My BM800 microphone isn't working

[FallenHero02]

Hi
I have a bm800 condenser microphone that I bought online from ali express. I have built my own 48V phantom power supply that puts out a constant voltage of 49V that I know should be okay. But the weird thing is that when I plug the microphone into the power supply and I check the voltage on the other side of the XLR input, the voltage drops to like 4.7V. I don't know why it isn't working with the 48V phantom power, but when I use it with a cheap sound card that it came with online it works. The sound card produces around 2.2V. I don't understand why this microphone isn't working with phantom power but its working with very little power.
I thought about trying to switch the XLR point in the microphone but I don't know if that will change anything.

Could you help me figure out how I can fix this?

[Whales]

The 48V power supply will be very current limited to avoid destroying equipment, this limit is probably being hit.  Computer mic inputs typically supply something like 5V phantom power through a resistor.

It's possible your bm800 is using a cheap electret capsule or dynamic style element instead of the promised condenser element.  Few people open their mics up to check and you bought it from a primarily greymarket site, so it's more than possible.  That and I think many sellers confuse "condenser" and "electret".

EDIT: don't turn it on, take it apart

[FallenHero02]

What am I not turning on and taking apart?

[FallenHero02]

I talked with a friend and they thought that the mic impedance could be an issue.
Here's the impedance values for one I found:

Output Impedance: 55Ohm more or less 30 percent
Load Impedance: More or equal than 1000 ohms

Here's another I found that seems to be the more common one:

Output impedance:150Ω ± 30% (1 khz)
Load impedance:≥1000Ω

Could this be why the microphone isn't working with the 48V phantom power?

[floobydust]

BM800 is not a true condenser microphone. It has a small diaphragm electret-condenser mic element - so it does not require high (polarizing) voltage. It's really a bland low-cost electret packaged in a pretty, large diaphragm looking kind of case. It has the cosmetics but not the internals, and best to gut it and put in a large 34mm diaphragm with matching electronics. This is what many do, for a high performance microphone that can be used in a studio. They just use the case.

There are a few BM800 variants, some with only basic electronics inside as a "5V" or "Doesn’t need phantom power" specs, so it works when connected directly to your PC soundcard.
Your voltage measurements show it is likely this - just a JFET with unbalanced output.

Or the PSU is not working. It's 48VDC with two 6.8k resistors for a feed? At 3mA load, still should be much higher voltage there. Or one XLR terminal is shorting to GND in your cable.
Take the mic apart and give us a pic of your pc board to figure out what you got.

edit: added pic of typical board and drawing a BM800 schematic...

[magic]

Is it possible that the mic contains a 4.7V zener to make it work with either PC jacks or phantom power? That number is suspiciously familiar.

edit
Yeah, I guess it's not because the zener would also clamp the AC-coupled signal. So is it possibe it has a zener to protect it from 48V phantom power?

[floobydust]

A funny BM800 pcb pic, notice the IC is not connected to the rest of the circuit 

[floobydust]

A bit more insight into these BM800/NW800 microphones.
Got a new NW800 from the NEEWER factory store and it does not work well, distorted audio.

I drew the schematic, measured voltages and conclude a transistor is defective and the circuit is plain silly. So much for copying a copy. The LPF is done wrong, there is an extra transistor that does nothing other than leak and a few parts that are just in the wrong spot.
I would say build quality is pretty good, but they got gypped on the semi's and circuit design, as a "brand". I think I'll remove the transistor-of-stupidity and see how it does.
Realistically you could just wire the electret mic module right to a PC soundcard, instead of all the phantom power drama.

My plan is to use the housing to build a real condenser mic with a large capsule and decent pcb inside similar to Pimped Alice but my own way.

NEEWER NW800 microphone schematic, NW-7-800 V2.0 20170506. SMT version.

edit: Removing Q2 and the mic voltage went up from 2.7V to 6.8V and no more distortion from the diode action. But - the mic still has so much hiss, and freq. response is tinny with thumpy low end.
The EQ needs to be fixed - change R24 to 2k7; C7, C11 to 1uF. C12, C13 to 0.1uF. Move R18 to R8.  C9, C8 roll off the high end, not sure if needed.
With 5V power (and one 6k8 to 5V, similar to a PC soundcard), the mic gets starved with only 0.5V and the EF amp clips easily. So it's low volume and overloads with louder sounds. Noise is much less.

Overall, NW800 is a poor design that is not worth fixing. The circuit board is just done wrong and a lot of work to mend. Better to just run the mic capsule direct to the PC in 5V use.

[Whales]

Q2 is indeed interesting.  Original some form of pull-up?  High-side of an AB amplifier?  Not sure.

If you change the RC above it to connect to a positive power rail (rather than GND) it might actually do something.  Remember that BJT devices can be used successfully backwards, they just have crappier specs (including lower breakdown voltages). 

For years I used my cheap 2N2222's backwards (and wondered why they would so often fail), but they did work

[floobydust]

I think Q2 (SOT-23) and Q4 (TO-92) are for some JFET option where you have a condenser (non-electret) capsule - but R11/C2 are wrong for that. So Q2 would then be a JFET but why they stuffed in a reversed leaking BJT there is kind of funny. Maybe it's cheaper than putting the parts in landfill.
They've put a capacitor where a resistor should go and vice versa, so the board is a bodge.
I hear it's common in china to copy designs resulting in no one really knowing how they work and so they can't support any problems or questions for production.

I'm going to pull Q2 and see if works, the LPF filter is still bungled. I'm glad I only bought it for the shell.

[magic]

What's the "capsule" net?

Operation of the transistor of stupidity is pretty obvious: it's a (reversed) PNP which dumps some current into the grounded RC circuit whenever the "capsule" net draws current from the 4k7 pullup. Now, I'm not saying that it makes any kind of sense

[floobydust]

It comes with an electret-condenser mic element, which has the JFET inside.  (my pcb is different than pics)
People rip that out and put in a large diaphragm capsule that is non-electret, copying the legendary Neumann microphone designs U47, U87 etc. for a fraction of their cost.
But you have to change the electronics and add a 60VDC bias generator. so the Capsule node option connects to I think a Transsound TSB-2555B or something like it, that has no JFET inside.

The mic seems to be the ultimate Millennial/blogger bait, lots of glitter and wires but doesn't perform. It was under $20.

[magic]

So it's not connected to anything here? I think there will be little difference from removing Q2. The whole section is a constant current sink from the 8.5V rail. If anything, pull out R18 to kill it entirely.

Nevermind, I missed that R18 and this contraption are connected to the signal path. Meh, have fun with that junk