Anybody know what the input impedace of the mic input of a Nikon is (assuming that they are all the same)? That would have a bearing on how big the capacitor should be. I would recommend against electrolytics for several reasons, including possible DC leakage.

If the Nikon ME-1 mic has about 2.2 kilohms output impedance, then presumably the input impedance of the camera is much higher than that. That would make life easier.

I tried a pair of stereo headphones as a microphone and didn't get any plop when I started to record. I don't understand why Laurence gets one with his device.

EDIT

I missed that Laurence is hearing the plop in his headphones connected to the Beachtek.

Does a plop get recorded by the Nikon?

If not, one explanation is that the DC is switched first and recording starts after the transient has died. I wish my meter would read the DC. It must be via a very high impedance and my meter loads it down.

I tried a different multimeter (Keithley) and I see 2.6 volts when the camera is recording, and 0 volts when not, so the DC is switched on only when recording.

I am happy to monitor just over the Beachtek so long as I see some sort of visual (flickering green and red LED) that shows me my levels are ok.

Really, the Beachtek is fine with the D5100. I have good levels that are easy to judge with the green/red LEDs, It works fine with or without AGC. The AGC on the Nikon is actually quite nice. It only limits the loud peaks. It does not bring up the soft parts. The only problem is that if you overshoot the levels too much, there is a slow decay that seems to cut the audio way back for a moment afterwards. If your levels are in in the ballpark (which is easy to do with the Beachtek LED) it works like a safety limiter. My Rode VideoMic Pro works nicely in the 1/8" aux in on the Beachtek. That leaves me a free XLR input for a wired or wireless lapel mic when I do interviews. When I am not using the second mic, I can just switch the audio to mono, just like I am used to.

My only complaint is the pop, but I am beginning to think I will just have to live with it.

There is still the DC offset you mentioned in your first post. The DC must be upsetting the Beachtek amplifier somehow. It must be a poor design if the DC has sufficient drive to upset the amp OUTPUT. (I was an electronics engineer in a former life.)

Laurence, does the DC offset affect both channels equally, or is one biased up and the other down?

The DC has positive on the left channel and negative on the right. Wait a minute, is the tip on the jack left or right? Anyway, the positive is on the tip.

@peter "It must be a poor design if the DC has sufficient drive to upset the amp OUTPUT. (I was an electronics engineer in a former life.)"

Agreed. exactly. (and me in my current one)

Good point about if the voltage is only present on one pin. Could simply use the other input with no voltage on it. Now that would be simple :) Wish i had more data on the nikon's mic socket. Anyone got any good data?

Paul.

Laurence,
It appears your Beachtek device has no set of balanced outputs, so my direct box solution (which works on its own), becomes less practical than I first thought.

However, I found this item from Jensen. It's an unbalanced->unbalanced passive stereo isolation box, quite small.
I don't like the RCA connectors, but it may give you some ideas.
The direct current draw across the primary (<0.05mA) would be far less than for a typical electret or condenser mic, and completely isolated from the output:
http://www.audioxpress.com/magsdirx/ax/addenda/media/hansen-galo2706.pdf

Best of luck, and let us know how it works out.
;?)

This isn't rocket science - just take the battery out of the Beachtek.

I am surprised that you can't turn off the phantom power on the camera. I have only owned one Beachtek, and it had a switch to turn off the phantom power.

The Iso-Max is not going to help, in fact it may make the problem worse.
The real problem here isn't isolating the steady state DC voltage at all, the problem is caused by the transient as the DC power is switched on by the camera.

As it is switched on current will flow in the secondary of the transformer. That will cause a change in the magnetic flux in the secondary. That is magnetically coupled to the primary and hence the output of the BeachTek. The dV/dT at the secondary will appear at the primary as a transient with a fast rise time and slow decay with a peak to peak voltage in the order of 4V.

The output of the device in the BeachTek will see this and probably doesn't have a low enough impedance to attenuate this low frequncy transient. Also connected to this output would be the amplifier that drives the headphone output of the unit.

In reality this isn't a DC problem at all, although the signal never goes negative and hemce isn't AC it is still a signal and no doubt with frequency components in the audio spectrum hence it is impossible to filter out or block.

