Dead Nemann U87i

[Angus Kerr]

Hi, my single FET U87 is dead. No output at all.

Any suggestion on common failure modes? Where do I start? Typical healthy voltages at various test points?

Appreciate some guidance. I've got to do vocals soon. I'd prefer the U87 over my only other option, Oktava MK219.

Angus

[geoff]

On 20/09/2015 5:02 p.m., Angus Kerr wrote:
> Hi, my single FET U87 is dead. No output at all.
>
> Any suggestion on common failure modes? Where do I start? Typical healthy voltages at various test points?
>
> Appreciate some guidance. I've got to do vocals soon. I'd prefer the U87 over my only other option, Oktava MK219.
>
> Angus
>


Probably capsule has expired (worst case). Or a wire off connector
(better). Or a dead aluminium electrolytic, or where they tantalums ?
(not so bad).

I loath my U87 on vocals, but if iit suit a particular voice or you aint
got an alternative I guess that's moot.

Sorry to hear your nation in mourning today, hope to watch a replay in
90 minutes here (NZ) ;-)

geoff

[Phil Allison[_4_]]

Angus Kerr wrote:
>
>
> Hi, my single FET U87 is dead. No output at all.
>
> Any suggestion on common failure modes?
> Where do I start?
> Typical healthy voltages at various test points?
>
> Appreciate some guidance.


** Here is a schem:

http://recordinghacks.com/images/mic_extras/neumann/U87-schematic.png

I assume you have checked the 48v phantom from the desk with a DMM, tried a new lead and checked all switch positions on the mic.

With the mic disconnected, across pins 2 and 3 you should see a reading of about 10 ohms. If it is much higher like 4400 ohms, the internal transformer is not connected or has failed.

With the mic plugged into the desk connected and open, find zener diode Gr1 on the PCB and see if there is +33V there compared to ground.

Proceed with great care, using headphones to monitor progress.


.... Phil





.... Phil

[[email protected]]

On Sunday, September 20, 2015 at 7:02:37 AM UTC+2, Angus Kerr wrote:
> Hi, my single FET U87 is dead. No output at all.
>
> Any suggestion on common failure modes? Where do I start? Typical healthy voltages at various test points?
>
> Appreciate some guidance. I've got to do vocals soon. I'd prefer the U87 over my only other option, Oktava MK219.
>
> Angus

Thanks Phil

First of all, thanks for the schematic. All the ones I saw were different with a 24V Zener and not having certain parts that mine does have. Mine is a gunmetal grey version (with the serial number filed off - hrmmm.) All good for the phantom power, it's getting there.

I checked the voltage across the zener diode, and it's only 5V. I've just removed it from circuit and checked it (47 kOhm resistor in series) from either pin 2 or 3 of the phantom power to ground, and it's good, holding 33.5V.. So something is pulling down the voltage.

I'll do some more poking around..

Angus

[[email protected]]

On Sunday, September 20, 2015 at 7:25:43 AM UTC+2, geoff wrote:
> On 20/09/2015 5:02 p.m., Angus Kerr wrote:
> > Hi, my single FET U87 is dead. No output at all.
> >
> > Any suggestion on common failure modes? Where do I start? Typical healthy voltages at various test points?
> >
> > Appreciate some guidance. I've got to do vocals soon. I'd prefer the U87 over my only other option, Oktava MK219.
> >
> > Angus
> >
>
>
> Probably capsule has expired (worst case). Or a wire off connector
> (better). Or a dead aluminium electrolytic, or where they tantalums ?
> (not so bad).
>
> I loath my U87 on vocals, but if iit suit a particular voice or you aint
> got an alternative I guess that's moot.
>
> Sorry to hear your nation in mourning today, hope to watch a replay in
> 90 minutes here (NZ) ;-)
>
> geoff

The Japanese played out of their socks. They tackled, and moved very fast and well with the ball.

We were very mediocre. It should never have been even close.

But for sure the better team won on the night.

