Finished Mic Preamp?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

Leek

Well-known member
Joined
Jul 11, 2016
Messages
105
I now have a working THAT1512/1646 mic preamp. I've breadboarded the preamp based on the THAT data sheets and I hope I have a little better understanding of things. I am now working on a Transformer mic input into a Transistor based based preamp stage and a THAT 1646 output stage. I am a little better at calculating resistor values and etc so hopefully by doing so I have eliminated NOISE and the preamp stage looks ok. I will be using the Neve style input transformer wired in series to feed the circuit. Could you care to look over it and tell me rather or not I'm in the right direction?

I have recently, what I hope perfected a decent preamp. I was hoping you could over look it and give your insight?
I'll provide schematic and transformer link as well. All other parts are commonly used and shouldn't require explanation.

The input stage (balanced) is a Neve style input transformer connected in series. 

Pin 1 = + XLR In
Pin 2 & 3 tied together
Pin 4 = - XLR In
Pin 5 = Ground
Pin 6 = Ground
Pin 7 & 8 tied together
Pin 9 = + signal to preamp stage
Pin 10 = Ground

The preamp stage is based around two low noise BC549 transistors.
Each power supply input (on both the preamp stage and THAT 1646 output stage) are decoupled using 0.1Uf Wima Caps.

The output stage is exactly configured on the THAT 1646 data sheet.

Here is the link for the schematic. (https://postimg.org/image/utc00rutt/)

The transformer was concluded by CJ to be wired correct. Anyone care to give some pointers on where to go next to get my attempt at a design up to snuff?

The Datasheet for the THAt 1646 says it requires a low input impedance. Since the output of my preamp stage is fed to a 10K linear pot as a volume control does that mean my output is 10K and the impedance is too high for the THAT 1646? If so, how can I go about dropping the output impedance to work better with the THAT 1646? Add a 0.1UF cap and 1k resistor after the volume control? Or would I be better off moving the volume control to the front of the Preamp stage or in between the two transistor stages? If I was to put in between, I could control how much the first transistor drives the 2nd. If I were to put it at the front of the preamp stage, I could control how much the transformer drives the entire preamp stage. I'm not certain what to do, but I know that I need the low impedance output for the THAT 1646 output stage. Help please?
 
I didn't dig that deep, but the first thing that struck me is that everything before the "gain" pot is a fixed gain amplifier, with an output 'volume control'. Typically, mike preamps are asked to handle a wide range of mike output levels, and a large part of their job is to allow the user to adjust the preamp to the specific recording situation to result in an output signal that has a reasonable level. I've used preamps that have gains as low as 5 or 10dB as well as preamps that have 45-50dB of gain - both are useful, depending upon what you're doing.

Sue, you can adjust the "gain" pot to change the output level, but since the transistor stage is powered by a single 15V supply, it will clip at a specific level, and that signal can be attenuated by the pot. This means that the "gain" pot isn't a gain pot at all, but a volume knob after a fixed gain amplifier. The problem is, if you turn the pot down, the front end amplifier still clips at the same level, rather than what a true gain pot would do: reduce the front end gain to allow larger signals to be handled without clipping.

There area number of variable gain amplifier circuits whose gain can be controlled by a single pot, and it'd be informative for you to look into those. As a starting point, look at some console input channel mike amp circuits, and see how it's been done over the years. What you've drawn is a solid state update of what was done with a dual triode around the early-mid 50s, when performers were shooed away from mikes to prevent overloads from the high signal levels that you get from close mike placement. The next several decades of recording (and the entire multitrack process) were made possible only after folks figured out how to make a mike amp that didn't have a 46dB fixed gain, and could therefore be used with a spot mike for a loud instrument without clipping the mike amp.

To encourage you, what you have drawn seems basically competent, but it could only be used for very specific applications, and doesn't seem to be ideal for very many uses. Dig deeper and re-design!
 
Leek said:
The Datasheet for the THAt 1646 says it requires a low input impedance. Since the output of my preamp stage is fed to a 10K linear pot as a volume control does that mean my output is 10K and the impedance is too high for the THAT 1646?

Yes, it's too high. What they mean by "requires low-impedance drive" is "use an op-amp output to drive the 1646 input."

