Mic Preamp Schematic Collection

[clintrubber]

[quote author="mediatechnology"]Thanks Peter. You hit the bull's eye.[/quote]
But files gone again... :wink:
(because of the WJ-update (pdf-version), see http://waltjung.org/whatsnew.html )

Regards,

Peter

 

[JohnRoberts]

How often do mic cables short while phantom power is turned on, but no phantom using mic is plugged in? OK, perhaps in cheap consoles with global phantom switch. If a mic is using phantom when this rare short occurs, the peak voltage step and thus peak current will be reduced perhaps in half. More -V rail voltage will also reduce spike severity (depending on topology of clamp).

The interest in this subject makes me wonder if there isn't another failure vector, like perhaps microphones with transformer outputs whose center tap it tied directly to a zener shunt or such. While not as severe as a cable short this could also generate a current impulse. [Basic rule of design... never expect the other guy to worry about you.]

Taking a step back, this sounds like a classic fuse application. A small 1/4A or 1/8A fuse in series would surely allow truly huge input caps and zero series impedance while protecting fast small signal clamp diodes.

Folks willing to spend a little more for uninterrupted service in the presence of cable or mic faults could use a PTC (resettable) fuse that would also protect and serve. Theres a trade off between cold R, trip I, and max voltage but this looks like it might be usable. http://catalog.tycoelectronics.com/...452&M=PPROP&BML=&LG=1&I=13&G=G&N=1&IDS=435522

FWIW I haven't had problems personally with this matter, but I routinely used smaller (22uf) caps, and didn't directly input to a IC (typically discrete bipolar front end), with appropriate small signal diode clamps to prevent zenering (noise or failure) or voltage stressing (breakdown failure) later IC stages.

JR

 

[JohnRoberts]

Agreed, most designers who understand thermal noise and such will not be enthusiastic about pushing input series resistance to 50 ohms, and if they've bother to look at LF CMR they'll want to push the cap values to reduce tolerance sensitivity @ 120hz.

I'm not sure about lamps but I guess DCR cold can be pretty low and any inductance might be used to form a RF filter :wink:

PTC fuse I linked to is something like 1 ohm cold for 1/2 amp trip, but I didn't bother to check price, because I, like you don't plan to use input capacitors, but I also prefer to roll discrete front ends that may be easier to protect. [Note: my current thinking is just flying the input pair up at phantom voltage. All opamps are happy down around ground. ]

Regarding protecting the 15120 in your app. Indeed clamping the two rails to each other and just letting the whole thing swing if insulted from outside is not crazy. Why sink a high current pulse if you can finesse it?

JR

 

[aortizjr]

I think I am going to dive in and breadboard Design G. It looks pretty simple.

Probably use a Cinemag CMMI-2C for the transformer.

How do you feel about putting the polarity reverse on the input instead of the output?

I noticed that a lot of the Jensen schematics do that. I would guess the impedance through the switch, but never really tested to hear any differences.

 

[Samuel Groner]

Properly implemented, there should be not much difference between the two version. I prefer to have switches at line level as the signal runs to the switch and back will be less susceptible to hum pickup and possible contact degradation.

Jensen probably shows the polarity switch at the input because most of their schematic don't have a fully balanced output.

Samuel

 

[mark bohrer]

Samuel:
I've quoted an application note contract for a design using Analog Devices' new dual audio op amp. It looks a lot like your design G with two op amps and Jensen 1:2 input transformer.

Did you ever build your circuit? If so, what distortion at max gain and EIN results did you get?

Thanks.

Mark

 

[Samuel Groner]

Did you ever build your circuit? If so, what distortion at max gain and EIN results did you get?

Didn't build it yet. EIN calculates to -128.72 dBu for a 150 ohm source impedance and a JT-16-A at 20 kHz BW (using typical 1 kHz noise figures from the datasheet, i.e. ignoring several frequency-dependent effects such as 1/f opamp noise). You might be able to estimate the distortion figures from the published AD797 data--check the according AES preprint as well (don't remember the number off-hand).

I've quoted an application note contract for a design using Analog Devices' new dual audio op amp.

So this new opamp is not released yet?

Samuel

 

[Keith Taylor]

Re input protection against phantom power accidents I think I have read all the posts in this thread but have seen no reference to ultra-fast rectifiers. In the British magazine Electronics World May 1996 author Simon Bateson uses 4 UF4002 rectifiers to protect an SSM2017. An additional 4 are used to generate 2 volts of reverse bias. This latter function is realised by having a series pair of diodes fed by a current limiting resistor connected to v+ and the cathodes connected to ground. Thus a reference is generated to reverse bias one pair of the clipping diodes. This is repeated for the other polarity. He uses 4.7ohm resistors in each leg of the audio "to limit the maximum fault current to 10A - well within the surge rating of the diodes"; to quote some of his text.

