Impedance Tests

[sr1200]

I'm putting together a video on impedance focusing mainly on dynamic and ribbon mics. I did a circuit test using a simulator first and when run in the real world, the simulations held up pretty well, almost dead on actually. I ran into an issue when I started using condenser mics though and I'm not really sure how to explain the results.

As mic pre impedance was increased on the mic pre, so was level when using dynamic or ribbon mic.
As mic pre impedance was increased on the mic pre, the level DECREASED using a condenser. My only guess here is that since the passive mics are just that, passive, the interaction between the load on the preamp and the mic is more "direct" (if thats a good word), whereas the condenser mics have a buffered output if you will which would work negatively with an increasing load on the pre. Again, guess... would love to know what the real answer is behind this so I dont sound like a complete nincompoop.
(im impressed that nincompoop did not trigger a spelling error!)
Thanks

 

[Bo Deadly]

As mic pre impedance was increased on the mic pre, so was level when using dynamic or ribbon mic.

How much drop in dB vaguely?

The source impedance of the mic and the load impedance of the pre make a voltage divider.
If the source impedance of a microphone is say 300 ohms and the load is say 1200, that's a few dB.
But otherwise, unless your making the mic impedance lower than that, I can't immediately think of a reason why the level should drop so much.

 

[sr1200]

It was a significant drop. The mic was 150 and load was being switched between 500 and 2k.

my peak/rms values in dbfs
500 ohms: pk -8.7 RMS -16.65
2K ohms: pk -13.12 RMS -22.43

My question though, is why does a condenser have an opposite reaction than a dynamic? The voltage goes UP for the dynamic and DOWN for the condenser.

 

[Bo Deadly]

It was a significant drop. The mic was 150 and load was being switched between 500 and 2k.

my peak/rms values in dbfs
500 ohms: pk -8.7 RMS -16.65
2K ohms: pk -13.12 RMS -22.43

My question though, is why does a condenser have an opposite reaction than a dynamic? The voltage goes UP for the dynamic and DOWN for the condenser.

Looks like you're not just switching the load, you're switching a transformer primary between series to parallel which is going to have a +-6dB-ish change in level?

 

[sr1200]

It is, however, when i do that with the dynamic mic's i get the opposite result. why?
EDIT:

NEVER MIND... that the key to it!! Thank you! I mis worded that. edit, im gonna try again. The coil tapping gain/loss off sets the amount of gain/loss from the impedance mismatch between the mic/pre. Begs the question, is it worth it though. My tests for this research has led me to believe that theres a lot of creative marketing going on out there thats making some claims that i can't reproduce in reality.

 

[emrr]

Similar reaction found with tube output stages with low/no NFB. Mixed in there is a change in distortion profile.

I haven't listed all the different ways this gets done on a mic pre, but there's enough to say there's no cut and paste result to add to marketing, but that's what happens.

 

[user 37518]

I'm putting together a video on impedance focusing mainly on dynamic and ribbon mics. I did a circuit test using a simulator first and when run in the real world, the simulations held up pretty well, almost dead on actually. I ran into an issue when I started using condenser mics though and I'm not really sure how to explain the results.

As mic pre impedance was increased on the mic pre, so was level when using dynamic or ribbon mic.
As mic pre impedance was increased on the mic pre, the level DECREASED using a condenser. My only guess here is that since the passive mics are just that, passive, the interaction between the load on the preamp and the mic is more "direct" (if thats a good word), whereas the condenser mics have a buffered output if you will which would work negatively with an increasing load on the pre. Again, guess... would love to know what the real answer is behind this so I dont sound like a complete nincompoop.
(im impressed that nincompoop did not trigger a spelling error!)
Thanks

What electrical equivalent model for the capsule did you use in the simulation? Or what exactly did you simulate?

 

[Matt Syson]

There may be considerable differences between the actual output impedances of dynamic and condenser mics and what the 'spec sheets' say. Feeding a dynamic (passive) mic into a lowish impedance (lower than recommended) will damp the diaphragm which of course affects the sound quality and level.. Mic amps with transformers are of course more reactive than a typical transformerless input which would tend to be resistive. Then of course if the preamp is switching input primaries it then presents a difference of 4 times (500 versus 2K) and the reactive effects will change things too depending on the nature of the mic. You also don't really know the impedance seen by the mic. Your 'model' needs to include the real impedances not just what the manufacturers say.

 

[sr1200]

The point of this is to use what the manufacturer gives us as far as a spec goes. Right or wrong. (usually wrong, being that a dynamic mic doesn't have a single impedance given a frequency the mic is picking up, but again, thats the point that the spec sheets are kind of BSish).
All im going to say is that my findings on this test are not lining up with manufacturers claims for impedance on mic pres in conjunction with mics.
Im testing with a royer 121, and sm57 and a u87 with a resistive input.

 

[Bo Deadly]

If you want to properly test the output impedance of a mircophone, you would have to use a different methodology entirely than what you're doing.

One method would be to use a scope and compare a high impedance load to a relatively low load such that the level drops by 6dB. Ideally this would be done in a very quiet space with no reflections and with care sufficient to get reproducible results.

Another method might be to actually drive the microphone through a well defined impedance (being very careful not to put too much signal into it which could damage it). Then, based on the voltage drop across that known impedance value, you can compute the impedance of the source. But the specifics of this test would need to be very carefully considered to get meaningful and accurate results (for example a condenser mic would have to be actually powered and operated under it's normal conditions and completely isolated from vibration and sound).