Look at the 1646 internal schematic in its data sheet. The input isn't high impedance like an op-amp, it's a 5k resistance into the diff amp, and there's a matched 5k resistance to ground on the other input to that diff-amp (which is internal). Thus the internal 5k on the input is a network with your volume control. At max output level, the pot is (almost) out of the circuit, and as you turn the pot to decrease the level, you add resistance in series with the internal 5k, thus affecting the 1646's op-amp gain balance.

So, really, drive the 1646 input with an op-amp.

Also: everything Monte says is on the mark. I'm sure there are plenty of examples on the web of variable-gain discrete mic input stages.
 
Leek said:
Here is the link for the schematic. (https://postimg.org/image/utc00rutt/)
This schemo seems to be taken from a beginner's textbook rather than from an audio-savvy source.
In addition to what's already been written:
The bias resistors set the input impedance at a lower-than-reasonable value. The 0.1uF input cap results in a 560Hz LF corner.
The 2nd stage input impedance is even lower, putting an indue load on the 1st stage.
 
Thanks for the tips and information guys. I'm still working everything out. I'm hoping to get a decent Mic Preamp together with all the information you guys are providing me with. I'm working out new bias resistor calculations which will hopefully get my input impedance to a more suitable level. I've replaced the 0.1Uf input capacitor with a larger 4.7Uf electro cap. My first thoughts is to use a unity gain op amp buffer after the transistor stage to bring the impedance to an acceptable level of the THAT 1646. But, based on the design flaw (fixed gain amplifier) I feel it may be easier eliminating the transistor staged and use a variable gain Op Amp stage. Though, that sort of eliminates the point of the design which was a discrete front end.

So, as I work out the unity gain Op amp Buffer, I'll also be working out a variable gain transistor preamp stage. Hopefully, with a little reading and help from you guys I can get this together shortly.
 
Does this look better you guys?

https://postimg.org/image/myq1kh3fn/
 
Leek said:
Does this look better you guys?

https://postimg.org/image/myq1kh3fn/
Not really; the pot adds 10 kohms to the impedance seen by the input transistor, so the noise factor is just about 10dB too high.  There is DC through the pot, which is bound to induce noise when changing gain. The available current is probably limited so the input transistor may not operate in its best region. the output stage drive capability is limited by the 10k resistor. You should take advantage of the bipolar power rails to increase the output capability. The bias resistors are somewhat low IMO.
 
abbey road d enfer said:
Not really; the pot adds 10 kohms to the impedance seen by the input transistor, so the noise factor is just about 10dB too high.  There is DC through the pot, which is bound to induce noise when changing gain. The available current is probably limited so the input transistor may not operate in its best region. the output stage drive capability is limited by the 10k resistor. You should take advantage of the bipolar power rails to increase the output capability. The bias resistors are somewhat low IMO.

How can I get the pot to control gain of the preamp stage but not increase impedance and have noise? I can't seem to find any information about how to do so. I can also increase the bias resistors. I used a simple formula to get them but it was based on 12v+ not 15v+. Perhaps putting the preamp stage back down to 12v would make them more suitable?  I've removed the 10k resistor in the buffer stage as well.

As for the bias resistors, I could bump them up by putting some resistors in series to increase resistance.

Say R1 = 15K and R2 = 3.3k? That would give me about 2.7v at the base of the BC549(positive terminal to the base and negative terminal to ground) according to my splice model. with 15v Supply. Does that seem more suitable?

Would the correct spot for a pot be before the entire preamp stage? That'd still adjust input impedance, correct? I don't quite understand how I can place a potentiometer anywhere to control gain without it effecting the input impedance. Any tips please?


With those changes, I get

14.8v = collector
2.7v = Base
2.1v = Emitter

of the BC549.
 
Does this appear to be more useable? I eliminated the discrete preamp stage and opted for a NE5532 preamp stage. Since the NE5532 is a dual op, I properly terminated the 2nd amp. If I'm not mistaken, with the placement of the potentiometer, it controls the gain of the preamp stage? Since the OP amp can handle high input impedance, the pot being there doesn't hurt, correct? Which shouldn't be an issue since the output of the NE5532 will properly feed the THAT 1646, correct? If all looks well, I'm going to breadboard then build this.