He avoided zeners because of the non-linear junction capacitance when reverse biased, which he viewed as a possible source of distortion. I have been using this circuitry in a pair of preamps I use for recording with no failures over 10 years. As I am new to this forum I am sorry to be using words where images (circuitry) would be better.

Keith Taylor Adelaide, South Australia

 

[amorris]

Excellent post "1" Keith, welcome! and thanks for your input.

 

[Keith Taylor]

Re my phantom protection post thanks for the welcoming words. I am an AES member and I often encounter Graeme Cohen at our meetings in Adelaide. He was quite chuffed to learn that he has aquired legend status on forums such as this. On another thread I noticed someone was modelling his preamp design and puzzling over certain aspects including a pair of LEDs. When I mentioned this to him he explained that they were gallium arsenide LEDs which were chosen for their forward voltage drop characteristics as they form a reference voltage. Apart from the voltage they may have had other desirable attributes such as temperature tracking to compensate for junctions elsewhere in the circuit. On more mundane matters he now realises that quite a few dollars have been made out of his design by others!

Perhaps I could act as a go between for anyone having queries on his preamp design. Another of my AES collegues is/was? a member of this forum. He designed a double sided PC board for Graeme's preamp which uses the last of very few hybrid amps that were left over from his original work at Phillips. Mediatech, I will PM you re the offer of the AES paper.

Keith

 

[bcarso]

The first I saw of it was a Precision Monolithics app note, where it asserted that the forward voltage of a GaAsP "standard red" LED matched well the tempco of a base-emitter junction, of order -2.2mV/degree C. I recall the claim was made that for the circuit shown the worst tempco mismatch was about 100uV/degree C.

I attempted to design a cheap interface to comply with a misguided Ford specification, using as part of the circuit a common-base stage, temp-compensated with an LED. What had happened was, I had been given a very sketchy spec and told to estimate cost for simply making an automotive amplifier trigger line be compatible with a logic level rather than the typical full battery voltage swing. When I confidently produced a simple circuit I then was given the full spec, which had been based on a Ford guy's misinterpretation of a CMOS schmitt-trigger input gate specification. But while Ford may not always be right, they are never wrong, so the specification held (and Ford themselves were not bound to it for the things they made, but we outside vendors were ).

So I was scrambling to do something cheap that met the spec over the full automotive temp range. The circuit worked, but there was a great deal of opposition from one guy in the org who insisted that LEDs were fundamentally unreliable.

Another obstacle was getting the nice but not-too-bright purchasing agent to understand that the part was being used as a bias source and not as an indicator lamp. She had reps flowing in the door telling her that their LEDs were better than anything, yada yada, and I would keep reminding her that brightness and color match were irrelevant, but material was critical.

Another objection to using LEDs for bias made by some is the light sensitivity of the parts. However, the sensitivity is very very low, such that even in a very strong light the photocurrents are tiny compared to even as low a bias as 100uA. If it really bothers you in principle, put a cover on the thing or a coating of opaque paint.

Beware of the very low melting point of the LED encapsulant and don't stress the leads before soldering. Otherwise the plastic when it softens transmits the stress to the bond wire and if you are very lucky it will break then and there. If you are very unlucky it will break at some indeterminate time later :evil: . This was learned the hard way during another manufacturing screwup with that same amp for Ford, which I believe I've related in here before.

 

[bcarso]

[quote author="mediatechnology"]
Remember which one or what part number? I'd like to post that one and have most of the old data to scan for it.[/quote]

Looks like I conflated the mention of a 100uV number with the passage about the GaAsP LED, as I found what was probably the reference, the data on the dual transistor SSM-2210 in the circa 1989 PMI Audio Handbook, pp. 104-114. The text reads in part:

"The difference between this voltage and the base-emitter voltage of a silicon transistor is predictable and constant (to within a few percent) over the military temperature range."

So maybe this isn't the earliest mention, just the first reference I saw. I remember I was hunting for some justification to use the technique and had a hard time finding anything. Citing semiconductor physics doesn't impress QA folk so there was little point in attempting to derive the numbers.

 

[pj]

Hopefully I'm not digressing too far OT with this Q but I've been reading the app notes and data sheet's for a a balanced mic pre and have bumped up against the various protection diodes configurations mentioned here.

I'm looking at building the AN-242 circuit with a THAT1510 replacing the SSM-2017. I'm planning to follow the THAT1510 datasheet in using 1n4004 protection diodes and dropping the current limiting R down to 10R or 4R7. The preamp will be used for ambient/nature recording, so the possibility if reducing resistor noise is attractive.

I was curious as to the pros/cons of using the 1N4004's, and whether I'm wise hacking a WJ design in this manner?

cheers
Paul

 

[bcarso]

[quote author="mediatechnology"]This is from J*ng 1975:

[/quote]

Note that with the MLED55, a single GaAsP LED, it looks like the circuit is somewhat under-compensated, if the goal is constant quiescent output Q current versus ambient temperature.