It just wasn't us :(

[Scott Dorsey]

Phil Allison > wrote:
>** Here is a schem:
>
>http://recordinghacks.com/images/mic_extras/neumann/U87-schematic.png
>
>I assume you have checked the 48v phantom from the desk with a DMM, tried a new lead and checked all switch positions on the mic.
>
>With the mic disconnected, across pins 2 and 3 you should see a reading of about 10 ohms. If it is much higher like 4400 ohms, the internal transformer is not connected or has failed.
>
>With the mic plugged into the desk connected and open, find zener diode Gr1 on the PCB and see if there is +33V there compared to ground.
>
>Proceed with great care, using headphones to monitor progress.

This is all good advice, but remember that all electrical problems are really
mechanical problems in disguise. Fixing anything that has switches and
connectors, the first thing to do is to check the switches and connectors.

Also... check the voltages on the FET... if the 33V supply is good, the next
step is knowing that the FET is good and again on some of the earlier U87s
there were occasionally FET failures.

The capsules do go bad, but usually they go bad in ways that create a lot
of noise rather than silence.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Scott Dorsey]

> wrote:
>I checked the voltage across the zener diode, and it's only 5V. I've just r=
>emoved it from circuit and checked it (47 kOhm resistor in series) from eit=
>her pin 2 or 3 of the phantom power to ground, and it's good, holding 33.5V=
>. So something is pulling down the voltage.

Is the FET shorted? Use the continuity test function on the meter to see.

Is your preamp actually producing 48V through a 6.81K resistor? There are
some preamps out there that do not provide specified power. They will work
with a lot of microphones... but not the U87.
--scott


--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[[email protected]]

On Sunday, September 20, 2015 at 2:15:21 PM UTC+2, Scott Dorsey wrote:
> Angus Kerr wrote:
> >I checked the voltage across the zener diode, and it's only 5V. I've just r=
> >emoved it from circuit and checked it (47 kOhm resistor in series) from eit=
> >her pin 2 or 3 of the phantom power to ground, and it's good, holding 33..5V=
> >. So something is pulling down the voltage.
>
> Is the FET shorted? Use the continuity test function on the meter to see..
>
> Is your preamp actually producing 48V through a 6.81K resistor? There are
> some preamps out there that do not provide specified power. They will work
> with a lot of microphones... but not the U87.
> --scott
>
>
> --
> "C'est un Nagra. C'est suisse, et tres, tres precis."

Thanks Scott (and all)

Turns out that C10 (4.7uF 35V Tantalum) had failed short circuit. Replaced it with an electrolytic (all I have right now) and it's back to normal. Will replace with a correct tantalum when I get a chance to buy one on Monday.

In my experience, 90% or more of failures are due to caps, but usually the failure mode for electrolytics is an increase in ESR and a loss of electrolyte where most of the microfarads are no longer in the can.

From the circuit diagram the only thing I could see pulling down the voltage would be either C10 or C12, so I measured the voltage across C10, and bang, no volts. Do tantalums normally fail short circuit?

Anyhow, it's now giving the Neumann sound, but the capsule is shorting due to breath condensation - it must be dirty. Hopefully if I use a pop filter it should stop the capsule from shorting out during vocal takes. I do know how to clean a capsule, but I'm really not feeling it right now.....

-Angus

[Mike Rivers[_2_]]

On 9/20/2015 8:59 AM, wrote:

> a loss of electrolyte where most of the microfarads are no longer in
> the can.

What a great and poetic way of putting it. Like all the vacuum ran out
of the tube.

> From the circuit diagram the only thing I could see pulling down the
> voltage would be either C10 or C12, so I measured the voltage across
> C10, and bang, no volts. Do tantalums normally fail short circuit?

Short circuit is the most common failure mode for one of the more common
ways that tantalum electrolytic capacitors are constructed.

--
For a good time, visit http://mikeriversaudio.wordpress.com

[[email protected]]

On Sunday, September 20, 2015 at 4:01:50 PM UTC+2, Mike Rivers wrote:
> On 9/20/2015 8:59 AM, Angus Kerr wrote:
>
> > a loss of electrolyte where most of the microfarads are no longer in
> > the can.
>
> What a great and poetic way of putting it. Like all the vacuum ran out
> of the tube.

Lol, that's funny. But you are right, when the vacuum goes out of the tube, it no longer works, no?

Or the magic smoke came out..

