Mic Preamp Design

[James UK]

Hi,

I'm working on a design for a mic preamp. The overall project is to build two different mic preamps, one with an input transformer and one using the THAT 1512 IC, mount them in the same rack enclosure with identical front panel controls and carry out blind tests between them. The aim is to see if people have a preference for either and indeed if they can tell the difference.

I've begun by working on the design for the transformer preamp. I would like it to

• Be as low noise as possible.
• Include phantom power, -20dB pad and phase reversal.
• Have a gain of 20dB to 64dB in 4dB steps (not including the pad.)
• Have an input impedance of approximately 2k.

The design at the moment has an 'impedance balanced' output which I believe is sufficient in a studio setting. I'm happy to be told otherwise!

The transformer is 1:6 so provides 15.6dB of gain and the op-amp provides the rest. The 75k resistor on the secondary should give an input impedance of 2.1k.

I would also like to include a clipping indicator (and possibly 'signal present') but haven't designed that yet. I'm a bit worried about being able to design it not to inject noise into the preamp.

I've attached a PDF of the current design. Any suggestions for changes and improvements are very welcome!

Thanks,
James

 

[rayma]

No decoupling capacitor(s).

 

[James UK]

Ah yes, I haven't included them yet because I didn't want to make the schematic too messy. I'm putting 2 x 220uF electrolytics on the board and 100nF at each power pin on the op-amp.

 

[tulsultant]

Hello James UK,

I think your plan is a terrific experiment, both for you and your audio aficionado’s audience. I am certain that the discerning ‘Golden-Ear” will quickly spot the difference and, maybe, even identify which unit has a transformer influence.

I think THAT Corp is exceptional for its quality and breadth of product offering as well as a voluminous library of applications information.

I was not familiar with the OEP transformer family and, for the price, seems reasonable. I’m concerned, however, about the low frequency 3% THD @ 0dBu callout at 30Hz. I am a big fan of Jensen transformers (as well as Cinemag & Lundahl) for its typical callout of 1% @up to +8dBu at 20Hz. Of course, this comes at a premium of about 4x$ (US) in price. Nevertheless, I think the improved quality of the transformer would give the THAT transformerless version a run for its money (so to speak lol).

Going a step further, one might consider a high-quality Jensen input transformer coupled with the classic Jensen 990 class of discrete preamp.

Both camps have typically utilized the NE5534. AD743, and MC33078 low noise opamps as core elements to various frontend designs.

Please keep us posted as to your progress.

Regards, cliff

 

[James UK]

Hi,

I've had some time to make a bit more progress on this and I've revised the design a bit.

• I've changed the output to an active balanced output using Douglas Self's 'quasi-floating output'.
• I've swapped the expensive Grayhill gain switch for a Lorlin CK, mounted on a separate board.
• The gain control has been altered so that it is wired like a reverse log pot and changes both gain resistances. This allows the gain of the NE5534 stage to be reduced lower than previously.
• I've removed the phase switch and indicator LEDs, since this is just a test circuit.
• The gain of the NE5534 stage is configured to give between 22dB and 66dB of overall gain, with 15.6dB coming from the input transformer and 6dB from the output stage

I've attached the schematics for both the preamp circuit and the gain switch, which will be mounted on a separate PCB.

I have a few questions:

• Douglas Self states that the electrolytics in the output stage should be non-polarised. However, non-polarised 35V SMD electrolytics seem extremely hard to find, and those are are available are very expensive. Is it not possible to use polarised types, as I've seen used in various other output stages?
• How much of a difference will thin film vs thick film SMD resistors make? With through-hole stuff, I always use metal film and Self says that thin film should be used for audio. However, thin film resistors seem massively more expensive than thick. Also, if I have the boards assembled by JLCPCB, there will be a 3$ extended component fee for each resistor, which will add up to a lot for the gain switch.
• Is it worth running separate grounds from signal and power and having two ground wires back to the PSU, which will be external?
• Finally, and I can't seem to find an answer to this online, on a mixing desk, is the gain labelled on the mic preamp measured from input to output or just of the preamp stage? I.e. does it include the +6dB provided by the output stage?

