Active ribbon-mic

[clintrubber]

Just catching up on all the new info here, an interesting read.

I must admit: I'm almost disappointed that various recording situations don't seem to be needing this nice additional circuit... :sad: :?

I mean: used the mic on a not that loud Twin Reverb this evening and even got by with a mic-pre that maxes out at 40 dB gain. But when using the mic in a fig-8 for M/S there definitely could have been more gain.

But that's all w.r.t. the added gain - - there's also the phantom-protection benefit of course.

 

[bcarso]

PRR: "Some kind of lame simulator? "

That reminds me of the joke about the dishwasher. "Yes, I have a dishwasher---my wife."

I think there is a VCVS block---I just didn't think to use it. I do have to periodically throw a cascaded series of notch filters in though, to estimate distortion, when the fourier analysis function chokes on something or other.

The original circuit does have lots of 2nd as is (note added in review: single-ended that is---it's third o/w). By replacing the 470 ohm R with an I source it drops dramatically.

And the fourier analyzer just choked too.

 

[PRR]

Erno will sell you 2SK170GR/BL for ?2.00; however minimum order is ?50.
http://www.borbelyaudio.com/audiophile_components.asp

Datasheet:
http://www.audiokits.com/pics/2SK170_datasheet.pdf {links expired}

I see what you mean about gate leakage.

I hacked-up a device model that approximately matches the 2SK170BL Vg/Id curve. (I did NOT fool with capacitance and other parameters.)

Performance with 2SK170BL:

Biased at 2mA per device, with 300Ω||2KΩ||2*6.7KΩ load, I get a gain of only 6dB with the Sources shorted together. Hardly worth it. 4.6mA (about the limit of Phantom) isn't a lot better. Slide-rule from the datasheet numbers gives slightly better gain, so my model may be skewed, but not enough to save it. We can't really short the Sources together because we can't match two devices so exactly that we won't get disturbing current in a subsequent input transformer. (If a Dual becomes available, that may help.)

 

[PRR]

> used the mic on a not that loud Twin Reverb this evening and even got by with a mic-pre that maxes out at 40 dB gain.

114dB SPL near the grille. (50 Watts or +17dBW, speaker 95dB SPL/Watt/4', 12dB gain for going close to a 1-foot cone, 10dB allowance for "not that loud".) 40dB over 0.050mV gives 5mV out of the mike, 500mV out of a 40dB preamp, which should be plenty. And even if your studio (and Fender) noise were as low as 14dB SPL, and your preamp noise was a slightly high 0.5uV, you'd get a full 120dB dynamic range from guitar to room/preamp noise floor.

.......I'm trying to remember the last time I put a mike in a field like that.

 

[bcarso]

PRR, I get, based on a model from Orcad (which I transcribed into my lower-octane simulator) for the part, fairly much higher gain just by substituting the parts into your original circuit (absent the 22 ohm R's). When I raise the bias voltage by changing the 220k to 100k, hence 3.845V on the gates and the common sources at about 4.1V, each Q pulls 4.3mA and the drains sit at 18V. I changed the coupling C's to 1u and the bias R's to 470k, so the source Z is not loaded to speak of. I get 11dB, nearly the same as with the 4401's and the 22 ohm R's, with the 300 ohm || 2k load. I think the 470 ohms and the bias up helps things a bit. I'll see if the Orcad model also has something like the gain you get.

On the other hand your latest schematic does remove any concerns about d.c. in the transformer, and saves some parts.

 

[bcarso]

...Also one would have to adjust the output d.c. values as you say to avoid significant d.c. in what might be a transformer. I can see a servo...but in that way lies madness, of a sort at least. The duals will help and maybe be enough---Linear Integrated Systems says their interdigitated construction will result in a better match than the old and now discontinued 2SK389s.

Totally off the subject: I just opened a bottle of Fess Parker Syrah 2000. God loves me---it is a dead ringer for a top-notch Cote-Rotie.

To celebrate, the first ten people who PM me wanting some 2SK364BL's can have a few for sending a SASE (the 364's are the same die as the 170's, just characterized for analog switching apps instead of amps).

Brad

 

[PRR]

> your latest schematic does remove any concerns about d.c. in the transformer, and saves some parts.

Saves parts (I'm a minimalist); but if the Idss is mismatched several mA (which can be expected), then there could be a LOT of DC in the following input transformer.

The nice thing about the BJT is that two peas from the same bag will match to a few mV, which with the necessary emitter resistors ensures tenth-mA balance.

I've never messed with SPICE's part-tolerance features. But I bet if you force "Beta"(*) to vary 1.5:1 (Idss from 8mA to 12mA), put a 10H choke with 10Ω DC resistance across the 2K (which is the following preamp input), and check choke current, it can be large. How large is too large? I've seen some tubby trannies that could eat DC, and others with hyper-alloy cores rated "zero", and designers trim-out uA bias currents to keep the iron very dry. So this may be the real flaw. (Of course "budget" recording is mostly transformerless inputs which will ignore the DC offset, but the next step is some old iron and that could be trouble.)