Or in the case of cars, all the horses escaped and galloped away from under the bonnet, leaving only a few tired nags behind...
>
> > From the circuit diagram the only thing I could see pulling down the
> > voltage would be either C10 or C12, so I measured the voltage across
> > C10, and bang, no volts. Do tantalums normally fail short circuit?
>
> Short circuit is the most common failure mode for one of the more common
> ways that tantalum electrolytic capacitors are constructed.

Thanks for that, I shall put that in my memory bank for future troubleshooting..
>
> --
> For a good time, visit http://mikeriversaudio.wordpress.com

[jason]

On Sun, 20 Sep 2015 10:01:44 -0400 "Mike Rivers" > wrote
in article >
> Short circuit is the most common failure mode for one of the more common
> ways that tantalum electrolytic capacitors are constructed.
>

At an IBM lab many moons ago, a colleague had designed a large, complex
graphics controller board. It had dozens of tantalum bypass caps. Before
he first applied power, he put on saftey goggles and told everybody to
stand clear. Fourth of July ensued. The vendor had apparently put ALL of
the caps in backwards.

[[email protected]]

On Sunday, September 20, 2015 at 4:44:07 PM UTC+2, Jason wrote:
> On Sun, 20 Sep 2015 10:01:44 -0400 "Mike Rivers" > wrote
> in article >
> > Short circuit is the most common failure mode for one of the more common
> > ways that tantalum electrolytic capacitors are constructed.
> >
>
> At an IBM lab many moons ago, a colleague had designed a large, complex
> graphics controller board. It had dozens of tantalum bypass caps. Before
> he first applied power, he put on saftey goggles and told everybody to
> stand clear. Fourth of July ensued. The vendor had apparently put ALL of
> the caps in backwards.

Lol, how big was this board that safety goggles were specified up front?

[geoff]

On 21/09/2015 2:17 a.m., wrote:
> On Sunday, September 20, 2015 at 4:01:50 PM UTC+2, Mike Rivers
> wrote:
>> On 9/20/2015 8:59 AM, Angus Kerr wrote:
>>
>>> a loss of electrolyte where most of the microfarads are no longer
>>> in the can.
>>
>> What a great and poetic way of putting it. Like all the vacuum ran
>> out of the tube.
>
> Lol, that's funny. But you are right, when the vacuum goes out of the
> tube, it no longer works, no?
>
> Or the magic smoke came out..
>
> Or in the case of cars, all the horses escaped and galloped away from
> under the bonnet, leaving only a few tired nags behind...
>

Or "her resistance gave up".

geoff

[PStamler]

You might consider leaving that aluminum electrolytic in there, if it's rated 105 degrees F., and replacing the other tantalums with aluminum equivalents. Make sure they're rated 105 (those are higher reliability).

I did that on my KM 84s, and the amplifiers have been nice and reliable for a few decades now. (Had a capsule that needed rebuilding after an Opera Theater performance in a very humid, very hot loft, but the head amps have been fine.)

I evwn replaced the 1uF cap between the FET and the output transformer with a Wima polycarbonate in a couple of U 87s, but the cramped conditions made for poor reliability.

Peace,
Paul

[Phil Allison[_4_]]

Angus Kerr wrote:
>
>
>
> Turns out that C10 (4.7uF 35V Tantalum) had failed short circuit.
> Replaced it with an electrolytic (all I have right now) and it's
> back to normal.
> Will replace with a correct tantalum when I get a chance to buy
> one on Monday.


** Nice going.


>
> From the circuit diagram the only thing I could see pulling
> down the voltage would be either C10 or C12, so I measured
> the voltage across C10, and bang, no volts.
> Do tantalums normally fail short circuit?


** Yes and for a number of reasons, old age being one.

Tantalum caps have advantages in terms of small size, low leakage and long life. But the disadvantages of being electrically fragile and failing short circuit have led to them getting a bad reputation.

C8, C11, C12 & C13 all appear to be tantalums too, might be good time to replace them all before another goes short and stops the mic again.

Also check the voltage across R12 (47k), should be about 18V.


