AT 4060 schematic?

[DrFrankencopter]

Hi Guys,

This weekend I decided to take apart my AT4060 microphone...

I thought it was going to be a pretty simple circuit considering it's just a cardiod condensor with a tube amp, and has a transformer for output balancing. But, there seems to be lots going on inside this little wonder, tons of diodes, and a couple of 0.22F caps. Does anyone have a schematic for this sucker, or am I stuck reverse engineering it in order to figure out how it works?

Cheers,

Kris

 

[Gus]

I have not seen a 4060. I do have two 4033s and they have more diodes than I would expect. I have not traced the 4033s but some diodes seem to be used at the input for the electret capsule. Is the 4060 electret?

I think I read the 4060 is a 6dj8 based cascode. I don't understand why anyone would want a 6dj8 base tube microphone. Maybe the answer is the 6dj8 has a russian tube that does not cost alot.

 

[DrFrankencopter]

Hi Gus,

I don't think the 4060 is electret, but I suppose it IS possible. I can't see why they would take that approach when they have B+ voltage (somewhere around 100-125V from what I can guess so far) available to be used for capsule polarization. But yes, there's a stream of about 8 diodes right at the capsule input (4 on each side of the cap).

Here's a quote from Bill Balmer (whose father was an engineer at AT) off rec.audio.pro

The AT4047 and AT4050 have similarly constructed capsules. They are both
externally polarized with double diaphragm construction. The diameter of
the 4050 diaphragm is 1" and that of the 4047 - 1.1". They have no
acoustic delay network around the diaphragm to create the cardioid
pattern. That is done by electrostatic tensioning of the 2nd diaphragm.
Switching of that tensioning is how the patterns are changed on the 4050.
This construction makes the off-axis response more accurate as there is
no delay network to contend with. The delay networks work well, but it
is very difficult to get the delay linear across the frequency band of
the mic. So, at very low and very high frequencies, there is more
variation from the desired pattern shape.

The AT4033a has a 16mm diameter self-polarized (electret) diaphragm with
a delay network around the diaphragm that brings the overall diameter up
to 1". It's the delay network in the 4033 that creates the cardioid pattern
without a second diaphragm.

The tube is a 6922, so yup its a 6dJ8. It's a Sovtek tube, and allegedly selected for noise.

Why not use a 6DJ8 in a mic? It's a pretty common tube for hi-fi guys, at least in line stages....

Cheers,

Kris

 

[gyraf]

ECC88/6DJ8 is a frame-grid tube, and I wouldn't expect it to do well at very-very high impedances as the ones at a capsule converter stage. But if we're talking electrect capsules, there may very well be a FET capsule converter - and in that case the ECC88/6DJ8 makes (some kind of) sense..

JAkob E.

 

[Flatpicker]

[quote author="DrFrankencopter"]This weekend I decided to take apart my AT4060 microphone...[/quote]Why, was there a problem with it? Just curious...

 

[Gus]

Kris

I have built a test microphone that I tried different tubes in the 6dj8 just did not sound right(140V, 100k plate, cap bypassed cathode bias R, 1uf poly cap to a jensen transformer. It was a "smooth" sound but seemed to lack definition.

FWIW I have rode K2 that I changed the china 6dj8 type tube to a JAN6922 (6dj8 type) and it did sound better I then changed it to a 6072a and it seems to sound even better(had to cut two traces and add a jumper).

Now if the 6040 is a cascode the microphone might sound good maybe the miller cap of a 6dj8 is what I don't like with the tube as a "simple" triode gain stage.

I wonder if the diodes are acting kind of like high value resistors? Different diode construction have different leakage currents.

 

[DrFrankencopter]

Why, was there a problem with it? Just curious...

So was I :green:

Really, I just wanted to see what makes it tick, and if there might be some easy component changes (especially caps) to improve the sound of the mic.

I'll try asking AT for the schematic, otherwise, I'll do my best to reverse it for the folks here. I'd be interested in your collective take on the design. I was shocked to see those .22F caps in there, they're like computer memory backup type caps. FWIW, the electolytic caps in this mic are all Nichicon MUSE parts...pretty good quality stuff for mass production. Polypropylene caps look like Wimas (square box construction), and one looks more like a polyester. I have some photos, but would like to take it apart again to get some better shots.

Cheers,

Kris

 

[Gus]

the sony c800g schematic shows supercaps (.22F) in the fil supply. I had a few .022F suer caps that I added to the fil supply of my 800 type microphone build and it seemed to drop the noise down a little.