(*) I had to look-up what SPICE model FET parameters do. FET "Beta" is not at all like BJT Beta. Seems to be roughly "area", though gate-voltage parameters interact a lot. I simply fudged til several points looked like the datasheet.

I think we are cross-posting..... this just in....

> I can see a servo...but in that way lies madness

Madness indeed. Jiminy Cricket, this is a budget project to use E129 mikes. And there are perfectly usable high-gain preamps as cheap as $100 2-channel.

In fact, rather than a servo, I could see (if enough air is available) putting a 24VAC power-transformer winding across the output. It has enough inductance, low enough capacitance, leakage inductance is non-issue, and the DCR is low enough to swamp unbalnced DC current.

 

[PRR]

> a gain of 10dB I thought (at least could be set to do so) so that looks just the amount we could use - bringing it in line with the other kind of mic-signals

Wait... the t.bone RB500 is rated -55dB below 1V/Pa, the Shure SM57 is rated -54.5dBV/Pa open-circuit. Assuming the t.bone is also rated open-circuit (reasonably likely), these are the same! (0.5dB is not meaningful, especially on the lumpy SM57.)

Are you saying the spec is made-up? I don't doubt you, but if this sensitivity estimate is way off, then overload needs to be higher.

Yes, the lumpy SM57 is rated 1KHz, and has 4-6dB bumps 3KHz-12KHz which will impress the ear. That still isn't 10dB difference.

 

[Marik]

[quote author="PRR"] We can't really short the Sources together because we can't match two devices so exactly that we won't get disturbing current in a subsequent input transformer. (If a Dual becomes available, that may help.)[/quote]

2SK389 is a dual.

 

[bcarso]

re 2SK389: Toshiba has discontinued them but there are a bunch of them still in and around.

 

[Rossi]

PRR, as I said: there's a big difference in output between an SM58 and the RB 500. As I said, the ribbon requires about 10 dB more gain than the Shure. I have two Thomann mics which I bought some weeks apart and probably not from the same batch. They're both the same sensitivity. Also, I recently had the brand new dynamic Neumann broadcast mic for review. Its sensitivity is rated at -55.4 dB (1.7 V/Pa). Interestingly enough it was inbetween the two; louder than the ribbon, but not as loud as the Shure. So I don't really trust the numbers. They're either incorrect or measured in different ways. But even if the numbers were correct, you'd still have differences in gain requirements, because an SM58 is typically positioned just few inches from the singer's mouth whereas a ribbon wants some distance to "breathe" and make up for the huge proximity effect.

Sure, you don't need a booster for a guitar cabinet, but when it comes to singers, many budget preamps will crap out either because of noise or because they don't deliver enough gain in the first place. Also keep in mind that many soundcards and ADCs want to see serious level. With even relatively inexpensive converters approaching 110-120 dB dynamic range, there seems to be a trend to place digital 0 dB FS at +15-20 dBu, sometimes even more.

Thanks PRR for the Erno tip, and also thanks to Brad for the 2SK364BL' offer. But it's probably easier for me to just order them at Reichelt. The last time I looked they also had some 2SK389 leftovers.

 

[clintrubber]

difference in output between an SM58 and the RB 500

While I didn't closely watch the gain-settings I can second that difference.

When doing a q&d ribbon-vs-'57 comparison on mentioned amp (as on the RoyerLabs-site for the Albini-ribbon-demo's) I placed them as close as possible to each other (say 1.. 1.5 meters from amp) and had to crank the ribbon-gain quite some more.
Indeed puzzling how they measured that ribbon-sensitivity, but OK, I must say I'm even more puzzled about how they arrived at the 165dB max SPL ! :wink:

All fine, these are fine mics despite the fuzzy quoted specs.

 

[Rossi]

This thread is going at quite some pace! You can't even edit your bad spelling without someone else posting at the same time :?

Yes 165 dB SPL seems like a lot and I, too, wonder what they used as a source to test it. Not a lady's fart, I suppose. :roll:

BTW: Even though it's kind of unusual for a dynamic mic, self noise is spec'd at 18 dB-A.

 

[clintrubber]

This thread is going at quite some pace! You can't even edit your bad spelling without someone else posting at the same time

Right ! :wink: And there are even (attempts at) related threads as well, like: http://www.groupdiy.com/index.php?topic=6277

Yes 165 dB SPL seems like a lot and I, too, wonder what they used as a source to test it. Not a lady's fart, I suppose.

I have no idea about the power, capabilities & determination of Chinese women in this respect... :grin:

BTW, must say I'm impressed with the effectiveness of the metal popscreen that I bought as well - lots of fun seeing the direction of airblasts changed so much. This of course not the very same as high SPL, but...

 

[bcarso]

clintrubber: "I have no idea about the power, capabilities & determination of Chinese women in this respect... "

Well, we've already heard, earlier in this thread, about the abundance of Chinese air...

Meanwhile, I simmed PRR's minimalist FET version with the Orcad 170 model and it gives about 11 dB of gain. So perhaps it is more optimistic as far as transconductance. The problems with unmatched parts will remain. BTW each half of the 2SK389 is not quite as high gm as the 170 from what I can see. Linear Integrated Systems (from whom I receive no promotional consideration, I swear) talks about the upcoming LSK389 as being a dual 170, but we shall see.

