Telephone with carbon mic using phantom power

[joro]

I would like to try connecting an old rotary telephone with a carbon button in the handset to a mic pre and powering it with phantom power. I've already configured just a handset with a 9v battery that works well. But, I thought it might be interesting to try building an interface box that I could plug any phone into and it would be powered by phantom power.

The phone company supplies 48v over the phone lines (actually, mine read 44.8v), so I guess phantom should be a good supply. The difference is the phone lines have 2 wires (basically + & - for the power) whereas mic lines have 3 wires, pin 2 hot, pin 3 cold, and pin 1 ground (+,+,- for the power). I figured I could use a few resistors and a couple capacitors to make the interface. I'm wondering what would be the best values to use and would like to understand how different values would effect the circuit.

Would someone please be able to explain how different values would effect the performance of the circuit?

I put resistors on pin 1, pin 2, and pin 3 and tied the pin 2 and pin 3 resistors together to combine the +48v and connected that and the pin 1 resistor to the phone. Then I put caps from the phone wires to pin 2 and pin 3 bypassing the resistors.

I know the value of the resistors will effect the amount of current for the carbon button, and the pin 2 and pin 3 resistors should be matched, right? But I'm wondering whether the pin 1 resistor should be less than, equal to, or greater than the pin 2 and 3 resistors and how each configuration will effect the circuit.

I tried the circuit with 1.5K ohms for each resistor (and 10uf caps) and it worked fine. It put the current at around 7mA (instead of the 23mA that I measured when connected to the phone company). By the way, the resistance of the phone measures about 330 ohms. I also tried the circuit with a 12K pot on pin 1 along with the 1.5K resistors on pins 2 and 3. It obviously effected the current and thus the volume, but I didn't really notice a change in the sound whether pin 1 resistance was less than or greater than pin 2 and 3. Does it not matter what values I pick, or are there ideal values that I should use? With no resistance on pin 1 then all the signal would go to ground, right? And with no resistance on pins 2 & 3 then the signal would cancel itself, right? Is there a minimum value that I would definitely want to be above for pin 2 and 3, and is there a minimum value that I want to be above for pin 1?

Thanks for any help understanding this,
Pete

 

[MagnetoSound]

There was a thread on this some time ago, if you do a search on carbon mic you should find it ... I remember it was established that you need a lot less current than you might think.

 

[joro]

Yes, I am aware of all the threads on carbon mic's. I was actually part of some of them. Sorry if my question was unclear. I'll try to restate it more clearly. My question is not about how much current the mic should have or what resistors values I should use to get a particular current. My question is about the ratio of resistor values between pin 1 and pins 2&3.

For example, lets say a total resistance of 10K gives me the current I want. I could choose 5K for pin 1 and 10K for pins 2 & 3 and that would give me 10K total, right? I could also choose 6.66K for pin 1 and 6.66K for pins 2 & 3 and that would give me 10K total, right? I could also choose 9.5K for pin 1 and 1K for pins 2 & 3. Or, I could choose 500 ohms for pin 1 and 19K for pins 2 & 3. I'm wondering how these choices may affect the circuit.

Thanks

 

[MagnetoSound]

Well ... this is feeding a balanced input, right?

So in the interests of good CMRR, you want to keep the source impedance as low as possible and keep the power scheme balanced if you can - so with that in mind, you would look at how small R2 can be made (and therefore R1) in terms of the loading across the mic, so that current can be kept to a minimum by use of a high value at R3. Then put a resistor of equal value to R3 from the hot side of the mic to ground to keep it balanced.

Telephone lines are typically around 600-ohms, so you could probably make that a starting point for R1 and R2.

 

[joro]

I'm sorry. I'm still confused. Let me see if I have this right. Wouldn't the source impedance for the audio signal be the impedance of the carbon mic, which is 330 ohms, since the capacitors bypass R1 and R2. Unless R1 and R2 are low values. (Isn't it 10x the value that you don't need to worry about the effect of the impedance?) So, if R1 and R2 added to 3300 ohms then the impedance would basically be 330 ohms? But, if R1 and R2 added to 330 ohms then the impedance would be 165 ohms? And, if R1 and R2 added to 33 ohms then the impedance would basically be 33 ohms? Do I have this right?

I don't really understand "loading across the mic". Does R1 and R2 create a load across the mic? Should R1 & R2 be 10x the carbon mic or is it okay to have their values be low like 100 ohms. How does this effect the signal? Is there a level drop? Does it effect the frequency response?

Since 330 ohms should be an okay impedance for feeding a mic preamp input, wouldn't I want R1 and R2 to be high values (10x 330 ohms)? Although, for the power, I guess the high resistor values would mean I would have less current if I used a high value at R3 as you suggest.

I guess I didn't think about how R3 could effect CMRR. I guess it would be best to have R3 as high as possible. Is that correct? (at least 10x 330 ohms, but maybe ideally higher?)

I don't really understand why you suggest putting a resistor from the hot side of the mic to ground. Wouldn't that allow the current to bypass the carbon mic? Can you help me understand that?

Is this an okay schematic to use, or would I be better off going with a transformer instead?

Thanks

 

[MagnetoSound]

Wouldn't the source impedance for the audio signal be the impedance of the carbon mic, which is 330 ohms, since the capacitors bypass R1 and R2. Unless R1 and R2 are low values. (Isn't it 10x the value that you don't need to worry about the effect of the impedance?)