 

[Flatpicker]

[quote author="DrFrankencopter"]I'll try asking AT for the schematic...[/quote]Good luck. I recently called them about my 4050 - in cardioid position it works fine for a short while and then fades down to a lower level as if the -10dB attenuation was turned on (but of course it isn't) and remains consistent at that level. Fig 8 and omni work fine.

I called and spoke to a repair tech who told me it was probably the switch and that tuner cleaner sometimes works, but if not I would have to send it in. I ask for the approx. cost but was only told that "it would be under $100". The switches are made by Alps but look really cheap - the switches on my MXL are MUCH better! Anyway, the cleaner didn't fix it, so I'll probably just send it to them. It really puzzles me though, how a bad switch would cause something like this. I ask about just getting a new switch which is thru-hole and would be no trouble to replace, but they didn't want to send me one.

 

[DrFrankencopter]

AT just sent me the schematic and bill of materials....if someone's willing to host it, PM me and I'll send it their way.

Cheers,

Kris

 

[Flatpicker]

Wow! :shock:

Who did you talk to at AT?

 

[rodabod]

Here we go, Kris.

Hopefully this should work.
Large BMP images:
at4060part1
at4060part2

Large GIF images (smaller file size), alternate host:
AT 4060 part 1
AT 4060 part 2

 

[Gus]

can't read all the numbers on the schematic page.

It looks like a cathode follower with an active load in the cathode circuit. Kind of like some gefell,oktava and neumann solid state transformerless circuits.

Is c4 a 220uf?

c5 a 470uf?

The two diodes in the top tubes cathode circuit look like bias diodes.

120V is kind of low for a 6922 but this is more of a follower circuit

I would guess the transformer is 2:1 to 4:1

Not sure what is going on with the input yet, need to look up the diode specs and leakage currents.

 

[PRR]

> It looks like a cathode follower with an actice load in the cathode circuit.

It is a White Cathode Follower.

WCF gives half the output impedance of a simple cathode follower, and much higher output in low-Z loads.

Without an analysis, I'd guess the output impedance could be 200 ohms.

It is odd in that the top tube is biased with diodes in the cathode, the bottom tube is zero-biased. That may be misunderstanding of tubes, or it may be a very careful attempt to put tube-sound in.

The grid diodes leak, and act as very-high resistors. The AT3035 seems to do the same with an FET.

The top plate resistor "can be shown" to be way too large, but this probably does not matter at mike levels.

 

[Gus]

looking at the circuit some more.

A few things I noted
the cathode bias diodes does not look like it sounds good I have tried bias diode circuits in microphones. When I use bias diodes I do something different.

Part of what a tube circuit lets you do is to use film caps to the transformer this one looks like it has electros because it is a follower.

Does the input section use diodes because they are cheaper than high value resistors?


Maybe why people like this microphone(I have not heard one) is the capsule and grill work together well?

I don't understand what the design is about. Maybe like PRR noted the way the bottom tube is used might have something to do with the sound.

 

[DrFrankencopter]

C4 is a Nichicon 220uFd. Cathode bypass I presume...

C5 is a Nichicon 47uFd. Output coupling I also presume....

Output impedance...per the schematic is 200 ohm+/- 30%

So, capsule polarization is happening through diode leakage? Wow....and Huh, isn't that a strange way of doing it?

What are the diodes from the cathode to the grid of the 1st side of the tube doing? I figure the first diodes in the chain, (the ones on the left of the schematic) set the capsule polarization through their leakage, but I just don't see what the ones to the right of the 2 coupling caps do (the ones connected from the cathode to the grid).

Cheers,

Kris

 

[Gus]

Kris

The left side 4 diodes have a connection to the caps C02 .1uf? C02 allows feedback from the node at the bottoms tube plate to increase the effective input Z from the diodes

Also the next 4 from the left diodes get feedback from the bottom tube plate node as well raise the effective Z

This is a form of bootstrapping there is a section in the big red book about it.

Some AKG CF tube schematics have this type of cathode to grid feedback with a resistor IIRC

The pot and two resistors on the left set the capsule voltage

The more I look at this circuit and think about it the more I don't like the bottom tubes grounded cathode the tube current will change with different tubes and how do you build a microphone that you have control of the series current thought the tubes and plate r and bias diodes. Will this only work with rusian tubes?

How is the current settup leakage bias?

The more current the greater the voltage drop across the plate R of the top tube and the greater the voltage supplied to the bottoms tube grid.