 

[FIX]

Technically, a mic transformer is supposed to follow the "bridging transformer" rule, which is that the load in the secondary isn't supposed to attenuate the mic by more than 1dB. Even if/though it's a step up. Also, the primary should be 10X the source (mic). Remember that any load on the secondary will be multiplied by the ratio. Also, for reference, a pad is also supposed to be a 1/10 ratio as well, so it's primary on a 150 ohm mic would be 1500 and it's output would be 150 looking into the 1500 impedance of the input transformer. With a simple balance L pad, you can only really do 20dB, but an H pad can be a little different.

 

[JohnRoberts]

I recall last century while working at Peavey, I dropped into the transducer engineering lab to chat with a few of the senior microphone design engineers. Some of these guys had decades of experience working at other major microphone manufacturers. When I tried to pick their brains about optimal microphone source impedance and terminations, I mainly got a puppy dog tilted head response. Rather than learning about firm design rules or metrics it was kind of, "doing whatever works best in typical applications".

As Paul has shared we have fairly well established rules for nominal terminations.

JR

PS: I feel like I should repeat, when trying to sell a SKU that sounds better, it helps to at least sound different. Using unconventional terminations can alter frequency response and output level for better or worse.

 

[FIX]

Those engineer went on to become politicians, because they were so good at not answering questions...

 

[sr1200]

HAHAHA. This is great.

My purpose is to use the device in a practical setting in relation to what is claimed by manufacturers.
My findings make mathematical sense in that as the impedance increases, so does voltage.

What i DONT see is any change in the "transient sharpness" between any impedance setting (as manufacturers claim), on any pre with any input transformer (across 4 different mic pres, and using the variable impedance device). Theres a level difference, but once compensated for, there is virtually NO change in the attack/transient being picked up.

As far as frequencies changing, I see no significant difference in the upper end, and see up to an 8db roll off on the bottom end between 150 ohms and 15k ohms. Between settings like 300 and 1200 its barely a db... nothing that can't be compensated for if really needed at mix time.

I also see a point of diminishing returns in regards to level somewhere above 3k. Between 3k and 15k level difference is barely 1dbfs. between 150ohms and 3k, the difference is HUGE.

 

[user 37518]

HAHAHA. This is great.

My purpose is to use the device in a practical setting in relation to what is claimed by manufacturers.
My findings make mathematical sense in that as the impedance increases, so does voltage.

What i DONT see is any change in the "transient sharpness" between any impedance setting (as manufacturers claim), on any pre with any input transformer (across 4 different mic pres, and using the variable impedance device). Theres a level difference, but once compensated for, there is virtually NO change in the attack/transient being picked up.

As far as frequencies changing, I see no significant difference in the upper end, and see up to an 8db roll off on the bottom end between 150 ohms and 15k ohms. Between settings like 300 and 1200 its barely a db... nothing that can't be compensated for if really needed at mix time.

I also see a point of diminishing returns in regards to level somewhere above 3k. Between 3k and 15k level difference is barely 1dbfs. between 150ohms and 3k, the difference is HUGE.

how are you testing the mics, and also, how did you simulate everything? Many manufacturers test their microphones in anechoic chambers, with omnidirectional sources, etc... what exactly are you doing to try and discredit or affirm what the manufacturers report?

 

[emrr]

One standout piece that's not acknowledged - many modern transformerless preamps are actually 5K or higher. They make a Cloudlifter look low at 3K. The lowest modern thing I have is in my portable Zoom F8n, which is 2K.

The 'classic' Mackie's are 5K. Some early '90's Yamahas I have are 4K. Current Yamaha QL series mixers are 7k5.

 

[JohnRoberts]

One standout piece that's not acknowledged - many modern transformerless preamps are actually 5K or higher. They make a Cloudlifter look low at 3K. The lowest modern thing I have is in my portable Zoom F8n, which is 2K.

The 'classic' Mackie's are 5K. Some early '90's Yamahas I have are 4K. Current Yamaha QL series mixers are 7k5.

Are you sure?

I have been responsible for many transformer-less preamps over the decades and always targeted 1.5k-2k input Z to be bridging to nominal 150-200 ohm microphones source impedance.

JR

 

[john12ax7]

10k+ input impedance is not uncommon for high end "clean" pres, for example Forssell and Gordon.

 

[emrr]

Are you sure?

I have been responsible for many transformer-less preamps over the decades and always targeted 1.5k-2k input Z to be bridging to nominal 150-200 ohm microphones source impedance.

JR

Certain.

 

[sr1200]

This was part of the research:

Scarlett / UA Volt interface pres = 3k
RME Fireface = 3.4K
SSL2 = 1.2k
Presonus XMAX 1.6K
Apollo X = 5.4k but is variable depending on the model loaded according to their tech support
AEA RPQ = 10K
AEA TRP2 = 63K
Neve 1073 300/1.2K
API312 = 1.5K
UA610 = 500/2K

Testing in a chamber is nice and gives you really nice clean neat numbers. Thats not what i was after. Im after what REALLY happens, in a studio, recording music with equipment commonly found in home and pro studios. If you can't recreate the perfect scenario results in the real world, then your test numbers aren't worth the paper they're printed on. Its just a spec for someone to oggle or some marketing guy to use as an edge over the competition.