AKG Perception P220 to Neumann u87 5 min mod ( p200, p100, p400, p420? )

[ccaudle]

It would be nice for us paint-by-number noobs if some expert would say EXACTLY what specs a capacitor should have, for this mod, like the type, voltage and tolerance.

Khron covered the tolerance question.
Regarding type, in general for audio path capacitors you want a dielectric which does not change capacitance with voltage or temperature, very stable. Film capacitors meet that requirement, but tend to be bulky. Of ceramic capacitors only C0G or NP0 dielectric meets that requirement (technically that is just a stability vs. temperature specification, but in practice ceramic dielectric formulations which are stable with temperature are also stable with voltage).
The biggest disadvantage of C0G capacitors is that only small values are available. In this application you need less than 1nF value, so C0G or film can satisfy the values needed.

Regarding voltage this design is phantom powered, so you know that the highest voltage any node in the mic will see is limited to less than 55V (48V plus worst case power supply tolerance), and most nodes downstream of the input pins will only see in the 35V range, so any cap rated 50V or more is fine.
Most film and C0G ceramic caps should meet that requirement as well.

 

[kingkorg]

Adding this here. My buddy Nate from the Time Preservation Society channel, who has never done a mic mod before being inspired by this thread and the conversation it has started off-site, is doing his part to spread the Mic DIY Virus and did a video on the mod as well.

Not that it matters much, but the original target u87 was a vintage, iirc 70's black one. Not an ai. Subsequent ai ones I measured turned out to be slightly brighter sounding. So 680pF is matching value for slightly warmer sounding vintage 87.

 

[Wordsushi]

You might want to let him know the link to this thread wasn't parsed (properly), in the description

Thanks for the heads up. I let him know and he has since fixed it and wanted me to pass along his apologizes for the confusion. I told him he's now banished from ever opening up another mic again because that's the mic law and mic law cannot be broken without consequence.

 

[micolas]

Honestly, on testing thus far in comparison with stock 220 and a 420? I am not hearing a significant difference in character.

OK Mike,

Let's make the comparison differently.
A good microphone must highlight its qualities when it is processed Eq, dynamic, etc.
We apply a little compression to it, in the real mix you will see even more, in different successive stages.
We use any compressor, hardware or plugin.
Let's set the parameters for the voice (for simplicity to some average starting values, in practice they will adjust in context):
Ratio: 2.5:1 - 3:1
Attack: 5...10ms
Release: 20...30ms (reduce the Release Time until you don't hear "pumping" anymore)
Thrershold: since the compression starts to work, we adjust the amount of Gain Reduction between 6...9dB (in practice we will have more by summation).
From Makeup Gain compensate for the reduction in output level.
With the same compressor settings, test the stock and modified microphone.
Because human auditive memory is a few seconds, it would be good to record for a more objective comparison.
Post the tests and settings so we can hear them too, it's educational
Let us know what you've come to, whether @kingkorg's proposed mod makes any noticeable difference, and how you'd describe it.

Thanks!

 

[mikes425]

OK Mike,

Let's make the comparison differently.
A good microphone must highlight its qualities when it is processed Eq, dynamic, etc.
We apply a little compression to it, in the real mix you will see even more, in different successive stages.
We use any compressor, hardware or plugin.
Let's set the parameters for the voice (for simplicity to some average starting values, in practice they will adjust in context):
Ratio: 2.5:1 - 3:1
Attack: 5...10ms
Release: 20...30ms (reduce the Release Time until you don't hear "pumping" anymore)
Thrershold: since the compression starts to work, we adjust the amount of Gain Reduction between 6...9dB (in practice we will have more by summation).
From Makeup Gain compensate for the reduction in output level.
With the same compressor settings, test the stock and modified microphone.
Because human auditive memory is a few seconds, it would be good to record for a more objective comparison.
Post the tests and settings so we can hear them too, it's educational
Let us know what you've come to, whether @kingkorg's proposed mod makes any noticeable difference, and how you'd describe it.

Thanks!

For sure, totally get that, and this was just preliminary notation on a quick comparison unprocessed... I am subsequently, with my typical processing chain, liking it a lot more (noticing the 'cleaner' high end) - especially on A/B'ing on different monitoring platforms, headphone v. speaker etc - I'd be glad to test per your chain when i have a chance and share the results. There's a definite smoother upper range response... and i'd be interested in modding another with a different cap value. But yep, appreciate the point and can say for now that this is a remarkably close replica of my 87Ai. Whatever i heard as far as the stock 420 vs 220...could have been 'colored' by whatever state of compression my ears were in or the cans i was using but do intend to check them again and will be happy to post further thoughts! Thanks again!

 

[Paul W]

Regarding voltage this design is phantom powered, so you know that the highest voltage any node in the mic will see is limited to less than 55V (48V plus worst case power supply tolerance), and most nodes downstream of the input pins will only see in the 35V range, so any cap rated 50V or more is fine.
Most film and C0G ceramic caps should meet that requirement as well.

