Aurycle Tube Mic qs (retitled)

[midwayfair]

Working on one of these:

http://www.aurycle.com/manual/a5500mp-schematic_v2.jpg

First thing is I'm interested in removing the input cap.
EDIT: Removing the input cap can't be done here without significant modifications, scrap this idea.

They don't have a schematic for the power supply, but I believe pin 6 is the half (or variable) voltage point for the polar patterns, right? (Because pin3 has to be the high voltage for the tube anode.)

If I'm analogizing this to the FET mics I've built, I would move the R10 connection to the rear capsule instead of the backplate and disconnect the rear capsule connection from ground. But I was wondering if it might be easier to instead connect the red wire currently going to the rear capsule directly to a pin2 connection (like the hole left by leaving off C4).

My other question was C8 and C6. C6 looks like overkill, but then again, I haven't build microphone stuff with tubes, only guitar amps. I read somewhere (don't remember where -- maybe on the Fox audio page?) that the mic can be a little rumbly. I'm most likely replacing the transformer with a lower ratio (a 1:8 cinemag that I have lying around) and was initially going to bump up C8, but now I'm wondering if I ought to decrease it to .47uF instead or just leave it alone.

My last change is to use a 12AY7 instead, to bring the output impedance back down. (The 12AX7 would have been a 600R output impedance with the 8:1. I think the 12AY7 is probably the right call even with the 9.5:1 that came with the kit.) Is there anything else I need to change to accommodate this? I know the ELA251 build I have has a 1.8K on the cathode, and that's a 12AY7.

 

[abbey road d enfer]

Working on one of these:

It would be nice if you rescaled the picture so that it comes readable. Just put it in an attachment so one can open it in their favorite app.

I would move the R10 connection to the rear capsule instead of the backplate and disconnect the rear capsule connection from ground. But I was wondering if it might be easier to instead connect the red wire currently going to the rear capsule directly to a pin2 connection (like the hole left by leaving off C4).

Could you post a schemo instead of ambiguous descriptions? A picture is worth a thousand words.

C6 looks like overkill,

Against what metric?

that the mic can be a little rumbly.

It can be caused by several different factors. First I would evaluate the veracity of this claim, then apply remedies, if needed.

I'm most likely replacing the transformer with a lower ratio (a 1:8 cinemag that I have lying around) and was initially going to bump up C8, but now I'm wondering if I ought to decrease it to .47uF instead or just leave it alone.

Don't apply treatment before knowing if there's an actual disease. Actually, increasing the coupling cap value can result in less "rumble".

My last change is to use a 12AY7 instead, to bring the output impedance back down. (The 12AX7 would have been a 600R output impedance with the 8:1. I think the 12AY7 is probably the right call even with the 9.5:1 that came with the kit.)

Since Rp=25k, the resulting output Z would be ca. 300r.

[/quote] Is there anything else I need to change to accommodate this?
[/quote] You would probably have to change the anode and cathode resistors. I would change R2 to about 50k and experiment with R7 for best operational compromise. Higher bias current gives better headroom but may impair noise. Now real optimization depends on what mic preamp it is connected to.

 

[midwayfair]

It would be nice if you rescaled the picture so that it comes readable. Just put it in an attachment so one can open it in their favorite app.

Sorry, other forums using this software usually resize the pictures automatically to 800W.

Could you post a schemo instead of ambiguous descriptions? A picture is worth a thousand words.

Edit: The schem I posted won't work without a negative voltage supply for pin6. I've removed it. (Link for posterity.)

(Also attached to the post.)

https://www.dropbox.com/s/n4j2hpdu9fpnm55/capsule%20wiring.png

Against what metric?

Well, usually the bypass cap is 22uF. I've seen a couple amps with 100uF. The cutoff frequency is 7Hz if I use a 22uF and a small bias resistor of 1K (it's 4 Hz if I use the typical 1k8). It's 1.5Hz if I use the 100uF, so it seems like using the larger cap isn't really necessary here. Increasing the size of the bypass cap also, IIRC, does something to the speed of the cap and affects the highs, which is probably why they've also bypassed it with a film cap, though with the stock capsule it might have been better to leave that out.

It can be caused by several different factors. First I would evaluate the veracity of this claim, then apply remedies, if needed.
Don't apply treatment before knowing if there's an actual disease. Actually, increasing the coupling cap value can result in less "rumble".

Well, I was planning on wiring it up stock first. But the reason I ask questions is because I don't haven't built one of these from scratch, and I would never have guessed that a potential remedy would be increasing the coupling caps like your last sentence ^ so I learned something there. (Why would that be the case?)

Since Rp=25k, the resulting output Z would be ca. 300r.

Just wondering because I used an online calculator that gave me 20K: Where did the 25K come from? I'm not doubting you, just trying to understand where I should be getting that number from.