10x the value is the ballpark for high quality audio ... but I assume that's not what you're after here.

In times past, power transfer was considered the objective for transmission and impedances were equal at either end of the line, so don't be afraid to try things out, even if they look weird to a recording engineer.

I don't really understand "loading across the mic". Does R1 and R2 create a load across the mic? Should R1 & R2 be 10x the carbon mic or is it okay to have their values be low like 100 ohms. How does this effect the signal? Is there a level drop? Does it effect the frequency response?

Look at R2/C1. They are connected in series across the mic element and present a load, an AC impedance, to the mic. The question is, how small can this be in order to allow for a high value at R3. That's for you to experiment, I suggested 600R as a starting point as it appears to represent the impedance of a standard telephone line. Cap values may be fine, or not. Since the impedance at the preamp is likely to be much higher, you don't need to worry about the loading effect of R1/C2 as they present via the transformer at the other end.

Since 330 ohms should be an okay impedance for feeding a mic preamp input, wouldn't I want R1 and R2 to be high values (10x 330 ohms)?

330 ohms is actually a little on the high side. Fairly typical, but still a little high. 600 ohms in parallel gives almost exactly 200 ohms, which is more like the value your pre will be designed for in any case.

I guess I didn't think about how R3 could effect CMRR. I guess it would be best to have R3 as high as possible. Is that correct? (at least 10x 330 ohms, but maybe ideally higher?)

You probably only need a volt or two across the element for it to work. That means a current of about 3 or 4mA. To drop 46 volts with 3mA, you need 15333 ohms, or just over 15k. You already have 3k4 at the preamp end (two 6k8 resistors in parallel) and 300R at the mic (two 600R resistors in parallel) so R3 can be about 11.6k. 10k should work fine.

I don't really understand why you suggest putting a resistor from the hot side of the mic to ground. Wouldn't that allow the current to bypass the carbon mic? Can you help me understand that?

You don't really need it. I figured it would just ensure that any noise coming up pin 3 from a dirty ground, comes up pin 2 as well and cancels at the input. Since 10k is a little bit of a steeper hill to climb than 330R, the current will pass through the mic.

 

[PRR]

> change in the sound whether pin 1 resistance was less than or greater than pin 2 and 3. Does it not matter

Does not matter. Why would it?

Your R1||R2 & R3 (plus 3.4K) set the mike current. (10K||10K)+10K+3.4K is 18.4K. 48V in 18K is almost 3mA.

These is no reason to run Telco 23mA through the mike. Telco runs "without amplifiers" on short lines, needs HIGH output from mike. You have the luxury of HIGH-gain amplifiers; also your lines don't run in the street next to electric trolley sparks and other crap.

I can see a theoretical reason to make (R1||R2) equal to R3. i.e. make R3=20K. However this does not change "sound". It "improves" output and balance, but perhaps not enough to matter, especially considering the hot output of a carbon mike.

 

[joro]

> I can see a theoretical reason to make (R1||R2) equal to R3. i.e. make R3=20K.

Did you mean R3=5K?

Thanks Magneto Sound and PRR for all the help!

 

[Whoops]

Hi Pete,
are you sure the circuit in the schematic you posted in the first post was exactly the same as you used?
is the schematic correct?

Is that I tried to use that circuit with my carbon capsules but it didn't work, no sound output.
I know the capsules are good, because they work with another circuit I made , but that circuit used a Transformer, I would like to use your idea because it doesn't need a transformer.

thanks

 

[Whoops]

Anyone tried this circuit?

thank you

 

[jordan s]

I did this a few years ago. As I recall, I used this exact circuit and it worked fine.

 

[revmonkey]

I think I have a variation on this design for connection any carbon button transmitter handset connected to a phantom power audio interface for use in Zoom meetings. At least, this is what I'm going to breadboard up and see if I can plug in an old payphone handset and get it working.

 

[Tubetec]

I rewired an old phone one time , not to pick up sound , but to take the the headphone output of a mobile phone and reproduce it over the ear piece , worked fine , the person who was using it simply sat the mobile underneath the phone with a real ring tone where it also picked up the sound of their voice . Come to think of it the usual wired handsfree kit with mic, headphones, and call answer button would be easy enough to wire into an old phone , where lifting the reciever would close the contacts on the call answer button and the usual mini jack for connection to the mobile . If the carbon mouthpiece was'nt suitable to run off the usual plug in power an electret mic could be used instead . Bit of a novelty item alright ,but might be useful instead of holding the mobile to the side of your head on longer calls .

 

[Chrisfromthepast]

I found an old bakelite western union telephone handset this week and tried the OP's circuit.
It worked just fine. tons of output level, and a bit of noise.
I connected the ear piece as well. Talking into that mouthpiece and listening through the headset was a trip down memory lane.
I talked into the earpiece too for reference. required less gain than I expected, maybe 40db.
The rf noise is probably due to the data cable I chose, and the fact that the speaker run inside the phone is not a twisted pair, so it became an antenna, and the plastic (bakelite?) enclosure that I didn't want to ruin.
Here's a test where I sound like someone from the song valley girl:

 

[Whoops]

I found an old bakelite western union telephone handset this week and tried the OP's circuit.
It worked just fine. tons of output level, and a bit of noise.
I connected the ear piece as well.
I talked into the earpiece too for reference. required less gain than I expected
Here's a test where I sound like someone from the song valley girl:

Both sound pretty good and usefull for Lo-Fi duties