EMRFD Message Archive 14171

Message Date From Subject
14171 2017-07-27 16:26:28 timshoppa Common-base differential mic preamp

I have been playing around with different "suggested" mic preamps in the shack.


I am starting, not with an electret mic, but from a dynamic mic that produces low hundred microvolts at an impedance of 200 ohms. I want to get to line level, about 1V pk.


The microphone and cable is inherently differential so I tried some differential designs.


I tried a INA217 based preamp but, despite putting CLC pi network low pass filter on the input and building it "RF-clean" dead bug style, it was horribly sensitive to RF in the shack. Not just RF that I was transmitting, but RF from the local 630kHz BCB transmitter. To be fair, I have to admit that this BCB transmitter makes enough RF that an open circuit scope probe has a few hundred millivolts on it. I'm not sure of all the reasons for this RF sensitivity but I kinda feel that the high-impedance of the input stage makes it more susceptible.


I then looked in EMRFD and ran across KK7B's 50-ohm input impedance common-base audio preamp he uses for the low-level output of the mixers in his R2 etc. This looked pretty good. Play around a little bit with values, and it becomes a 100 ohm input impedance. The R2 has two of them (one for I and one for Q), I use two of them one for each of the differential wires from the microphone - thus a differential input impedance of 200 ohms.


The microphone is hooked up straight to the emitters without a capacitor, so the microphone is at about 2VDC. See below for some notes on DC balance.


The common-base input stage takes the low hundred microvolts input and turns it into tens of millivolts.


The Common base input stage then feeds an op-amp differential amplifier with modest voltage gain of about 20.


You can see my LTSpice schematic here: http://n3qe.org/mic-preamp.png


Note that the two 100 ohm resistors on the microphone are not real parts in the preamp, they are just there to represent the 200 ohm impedance of the microphone. The "10V" supply is actually a zener in real life. The 6V supply is actually a virtual ground made by an op amp and a voltage divider from 12V. Also I use NE5532's in real life but replace them with a LT op-amp for LTSpice modeling.


I discovered some interesting things. The "super decoupler" commonly used in KK7B's designs seems unnecessary. Maybe battery-powered equipment benefits more from the super decoupler. I use a simple single zener instead, to keep a constant 10V on the resistors common base preamp stage (my actual VCC varies between 12V and 15V). I know zeners can be noisy but I hear nothing from it, maybe because both channels get the same noise and it subtracts out?


I was having a hard time getting the DC voltage balance at the collectors of the two transistors especially when I hooked up the microphone (which is a low DC impedance) between the emitters. I tried a few things and discovered that if I just hooked the bases directly to each other, the DC voltage balance seems to follow automatically. I do not fully understand the deep logic behind this but it works.


I am very happy with the result.I built it for ham use but have been sampling it hooked to some real musical mics. The gain as I show it is just about right for driving headphones if you want to listen to the "raw" output directly.


Critiques? Suggestions? Did I work too hard to make what is just a microphone preamp?


Tim N3QE

14173 2017-07-29 23:12:07 Gary Johnson Re: Common-base differential mic preamp
Since I�m an old audio guy, here are a some comments about your interesting mic preamp project. Forgive my rambling but I do tend to get carried away when something like this comes up :-)

The likely reason you ran into RFI problems with the INA217 instrumentation amp is because of a shielding problem and the attendant common-mode conducted RF energy. If you fix that problem, it will make an excellent mic preamp and these days would be my first choice. Here are the standard procedures I used to teach in my EMC classes during a segment on signal conditioning. These steps apply to nearly any low-level, low-frequency signal input.

1. Congratulations on using shielded twisted pair and a differential input. The entire world of pro audio and commercial broadcast uses this exclusively for many good engineering reasons.
2. Connect the shield to the OUTSIDE of the preamp enclosure. Do NOT run it inside thru a connector pin. This is the infamous Pin One Problem from the pro audio world. The shield has two sides: Keep the RF on the outside! BTW, many commercial ham rigs have this problem. See http://audiosystemsgroup.com/RFI-Ham.pdf
3. If you can�t connect the shield to the outside, add a common-mode choke to the shielded cable, i.e, wrap it around a toroid.
4. Add ferrite beads and a common-mode choke to the signal leads immediately after they enter the enclosure. The choke could be something commercial or perhaps 15 bifilar turns on a ferrite core such as an FT50-43.
5. Apply a matched pair of RF bypass capacitors to the signal ground right at the amplifier input pins.
6. You may also build an LC filter but try to maintain balance by making the common-mode capacitor on the order of 10 times the value of the caps that go to ground.
7. Do not ignore the rest of your system with regards to ingress paths for RFI via other conductors. This includes all signal lines, power connections, etc. Maintaining shield integrity is great, but can be for naught if unfiltered conductors are penetrating it.

One other thing you may have already heard is that it�s important to avoid dc current thru dynamic mics as it may cause frequency response anomalies or distortion. I could call my old friend who�s a VP at Shure to find out exactly what the threshold of pain is, but I�m pretty sure it�s at the microamp level.