..... Phil

[Jason[_15_]]

In article >,
says...
>
> Lol, how big was this board that safety goggles were specified up front?
>

It was a large board. It was to drive a 2k x 2k raster display which
then cost as much as a small house, and the display buffer RAM chips
were, I think, 16k bits, hence there were a LOT of them. Corky, the
engineer, told me that he -always- donned the goggles on first applying
power to anything. Good move!

[geoff]

On 21/09/2015 2:13 p.m., PStamler wrote:
> You might consider leaving that aluminum electrolytic in there, if it's rated 105 degrees F., and replacing the other tantalums with aluminum equivalents. Make sure they're rated 105 (those are higher reliability).


105°C caps generally better than 85°C ones, but temperature itself isn't
a factor in a FET mic. Or shouldn't be !

Better still use 'RBLL' low-leakage electrolytics.

geoff

[Phil Allison[_4_]]

Scott Dorsey wrote:

>
>
> Is your preamp actually producing 48V through a 6.81K resistor? There are
> some preamps out there that do not provide specified power. They will work
> with a lot of microphones... but not the U87.


** The published specs for the U87i say the current draw at 48V is a miserly 0.4mA - and according to the schem, nearly all of which goes to the FET.

So, it should work as advertised with any pre-amp that has the standard 48V plus 2 x 6.81k ohms feed resistors.

BTW:

Are mixing desks still around that use balanced transformer feed for phantom power ?

Peavey used to make a live sound desks ( called Mk IIIs ) that were loaded with transformers, one for each XLR socket. All the mic input transformers had centre taps connected to a common regulated 48V supply that could deliver over 250mA.

Considerable overkill and a *bit dodgy* if pins 2 or 3 ever got shorted to ground ...



.... Phil

[Scott Dorsey]

PStamler > wrote:
>You might consider leaving that aluminum electrolytic in there, if it's rated 105 degrees F., and replacing the other tantalums with aluminum equivalents. Make sure they're rated 105 (those are higher reliability).

The tantalums were used originally because way back when that design was new,
it wasn't possible to get aluminiums with good high frequency response. The
demands of switching power supplies has caused a huge improvement in series
inductance of aluminum electrolytics today, and you can get something from
Panasonic for a buck that outperforms any of the old tantalums.

That said... dry slug tantalums fail often, wet slug military type tantalums
hardly ever fail... but both fail into shorts when they fail, sometimes
leading to collateral damage.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Scott Dorsey]

geoff > wrote:
>On 21/09/2015 2:13 p.m., PStamler wrote:
>> You might consider leaving that aluminum electrolytic in there, if it's rated 105 degrees F., and replacing the other tantalums with aluminum equivalents. Make sure they're rated 105 (those are higher reliability).
>
>105°C caps generally better than 85°C ones, but temperature itself isn't
>a factor in a FET mic. Or shouldn't be !

Temperature is the main factor in electrolytics failing. The 105C ones will
last longer than the 85C ones if you use them at 35C.

But they'll last longer at 25C than they did at 35C. The main failure mode
is loss of electrolyte.

>Better still use 'RBLL' low-leakage electrolytics.

Do they actually last longer? For a decoupling cap I don't really care about
a little leakage.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Scott Dorsey]

Phil Allison > wrote:
>Scott Dorsey wrote:
>
>> Is your preamp actually producing 48V through a 6.81K resistor? There are
>> some preamps out there that do not provide specified power. They will work
>> with a lot of microphones... but not the U87.
>
>
>** The published specs for the U87i say the current draw at 48V is a miserly 0.4mA - and according to the schem, nearly all of which goes to the FET.
>
>So, it should work as advertised with any pre-amp that has the standard 48V plus 2 x 6.81k ohms feed resistors.

Right.. But they will not work from a preamp that has a 12V rail being used
to provide phantom, like some inexpensive consoles.

What's worse is that the KM105 will break out into oscillation when powered
that way.

>BTW:
>
>Are mixing desks still around that use balanced transformer feed for phantom power ?
>
>Peavey used to make a live sound desks ( called Mk IIIs ) that were loaded with transformers, one for each XLR socket. All the mic input transformers had centre taps connected to a common regulated 48V supply that could deliver over 250mA.
>
>Considerable overkill and a *bit dodgy* if pins 2 or 3 ever got shorted to ground ...

