Can I convert a mic input to line level?

[Katch]

Hiya,

Just wondering if anyone can help. I have a pair of Roland MA12C monitors that have 3 inputs on the back. 2 are at line level and one is at mic level. Specs below.

Line input -15dBm at 50kohms
Mic input -45dBm at 1kohms

Is there anything I could run inline to feed a line level into that mic input? Or do you think it might be possible to mod the input itself to bring it to line level specs?

 

[MAAC0]

In order to mod the input You need to change the feedback network of the op-amp. A schematic would help a lot.

You will need about 40-60dB attenuation. Check this passive circuits.

Uneeda Audio - Build your own attenuator pads

 

[sgrossklass]

Since you apparently need about 30 dB of attenuation, you could take advantage of the existing 1 kOhm input impedance to make the simplest unbalanced L-pad ever, with just a 33 kOhm resistor in series.

Since all 3 inputs are mixed together, however, noise floor at higher volume settings may benefit from modifying the actual circuit responsible for the mic input for lower gain. In this case some reverse-engineering may be required.

That said, if you have this many sources, you should really contemplate using some sort of mixer, which would have the added benefit of more flexibility.

BTW, an input sensitivity given in dBm is very quirky. I hope the good folks at Roland know what they mean by that, 'cause normally it's a unit used in impedance-matched systems (as common in RF) and does not apply to voltage matching "short into open" scenarios like audio connections. It's dB relative to 1 mW into nominal impedance. Possibly someone was using a signal generator from the 50 ohm world, in which case 0 dBm would be an unloaded output voltage of 447 mVrms, -15 dBm would be about 80 mVrms and -45dBm would be about 2.5 mVrms.

Given the output power spec of 10 W @ 6 ohms, that translates to a gain of 39-40 dB for the line inputs and 30 dB more for the mic input. There are several older power amp ICs with 40 dB of gain (TDA2822(M) comes to mind), so that sounds plausible enough.

 

[MarcelvdG]

More likely 600 ohm, which used to be the normal termination impedance for professional audio connections. 0 dBm into 600 ohm corresponds to 0.7746 V RMS. Nowadays it is usually called 0 dBu, meaning the voltage that would have resulted in a 1 mW power if the load impedance had been 600 ohm, no matter what the actual impedance may be.

I have the impression that those numbers are sensitivity specs, indicating the minimum level to fully drive the loudspeakers. Are there also signal handling specifications, showing how much signal the microphone input can handle without clipping? Does the microphone input have a phantom supply and if so, is it easy to turn it on by mistake?

 

[sgrossklass]

PS -
Assuming that the opamp used in the mic circuit is some sort of 4558 or TL07x, mic preamp output noise would be dominating total noise output and give about 5-6 mVrms of power amplifier output noise at full volume, or a SNR of 62-64 dB at full tilt... I'd estimate that an improvement by almost 20 dB may be possible here.

Obviously, if you are already operating in the volume control range where the power amp IC's basic noise floor dominates (these can contribute 200, 300 or 500 µV all by themselves), this improvement would go unheard.

EDIT:

More likely 600 ohm, which used to be the normal termination impedance for professional audio connections. 0 dBm into 600 ohm corresponds to 0.7746 V RMS.

That would complicate things a bit.
Then we'd be talking about 278 mV for the line inputs and 5.4 mV for the mic input (600R/1k divider included), and a power amp gain of 29 dB.

I think the power amp might be a car audio kind of job, 10 W / 6 ohms (or +20 dBu or ~22 Vpp) is just about what I'd expect from a BTL amp on a 12 V supply.

I have the impression that those numbers are sensitivity specs, indicating the minimum level to fully drive the loudspeakers. Are there also signal handling specifications, showing how much signal the microphone input can handle without clipping?

Nope. This is all you get.

Does the microphone input have a phantom supply and if so, is it easy to turn it on by mistake?

As basic and old as this thing is, I wouldn't be surprised if it didn't even have electret bias and the (unbalanced) mic input were intended solely for karaoke type applications with a dynamic mic. Mic amplifier might be just an inverting opamp stage with a cheap and cheerful audio opamp, either 4558 class or TL07x. I think you can do the entire preamp section on one quad OP.

 

[simon7000]

A 30K resistor in series with a 100 ohm one will give you the attenuation you need.

For the confused the unbalanced output goes through the 30K resistor to the 100 ohm unit. The output is across the 100 ohm resistor. Both shields/grounds combine at the return end of the 100 ohm unit. If it makes a loud pop when you connect the pad there may be a DC bias on the input in that case add a 10 uF 25 volt electrolytic capacitor in series with the output. The positive lead goes to the input.

To do it balanced the shields are connected together and each output is fed via a 15K resistor to the single 100 ohm unit. The outputs are from each end of the 100 ohm resistor.

The slightly greater gain is reduced by the input impedance.

 

[PRR]

> A 30K resistor in series with a 100 ohm one will give you the attenuation you need.

That's 50dB.

> Line input -15dBm
> Mic input -45dBm

That's a 30dB difference, no?

Foo. Get a handful of 10k resistors and series them until stock-line and new-line sound about the same. This assumes the mic "is" 1k, which is dubious, but experimentation may lead to a workable answer.

"dBm" may be obsolete but many standard audio voltmeters are calibrated in dBm re:600r. They just reading the numbers off the scale.

 

[sgrossklass]

"dBm" may be obsolete but many standard audio voltmeters are calibrated in dBm re:600r. They just reading the numbers off the scale.

So ultimately really just an old-fashioned way to say dBu then, as Marcel already stated earlier. I am well familiar with those.

In that case I'd have to revise my input sensitivity figures to 275 mV / 8.7 mV rms.

 

[simon7000]

Slipped the decimal point. 3K! Oops. That is what I get for doing math in my head.

20xLog(R2/(R1+R2))

 

[MarcelvdG]

As basic and old as this thing is, I wouldn't be surprised if it didn't even have electret bias and the (unbalanced) mic input were intended solely for karaoke type applications with a dynamic mic. Mic amplifier might be just an inverting opamp stage with a cheap and cheerful audio opamp, either 4558 class or TL07x. I think you can do the entire preamp section on one quad OP.

If that is true, then the lowest-noise option to get 30 dB of attenuation without modifying the internal circuitry is a series resistor of 30 kohm or 33 kohm placed in the plug that goes into the microphone input.

If there were a phantom supply, it might be necessary to protect the line level source from it, particularly if it were a 48 V phantom supply. But that's indeed unlikely with a simple unbalanced input.