Unexpected sub-bass curve

[xmag]

After the fall to lower frequencies, a sub-frequency curve (or post-frequency curve, see red arrow ) reappears. Anyone know why this is so? It started suddenly, before it was not there. But I have seen the same sub-frequency in measurements I made more than 10 years ago. I have also seen that other measurements online sometimes have the same sub-frequency. I have measured 4 inch and 12 inch elements without a box in a 1 cm near field. If I move the microphone to 2 cm, the level of the frequency curve decreases but the sub-frequency curve does not change. Anyone can explain the reason for this and why some measurements do not turn out that way?

 

[xmag]

Fler med samma sub-kurva:

Do not know how it is measured:

https://imgur.com/Z9s0cxt

Measured by free element only:

http://www.troelsgravesen.dk/HES/215GMFvsRTF.jpg

Do not know how it is measured:

https://www.diyaudio.com/community/attachments/er18_raw_nearfield-gif.690241/

Measured on tweeter horn driver (without horn):

https://greatplainsaudio.com/wp-content/uploads/2017/07/604espl.gif

The curve on the tweet goes down to 20hz. you can see that it most likely goes down to below 10hz. It is extremely doubtful that it can handle such low frequencies. So there must be some kind of measurement anomaly?

 

[John Mulcahy]

Difficult to say just from the response. It could be background noise, if you look at the distortion graph there is a trace for the noise floor.

 

[xmag]

Difficult to say just from the response. It could be background noise, if you look at the distortion graph there is a trace for the noise floor.

The noise floor is far below the base level. However, the distortion follows the sub frequency curve exactly and is approximately 25dB below (THD = approximately 20%). I was in that it could be the distortion as the cause but have now measured in several different ways and it seems (I am far from sure this) it is due to the microphone's ability to pick up sound from all directions for low tones. I measure in the near field and the sound goes directly to the microphone. Very little of these frequencies come from reflexes because near fields are much higher level than any reflexes. The low-frequency sound is significantly lower level than other sound but spreads across the room and is received by the microphone from all directions. The radiation surface thus becomes larger for low frequencies into the microphone. It is still low level but enough to end the falling frequency and rise again a bit to fall when the element can no longer generate enough energy. Could it be so or am I completely lost among the frequencies ..?

Frequency measurement with THD:

 

[xmag]

Difficult to say just from the response. It could be background noise, if you look at the distortion graph there is a trace for the noise floor.

I have solved the problem with the low frequencies. It was the power supply's current filter that gave the signal in the low base. It was not visible in the noise measurements !?
This is what the measurement now looks like 30 cm distance:

 

[John Mulcahy]

Good detective work. The noise is captured before the measurement, so it won't capture anything that is signal dependent.

 

[xmag]

Good detective work. The noise is captured before the measurement, so it won't capture anything that is signal dependent.

THANKS!!
Yes, of course it is. Why did I not think about it? Probably because I was excited to find the error

 

[xmag]

Have continued to test. I have located the problem with the DC filter to the microphone amplifier. With filters for the whole system, I get a frequency arc with the middle at 30hz but low distortion. If I connect the microphone amplifier past the filter, the frequency arc disappears but now the distortion increases.

It is good with the DC filter because without it there will be a lot of noise on the impedance curve from 3khz and up and lower distortion so I want to keep the filter.

The next step is to adapt a filter for the microphone amplifier or use a non-switched power supply, which is not so many to buy cheap.

The error is from the beginning because I use switched power supply. Chose it because it is cheap and many have it at home in the scrap box. May I think about whether I should recommend a regular power supply.

 

[sm52]

The next step is to adapt a filter for the microphone amplifier or use a non-switched power supply, which is not so many to buy cheap.

You can use a battery for a screwdriver or similar tool. Or a powerbank to charge your phone.

 

[xmag]

You can use a battery for a screwdriver or similar tool. Or a powerbank to charge your phone.

Yes, I'm on the to-do list.

 

[xmag]

Have now rebuilt my test-Matrojig to battery operation. 10 AA batteries 1.2v at 2800 mAh becomes 28Ah. A few days should be enough to measure before the batteries need to be charged. It became cheaper than building a line power supply. Will return with measurement results later.

 

[xmag]

Now I am really confused, it will be the same result with batteries as with power supplies and DC filters.

If I remove the DC filter for the power supply , the measurements will be good (but higher distortion over 3kHz).

But with batteries, I can not solve the problem of removing DC filters because I do not have one there.

The strange thing is that if I use different batteries for the power amplifier and the microphone amplifier, it works.

My not important questions are:
1. Why do I get no low frequency interference when using power supplies?
2. Why do I get low frequency interference when I used power supplies with DC filters?