The Perception 200/220 (and I assume the related others) has a DC-DC converter, as discussed in this thread: https://groupdiy.com/threads/perception-200-polarization-voltage.54162/

Post 5 from Khron links to Henry at audioimprov http://www.audioimprov.com/AudioImprov/Mics/Entries/2013/2/26_AKG_Perception_220.html

Henry measured only 52 volts there which seems low for having a voltage boost, but there was some speculation as to whether that was a duff measurement given the difficulty of correctly measuring voltage in the high impedance section with a normal voltmeter. (All of that is above my pay grade.)

 

[MicMaven]

I told him he's now banished from ever opening up another mic again because that's the mic law and mic law cannot be broken without consequence.

Not true, Old Man. Every DIY enthusiast is entitled to to claim ONE MULLIGAN PER DAY, exempting one screw up of his choice without penalty or hounding per day. Shoot, even my young bride grants me this relief! James

 

[Khron]

Henry measured only 52 volts there which seems low for having a voltage boost, but there was some speculation as to whether that was a duff measurement given the difficulty of correctly measuring voltage in the high impedance section with a normal voltmeter. (All of that is above my pay grade.)

Not 100% sure where and how Henry measured that, but for what it's worth, the positive bias voltage can be measured, with just about any multimeter, in the spot indicated here:



That is, BEFORE the last series 1meg resistor. The giveaway is the trace on the other side of that resistor, going off to the socket-header, going to the horizontal board. I don't recall off the top of my head which other node is for the "big" negative voltage, but that's not relevant for the 200 / 220 anyway.

 

[micolas]

Not 100% sure where and how Henry measured that, but for what it's worth, the positive bias voltage can be measured, with just about any multimeter, in the spot indicated here:

View attachment 130413

That is, BEFORE the last series 1meg resistor. The giveaway is the trace on the other side of that resistor, going off to the socket-header, going to the horizontal board. I don't recall off the top of my head which other node is for the "big" negative voltage, but that's not relevant for the 200 / 220 anyway.

@Khron , you are absolutely right
I usually measure directly on the voltage doubler diode(+), to know the maximum value provided by DC/DC.
In the case of P420 black, the 3rd one from the right, practically at the same point indicated by you.

I powered the microphone from the mixer that I use most frequently for work, it has Uphantom slightly lower than 48v, but still within the allowed range.

So without load (microphone not connected) I measured
Uphantom = 47.5v
I connected the microphone and let the working mode stabilize (honestly, I didn't draw the diagram, I don't know what time constant it is)
After a few minutes I measured:
Upol+ = 58.7v
(At Uphantom=48v, Upol+ does not exceed 59.2v...59.3v)
*Note: since the polarization voltage of the capsule has a more static role, when measuring with a regular DMM with an input impedance of the order of a few Mohm, (but which still has some nanocurrent consumption unlike the capsule), the read value(on this node) of this Upol+ decreases , because they didn't used a large unnecessary filter (unprofitable).
Returning to the original problem, I use 63v capacitors in the audio circuit and converter.
If I decide on a higher Upol of the capsule, or another arrangement for the polar patterns, for example >120v/60v then I use capacitors of the appropriate voltage.
Some use 250v, 630v capacitors for coupling and more, for some extra mojo. Everyone does what they want, the important thing is to be happy.

Edit: The measurements were made with the microphone set in Cardioid mode.

 

[micolas]

I don't recall off the top of my head which other node is for the "big" negative voltage, but that's not relevant for the 200 / 220 anyway.

For those who have P420, in the situation when they have problems with the polar patterns and suspect a malfunction of the DC/DC, I opened the microphone once more for measurements.I'm annoyed by the threaded nut that requires a dedicated tool for screwing/unscrewing.
I have suffered several times, due to the painted body, if it was not screwed down enough, to have a very pronounced ground hum of 50Hz.

The negative polarization voltage is equal to the positive one (+/- 0.1v due to the tolerances of the components) and can be measured at the point (Vpol-) highlighted in the picture.

 

[hifizen]

In his video review of the cap mod, Nate was rightly criticizing the P420's odd choice of low cut filter frequency, which is listed in the manual and spec sheet as 300 Hz.

So, I decided to have a look at the circuit while inside my mic to see if it could be changed to something more useful. I quickly discovered that while the pattern selection and capsule EQ circuitry is all cloned nearly verbatim from the U87ai, the low cut is not. On the transformer board, there is what looks like a 2nd order low cut filter circuit, using what seems to be a PNP (type unknown - mine's covered with goo) emitter follower buffer for the filter & output transformer.

I traced out the circuit and measured the necessary part values, and though I don't have a test setup to confirm with an acoustic sweep, electrically it should behave as about a 100 Hz 2nd order low cut (assuming I didn't bugger up somewhere in tracing the circuit or the ltspice transcription). So, much more sensible than a 300 Hz cutoff.

Maybe the AK.. err, Samsung people got the spec sheet wrong??

Here's the circuit as traced out & LTspice result, for those nerdy enough to care...