You would probably have to change the anode and cathode resistors. I would change R2 to about 50k and experiment with R7 for best operational compromise. Higher bias current gives better headroom but may impair noise. Now real optimization depends on what mic preamp it is connected to.

I'll try these. I suppose I can put in a trimpot to find the cathode bias temporarily. I may also revisit the thread on here about parallel triodes in a 12AY7 mic, I remember the actual results being somewhat ambiguous last time I read it.

 

[abbey road d enfer]

R1 cannot be 22Meg. Too noisy. You don't need a 1Gig to feed the left diaphragm, but it must be decoupled to ground with another 22nF cap.

Well, usually the bypass cap is 22uF. I've seen a couple amps with 100uF. The cutoff frequency is 7Hz if I use a 22uF and a small bias resistor of 1K (it's 4 Hz if I use the typical 1k8). It's 1.5Hz if I use the 100uF, so it seems like using the larger cap isn't really necessary here.

22uF instead of 100 would still result in infrasonic cut-off so there's not much operational difference. It matters in guitar amps because the 1st stages have to deal with very strong LF peaks due to plectrum attack. No such thing here.
Basically, my comment was don't sweat about it.

Increasing the size of the bypass cap also, IIRC, does something to the speed of the cap and affects the highs,

Typical audio myth.

I would never have guessed that a potential remedy would be increasing the coupling caps like your last sentence ^ so I learned something there. (Why would that be the case?)

Sometimes, the resonance created by the interaction of the coupling cap and the primary's inductance falls in the lower octave. Increasing teh cap shifts the resonance towards infrasonics and generally is more damped.

Just wondering because I used an online calculator that gave me 20K: Where did the 25K come from?

Tung-Sol datasheet.

I'm not doubting you, just trying to understand where I should be getting that number from.

Rout is Rp in parallels with Ra. 25k in // with 100k is 20k. It's that value I took in the calculation. 20k/(8)²=312.5 but there are losses and parasitics, so I would approximate at 350.

I'll try these. I suppose I can put in a trimpot to find the cathode bias temporarily.

I would recommend that. Beware that noise increases with gain, so you must assess both output level and noise level.

 

[midwayfair]

R1 R6 cannot be 22Meg. Too noisy.

Oh woops, I copy and pasted the resistor. It's 200M in their schem (also R6). Thanks for the catch and I updated the picture. Does it need to be as high as possible? The ELA251 has a 30M in that spot, some other mics have a 1G, and I'm not sure if it's applicable given that it's a completely different tube, but the u47 has a 60M. I have 22M, 100M, 200M, and 1G lying around.

You don't need a 1Gig to feed the left diaphragm, but it must be decoupled to ground with another 22nF cap.

Ah, I knew I would forget something! Thanks!

 

[gyraf]

..you probably don't want high DC voltage on your front membrane, this attracts lots of environmental dirt and dust.. Also, rejection of outside electrostatic noises gets weakened when going this route.

What is your reason for wanting to get rid of that input capacitor?

/Jakob E.

 

[midwayfair]

..you probably don't want high DC voltage on your front membrane, this attracts lots of environmental dirt and dust.. Also, rejection of outside electrostatic noises gets weakened when going this route.

In the aurycle schem, the variable voltage is on the front diaphragm, so in some patterns that's going to be polarized (even if only partially when in cardioid). I put the high voltage on the rear diaphragm in my diagram. I don't see how I can get around this though I can see the argument for grounding the front capsule and putting the high voltage on the backplate.

What is your reason for wanting to get rid of that input capacitor?

"It seemed like a good idea at the time."

Honestly just minimizing the parts connected to the capsule, but also because the U47 didn't have one. (The U47 also didn't have Figure 8 though)

I'm getting a sense that maybe I should be leaving well enough alone on a lot of this schem ... but there are arguments floating around the forum about the virtues of removing the input cap and for not polarizing the membranes, but I can't do both.

 

[midwayfair]

Built it up stock and it's mostly fine, except that the supply voltage is messed up. I'm seeing 120V at the junction of R4 and R5, and 0 to a whopping 220V on pin6! It's way more than enough to collapse the back membrane (I'm honestly surprised both aren't collapsing).

Unfortunately I can't find a schematic for the power supply, so I'm a little nervous about poking around with the parts in it, but I suppose I could find a convenient spot with a mutual connection between the pattern pot and pin3 and add a voltage divider or zener regulator or something.

EDIT: The voltages match the instructions:
http://www.diymic.com/education/tubeInstructions.html

But I didn't know there was any capsule that could handle 120V, I though 80V was about the max. This is crazy, right?