One solution that would work but will be difficult to implement is a mod to the BeachTek that disables the signal to the headphone amps until it detects the DC voltage coming from the camera. A simple RC network could delay turnon until after the transient has passed. This would save Laurence's ears but on the downside would mean he could only monitor when the camera was in record.

Bob.

Bob, of course you are correct. This is a transient produced when the camera goes into record. However my thinking with the simple capacitor DC filter is this: it depends on the transient rise time. If this is very fast, then yes, i doubt it could be filtered out due to the wide spectrum of frequecies produced by the low to high transient. However, i suspect the rise time is not that fast. From a design point of view, i would have personally made the rise time gradual rather than sudden to reduce 'thump' when the camera starts to record. Well, thats what i would do but this does not mean Nikon has done this. Measuring the rise time from mic plug would be great data to have but unfortunatly we don't have that. So, if this rise time is gradual enough then it can be filtered out because it would have no high frequency components.i.e. the simple capacitor filter would work.
Worst case is the rise time is very fast and cant be filtered, then my suggestion earlier of delayed audio connection would be the only other way. Basically a voltage controlled audio switch with a small delay. RC network used to set the delay.
I find it really heard to understand why the preamp unit is doing this in the first place. Have the designers not tested this preamp by connecting it to cameras with a DC bias mic socket?
Laurence, i do hope you find an answer to this. Its an interesting problem which ulimately is an incomapabilty between camera and preamp. Filtering with a capacitor may or may not work for the reasons above. Lack of data to say for certain. good luck.

Paul.

"Good point about if the voltage is only present on one pin"

No, the voltage is floating and appears between the left and right audio hot pins. If I measure between either pin and ground I get zero (very high impedance compared to my meter).

EDIT

The mic jack on my camera is very poor. The plug keeps slipping out. Now I am getting +2.6 V on both audio pins relative to ground.

I'm not sure exactly what is going on. I do know that my D5100 powers both sides of a stereo pair like the ones shown on this page:

http://microphonemadness.com/categories/mini_binaural_stereo.html

It seems to me that this must mean that there is a stereo pair of voltages, one going to the tip and sleeve, and the other going to the ring and sleeve. My volt ohm meter tests however where similar to yours in spite of this.

Yes. See my edited post above.

Where people are getting confused here is this "plug-in power" is NOT the same as phantom power.
Phantom power is a common mode DC voltage used to power a microphone. In one of my jobs years ago we used 48V phantom power to operate relays over telephone lines that also carried audio. The phantom power was "injected" into the line between the centretap of the line isolation transformer and ground and "extracted" from the line at the other end again between the centretap of the transformer and ground to operate a relay. Despite switching 48V DC at quite a few milliamps on and off not a pip could be heard on the line. The simple joys of the old days of audio :)

Plugin-on Power is quite different. What happens here is part of the preamp circuit is inside the device with the plug-in power socket. The device contains the drain resistor (typically 10K) for the FET amplifier that is inside the mic capsule. Yes, you need one for each channel.

The page here explains how it works.

There's two different schemes in use on the same friggin socket just to confound all this. A mono mic input does supply clean DC on one contact and gets audio back on another. A stereo input does it differently..The page here shows all the different configurations in use.

Bob.

"Now I am getting +2.6 V on both audio pins relative to ground."

Peter, that makes sense. It means you can connect a mono mic to either the left or the right and still you have power to it. And of course it also means you can power a stereo mic.
Well, the downside is my earlier comment about one input possibly not having a voltage on it is unfortunatly, not the case. Back to square one really. Just means there are two inputs with voltage rather than just one.

Paul

Laurence, getting back to the asymetry you see. I wondered whether what you are seeing is the natural asymetry of voiced speech. If you record close speaking with good low frequency response you will see that each time the glottis snaps shut you will get a high peak which oscillates with decreasing amplitude (more or less) until the next glottal event. When recorded at some distance, however, reverberation will smudge this, while poor low frequency response will introduce phase shift which also hides this.

Do you get any noticeable asymetry when you record distant music played through load speakers?