 

[micolas]

In his video review of the cap mod, Nate was rightly criticizing the P420's odd choice of low cut filter frequency, which is listed in the manual and spec sheet as 300 Hz.

So, I decided to have a look at the circuit while inside my mic to see if it could be changed to something more useful. I quickly discovered that while the pattern selection and capsule EQ circuitry is all cloned nearly verbatim from the U87ai, the low cut is not. On the transformer board, there is what looks like a 2nd order low cut filter circuit, using what seems to be a PNP (type unknown - mine's covered with goo) emitter follower buffer for the filter & output transformer.

I traced out the circuit and measured the necessary part values, and though I don't have a test setup to confirm with an acoustic sweep, electrically it should behave as about a 100 Hz 2nd order low cut (assuming I didn't bugger up somewhere in tracing the circuit or the ltspice transcription). So, much more sensible than a 300 Hz cutoff.

Maybe the AK.. err, Samsung people got the spec sheet wrong??

Here's the circuit as traced out & LTspice result, for those nerdy enough to care...

Honestly, to my ears, the hi-pass on the P420 never sounded like a cut at 300Hz, but rather at 120...130Hz. I didn't take measurements I just recorded my voice years ago with and without HP, I listened and decided not to make a mod here (although when I saw the published specs I thought I would change for HP to 70 ...80Hz). I actually like the way 'flat' sounds.
Those who drew up the spec sheet for the P420 are probably in a nice office in the marketing department far away from the people who actually work, test, measure, etc.
Thanks for your contribution to the knowledge base.

 

[GhostlyStudio]

Just performed my first microphone mod on an AKG p420

Not my prettiest work, but just reassembled and given a quick test and really love the change just narrating some poems through it.

Because I've soldered on a through hole capacitor onto a pad, would anyone consider adding anything on top like glue or resin to help keep the component in place over time? Its a reasonably strong bond but didn't want to really give it a tug incase it came off taking a pad with it.

I also need to clean the remaining flux off.

 

[micolas]

Just performed my first microphone mod on an AKG p420

Not my prettiest work, but just reassembled and given a quick test and really love the change just narrating some poems through it.

Because I've soldered on a through hole capacitor onto a pad, would anyone consider adding anything on top like glue or resin to help keep the component in place over time? Its a reasonably strong bond but didn't want to really give it a tug incase it came off taking a pad with it.

I also need to clean the remaining flux off.

View attachment 131251

I'm sure that @Khron would recommend you to fix it preventively with some glue, against shocks, vibrations, etc.

Congratulations! The important thing is to be satisfied with your work.

 

[micolas]

@GhostlyStudio
Did you use a 680pF capacitor?

 

[GhostlyStudio]

Thanks @micolas!

Yeah, it's 680pF.

 

[kingkorg]

Just performed my first microphone mod on an AKG p420

Not my prettiest work, but just reassembled and given a quick test and really love the change just narrating some poems through it.

Because I've soldered on a through hole capacitor onto a pad, would anyone consider adding anything on top like glue or resin to help keep the component in place over time? Its a reasonably strong bond but didn't want to really give it a tug incase it came off taking a pad with it.

I also need to clean the remaining flux off.

View attachment 131251

I'm personally hesitant adding glue. Some types could leak over time, or disintegrate, or who knows what. For anyone else in future, please, the capacitor type is not critical here, use something that will fit easily, and it doesn't have to be like 3KV rated. C0G are great, cheap, small.

Polystyrene type can change value while soldering, especially if one is inexperienced and overheats it, so what you are left with could be a random value.

One last time, the original cap doesn't have to be removed, cap value adds in parallel.

 

[GhostlyStudio]

I'm personally hesitant adding glue. Some types could leak over time, or disintegrate, or who knows what. For anyone else in future, please, the capacitor type is not critical here, use something that will fit easily, and it doesn't have to be like 3KV rated. C0G are great, cheap, small.

Polystyrene type can change value while soldering, especially if one is inexperienced and overheats it, so what you are left with could be a random value.

One last time, the original cap doesn't have to be removed, cap value adds in parallel.

Ah, apologies that was me trying to be clever... It didn't occur to me that was a risk and given how fiddly the process is, I can see why it's best to use other caps.

@micolas : it might be 680pF.

 

[s wave]

Just performed my first microphone mod on an AKG p420

Not my prettiest work, but just reassembled and given a quick test and really love the change just narrating some poems through it.

Because I've soldered on a through hole capacitor onto a pad, would anyone consider adding anything on top like glue or resin to help keep the component in place over time? Its a reasonably strong bond but didn't want to really give it a tug incase it came off taking a pad with it.

I also need to clean the remaining flux off.

View attachment 131251

Nice what kind of soldering gun/station you use?

 

[GhostlyStudio]

Nice what kind of soldering gun/station you use?

I've got a Hakko FX-888D and used the really fine solder tip that came with the original kit?

I used a method seen elsewhere in this forum to remove the original capacitor and added some extra solder to the ends of the ceramic capacitor and then covered both ends with the fine tip. The original lifted away quite quickly leaving the remaining pads for me to solder the replacement part to.