OPEN SOURCE DIY Mic Project - ORS 87 - Stripped Down u87

[Wordsushi]

FWIW, though I didn't write it down, when I measured the resistance of each side of the T-13, because the wires are not marked, I noticed the DC resistance was around 440/20, very very similar to what the AMI literature says.

You see so many of these boutique companies offering very similar T-style tranformers and now, I'm just thinking they're all made in the same couple of factories and the boutiques put their own wrapper on them. It's not like everyone's making their own.
It seems to be about the same exact size as the 3u, and given that Guosheng doesn't seem to push a narrative that they are also making original transformers alongside their capsules, I'm inclined to believe it's possible he's outsourcing them.

It's also possible that these T-13's on Aliexpress are B-stock, I suppose.

 

[micolas]

FWIW, though I didn't write it down, when I measured the resistance of each side of the T-13, because the wires are not marked, I noticed the DC resistance was around 440/20, very very similar to what the AMI literature says.

You see so many of these boutique companies offering very similar T-style tranformers and now, I'm just thinking they're all made in the same couple of factories and the boutiques put their own wrapper on them. It's not like everyone's making their own.
It seems to be about the same exact size as the 3u, and given that Guosheng doesn't seem to push a narrative that they are also making original transformers alongside their capsules, I'm inclined to believe it's possible he's outsourcing them.

It's also possible that these T-13's on Aliexpress are B-stock, I suppose.

I had the same suspicion, a comparison of 3u gzt-87 with Ali T13
(and AMI T13 ) would provide valuable information

 

[Wordsushi]

Also worth noting, if you do have problems with overloading the FET, you could use a 33V zener to increase the FET headroom relative to the capsule voltage. You should have plenty of headroom for all but the loudest sources.

Question. I built one with a 33v zener (47k resistor) and the levels on the mic were significantly lower than the same build with a 24v/56k so I thought I borked it somehow, but I woke up thinking of this comment in the thread, so is this the 33v zener increasing the headroom? What exactly does using the 33v zener do to the circuit?

 

[OneRoomStudio]

Question. I built one with a 33v zener (47k resistor) and the levels on the mic were significantly lower than the same build with a 24v/56k so I thought I borked it somehow, but I woke up thinking of this comment in the thread, so is this the 33v zener increasing the headroom? What exactly does using the 33v zener do to the circuit?

Did you re-bias the FET after the change? You have to.

The 33V zener changes the FET "B+" from 24V to 33V. This gives the FET more "swing" capability, which increases the headroom (not by a ton, but should be noticeable). The amount of gain is a function of the FET used and how it's biased, along with the B+. If you haven't already, try biasing your FET for the new voltage, and you should be all set.


If you still have less gain, I would check the voltages to make sure nothing else is wrong, and then I would suspect the FET as not having as much gain (could be bad or out of spec).

 

[Khron]

Regardless of schematic used, the ACTUAL supply voltage that ends up AT the JFET is what you measure between ground and the "top" of the drain resistor (and nowhere "before" that).

For example in the U87A, after the 33v zener, the supply rail still goes through a 10k resistor, which drops another 5v (let's assume a 0.5mA quiescent current through the JFET), so the supply ends up as actually "only" 28v or so. That current also depends on the source resistor though, of course...

Side note: Couldn't tell you why Neumann chose that particular supply arrangement, by the way...

 

[Wordsushi]

Did you re-bias the FET after the change? You have to.

The 33V zener changes the FET "B+" from 24V to 33V. This gives the FET more "swing" capability, which increases the headroom (not by a ton, but should be noticeable). The amount of gain is a function of the FET used and how it's biased, along with the B+. If you haven't already, try biasing your FET for the new voltage, and you should be all set.


If you still have less gain, I would check the voltages to make sure nothing else is wrong, and then I would suspect the FET as not having as much gain (could be bad or out of spec).

Thanks guys! This was a new build so I biased it based on it. Got 10v at the source to start and otherwise the biasing seemed to be normal, tried adjusting the bias with no effect on the level, so I have a feeling a FET swap may be in order.

I'll check the drain resistor again, too.