I haven't seen one in a good while... you need to have that center tap in
exactly the right place or you get DC flux in the core. It's one of those
things that is probably fine most of the time but when it's not fine it will
be at the worst possible moment.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Angus Kerr]

Interesting. I was always under the impression that electrolytics were noisy
, and thus were kept out of the sigal path,
mainly being used for smoothing on power supplies.

Am I wrong?

Also, I would love to know what that strange diode / half fet looking component in series with the battery meter is: D1, TCR503 on the schmatic. Google search reveals nothing.

--
- Angus

[Rusty Gear]

"Angus Kerr" wrote in message
...

Interesting. I was always under the impression that electrolytics were noisy
, and thus were kept out of the sigal path,
mainly being used for smoothing on power supplies.

Am I wrong?

Also, I would love to know what that strange diode / half fet looking
component in series with the battery meter is: D1, TCR503 on the schmatic.
Google search reveals nothing.

--
- Angus

It is a current regulator diode..... really just a JFET with the gate
connected to the source internally. I believe this part is a 560 microamp
nominal rating diode.

Best regards,
John

[None]

"Angus Kerr" > wrote in message
...
> Also, I would love to know what that strange diode / half fet
> looking component in series with the battery meter is: D1, TCR503
> on the schmatic. Google search reveals nothing.

constant current diode

[Phil Allison[_4_]]

Angus Kerr wrote:
>
>
> Interesting. I was always under the impression that
> electrolytics were noisy, and thus were kept out of
> the sigal path, mainly being used for smoothing on
> power supplies.
>
> Am I wrong?


** Fraid you are, regular aluminium electros are very widely used for signal coupling in pro and consumer audio.

Of course, they are also used for smoothing and de-coupling supply noise as well, so most products have a fair number of them doted around the PCBs.



..... Phil

[Angus Kerr]

On Tuesday, September 22, 2015 at 5:30:22 AM UTC+2, Phil Allison wrote:
> Angus Kerr wrote:
> >
> >
> > Interesting. I was always under the impression that
> > electrolytics were noisy, and thus were kept out of
> > the sigal path, mainly being used for smoothing on
> > power supplies.
> >
> > Am I wrong?
>
>
> ** Fraid you are, regular aluminium electros are very widely used for signal coupling in pro and consumer audio.
>
> Of course, they are also used for smoothing and de-coupling supply noise as well, so most products have a fair number of them doted around the PCBs.
>
>
>
> .... Phil

Are they widely used because they are good, or because they are not bad, and cheap?

I mean every console probably uses them to decouple the phantom power on the mic pre.

I'm just wondering, because half of the caps in the U87 that are tantalum, like the one that failed, doesn't seem to me to be in the signal path anyway...

I mean, if I replace the caps in this mic, I would be inclined to replace them with the same type as before, tantalums with tantalums, and electrolytics with electrolytics.

Any idea of the reason for the choices in the circuits? I mean, a 4.7uF 63V 105 degree cap is about the same size as a tantalum anyway.

-Angus

[Phil Allison[_4_]]

Angus Kerr wrote:
>
> >
> > ** Fraid you are, regular aluminium electros are very
> > widely used for signal coupling in pro and consumer audio.
> >
> > Of course, they are also used for smoothing and de-coupling
> > supply noise as well, so most products have a fair number
> > of them doted around the PCBs.
> >
>
> Are they widely used because they are good, or because they are
> not bad, and cheap?


** Being much cheaper than the alternatives is certainly a reason to use them.

>
> I mean every console probably uses them to decouple the phantom
> power on the mic pre.

** There is no practical alternative capacitor for that job.

>
> I'm just wondering, because half of the caps in the U87 that
> are tantalum, like the one that failed, doesn't seem to me
> to be in the signal path anyway...
>


** Most of them are used for filtering the high impedance DC supply for the capsule - where having low leakage current is the advantage.


> Any idea of the reason for the choices in the circuits?
> I mean, a 4.7uF 63V 105 degree cap is about the same size
> as a tantalum anyway.
>

** Some designers like using tantalums while others avoid them completely.

The schem I posted is dated 1980, about the peak in popularity, after which their propensity to sudden failure became better known to manufacturers and designers.


