Active ribbon-mic

[sellch]

please have an opinion on which one might work best.
thank you very much

 

[Tubetec]

Its an interesting line of enquirey , thanks for all the help Abbey .

I do fancy the idea of a powered 58 , in something like front of house sound you might have long cable lenghts typically on multicore with foil and drain wire screen and other sources of noise in the cabling etc .

The typical inline fet boosters give 20-25db of gain , why not make gain trimable over a wider range say 20 to 50 db .
They still leave most of the heavy lifting to the mic input on the console , why not make the inline amp full line level output, can a conventional dual rail op amps be made to run from 48 volts supply ?

 

[sellch]

please have an opinion on which one might work best.
thank you very much

 

[abbey road d enfer]

Its an interesting line of enquirey , thanks for all the help Abbey .

I do fancy the idea of a powered 58 , in something like front of house sound you might have long cable lenghts typically on multicore with foil and drain wire screen and other sources of noise in the cabling etc .

The typical inline fet boosters give 20-25db of gain , why not make gain trimable over a wider range say 20 to 50 db .
They still leave most of the heavy lifting to the mic input on the console , why not make the inline amp full line level output, can a conventional dual rail op amps be made to run from 48 volts supply ?

Making a phantom powered preamp with performance similar to a decent console preamp is bound to exceed the resources of phantom power.
You need about 10mA, which is the maximum allowed by P48. In these conditions, the available voltage drops down to 14 V, which does limit headroom.
Then there is the issue of gain control. There is nothing faster than a gain knob on a mixer. The same knob on a mic is a receipe for disaster.

 

[abbey road d enfer]

please have an opinion on which one might work best.
thank you very much

View attachment 86111

North-West needs a larger input cap, probably 5-10uF. Probably rather high THD.
North-East looks OK, except the 4.7k should be deleted. Pobably too much gain for guitar. Delete 22uF cap.
South-West has too low input impedance for guitar, needs 500k-1Meg. 2SK170 may run out of juice.
South-East has way too low input impedance for guitar. Delete 4.7k.

But don't trust me on that. You need to experiment.

 

[sellch]

North-West needs a larger input cap, probably 5-10uF. Probably rather high THD.
North-East looks OK, except the 4.7k should be deleted. Pobably too much gain for guitar. Delete 22uF cap.
South-West has too low input impedance for guitar, needs 500k-1Meg. 2SK170 may run out of juice.
South-East has way too low input impedance for guitar. Delete 4.7k.

But don't trust me on that. You need to experiment.

good morning
Basically I will not use it for guitar but I think of all this for the ribbon microphone as a simple idea without many quality requirements for connection as I said on a mobile phone or tablet or why not a desktop computer.
ok i will try some of them and transfer my experience.

Abbey thanks very much
good morning

 

[Tubetec]

How about using a dedicated phantom power supply that allows more current than a standard console channel . There many variations on the theme, 2x9v battery powered, 5v usb power or other external DC or AC supplies . Some dont give out true 48 v .
The rockville RNDI-1 promises upto 100ma at 48 volts .

 

[abbey road d enfer]

How about using a dedicated phantom power supply that allows more current than a standard console channel .

Current limitation in P48 is essentially due to the 6.8k feeder resistors. Short-circuit current is thus limited to 14mA. The 10mA figure is not the current limit; it's the current that the phantom circuit should be capable of delivering. Nothing in the standard says that the supply should be limited to 10mA.
Any power scheme that is capable of delivering more than 14mA to a single mic is not in accordance with the P48 standard.
For example, one could use 1k feeder resistors, which would result in about 100mA short-circuit current, but wouldn't be P48.

The rockville RNDI-1 promises upto 100ma at 48 volts .

This is the max current the wall wart is capable of delivering. It means that it can power up to 10 mics at 10mA, but it could also power about 100 U87's.
RND1 is a DI box.

 

[turnerk6]

Any tips simulating a circuit like this in LTSpice? I can't figure out if my issue is in the phantom power simulation.

 

[turnerk6]

Any tips simulating a circuit like this in LTSpice? I can't figure out if my issue is in the phantom power simulation.

Nevermind! LTSpice's LSK389 wasn't working for me. Replaced them with plain jfets and it works! Hey abbey road d enfer, can you guide me in the right direction for adjusting gain level in your design?

 

[abbey road d enfer]

Any tips simulating a circuit like this in LTSpice? I can't figure out if my issue is in the phantom power simulation.

Post your circuit, I'll look at it.
You must rename it as .txt before.

 

[abbey road d enfer]

Nevermind! LTSpice's LSK389 wasn't working for me. Replaced them with plain jfets and it works! Hey abbey road d enfer, can you guide me in the right direction for adjusting gain level in your design?

Gain is controlled by the ratio of drain resistors vs. source resistors.

 

[turnerk6]

Let me know if that works. My simulation is also giving me no positive gain, it is attenuating the signal. I'm pretty new to LTSpice so it's probably a silly mistake.

 

[turnerk6]

When I simulate the circuit, I can only get a 25dB boost by setting the resistors going to the LSK389 drain pins to 27k instead of the 1k from your schematic. What am I doing wrong?

Note: when I convert the asc file to txt, the polarized capacitors change from microamps to AuA? I had to manually change them back once back in LTSpice.

 

[abbey road d enfer]

You can't use the basic NJF model, which is inadequate, nor the LSK389 model, which is wrong.
I used the LSK170A model, which is correct. At least the sim results are confirmed by experience.

EDIT: I was wrong, the new models for LSK389 are good (and actually identical to the LSK170's).

 

[turnerk6]

Thanks for the help. I switched the jfets to LSK170A. On your schematic, R7 and R8 are labeled as 470k. Is it possible those were simulated as 47k? At 470k, I'm getting attenuation. When I switched those to 47k, I got about 25dB of gain.

 

[abbey road d enfer]

Thanks for the help. I switched the jfets to LSK170A. On your schematic, R7 and R8 are labeled as 470k. Is it possible those were simulated as 47k? At 470k, I'm getting attenuation. When I switched those to 47k, I got about 25dB of gain.

Got it! You use the generic PNP transistors, that have a lower h21e, so with 470k, there is not enough current for the FET's. If you replace them with the recommended BC327-40, gain is correct.

 

[turnerk6]

Wow thanks for your help! What is the max gain for this circuit with the LSK389 and BC327-40? My simulation seems to max out around 31dB.
Also, is there any reason to be particular about which FET to choose between LSK389A, B, C, or D?

 

[abbey road d enfer]

As I wrote earlier, gain depends on the ratio of drain and source resistors.Only the "A" version of the LSK389 works adequately in this cicuit.
The "B" version works if only one is used. Other versions don't work because they draw more current than P48 is capable of delivering.

 

[turnerk6]

Thanks, I really enjoy little projects like this so thanks for your help and for sharing the circuit. Do you typically choose the surface mount or metal can packages? Any noise/performance differences?