 

[JohnRoberts]

[quote author="mediatechnology"]



Do you suppose that it was used more for 3*Vbe than temp comp? The text says nothing about thermal coupling.[/quote]

Could very well be. I have used LEDs a few times just as convenient if not somewhat exotic voltage references. It worked out conveniently for generating one of the bias voltages I needed for an old BBD delay line kit. I think I also used one as a crude voltage reference in a very low noise power supply for a phono pre. Precise voltage or tempco wasn't an issue in either case.

Easier than stacking several diodes in series or bringing in a special part when you already have LEDs in use elsewhere.

JR

 

[bcarso]

Sze, Physics of Semiconductor Devices, 1st Ed., has a good discussion of the emission wavelength vs. bandgap (wish I could find my second edition ).

As to the application of the Shockley equation to the forward voltage, Sze remarks: "The Shockley equation adequately predicts the current-voltage characteristics of germanium p-n junctions at low-current densities. For Si and GaAs p-n junctions , however, the ideal equation can only give qualitiative agreement. The departures from thje ideal are mainly due to the following effects: ......" (Sze, op. cit. pg. 102).

After a few hairy pages we get to "The experimental results in general can be represented by the following empirical form Jf ~ exp(qV/nkT), where the factor n = 2 when the recombination current dominates...and n = 1 when the diffusion current dominates...When both current are comparable, n has a value between 1 and 2."

I recall a brief mention of this in Horowitz and Hill as well.

Note that this means the temperature coefficient also varies with current. It also explains why a particular circuit I saw, that "linearized" a transistor stage using a diode worked well, whereas replacing the diode with a diode-connected transistor did not.

The bandgap enters in to the saturation current density and its temperature coefficient, and Sze shows the expression Js = T^([3 + γ]/2)*exp (-Eg/kT), and remarks that the first term is not important compared to the exponential term. Actually, you rarely see this first term at all, although it actually does get significant at very low temperatures.

When I had a brief email correspondence going with Barrie Gilbert and asked how some experimental heterojunction bipolars they were developing at Analog Devices got betas of 100,000 or more, he got impatient and told me to go and read Sze :roll: I think that's a little along the lines of telling a person interested in chemistry to go and study the Schrödinger equation, but maybe he was having a bad day :razz:

 

[JohnRoberts]

Yeah I was going to say that but then I woke up..

I don't even recognize what all the letters mean. I hope there isn't a test at the end of this thread.

JR

 

[bcarso]

[quote author="JohnRoberts"]Yeah I was going to say that but then I woke up..

I don't even recognize what all the letters mean. I hope there isn't a test at the end of this thread.

JR[/quote]

They say the best way to learn something is to teach it. I think the second-best way is to try to come up with a post on a forum :green:

I have this problem whenever I try to write a paper or even just an article---supporting one seemingly simple assertion takes digging out volumes of information. That's one excuse for the paucity of my publications list.

I'd have a nervous breakdown if I ever agreed to write a book.

 

[JohnRoberts]

Indeed, I used to enjoy writing my audio mythology column back in the '80s because it's one thing to have an opinion about something and quite another to debunk it in print. You tend to refine your though process while getting all your ducks in a row. There was often an ounce of truth in the audio myths to keep it interesting. Plus if you do something stupid like misspell bus it'll come back to haunt you decades later.

I've seen some teachers/lecturer's that didn't appear to have a good grasp of all their material, and when they get asked a tough question would try to overwhelm the student with technobabble. I speculated to myself a little of that might be going on with Gilbert's dismissive response to your query, while it's obvious he knows more than the average bear about semiconductor junctions..

JR

 

[bcarso]

[quote author="JohnRoberts"]... I speculated to myself a little of that might be going on with Gilbert's dismissive response to your query, while it's obvious he knows more than the average bear about semiconductor junctions..

JR[/quote]

Barrie is a justifiably celebrated guy and great circuit inventor.

However, the superbright people (like Feynman for example) usually love to find ways of making difficult ideas comprehensible. Others often seek the refuge of babble as you say, especially if they are caught off guard.

One hilarious case: there is this sort-of optical illusion puzzle, where you take an alleged right triangle apart and put it back together as a rectangle, somehow in the process adding a unit of area. The added area is there, but unless the image is quite large you can't readily see it---it's a long sliver along the diagonal of the rectangle.

In a chatroom (deoxy.org) a guy presented it, and I explained where the fallacy was, and he said "Oh no that's what everyone says" ( ). "But my friend the mathematician explained that it was an example of tesselations of the plane..." I said your math friend didn't get it and threw out some high-sounding words to baffle you and support his insecure ego, and you were taken in, fool. It is simply straightforward high school geometry---shame on him.

Needless to say this wan't well-received (I was really nicer than that paraphrase, and didn't use the word fool at all).

I am really a nice guy most of the time.