..... Phil

[Angus Kerr]

On Tuesday, September 22, 2015 at 2:54:39 AM UTC+2, None wrote:
> "Angus Kerr" > wrote in message
> ...
> > Also, I would love to know what that strange diode / half fet
> > looking component in series with the battery meter is: D1, TCR503
> > on the schmatic. Google search reveals nothing.
>
> constant current diode

Thanks for the info. I find the battery voltage meter circuit bizarre to say the least.

[geoff]

On 22/09/2015 9:11 p.m., Angus Kerr wrote:

>
> Are they widely used because they are good, or because they are not
> bad, and cheap?

Because 10uF in any film cap would be physically enormous.

geoff
>

[Mike Rivers[_2_]]

On 9/22/2015 5:11 AM, Angus Kerr wrote:

> Are they [aluminum electrolytic capacitors] widely used because they are good, or because they are not
> bad, and cheap?

> I'm just wondering, because half of the caps in the U87 that are
> tantalum, like the one that failed, doesn't seem to me to be in the
> signal path anyway...

Tantalum used to be the way to get higher capacitance in a smaller
package and with narrower tolerance.

> Any idea of the reason for the choices in the circuits? I mean, a
> 4.7uF 63V 105 degree cap is about the same size as a tantalum
> anyway.

It is now, but when the U87 was designed, it would have been a lot
bigger. There's been a lot of progress with aluminum electrolytics with
the prevalence of relatively low voltage solid stage gear. Of course
with progress to make them better and smaller, there's also
opportunities to make them cheaper and less reliable or stable, too, but
today a small form factor aluminum electrolytic capacitor can be very
good and very compact, and may even be better than the original tantalum
capacitor. If you get a good quality one, it's not likely to fail in
your lifetime.


--
For a good time, visit http://mikeriversaudio.wordpress.com

[Scott Dorsey]

Angus Kerr > wrote:
>Interesting. I was always under the impression that electrolytics were noisy
>, and thus were kept out of the sigal path,
>mainly being used for smoothing on power supplies.

Well, it IS being used for smoothing a power supply in this case.

But..... if you have a single-ended circuit... the power supply (and thus
the supply decoupling capacitor) IS directly in the signal path! It's in
series with the load!


The problems with electrolytics are mostly that they aren't linear at high
frequencies or with low voltages across them. So the key to using them for
signal coupling is to keep a good couple volts away from the zero crossing
region, and _either_ bypass it with a film capacitor which has better high
frequency performance OR severely oversize the capacitor so that the ripple
current across it is much smaller.

There is a really good discussion of this in Doug Self's book on power
amplifier design, which I really really recommend to everyone. He identifies
coupling electrolytics as a major distortion source and demonstrates how much
better the numbers get with severely oversized capacitors.

>Also, I would love to know what that strange diode / half fet looking component in series with the battery meter is: D1, TCR503 on the schmatic. Google search reveals nothing.

It's a constant current diode, it basically extends the range of the battery
meter.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Scott Dorsey]

Angus Kerr > wrote:
>
>I mean, if I replace the caps in this mic, I would be inclined to replace them with the same type as before, tantalums with tantalums, and electrolytics with electrolytics.

Tantalums ARE electrolytics.

>Any idea of the reason for the choices in the circuits? I mean, a 4.7uF 63V 105 degree cap is about the same size as a tantalum anyway.

Back in the seventies, you couldn't get good aluminum electrolytics so
everybody used tantalums for things that needed high density but also good
high frequency response. People would use aluminums for this today.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[[email protected]]

>
> Back in the seventies, you couldn't get good aluminum electrolytics so
> everybody used tantalums for things that needed high density but also good
> high frequency response. People would use aluminums for this today.
> --scott
>

Tantalums usually have lower leakage current compared to aluminum and this can be important in Hi Z coupling applications, but not so important in bypass cap applications.

