bass pedal phantom power

[reinerterig]

Hi everyone,

I used to own an MXR Bass DI+ and it was super convenient because I didn't have to lug a bass amp around where ever I went and the pedal was phantom powered.

I am trying to build a new bass pedal starting with the balanced out. I successfully split the P48 into a dual rail supply for my opamps and signal is passing right until R9 and R10 at which point it disappears. I assume I am sending via R11 and R12 through ph and down to ground, but this is taken pretty much exactly from a schematic I found on
 http://sound.whsites.net/project35.htm

Does anyone have any advice on how to separate my signal and power in this application, and can phantom power provide me with enough current for the rest of my design? ( it must be possible if MXR and Tech21 SansAmp can do it. )

[JS]

How are you feeding the 48V? directly to where it says 48V? If so you should add a 6k8 resistor in series with each one of them. So you have 48V->6k8(hot wire)6k8->VCC and same thing for cold wire.

In any case 100k seems a bit high for R9 and R10, in fact the circuit you are quoting use 100Ω.

Now, on the current availability, you can get up to 3.5mA at 24V with the pruposed DC circuit. In which case your current is already chewed up almost entirely for you TL072. Now, you don't NEED 24V, you don't NEED to use 6k8 resistors (for R10 and R11 on your schem) and you don't NEED to use TL072. Other thing you don't need is two opamps driving the output, so you can have one available there. Now, consider the phantom in the mic preamp is pretty standard 6k8 48V while it might not be I dear you to find a console that makes it different, entry level sound card might, let's say you won't be connecting to those.

You could start using 9V VCC, good enough to work with ton of guitar and bass stomp boxes, good enough for this. Certainly good enough to drive the required small signal for a mic preamp it will be driving. One or two Volts peak to peak is enough. Now, the input impedance of a mic preamp is over 2kΩ, we wouldn't want to waste all the current driving our internal resistors nor the mic preamp, so at most we would use 1kΩ resistors instead of the 6k8, even higher if possible. But with the 1kΩ resistors at 9V now we have 10mA to play with. 1 of which might be going to the load, so 9 to feed opamps. Are 3 TL072 enough for you? if not use a lower power ones. TECH21 uses TLC226X opamps, and use a buch of opamps in a single box. They will take under a third of what a TL07x will do for each OA, just be careful as they can only be suplied from ±8V, so not 29V batteries unless you drain them quite a bit before using them! They can run perfectly fine from a single 9V rail (or ±4.5V) and the output does swing from rail to rail so better headroom than with other things, good for low voltage applications.

JS

[reinerterig]

How are you feeding the 48V? directly to where it says 48V? If so you should add a 6k8 resistor in series with each one of them. So you have 48V->6k8(hot wire)6k8->VCC and same thing for cold wire.

Yes, the 48V label on my schem is from any preamp ( in my case a Focusrite OctoPre ). I am aware that most preamps have 6k8 resistors to pin 2 and 3 so I left them out of my schem. I have added the 6k8 resistors in series (R11 and R12) which goes to my voltage divider, the problem is I think they are also sending my signal to ground.

Other thing you don't need is two opamps driving the output, so you can have one available there.

I've seen that I can use only one inverting opamp and impedance match the non-inverting side. that shouldn't be too hard?

1kΩ resistors at 9V now we have 10mA to play with.

So I should basically get my 48v down do +9V and -9V using minimal resistance?

Are 3 TL072 enough for you?

Do Transistors consume too much power? I was thinking about using 3 or 4 transistors and an eq section.

[JS]

I've seen that I can use only one inverting opamp and impedance match the non-inverting side. that shouldn't be too hard?

That's right, just that. Same series resistor and cap from your reference or ground to the output.

So I should basically get my 48v down do +9V and -9V using minimal resistance?

I'm saying just 9V and divide to get +4.5V and -4.5V. Not minimal resistance, there's a balance in there, I'd say at least 1k each. In any case you are never getting negative voltages if we keep the ground as reference (you might with a voltage converter but no need and not a real application here)
In fact. max rating for the TLC226X is +8V to -8V or 16V total, so your ±9V rails would cook the thing.

Do Transistors consume too much power? I was thinking about using 3 or 4 transistors and an eq section.

That depends greatly on the circuit, is not like the opamps that have an idle current consumption (power will depend on voltage, here we are concerned about current)
In any case,  depending on the circuit you could easily use 3 or 4 transistors, each opamp has dozens of them!

