaching about balanced interfaces and CMRR that the subject is widely misunderstood by professors and in many textbooks. In the example cited earlier, equal induced common-mode voltages (presumably from an external magnetic or electric field) cannot be assured by simply matching impedances - cable construction (such as twisting conductors) is also a major factor - as is the form of shielding, if used. A good example is foil shielding with a "drain" wire - which has been shown to be a serious problem at audio frequencies when current flows in the shield.
 
Do you have an opinion about star-quad....? I did some crude comparisons back in the 90s(?) between star quad and normal*** mic cables wrapped around a fluorescent light fixture. The star quad wrap made a noticeable improvement.

JR

**** I don't recall specifics about the normal cable I used, whatever Peavey was making at the time, and I had laying around my test bench. FWIW the star quad wasn't even official branded star quad but a several hundred foot run our cable vendor made up for us to check out. That said it really seemed to work. I seriously considered selling it as an accessory product for our AMR (recording products) division but lost interest. Some guy working in purchasing ended up with that test reel of good mic cable. I recall looking for it later and nobody knew where it went (I figured it out).
 
"Star-Quad" typically has about 40 dB (100x) better immunity to external AC magnetic fields, based on measurements I made using a mic preamp, a 150 Ω dummy mic, and a hand-held tape demagnetizer as field source. This performance is rooted in actual science and will apply to any 4-conductor cable that's twisted with reasonable precision. We know that, if the two conductors in a cable occupied exactly the same space, there would be zero induced voltage difference between them from any external field - but that's literally impossible. But there are ways to indirectly do the same thing. For instance, the effective center of a uniform cylindrical shield is it's center-line. Since this is also the center-line of the inner conductor, the two conductors are effectively in the same place. Similarly, if opposing pairs of a 4-concuctor cable are tied in parallel, the effective center-line is a line midway between them. Now parallel the other pair and we effectively have a "pair" of conductors whose center-lines coincide. This, along with twisting, which effectively makes the average position of the two conductors the same, makes the "loop area" effectively zero - limited only by the accuracy of physical wire positioning. The only downsides to using it are increased cable capacitance and, depending on gauge and insulation thickness, increased cable diameter. It's too bad that so many outrageous claims are made for "magic" audio cables ... but }star-quad" really does perform magnetic magic!
 
I never saw star quad used in a music recording situation , TV and radio stations is the usual application which could be a hostile environment for audio with all kinds of mains and RF signals routed through conduits etc . Interesting explanation of the theory behind it though . I see some Van-Damme tour grade starquad up on ebay , 1.68 euros per meter , I might give it a try . Now with modern ADC and DAC capable of -130db or more the extra freedom from induced noise might be worth having .

I still have a couple of those old Peavey mic cables , they have stood the test of time , it has cotton in-fill between the conductors to preserve twist geometry ,proper braided sheild and neutric xlr plugs . I did up a bunch of mic cables years ago from Belden 8412 , with careful usage they last a lifetime .
 
Sadly star-quad (that actually works) was competing with snake oil funny wire sellers with bigger ad budgets.

As Bill shared it has higher capacitance but that is not an issued for line level I/O interfaces.

Perhaps don't use it for guitar cables.

JR
 
I did up a few guitar cables with the Belden 2 core mic cable for friends one time ,
I denoted the instrument end with a right angled jack connecting the screen and one of the central conductors together , at the amp end I left screen disconnected and made ground through one of the signal wires . I might have that the wrong way round and screen was connected at the amp end , there was a theory behind it I read somewhere but I forgot now . They worked great and never broke , but most likely got left behind or lost along the way .
 
Sadly it is a 'feature' of the internet that all too often the reason WHY something is good (or bad) gets lost somewhere, either deliberately or accidentally. So many 'modern' technical issues observed by the early electrical pioneers (Michael Faraday, Maxwell and so on with applications beyond the wildest dreams of the time. Imagine trying to receive (and resolve) a message from a 'spark transmitter' across the Atlantic now?
Having been involved with TV and radio installations in the past the use of Star quad for MIC lines was the usual application where many hundreds of feet sometimes past dimmer racks and lamp cabling with 120 Amp power feeds being pretty normal. The explanation WHY it star quad should be best in THOSE situations (low impedance mics being 'relatively' less affected by cable capacitance. Also the quality of the cable construction so good geometry and use of fillers to maintain it.