 
Leek said:
Does this appear to be more useable? I eliminated the discrete preamp stage and opted for a NE5532 preamp stage. Since the NE5532 is a dual op, I properly terminated the 2nd amp. If I'm not mistaken, with the placement of the potentiometer, it controls the gain of the preamp stage? Since the OP amp can handle high input impedance, the pot being there doesn't hurt, correct? Which shouldn't be an issue since the output of the NE5532 will properly feed the THAT 1646, correct? If all looks well, I'm going to breadboard then build this.


That wouildn't be right because the preamp is always at maximum gain, as is noise. Check how it's done there
https://groupdiy.com/index.php?topic=45604.msg775065#msg775065
Take advantage of the incredible wealth of info in the Technical documents section of this group.
 
Please tell me this is getting closer. I looked over the SSL sheet and tried to replicate the Gain control into my circuit as best I could. Since the SSL circuit uses NE5534 chips I figured it'd make perfect sense to use the 2nd amp of my NE5532 chip to work my gain control into my circuit as the E-series did. Hopefully things look a little better and I'm getting closer to what I'm after.

Here's the schematic



Also again, thank you so much for all your help. I wasn't sure what to do with the output of the 2nd op amp. So, maybe I have things drew out correctly. If not, possibly you could provide some more insight as to what to change and how to incorporate those changes.
 
Leek said:
Please tell me this is getting closer. I looked over the SSL sheet and tried to replicate the Gain control into my circuit as best I could. Since the SSL circuit uses NE5534 chips I figured it'd make perfect sense to use the 2nd amp of my NE5532 chip to work my gain control into my circuit as the E-series did. Hopefully things look a little better and I'm getting closer to what I'm after.

Here's the schematic



Also again, thank you so much for all your help. I wasn't sure what to do with the output of the 2nd op amp. So, maybe I have things drew out correctly. If not, possibly you could provide some more insight as to what to change and how to incorporate those changes.
You need DC path between the output and the inverting input of the 1st opamp. Leave the 2nd opamp alone.
 
abbey road d enfer said:
You need DC path between the output and the inverting input of the 1st opamp. Leave the 2nd opamp alone.

DC path? Not to sound like a novice, but I don't understand Abbey. Could you perhaps draw it so I understand?  Also, by leave the 2nd op amp alone did you mean it being properly terminated or the way it is connected now?
 
Leek said:
DC path? Not to sound like a novice, but I don't understand Abbey. Could you perhaps draw it so I understand?  Also, by leave the 2nd op amp alone did you mean it being properly terminated or the way it is connected now?
R6 should connect directly to the output of U1, not via a capacitor. The 2nd opamp, as it is, does nothing else than injecting noise in the 1st one.
 
abbey road d enfer said:
R6 should connect directly to the output of U1, not via a capacitor. The 2nd opamp, as it is, does nothing else than injecting noise in the 1st one.

So, if I understand correctly, get rid of C10 between R5 and R6. Connect R6 to the output of the NE5532 before C3. Eliminate the 750 ohm resistor, C9 and R4. Eliminate all other components of NE5532B and terminate it properly? By doing so, what do I do with the terminal of the pot nearest the 750 ohm resistor?  Leave it unconnected, ground it or something else?

If I connect to NE5532B - input via a voltage divider (10k pot & 1k resistor) but leave the op amp properly terminated (+ grounded and - connected to output as a unity gain buffer) would that be ok?
 
Leek said:
So, if I understand correctly, get rid of C10 between R5 and R6. Connect R6 to the output of the NE5532 before C3. Eliminate the 750 ohm resistor, C9 and R4. Eliminate all other components of NE5532B and terminate it properly? By doing so, what do I do with the terminal of the pot nearest the 750 ohm resistor?  Leave it unconnected, ground it or something else?
Leave unconnected.
If I connect to NE5532B - input via a voltage divider (10k pot & 1k resistor) but leave the op amp properly terminated (+ grounded and - connected to output as a unity gain buffer) would that be ok?
You don't need to make it complicated; unused opamps are connected as voltage-followers (output connected to inverting input) with non-inverting input grounded, preferrably via a resistor (value not critical, 1-10k).
 
abbey road d enfer said:
Leave unconnected.  You don't need to make it complicated; unused opamps are connected as voltage-followers (output connected to inverting input) with non-inverting input grounded, preferrably via a resistor (value not critical, 1-10k).

Thank you so much.
 
Back
Top