 

[Khron]

Thanks guys! This was a new build so I biased it based on it. Got 10v at the source to start and otherwise the biasing seemed to be normal, tried adjusting the bias with no effect on the level, so I have a feeling a FET swap may be in order.

I'll check the drain resistor again, too.

Ok, 10v on the drain, and what about at the other end of the drain resistor?

Side note: Couldn't tell you why Neumann chose that particular supply arrangement, by the way...

Only by rearranging some components and changing a zener value, drain biased at half-supply, at least in simulation i'm getting 3.5Vrms @ 0.7% THD versus 3Vrms @ 1% THD stock...

 

[OneRoomStudio]

10-11V on the drain should be about right using a 24V zener (actual voltage before the drain resistor is ~22V):


With a 33V zener, the actual voltage is just below 29V, and you should have closer to 14.5V on the drain:

 

[soliloqueen]

Ok, 10v on the drain, and what about at the other end of the drain resistor?



Only by rearranging some components and changing a zener value, drain biased at half-supply, at least in simulation i'm getting 3.5Vrms @ 0.7% THD versus 3Vrms @ 1% THD stock...

Not only did they choose this bizarre supply arrangement but they even went as far as to choose a fet where only like one in five units could function under these conditions. They did try to replace it several times but people complained

 

[LevinGuitar]

Not only did they choose this bizarre supply arrangement but they even went as far as to choose a fet where only like one in five units could function under these conditions. They did try to replace it several times but people complained

My god! And I thought I was so useless because I couldn't make them all work. Afraid because, as I have read several times, the body static can burn out a fet and so...

 

[abbey road d enfer]

Not only did they choose this bizarre supply arrangement but they even went as far as to choose a fet where only like one in five units could function under these conditions.

That's what happens when using a common source stage with an FET which characteristics are all over the place, like it is when it comes from the foundry.
All FET's need sorting/grading.
The only way to avoid it is having a feedback-controlled self-biasing circuit.

 

[soliloqueen]

My god! And I thought I was so useless because I couldn't make them all work. Afraid because, as I have read several times, the body static can burn out a fet and so...

Only one in several 2819 fets will work here. The rest don't have the right power specs and will cut out at high levels or not work at all

 

[soliloqueen]

My god! And I thought I was so useless because I couldn't make them all work. Afraid because, as I have read several times, the body static can burn out a fet and so...

No, this is common and expected with the layout of the circuit. You always have to be very selective with fets if the circuit doesn't automatically manage their operation. That's why it's best to buy a crap ton of them

 

[OneRoomStudio]

Only one in several 2819 fets will work here. The rest don't have the right power specs and will cut out at high levels or not work at all

Just so I understand your meaning, are you saying only a handful of 2N3819 FETs will work with the stock source resistor, or only a few will be able to be biased up at all with a 47K drain resistor and 22-29V B+?

 

[soliloqueen]

Just so I understand your meaning, are you saying only a handful of 2N3819 FETs will work with the stock source resistor, or only a few will be able to be biased up at all with a 47K drain resistor and 22-29V B+?

with the stock setup. no experience with the modified setup

 

[Wordsushi]

The 33v/47k combo works great. Signal problem was user error, a solder blob that shorted c6 to a resistor. Once that was fixed and I re-biased it to 14.5v, it sounds good.

 

[OneRoomStudio]

The 33v/47k combo works great. Signal problem was user error, a solder blob that shorted c6 to a resistor. Once that was fixed and I re-biased it to 14.5v, it sounds good.

Good to hear. It’s been my experience that most 2N3819’s can be used, as long as the source resistor is adjusted for the individual FET. Of the dozens I’ve tried in mic circuits over the years, I only had one actual “bad” one.

 

[Wordsushi]

Here's one.
Is a .5w 33v zener robust enough?
Should I order 1w?

 

[kingkorg]

Here's one.
Is a .5w 33v zener robust enough?
Should I order 1w?

.5 is plenty.

 

[OneRoomStudio]

Here's one.
Is a .5w 33v zener robust enough?
Should I order 1w?

The FET is pulling less than a half a milliamp of current, which means at this voltage the zener is dissipating less than 20mW - so .5W (500mW) is plenty!