开云体育

@Scott:

Even ignoring the Sony 7K complication, I don't think your diagram is correct.
The swing seen at the drain ("to audio input") would be some 20 dB lower than the swing seen at the source, basically wasting all the current gain (transconductance) of the JFET on the 33K to generate a voltage drop between source and ground that the bodypack would not see. The bodypack only sees the voltage drop between 5V and drain.
The bias would be very weird as well, as most of the voltage drop would be across the 33K and the audio input pin would sit really close to 5V.
I think either the connections are different from your depiction, or the assumption that the plastic package is a JFET is incorrect and even the non-CORE DPAs have something less conventional than a single JFET.

My misunderstanding of what's going on is also always a possibility.

If you agree that it can't work as depicted, is there any chance you could measure the U(I) curve? There could be some clues there. You can use and identical circuit and just vary the 5V between 0 and 5V and record this voltage (VSRC), as well as the voltage at "to audio input" (VD) in a spreadsheet. The voltage of the mic is VD and the current is (VSRC-VD)/3K6.

Hi Sergio,

I think I included the word "assume" somewhere when talking about the
FET-shaped 3-pin package in that mic. It's entirely possible there's
more in that package, especially since there isn't any other obvious
provision for adjusting the bias on the gate side.? But these mics do
put out a fairly strong signal, and this whole electret lavalier stuff
is definitely not new tech - most all such microphones put out a similar
signal strength.? And the Shure pinout includes a 20k resistor to ground
on one of the pins, like the DPA capsule might higher than the resistor
that adds the DC bias voltage.

Not sure when I'll get to it, but I do believe I have the means* to
record a voltage vs. current curve and it might be fun to test that out.

Thanks,
-Scott

*I have one of those Mooshimeter things, which sadly is no longer
supported.? Anybody else have one?

On 6/25/24 02:36, sergio_logic via wrote:
> @Scott:
>
> Even ignoring the Sony 7K complication, I don't think your diagram is
> correct.
> The swing seen at the drain ("to audio input") would be some 20 dB
> lower than the swing seen at the source, basically wasting all the
> current gain (transconductance) of the JFET on the 33K to generate a
> voltage drop between source and ground that the bodypack would not
> see. The bodypack only sees the voltage drop between 5V and drain.
> The bias would be very weird as well, as most of the voltage drop
> would be across the 33K and the audio input pin would sit really close
> to 5V.
> I think either the connections are different from your depiction, or
> the assumption that the plastic package is a JFET is incorrect and
> even the non-CORE DPAs have something less conventional than a single
> JFET.
>
> My misunderstanding of what's going on is also always a possibility.
>
> If you agree that it can't work as depicted, is there any chance you
> could measure the U(I) curve? There could be some clues there. You can
> use and identical circuit and just vary the 5V between 0 and 5V and
> record this voltage (VSRC), as well as the voltage at "to audio input"
> (VD) in a spreadsheet. The voltage of the mic is VD and the current is
> (VSRC-VD)/3K6.
>
>
>

--
---- Scott Helmke ---- scott@... ---- (734) 604-9340 ----
"I have ceased distinguishing between the religious and the secular,
for everything is holy"? - Joe Henry

The reason I am puzzled by the 33k on the DPA is not the value, but where it's connected.
Summarized by Sanken here <>, there are two topologies for the JFET

1. common drain AKA source follower AKA "3 wire method" (Shure uses this)
2. common source AKA "2 wire method"

In both topologies, the resistor is connected at the node where the audio signal is coupled because it's the load resistor of the amplifier.
This is why it would be "abnormal" for the DPA, in a "2 wire" connection, to have a 33k resistor between source and ground while coupling the audio signal from the drain: Sanken shows a dead short there.
"Abnormal" assuming it's just a JFET, of course.

FWIW I've connected WL185s in the 2-wire topology (black wire to shield, red wire as bias and audio) to a TRS jack, and they work fine with consumer grade mic inputs (Rode Wireless GO, Zoom H1n, etc.). In fact, Shure presents <>this solution for use with Sennheiser bodypacks. As long as it's just a JFET with reasonably low Idss, either way works.

I think I've figured it out: pre-CORE it was probably a topology like the one from the Audio-Technica patent <>(which would be later than DPA doing this).


Why do I say this? I've found old manuals <>, years before 2017 (when CORE was introduced), listing the output impedance as 30-40 ohms (JFET topologies have output impedances of 1000 ohms or more).
So this was at least an emitter follower even then, and 33K makes sense as a value for R1!

In this topology, VDS is "regulated" by the VBE of Q1, which means that for large input SPL generating e.g. ~2V pk-pk at Q1 base, it will still vary by 40-50mV.
In what I suspect is the CORE topology, VDS is regulated by feedback from the error amplifier and under the same input conditions will vary by 4-5mV, extending the linear range and improving the max SPL at 1% THD spec.

Are you in the USA? ?I'd be happy to mail that capsule to you for
further inspection.

Thanks for the kind offer, I live in Europe and I've already ordered one from the UK a couple of days ago, listed on eBay as defective with "thin tinny sound" for the equivalent of $38, shipping included. I don't know if it's CORE or pre-CORE, I guess I'll find out when I do the teardown.
Like I said, I own a 4080 CORE bought new from Thomann a couple of years ago, but I won't go past measuring the U(I) curve in my investigations on that one.

So here is the full non-CORE DPA 406x teardown.