If I don't get some real work soon I may have to build some of this stuff.

Brad

 

[PRR]

> Yes 165 dB SPL seems like a lot

It is quite believable that a ribbon will work at such levels. Remember the reason a ribbon makes such low output is that it has small surface area and high mass. At normal levels, the ribbon motion is sub-microscopic. And the typical magnet structure is linear to about 3mm p-p, more before distortion gets gross. Even if ribbon-slap against damping silk limits the motion, it is likely to be up above 150dB SPL.

> wonder what they used as a source to test it.

Compression driver on a tuned pipe. JBL 2440 on a non-tuned pipe can make 148dB SPL steady-state, easily 10dB higher on bursts, and tuning the pipe can approach another 10dB. So the 165dB SPL rating, if really tested, may be a measurement limit and not the mike's true limit.

There are more exotic methods. The near-field of a large pistol is well over 150dB SPL. However calibration is a pain because a pistol does not have a linear level control like an amp/speaker rig. Some of the very small stiff condensers have overload levels well above 160dB SPL and could calibrate the near-field. Of course a pistol's air-blast may be harmful to a ribbon.

There is also the pistonophone. Take the spark plug out of an engine, jam your mike in the hole, and crank. Same idea (without fuel!) will give pressures as high as you want, up to the 190+dB SPL overload level of the air itself (sound waves ccan't make negative absolute pressure), and if the parts (piston size and stroke, chamber volume) are carefully measured you can compute the pressure to sub-dB accuracy. But only practical at low (not tooo low) frequency. And not directly applicable to velocity sensitive mikes like a ribbon (though I suspect it could be adapted).

Electrostatic actuation should also work on a ribbon. To be linear over that distance it may have to be push-pull, a trivial matter. Considering the large distance and the high diaphragm mass, it might take insane drive voltages to approach ribbon overload.

Most dynamics have overload levels well above 150dB SPL. The ribbon may have an edge because of poor mechanical efficiency, but we don't worry about overloading dynamics. (We may splatt the mike-amp input, but not the mike.)

Condensers have a naturally non-linear electrical transducer, and an underpowered head-amp, and this normally (for music-racket mikes) raises the THD to ugly numbers before any dynamic or ribbon would show any distress. Obviously good condensers can handle any sound field that we really meet in musical recording, at least with the pad to reduce transducer and head-amp (and board-input) levels to tolerable size.

 

[Rossi]

Thanks, thats a lot of cool info, PRR. Although I seriously doubt the guys who came up with the specs use any really sophisticated methods of measurement. Given the somewhat unrealistic sensitivity claims.

Anyway. I did some more testing last weekend. First, I bought a bunch of transistors that were easily availabe and looked more or less suitable on paper or because I saw them used in preamps. I rarely found values for Rbb, though. These are the ones I tried: BD139-10, BC637, BC546B, BC337-40, BCY59-8.

Test with the "old" battery driven circuit: The germanium AC187 still ruled in terms of noise. BD139-10 and BC637 came close (but no cigar).

Test with the new phantom powered circuit (bipolar version): Here the BD139 was the lowest noise. BC637 was close, but the BD139 is both lower noise and cheaper. The AC187 is only 25V. It did work, but seemed a little stressed out. Noise was a little higher than BD139 but still better than BC546, BC337 and BCY59. The latter being a "vintage" part found in many old Siemens preamps.

All in all, the best noise performance was the battery circuit with AC187. The phantom powered circuit with BD139 is very close, though, and totally usable as is. The noise level is actually pretty similar, it's just that the AC187 noise sounds a little smoother. On the other hand, the phantom powered circuit performed equally well with all preamps including the Trident 4T which was noisy with the battery driven one.

Next, I'm going to order some transistors that I couldn't get locally such as the SA-parts Brad mentioned. I'll also order some 2SK170 for the FET-circuit. Given my good experiences with the BD139, other BD-transistors may be worth a try with the bipolar cirucit.

 

[bcarso]

My offer of some SK170 equivalents is still good btw---only one person has taken me up on it to date. There is no obligation---no salesman will call, no spam will be proferred.

I would like them to go to people that will really play with them and not just add them to stock, but that's just a preference.

 

[Rossi]

Thanks, Brad. Your offer sure looks attractive. But I can't buy US stamps for the return envelope, and I wouldn't want you to send me free stuff and pay for the postage, too.

 

[clintrubber]

I had the same 'problem', but there exist those small papers that can be bought here in EU and exchanged for real stamps anywhere in the world, as I understood.
That would be less suited thought - requires Brad to go to the postoffice etc.
So what we did: I sent him the SASE (it left yesterday) and send him some dollars in a second envelope as soon as he knows the shipping-costs. I don't mind the two-step approach.

Who knows, an alternative might be to swap some parts (everybody uses five 5534's a day, not ? :wink: ), but I can't decide on that of course.

Bye,

Peter