Check the specs for your mic (if you can), and you will find that the recommended load impedance is several times that of the mic itself. Most commercial mixers have Zin between 1k and 2k for low-Z mics. If Zin is too low, of course the level is reduced below spec and the frequency response is also affected. So your efforts to make the preamp look like 200 ohms are actually counterproductive.

The reason that you didn�t need such heavy filtering on the power supply is that the same supply is feeding both sides of the differential amplifier. As long as the CMRR of the next stage is sufficiently high, that noise will cancel. A good experiment is to inject a signal into the power supply line and see how much shows up at the output, thus directly measuring CMRR.

Your offset problem at the collectors has to do with transistor matching. In a differential amplifier, it�s best to use matched pairs, normally in a common package. In addition, the bias resistors need to be matched, usually to 1% or better. This matching improves even-order distortion and stabilizes the bias over temperature. Matched pairs (BJT or JFET) are now hard to come by. One of the few companies who make them anymore is http://www.linearsystems.com/

Attached is an article by the late great W. Marshall Leach, who was perhaps best-known as an EE professor who taught his students at Georgia Tech about high-performance audio. It�s a very low-noise, common-base preamp for moving coil cartridges (not differential, however). This preamp is unusual in that it calls for a floating power supply to guarantee zero dc current thru the transducer. He also goes on to show some different topologies. Excellent article.

Another good way to build a mic preamp is to use a high-quality transformer such as those from Jensen or Lundahl. They can give you essentially noise-free voltage gain (up to 20 dB), high CMRR, pleasingly low distortion, and excellent RF rejection due to limited bandwidth and a Faraday shield. They work best with FET preamps. They should have a mu metal shield and be located away from big power transformers to avoid hum pickup. Big drawbacks are cost and size/weight. Still, transformer-based preamps make a really fun project.

Gary NA6O

[see attached file: Signal conditioner.pdf]
[see attached file: Leach Moving Coil Cartridge Head Amps.pdf]





[Non-text portions of this message have been removed]
14174 2017-07-30 06:20:29 Graham / KE9H Re: Common-base differential mic preamp
Hi Gary:

Thanks for the discussion on differential audio.
I would like to read the articles you mentioned.
But, Yahoo stripped them off.
Could you post them to the group files section, or I could send you my email address off the group.

Thanks,
--- Graham / KE9H

==

==

On Sun, Jul 30, 2017 at 1:11 AM, Gary Johns
14175 2017-07-30 07:55:52 arfghans Re: Common-base differential mic preamp
The attachments are now in the new NA6O folder.

There is one other thing I should have emphasized that could be a root cause of the problems that Tim initially saw with the instrumentation amplifier IC. Since a microphone is a floating source, there is no path for the amplifier's input bias current to return to signal common. In that case, the inputs will float toward the supply rails, leaving the amplifier at or near the limits of its common-mode voltage range. Then it only takes a small signal to cause distortion, and RF signals may be demodulated by activating the input protection diodes. The cure is simple: A resistor to ground. This is shown in the INA217 applicati
14176 2017-07-30 12:22:34 timshoppa Re: Common-base differential mic preamp
As an update - I have revisited the INA217 microphone preamp prototype, taking into account some of the grounding/shielding issues mentioned by NA6O, as well as some of the wiring techniques I've seen in phono preamps. It seems remarkably less susceptible to BCB RFI today than when I last checked it out but I made several changes at once so I'm not sure exactly which one was the biggest improvement.

I am using a virtual ground, not a "real ground", and while I always had 1K resistors to virtual ground on the microphone leads it seems possible that a wiring error or accidental short circuit had let them reach one rail or the other.

I think it was either KK7B or W7ZOI who wrote a remark in the past, about how low-level audio techniques are not necessarily the same as RF techniques.

I still really am captivated by the common-base approach. Thanks so much for posting that Leach article. In it he mentions a clone where they call it a "current to voltage converter". My thoughts about the common-base stage were kinda similar, but in addition to the impedance conversion/voltage step-up, I never left out how it also provides actual power gain along the way HI HI.

NA6O - great to meet you in EMRFD group after over a hundred contest and CWOps QSO's :-). I see I've also worked Wes W7ZOI a couple times over the years.

Tim N3QE
14177 2017-07-30 19:59:02 AD7ZU Re: Common-base differential mic preamp
I am using a virtual ground, not a "real ground", and while I always had 1K resistors to virtual ground on the microphone leads it seems possible that a wiring error or accidental short circuit had let them reach one rail or the other.

There are inherent problems that are far from obvious using a single supply / split rail design.  I recently cured a number of issues using a high gain instrumentation amplifier in a DC receiver by reluctantly going to a dual supply.  It is not convenient when designing for a single 12v input supply but if you have an AC supply or are not designing anything portable it is easier to implement.  I will be using a charge pump device (LTC3260) or a dual battery in the final design.. I am working
14178 2017-07-31 18:11:26 Bob Mix Re: Common-base differential mic preamp

Interesting that Charles Regen Kitchin is the author.

Bob