Here is a conventional phantom circuit for a condenser mic capsule. That output stage is quite convencional for discrete phantom powered circuits, it drives the output correctly and it provides a 12V supply for the rest of the circuit, plus a 35V polarization voltage that you won't be using (no current at all from this one)

JS

[reinerterig]

Ok, I see that circuit is trying to do the same thing, using a "push-pull" configuration do drive the differential amplifier on the mixer side.

Same series resistor and cap from your reference or ground to the output.

I have made some changes based on what I understand you saying.





First I took my 48V from the Mixer and dropped it just below 8V using a diode rectifier. I divided that into about +4V and -4V.



I removed my Non-Inverting opamp and I now just have one opamp inverting my signal. This Gives me exact signals on pin 2 and 3 but
180' out of phase to drive the differential amplifier on the mixer side.



I am getting the correct signals after C2 and C3 showing the push-pull circuit is working as expected but once we go past R9 and R10 the output shoots up to a steady 48v killing my signal. I expect this to be the case because my signal is traveling on the same line as my power( the tracks after R9 and R10 is basically the XLR cable ), But that is how phantom power works, power and signal on the same line. I just don't know how to separate them so my signal can reach the differential amp on the mixer side.

here is the new circuit as a whole

 

[JS]

You have to put the 6k8 resistors that go inside the mixer in your simulation, otherwise it won't work.

Go back to your original design, just change the 100k for 100Ω and add the 6k8 resistors. Once you got that running we go to the next step.

Then, in order to emulate the circuit driving a Mic preamp you should load it as it was driving a Mic preamp. Let's say 2k resistor from hot to cold (hot and cold are pin 2 and 3 of the xlr)

Next, to remove the opamp, remove the one on the cold side and connect to ground what was connected to its output.

Build up step by step, look at the simulations so you identify first what each change is making to the signal and second you can know when it when wrong and just make one step back.

JS

[reinerterig]

Ok, there we go! everything is working as expected. I think wired the phantom power incorrectly so I redid it as it should look coming from a mixer.

[JS]

Now we are getting somewhere, your output circuit is not what I'd do it. Now I'm on the phone but I could draw it later.

Also, short u3 inputs and lose that op amp, just a resistor would make a preamp model for your purpose.

JS

[SiliconWizard]

Take a look at the OPA2376 datasheet: http://www.ti.com/product/OPA2376 (section: "Phantom-Powered Microphone").

[JS]

  Here is the circuit I'm describing, for R9 I say 2k for worst case load, most mic preamps would be higher. Even then you only need to put 2mV onto it to be good to go, an SM8 puts 1.85mV at 94dB SPL, and mic preamp takes it from there. Now you can get to that level values by using low gains or attenuating at the end, if you push 10Vpp to the output the mic preamp won't be able to handle it.

  Now, as your signal is pretty small you don't need much voltage at all, is more a SNR matter on how low you can go, that then will depend on the choosen opamps to set the noise floor you have and the maximum excursion of the signal at each stage. In any case there's no much point to aim to a SNR higher than 80 or 90 dB as the signal probably won't get there nor the mixer, let alone the live mix.

JS

[reinerterig]

Here is the circuit I'm describing,

I redrew this circuit in a new LTspice project. I just put the power section to one side so I have a better idea of what is going on.
Is this more or less correct?

Even then you only need to put 2mV onto it to be good to go

MXR expects +-250mV signal from instruments and it has a 2.2Meg Zin and 300ohm Balanced out.

there's no much point to aim to a SNR higher than 80 or 90 dB

90dB is perfect!

P.S Thank you for all the help so far!!

[JS]

R11 should be higher, it's value will define the input Z, 2.2M if you like, which is not fine for any opamp but no worries for TL071 or TLC2262. The node between R3 and R4 should go to vref.

I said 90dB as a ceiling you should aim for, no worries if you don't get there. If it's just a stage thing you can probably live fine with much less. To account for it you should read the datasheets and do the math for your opamps. Discrete transistors is a bit harder. At the end you need a certain level to feed the Mic preamp. Let's say 20mv peak. Your noise should be 20μV peak, if you end with a higher noise and you can have higher signal you should add an attenuator at the output using a resistive divider. I think that will be the case.

JS

[reinerterig]

Does This look good? I think I might be happy with it, don't know for sure until I can get it on a breadboard. But the basics are there...pin 2 lowZ signal(or +) pin 3 "GND" (or -) and pin 1 shield.

[JS]

Go for the breadboard! Do not populate R3, it's in parallel with the 2M2.