The other thing that I don't understand is the cathode bias diodes seem to be the same PN as the grid leakage diodes and clamps(if I understand this circuit correctly)the bias voltage drop has to be very close to two of the diodes in the grid circuit to start to conduct?????????????????????

The output Z must be low because there is an added 56 ohm resistor

Maybe the circuit was designed more as an effect than a clean CF circuit??????? 120V, diode bias vs input grid diode question, grounded cathode bottom tube

Thanks for the schematic it has been entertaining

 

[PRR]

> Is the 4060 electret?

The schematic we have shows a bias voltage to the capsule.

> supercaps (.22F) in the fil supply

Here too. 2x0.22F, effective 0.11F. About 1.5 second time constant against the heater resistance. Since this is less than the heater thermal time constant, it's a puzzle.

> use diodes because they are cheaper than high value resistors?

Can you even get 100Meg in surface-mount?

SMT diodes sure are low-price.

Because the diodes are back-back, some of the sins cancel. I hope bcarso sees this and elucidates zero-bias diode leakage.

> The more current the greater the voltage drop across the plate R of the top tube and the greater the voltage supplied to the bottoms tube grid.

Not for DC analysis. The bottom tube grid stays at zero volts.

You may be right for signal conditions, but the WCF is a paradoxical plan and I'm not ready to commit.

> the cathode bias diodes seem to be the same PN as the grid leakage diodes and clamps(if I understand this circuit correctly)the bias voltage drop has to be very close to two of the diodes in the grid circuit to start to conduct ?????????????????????

No, it conducts. These are triodes. Plate voltage has an effect on current. Only 1/Mu or 1/33 of the effect of the grid, but with low/no grid voltage and high plate supply voltage the plate voltage has significant effect.

I asked the PC for an answer, but let's see how I would rough it out in the martini-bar (or saki-joint).

The bottom, zero-bias, triode acts (for DC) a lot like a 3K resistor.

The top tube passes the same current, but has -1.2V bias to the grid. The plate-cathode voltage must be 1.2V*33= 40V higher than the bottom tube.

We start from 120V supply. Top tube has 40V excess drop, we have 80V left to split across two 3K tubes and one 22K resistor. 3K+3K+22K is 28K. 80V/28K= 2.8mA. The bottom tube plate drop is 3K*2.8mA= 8.4V. We have 1.2V of diodes. The top tube drops the same as the bottom plus 40V, 48V. 8.4V+1.2V+48V= 58V at top tube plate.

SPICE's lie is 12.8V at bottom plate, 71.8V at top plate, 2.2mA current. I actually like that better than my napkin analysis, which neglects the steepening of plate curves at very low voltages. I should not have extrapolated tube characteristics taken at 50V-100V down to a predicted 8V. However it was pretty close, and a little peek at curves would suggest that "3K" might be more like 4K or 5K as we get down to 12V and 8V.

 

[PRR]

AT 4060

Sensitivity: 20mV at 94dB SPL
Max level: 150dB SPL for 1% THD
150-94= 56dB= 631:1
631*20mV= 12.6V!!!

While other AT-40xx-series mikes list "-10dB pad" in their specs, this one does not?

And there is no pad on the schematic.

What am I missing? How can you even use a mike with 12V output? Is there a pad on the power box?

Even if it is padded at the box... ~18v peak is a big signal voltage for a 110V bias. Condenser transducer is square-law, so 18V peak on 110V bias should be over 2% at the capsule.

Oh, and the 12V is specified at the output. I've only assumed a 1:1 transformer. If the tranny is 3:1, that's 44V peak at the capsule!

Looking at another angle: the standing current in the tube is about 2mA. We could maybe make 4mA peak (but not with these values). ASSuming that 1:1 transformer, 18V peak and 4mA peak implies 4.5K load impedance. Assuming a 3:1 transformer, we could get 12mA in the load, and 18V/12mA= 1.5K load.

OK, many mike preamps are 2K impedance. But 300 feet of cable will be ~1K at the top of the audio band, we may have to drive that. And with this whopping output we may want a pad: pads are not high impedance, sometimes much less than 2K.

 

[DrFrankencopter]

A few things to note:

1. The AT4060 contains no surface mount parts...it's all hole-thru construction. The 8 diodes that connect to the capsule are mounted off the PCB bridging from the capsule to one of the capacitors.

2. There is no pad available on the power box...

3. 150dB is ridiculously loud....perhaps the 4060 was designed to record space shuttle launches!

Next time I have it apart perhaps I can do some transformer measurements.

Thanks for the discussion guys...keep it up!

Cheers,

Kris