Pro Audio Design Forum

....another lucky surplus store find

So you're not revelaing your secret?

The 47 uF is probably best left out between the 1512 and the 1646. Which drawing shows that?

Thanks Roger. Maybe we should do something. I've got some more basic 1510 stuff I'm working on as well.

Randy - So that's Apex "Sr?" Or is it our own Apex Jr"

As to the 47 uF if you're using a Cgain the Vos of the 1510 is going to be pretty low. Eliminating the 47 uF makes the 1646 work mo better. It will be more DC accurate, maybe a tad quieter at VLF and maintain AC balance.

If Cgain is in series with Rgain and pin 5 is at ground, the output of the 1510/1512 should be less than +/-5 mV.

Does the output offset vary with gain?

EDIT: Randy - the more I think about it pin 5 may not be grounded.

Good call....have you been snooping around my lab ???

Yeah. Where do you keep the coffee filters?

Those DC values look spot-on typical.

FWIW The 1510 datasheet has just been updated with the protection and T-bias changes:

http://www.thatcorp.com/datashts/1500data.pdf

Randy - So that's Apex "Sr?" Or is it our own Apex Jr"

There are 2 of us, I started working for my Dad on the books in 1968 and as I call it
Apex "Senior" and left to venture on my own in 1992 ( Hence my name Apex "Junior")
I use to answer the phone as Apex Jr and just kept the moniker....
My Dad passed away in 2002 and my brother is running Apex "Senior" now.

Steve @ Apex Jr.
http://www.apexjr.com

Using the T-bias scheme, what impedance will a mic see? I'm concerned about loading dynamic and ribbon mics. I have built a preamp around the 1510 using two 3.3k resistors to ground using the rule of thumb that the input impedance should be 10x the mic impedance (I do have a 600-ohm mic, AT804, and built the preamp to have a 6.6k input impedance). Rane builds a preamp around the 1512 with two 10k resistors to ground. What value should the resistors in this network be for a higher input impedance, say around 6k? Am I correct in saying that the circuit as drawn will present a 3k impedance to the mic?

Using the T-bias scheme, what impedance will a mic see? I'm concerned about loading dynamic and ribbon mics. I have built a preamp around the 1510 using two 3.3k resistors to ground using the rule of thumb that the input impedance should be 10x the mic impedance (I do have a 600-ohm mic, AT804, and built the preamp to have a 6.6k input impedance). Rane builds a preamp around the 1512 with two 10k resistors to ground. What value should the resistors in this network be for a higher input impedance, say around 6k? Am I correct in saying that the circuit as drawn will present a 3k impedance to the mic?

Hi Chris. Thanks for joining us and posting here. A very good question.

Yes, in the circuit I showed earlier the differential load (mic load) is 3k. The mic load is Rbias*2. You can indeed raise the bias resistors to 3K3 each to get a 6k-ish impedance. The trade off is open circuit noise caused primarily by the 1/f noise current times the bias resistor values. With a source connected, the source impedance (plus the reactance of the input caps) appears in parallel with the bias resistors and the mic (and caps) will usually dominate. With a ribbon mic in particular it's quite likely that the input capacitors will provide the highest overall impedance contributor at 1/f frequencies.

As to the T-bias and it's effect on loading, the bias resistor raises the CM impedance but not the differential load the mic sees. The T-bias value, "Rcm," can be made relatively high without noise penalty and it reduces capacitor matching requirements for improved LF common mode rejection. The T-bias resistor, which is shared with both bias resistors, can be made much higher because the noise current across the T-bias resistor develops in common mode and is canceled by the 1510/1512. Very high values >>22k do cause a common mode offset to develop internally to the 1510. This CM offset is canceled by the 1510's differential to single-ended converter but it does reduce, with low voltage supplies, input headroom slightly at a worst case amount of 28uA * Rcm.

Using an example where the bias resistor is 1k5 and the common mode T-bias resistor is 22k, the common mode impedance (per leg) is 44k+1k5. The equivalent resistance of the 22k T-bias resistor is 44k/leg. (This is due to Wye-delta transformation where the 1k5+22k+1k5 Wye becomes a 44k+3k+44k delta.) The input coupling caps are loaded in 1k5 without T-bias and 45.5k with T-bias. The CMR from capacitor tolerance mismatch improves by a factor of about 30 with T-bias. The low frequency cutoff of the input capacitors is determined by Cin/2 and Rdiff which, in the example is Cin/2 and 3k.

I came to the 3k figure (Rbias=1k5) as a trade off based on 1/f noise and the requirement to allow the preamp to bridge another preamp. The following is from "Mic Splitters" by Jim Brown where he felt that 3k was the minimum desired impedance. http://www.ka-electronics.com/Images/pd ... itters.pdf

Item #2 (loading of the mic’s output stage by multiple input stages and cable capacitance) can be quite significant, especially with poorly designed condenser mics whose output stages are current starved. Ray Rayburn observes, “Take the example of a three way split where each console has a 1500 ohm input impedance (1,000 – 2,000 ohms is typical). The resulting resistive load is 500 ohms which will drive these mics into distortion. One popular podium mic loses 15 dB of headroom as the resistive load goes from 1000 ohms to 700 ohms.” For a three way split, each console would need an input impedance of 3000 ohms or higher to work with that microphone. David Josephson (Josephson mics, and chair of the AES Standards Committee Working Group on microphones) makes it his business to know what is going on with his competitors. He says that his mics and the better pro mics of his major competitors are designed with considerably more robust output stages. One clue that a mic can’t drive multiple consoles is its phantom current - in that it doesn't’t draw much phantom current.

In your situation try using ~3k bias resistors for a 6k load. I don't think you'll have a problem but you may notice a slightly higher open-circuit noise, not due to T-bias, but simply larger bias resistors. With the source connected there may not be any discernible difference.

Thanks for the very informative reply.

So I'll be OK using the two 3.3k resistors and 22k for the center resistor?

I should mention that my design omits the phantom-power blocking capacitors. Strange, I know, but if you want phantom power you use an external supply. My preamp is used for testing mics and I don't want the audio from ribbons and dynamics going through capacitors. You can hear the prototype of my preamp here:

www.miclisteningroom.org

So I'll be OK using the two 3.3k resistors and 22k for the center resistor?

Yes, to make my rather long story short that should work fine.

Without coupling caps however that extra resistor doesn't bring much to the party.

Have you considered using a 1512? It would drop in the same socket and have 6 dB less gain but better noise performance at lower gains.