" In one of my jobs years ago we used 48V phantom power to operate relays over telephone lines that also carried audio. The phantom power was "injected" into the line between the centretap of the line isolation transformer and ground and "extracted" from the line at the other end again between the centretap of the transformer and ground to operate a relay."

When I was a lad with the PMG, telegraph signals were sometimes sent on the derived "wire" obtained from centre tapped transformers at each end of an open wire trunk circuit. Ground was used for the return path. Such a circuit was called a cailho circuit. You could use two derived "wires" to form a balanced pair for audio, and this was called a phantom circuit, which in turn could have a cailho circuit.

In theory you could take two phantom circuits and derive another balanced pair called a ghost circuit, and two ghosts to make a spectre and so on, but this was never done because any slight unbalance in one of the balanced pairs would cause crosstalk.

This site talks about cailho and phantom.

http://www.qsl.net/vk5br/StrayNoise.htm

I know the audio scene these days refers to a cailho as a phantom. Tut tut!

I just got the following response from Beachcom:

Hi Laurence,

It looks like I need one of these:



http://www.bhphotovideo.com/c/product/746645-REG/Sescom_LN2MIC_ZOOMH4N_LN2MIC_ZOOMH4N_Line_to.html

No I don't think so. Beachtek seems under the impression you are concerned about plops in the camera, not their device.

You want them to address the issue of why switching the DC on the mic port and hence amplifier output port of their device causes a plop in its monitoring circuit.

The Beachtek response is correct in one respect. The DC blocking capacitor alone will not solve the problem. By installing a capacitor on the camera side of the network and terminating it thru say a 10k resistor, the capacitor will remain charged. Then, install another capacitor on the Beachtek side of the network. When the camera is started and the DC bias turned on, the voltage is shunted to ground thru the 10k resistor) and the capacitor will be effective. This is the way noiseless switching is achieved in analog equipment. I would suggest use of 50mf tantalum capacitors with the resistor and caps installed in a small metal enclosure such as a 1/4" phono plug shell or XLR adaptor and the shield on the cable used to fabricate this as well as the ground side of the resistor be tied to the shell to eliminate hum and noise. So you end up with two capacitors in series with the center tap connected to ground thru a resistor. Feel free to email me thru the forum for schematic and fabrication suggestions.

Kent

Switching on causes a transient which will pass through the capacitor. If you make the capacitor small to reduce the transient you will reduce the low frequency response. Actually, reducing the capacitance would turn a thump into a click.

EDIT

I think the only solution other than getting Beachtek to fix their box is to place a unity gain amplifier on each audio lead between the Beachtek and the Nikon. It will block the transient going backwards but allow the signal to go forwards.

If you can't buy something suitable, you could buy a dual operational amplifier integrated circuit (are they still available?) and apply full negative feedback around each to give unity gain. Add a series capacitor after op amp out for good measure, but may not be necessary. Then you would need a power supply (batteries?) and a metal box for shielding.

Come on Opampman, why didn't you suggest the op amp solution?

Actually, a pair of transistors in emitter follower configuration would also do the trick, but can you still buy discrete transistors?

Then there is the cathode follower.....

I don't fully understand, without a schematic, what may be going on, but Peter has hit the nail on the head: a transient will pass right through a blocking capacitor.

The solution to turn-on transients usually involves a switch of some sort which stays open until the transient has passed. I'm sure everyone is familiar with this on stereo amplifiers which typically mute the speakers until circuits have stabilized.

If the pop is really caused by instantly going from zero volts up to some constant DC voltage then, if a switch won't work, it seems to me that you'd instead want to simply put a small bypass capacitor from the input to ground with, perhaps, a small resistor in series. The resistor would have to be small compared to the input impedance, and the capacitor would have to be small enough so that it would roll off well above 20,000 Hz.

Another piece of information that would be useful is what sort of energy can be delivered from the source. I say this because another way to think about this problem -- instead of viewing it as a steady-state frequency response problem -- is to think about the energy transferred from the pulse. You want to pick an RC time constant that is considerably less than the transient time, but with a capacitive value that is large enough to store the energy which the source is capable of delivering during that transient.