My important questions are:
1. Why do I get the exact same low frequency interference when I use batteries as power supplies with DC filters?
2. Why does low frequency interference disappear when I use different batteries?

How to solve this problem?

 

[John Mulcahy]

A ground loop somewhere, perhaps?

 

[xmag]

A ground loop somewhere, perhaps?

I first thought it might be so. Has connected signal earth and power lines in different ways to test. No difference. Signal ground has no closed circuits, everything starts at central signal ground and ends where it is needed without going anywhere. The power wires go straight from the battery to the on / off button and directly to the power amplifier and on to the microphone amplifier and end there.

 

[John Mulcahy]

The only thing you are really changing when supplying the power amp and mic preamp separately (given they have a common ground somewhere) is the path the supply current follows for each. Would generally want separate supply and return connections to each, with the returns going back to the common star ground for the system, and not to daisy chain anything.

 

[xmag]

The only thing you are really changing when supplying the power amp and mic preamp separately (given they have a common ground somewhere) is the path the supply current follows for each. Would generally want separate supply and return connections to each, with the returns going back to the common star ground for the system, and not to daisy chain anything.

Yes, I will try to pull completely separate power supply wires when I have time. When I connect the power supply and disconnect the batteries, it works perfectly. But switched power supply gives disturbances over 3 khz. The power supply is connected from the outside through a power socket. I will try to connect the batteries the same way and hope it gets better. With these problems, I have actually thought about removing the microphone amplifier and placing it close to the microphone with a 9v battery supply. But first I will test until I solve the problem with the noise in the signal that raises the level in low frequencies.

 

[xmag]

The only thing you are really changing when supplying the power amp and mic preamp separately (given they have a common ground somewhere) is the path the supply current follows for each. Would generally want separate supply and return connections to each, with the returns going back to the common star ground for the system, and not to daisy chain anything.

Now I know what the problem is: It is known especially in tube amplifier builders. The phenomenon is called "motorboating" https://www.electronicspoint.com/forums ... it.267151 / In addition to the fact that tube amplifiers can suffer from the problem, ordinary amplifiers with battery operation that drive both preamplifiers and power amplifiers can suffer from it. This is because the batteries have a high impedance which, together with the final stage, creates low-frequency pulsating waves (sounds like a motorboat) in the power cables. The power amplifier does not normally detect it, but low-level amplifiers such as preamps or microphone amplifiers connected to the same battery as the power amplifier do. The problem is solved by inserting a capacitor of a few thousand uF on the current and possibly a resistor to the current supply of the microphone amplifier (in my case). You must also run separate power lines to the preamplifier etc.

 

[xmag]

The only thing you are really changing when supplying the power amp and mic preamp separately (given they have a common ground somewhere) is the path the supply current follows for each. Would generally want separate supply and return connections to each, with the returns going back to the common star ground for the system, and not to daisy chain anything.

Now I know what the error is:
It is known especially in tube amplifier builders. The phenomenon is called "motorboating" = https://www.electronicspoint.com/forums/threads/motorboating-what-exactly-is-it.267151/
In addition to the fact that tube amplifiers can suffer from the problem, ordinary amplifiers with battery operation that drive both preamplifiers and power amplifiers can suffer from it. This is because the batteries have a high impedance which, together with the the power amplifier , creates low-frequency pulsating waves (sounds like a motorboat) in the power cables. The power amplifier does not normally detect it, but low-level amplifiers such as preamps or microphone amplifiers connected to the same battery as the power amplifier do. The problem is solved by inserting a capacitor of a few thousand uF on the current and possibly a resistor to the current supply of the microphone amplifier (in my case). You must also run separate power lines to the preamplifier etc.

I experimented a lot and I get the best result when I mount a resistor in series with the power cable near the microphone amplifier, between the resistor and the microphone amplifier I put one or more capacitors of several thousand uF. I only had up to 7,000uF but the difference was big:

The bottom measurement looks good. Just a little worse than with power supplies that go down to 43dB with and without capacitors. Will get more capacitors and test with higher values. What I learned from this is that many who measure have a little "motorboating" because I often see a small shelf at the bottom of the measurements.

Being stubborn pays off

 

[xmag]

How to remove both ripple from switched power supplies and "motorboating" from battery operation and switched power supplies with ripple filter:

Before ripple filter:

After ripple filter:

I have run an RTA in REW with a microphone to include the microphone amplifier in the measurements, that is why the spectrum is not a straight line. I could not get completely quiet, my wife was watching TV in the room next door and children were playing outside.




Frequency response with only ripple filter but before motorboat filter. That filter creates motorboat low frequency:

With also motorboat filter:

You can gain a lot from not giving up!