Edit #2: I think if I halved R3 and put a resistor equal to R9 in parallel with C5 I could cut both voltages in half and get to 0/60/120, and all of that can be done inside the mic instead of in the PSU.

 

[RuudNL]

In order to realize a figure-8 characteristic, you need to have a 'positive' voltage on one diaphragm, and a 'negative' voltage on the other diaphragm.
But since there is only one positive voltage, they realized this in the original design to put a 'half way' voltage on the backplate.
Because the backplate is on a positive voltage, you will need a capacitor to separate this DC voltage from the grid.
Now, seen from the perspective of the capsule, the grounded diaphragm is the 'negative' voltage' (=less positive than the backplate) and the full voltage is the 'positive voltage'.
By varying the voltage on the rear diaphragm, you can get omni (0 volts on the rear), cardioid ('half way' voltage on the rear) or figure-8 (full voltage on the rear).

What you are suggesting now is full (dangerous!) voltage on the front and variable (also dangerous!) voltage on the rear.
This means: cardioid to omni, but never figure-8.

 

[midwayfair]

In order to realize a figure-8 characteristic, you need to have a 'positive' voltage on one diaphragm, and a 'negative' voltage on the other diaphragm.
But since there is only one positive voltage, they realized this in the original design to put a 'half way' voltage on the backplate.
Because the backplate is on a positive voltage, you will need a capacitor to separate this DC voltage from the grid.
Now, seen from the perspective of the capsule, the grounded diaphragm is the 'negative' voltage' (=less positive than the backplate) and the full voltage is the 'positive voltage'.
By varying the voltage on the rear diaphragm, you can get omni (0 volts on the rear), cardioid ('half way' voltage on the rear) or figure-8 (full voltage on the rear).

What you are suggesting now is full (dangerous!) voltage on the front and variable (also dangerous!) voltage on the rear.
This means: cardioid to omni, but never figure-8.

Yeah, I realized this last night when I was actually soldering it up and should have edited my first couple posts (which I'll go do so no one else tries to follow my bad example). My FET mics have a negative voltage, that's why I was able to ground the backplate. It's just been a while since I did those and I forgot the details.

But the mic, as designed, puts 120V on both capsules (0-120V between the rear and the backplate, 120-240V on the front). This is definitely way more than I've seen on ... well, basically anything. The Apex 460 has Zener diodes in the power supply to knock the B+ (pin3) down to 200V (less once the tube is on, I think 160V or something like that) and then the inside of the mic is basically the same schematic I posted ... in fact, I'm not even sure they redrew the 460's schematic, because there are mistakes on the Aurycle board compared to their schematic. R3 on the schem here is actually R4 on the board.

 

[RuudNL]

A microphone that puts 120/240 volts on the capsule is poorly designed in my opinion.
(Unless there is a very special capsule design, which I doubt.)

 

[midwayfair]

A microphone that puts 120/240 volts on the capsule is poorly designed in my opinion.

Lol, I don't disagree. I wasn't really expecting the design to be great on a kit that's under $200.

I adjusted the resistors for the polarization voltages and they're now sitting at 0, ~60, and 0~120.

 

[RuudNL]

The problem with such high polarization voltages (apart from the risk of membranes collapsing), is that the diaphragms get attracted to the backplate. This can change the behaviour of the capsule.
In general I never go higher than 60 Volts, most of the time even a bit lower.

By the way: Rode also uses a pretty ridiculous high polarization voltage in their NT1 microphones.
This will of course improve the signal to noise ratio, because the capsule is delivering a higher output signal.
(And it is probaly also the reason that so many SMD capacitors at the end of the voltage multiplier chain fail...)

 

[Khron]

I've had to replace one in a friend's NT5, even.

By the way: Rode also uses a pretty ridiculous high polarization voltage in their NT1 microphones.
This will of course improve the signal to noise ratio, because the capsule is delivering a higher output signal.
(And it is probaly also the reason that so many SMD capacitors at the end of the voltage multiplier chain fail...)

 

[kingkorg]

The capsule wouldn't have to be that special, all you need is a thicker spacer between the membrane and backplate, which might be the case here.

Measure the capsule capacitance, if it's atypical that might be giveaway.

I did this once with a k67 and 120v polarisation. It gave me very different response from the k67 capsule, but still good snr.

A microphone that puts 120/240 volts on the capsule is poorly designed in my opinion.
(Unless there is a very special capsule design, which I doubt.)

 

[abbey road d enfer]

A microphone that puts 120/240 volts on the capsule is poorly designed in my opinion.
(Unless there is a very special capsule design, which I doubt.)

So B&K 200v mics are poorly designed? 

 

[RuudNL]

That's not what I was saying!
"(Unless there is a very special capsule design, which I doubt.)"

I think we agree that it is not wise to put 120 volts on a generic Chinese microphone capsule, isn't it?