Thank you in advance for any suggestions/ criticisms. Once I've got the transformer version sorted, I can move onto designing the THAT1512 version.

Thanks,
James

 

[bucks bunny]

Considering 75k input impedance, the NE5534 is not your first choice due to input noise current.
But when connected to a 200 Ohm dyn capsule, with the input impedance dropping to 7k, it should be fine.

 

[James UK]

Would you suggested changing the value of that 75k resistor? I went for that to give an input impedance of around 2k.

 

[MarcelvdG]

Do you care about the noise level with open input or only with a microphone? In the latter case, reducing that 75 kohm would only make matters (slightly) worse.

Regarding thin film resistors, there are relatively large price differences. Some Susumu 0.5 % tolerance thin film resistors are fairly cheap, although not as cheap as thick film. Thin film resistors are easy to damage by hand-soldering them with a much too hot soldering iron.

 

[James UK]

I'm not bothered about the noise with an open input as the channel would just be muted in that scenario.

I will be hand soldering this so maybe I should just go with thick film? What do you think about the differences regarding noise etc?

Thanks

 

[bucks bunny]

I have been hand soldering lots of SMD0805 thin film resistors without problems.
Additional noise may occur in case there is a substantial voltage drop along the resistor.
Which does not apply in case of the 75k input res, so I would expect no degradation using a thickfilm type in that position.

 

[MarcelvdG]

• I've swapped the expensive Grayhill gain switch for a Lorlin CK, mounted on a separate board.

With gold-plated contacts? Silver plated contacts tend to make cracking sounds after a while, when you use them for low-level signals.

I have a few questions:

• Douglas Self states that the electrolytics in the output stage should be non-polarised. However, non-polarised 35V SMD electrolytics seem extremely hard to find, and those are are available are very expensive. Is it not possible to use polarised types, as I've seen used in various other output stages?

Yes, but you get a bit more capacitor distortion.

I usually use 63 V working voltage for the output capacitors and connect the positive side (if applicable) to the output connector side, to give the circuit a better chance of survival if someone should connect its output to an input with 48 V (plus 4 V tolerance) phantom supply. In your case, 35 V should do because of the resistor network at the output that would attenuate the unintendedly connected phantom supply.

• How much of a difference will thin film vs thick film SMD resistors make? With through-hole stuff, I always use metal film and Self says that thin film should be used for audio. However, thin film resistors seem massively more expensive than thick. Also, if I have the boards assembled by JLCPCB, there will be a 3$ extended component fee for each resistor, which will add up to a lot for the gain switch.

It will not have much impact on the noise floor, as there are no big DC bias voltages across the resistors. With signal, you get some 1/f noise sidebands around the signal with thick film.

• Finally, and I can't seem to find an answer to this online, on a mixing desk, is the gain labelled on the mic preamp measured from input to output or just of the preamp stage? I.e. does it include the +6dB provided by the output stage?

It would seem logical to include it.

 

[MarcelvdG]

I have been hand soldering lots of SMD0805 thin film resistors without problems.

So have I, since I learned that I have to be careful with the temperature setting.

 

[James UK]

Thanks for the replies.

With gold-plated contacts? Silver plated contacts tend to make cracking sounds after a while, when you use them for low-level signals.

They are available with gold contacts so I'll get that.

I usually use 63 V working voltage for the output capacitors and connect the positive side (if applicable) to the output connector side, to give the circuit a better chance of survival if someone should connect its output to an input with 48 V (plus 4 V tolerance) phantom supply. In your case, 35 V should do because of the resistor network at the output that would attenuate the unintendedly connected phantom supply.

Okay, I'll try this option. My long-term goal is to build a complete mixing desk so keeping the cost-per-channel down is important.

I'll use thick film for this prototype at least, then, and see how it goes.

Does the rest of the circuit look sensible?

 

[James UK]

I've made the design for the THAT1512 version now, too. The resistors in the gain switch will be different values to the transformer version but it will use the same separate PCB.