Mark

[JackA]

On Tuesday, September 22, 2015 at 7:20:47 AM UTC-4, Mike Rivers wrote:
> On 9/22/2015 5:11 AM, Angus Kerr wrote:
>
> > Are they [aluminum electrolytic capacitors] widely used because they are good, or because they are not
> > bad, and cheap?
>
> > I'm just wondering, because half of the caps in the U87 that are
> > tantalum, like the one that failed, doesn't seem to me to be in the
> > signal path anyway...
>
> Tantalum used to be the way to get higher capacitance in a smaller
> package and with narrower tolerance.

They enjoy reverse polarity! Makes neat rocket-ship!!

Jack

>
> > Any idea of the reason for the choices in the circuits? I mean, a
> > 4.7uF 63V 105 degree cap is about the same size as a tantalum
> > anyway.
>
> It is now, but when the U87 was designed, it would have been a lot
> bigger. There's been a lot of progress with aluminum electrolytics with
> the prevalence of relatively low voltage solid stage gear. Of course
> with progress to make them better and smaller, there's also
> opportunities to make them cheaper and less reliable or stable, too, but
> today a small form factor aluminum electrolytic capacitor can be very
> good and very compact, and may even be better than the original tantalum
> capacitor. If you get a good quality one, it's not likely to fail in
> your lifetime.
>
>
> --
> For a good time, visit http://mikeriversaudio.wordpress.com

[Phil Allison[_4_]]

Scott Dorsey wrote:

>
>
> The problems with electrolytics are mostly that they aren't linear at high
> frequencies or with low voltages across them. So the key to using them for
> signal coupling is to keep a good couple volts away from the zero crossing
> region, and _either_ bypass it with a film capacitor which has better high
> frequency performance OR severely oversize the capacitor so that the ripple
> current across it is much smaller.


** Electro caps have to be the most maligned of all passive electronic components. Certainly, reliability can be poor if not sensibly used but most of the other complaints do not stand up to careful scrutiny or else are easily avoided during design.

Connecting a film cap in parallel with an electro is rarely seen and offers no real advantage outside of radio frequency circuits. Coupling electros operating without any DC bias are very common in audio - and this works out fine providing the value used is large enough that the lowest frequencies do not cause polarity reversal by more than about 0.3V. Bi-polar electros can be used when this is unavoidable.

BTW:

The issue can be entirely avoided by using only high impedance circuitry or valves.


..... Phil

[[email protected]]

On Tuesday, September 22, 2015 at 3:43:28 PM UTC+2, Scott Dorsey wrote:
> Angus Kerr > wrote:
> >
> >I mean, if I replace the caps in this mic, I would be inclined to replace them with the same type as before, tantalums with tantalums, and electrolytics with electrolytics.
>
> Tantalums ARE electrolytics.

Well, you know what I mean <g>. I mean aliminium cans vs tantalum, if we want to split hairs.


>
> >Any idea of the reason for the choices in the circuits? I mean, a 4.7uF 63V 105 degree cap is about the same size as a tantalum anyway.
>
> Back in the seventies, you couldn't get good aluminum electrolytics so
> everybody used tantalums for things that needed high density but also good
> high frequency response. People would use aluminums for this today.

As I pressed send on that last post, I realised that the 20uF 6V Cap (the orange one on its side) is HUGE.

So I can replace all with aluminiums? Even though it will make me feel guilty if I do?


> --scott
>
> --
> "C'est un Nagra. C'est suisse, et tres, tres precis."

[Scott Dorsey]

> wrote:
>> Back in the seventies, you couldn't get good aluminum electrolytics so
>> everybody used tantalums for things that needed high density but also good
>> high frequency response. People would use aluminums for this today.
>
>Tantalums usually have lower leakage current compared to aluminum and this can be important in Hi Z coupling applications, but not so important in bypass cap applications.

This is absolutely true! It's less true than it was a decade ago, though.
Aluminums are getting better by leaps and bounds.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Scott Dorsey]

> wrote:
>
>As I pressed send on that last post, I realised that the 20uF 6V Cap (the orange one on its side) is HUGE.
>
>So I can replace all with aluminiums? Even though it will make me feel guilty if I do?

You probably can with modern aluminums. It might sound different, but not
a lot different. My inclination personally would be to replace them with
wet-slug tantalums, just because they are so much more reliable than anything
else electrolytic. Expensive, though.
--scott

--
"C'est un Nagra. C'est suisse, et tres, tres precis."