I heard an anecdote about a multicore cable that had been installed where ONE pair out of 12 HUMMED when the other 11 pairs were swapped around (same source/destination etc) and it was eventually found that that one pair had been at the 'crossover' point where the cable manufacturers sneakily join on a new bit of inner conductors and there was a resulting few feet where they were inadequately twisted. (Or it had been squashed which had resulted in an area of no twists).
If course the 'old hands' at valve gear assembly knew about TIGHTLY twisted heater supplies, tucked into corners of the chassis to reduce 'radiation' of their magnetic hum field.
Matt S
 
John, are you saying that painting my ICs with magic silver audio paint doesn't keep the electron in the signal path and doesn't prevent errors? All that money... Now, I'll have to buy mercury filled speaker cables.
 
One thing to remember about CMMR, if you don't have a hostile environment, most things will work fine. With the original 512b and c mic pres, I criss-crossed the mic line coming from the rear connector to the front one, jumping from top to bottom so it looked like a twisted pair, but currently I just run them close together and make sure they are always the same path and length, on the bottom side of the PCB. both worked fine. Like I said before, when designing the 318a distribution amp, that was designed for hostile places like a remote truck or station under a transmitter, but as long as the audio cables are away from the AC cables and digital cables, everything seems fine. On a live stage or theater where you have a huge amount of AC fields, then it's more important. But then you have to deal with Eddie currents as well, so your ground plane and chassis design may be more important than the super nulled CMMR input. So many things to consider...
 
John, are you saying that painting my ICs with magic silver audio paint doesn't keep the electron in the signal path and doesn't prevent errors? All that money... Now, I'll have to buy mercury filled speaker cables.
I am not aware of any benefit from silver paint, it may kill germs. ;)

I've shared this before but fun is designing inexpensive fixed install background music mixer/amps with 70/100V audio output transformers and mic input transformers inside the same modest sized metal chassis. The lower wattage SKUs were not any easier, in fact the smaller chassis made them harder with same voltages just closer together.
===
Another challenge was designing topbox powered mixers with hundreds of watts of amplifier squeezed into a package barely larger than a breadbox. The old school designs also had to accommodate spring reverbs. The evolution of digital reverbs cheap enough to use inside powered mixers was a relief to eliminate one of the several design difficulties, while some of the early digital efx were barely usable for even that lo-fi application.

I pretty much had to use whatever the digital group gave me. They were inclined to cram in too many features not having to deal with the customer complaints. A popular practice with early digital efx was to use a lot of HF pre/de-emphasis to improve perceived noise floor but some of the efx (like reverbs) using too much recirculation could saturate accumulators (?) and sound nasty even with modest level HF content in audio signals. The first version they gave my group to use, had a red LED to indicate HF overload, but it was impossible to establish a usable gain structure that could avoid overload, and still deliver a decent noise floor. I brought it back to the digital group and told them to add a simple JFET limiter in front of the EFX that would limit audio signal level, only when HF OL flag was indicating overload. This worked like a charm at least for that product series.

JR

PS; When I had to demo these marginal digital efx inside my studio products area at trade shows I would patch a de-esser in series with the efx to reduce damage from HF content overloading the efx.
 
John, are you saying that painting my ICs with magic silver audio paint doesn't keep the electron in the signal path and doesn't prevent errors? All that money... Now, I'll have to buy mercury filled speaker cables.
Exactly. What you need is the special green Sharpie that you run all around the edge of your CDs to stop bad photons getting in.

Cheers

Ian
 
..................
I know from my experience teaching about balanced interfaces and CMRR that the subject is widely misunderstood by professors and in many textbooks. In the example cited earlier, equal induced common-mode voltages (presumably from an external magnetic or electric field) cannot be assured by simply matching impedances - cable construction (such as twisting conductors) is also a major factor - as is the form of shielding, if used. A good example is foil shielding with a "drain" wire - which has been shown to be a serious problem at audio frequencies when current flows in the shield.
After re-visiting my post it seems I have been a bit economical with what's been said by the authors and missed an important point. They do mention that the differential scheme works on twisted pair lines like RS232, but they also include audio which is not correct, unless of course, and as you mentioned, one decides to transmit audio over 35,000ft using wires at 20kHz. At the same time I am also partially wrong with my comments.

However, what I am not entirely sure about is how terminating the line with an impedance equal to the source can assure equal induction of common mode voltage on a say twisted pair. I have dug in a bit deeper on this and consulted a couple of good books, and I can not see any reference to it. So, it would be great if you could comment.

My thinking is that, and as you also rightly commented, as long as the wire pair is twisted properly (evenly) then the common mode voltage should induce (theoretically) equally on both wires. So, I see no effect of the termination resistor on this in bridge termination. In fact for single ended (on each line individually) and AC termination the resistors would require good matching to minimise the common mode voltage developed across each of them.
 
Back
Top