Jump to content

* * * * *

Audio Physics 101 - Transients


This article is by Black Cat who is an avid user of the Talk Audio boards and works in our field...as you can tell. Part One of two.


After seeing yet another post about power caps and someone posted a graph showing current instantly changing I thought it's about time to do some measurements and see how things work in reality. This test shows how transients behave in reality and that common perception is fundamentally flawed.

The missunderstanding comes from the fact that all moving coil speakers are inductive and that V=IR doesn't apply to AC circuits with reactive loads. Inductors dislike change in current. The voltage can change instantly but the current will lag behind. This is why inductors are used in crossovers as they can filter out high frequencies. The very same effect causes transients to behave in an unexpected way.

The test equipment used for the test:
* 100MHz 2 channel digital storage scope
* LCR meter
* hall effect current probe (not a clamp meter)
* 12V car battery (superstart jobbie I had to hand)
* 10W car bulb
* Infinity kappa perfect 10 sub 4ohm single coil

Test 1 - Resistive load

The first test used the 10W bulb to simulate a resistive load. The test setup is very simple, measure the voltage accross the bulb and the current flowing in the wire.

the bulb was connected to the battery and the results recorded:

Posted Image

As you can see the voltage (yellow line) and current (blue line) rise as close to instantly as the scope could measure. The current can be seen to fall due to the bulb's resistance increasing with the heating of the filament. A small voltage droop can be observed due to the resistance of the wiring and ESR of the battery. The droop lasted about 5ms.

This test proves the test setup is working as expected and that the current probe can react quickly enough to give valid results.

Test 2 - inductive load

The second test replaced the bulb with the kappa 10 speaker. The speaker was measured as having an Re of 4ohms and an Le of 2.4mH. This gives an effective upper f3 of 265Hz. A fair number for a sub.

On to the test:

Posted Image

Straight away you can see the difference, the current (blue line) is climbing very slowly compared to the first test. In fact it takes 2ms for the current to peak. The following reduction in current is caused by the back EMF of the coil moving. A very small voltage droop can be seen but you can also see how it comes after the initial rise in voltage.

If you do the maths, 2ms for a quarter cycle, you have a fundamental max frequency of 125Hz.

So what does this all mean? Well for starters, it means your amp isn't going to take a huge gulp of current when you get a transient, as the speaker just won't allow the current to change fast enough. Secondly, it shows that long term current supply is the more important factor rather than 'speed' of supply. And finally it shows that bulbs are really quite nasty things! Very low cold resistance means a big inrush of current. That is what kills switches and relays.

I think I have a 1F doorstop somewhere... but that can be for another day.


3 Comments

Blackcat,
Love the demonstration and the fact that you made it accessible to the lay person. It is well accepted that load characteristics can result in phase shifts between voltage and current; so I also agree with the concepts.
Nonetheless, I am still missing how these ideas negate the ripple-reducing effect of caps on a power source.
Again, good job!
Alberto Lopez
BlackCat Since I could not comment on the forum for some sort of technical glitch, I thought to try here.
I just read your post on Transients and all the subsequent answers. I am afraid that JoyLove has understood the essence of the video showing the cap being used to reduce ripple noise in the supply but that you some how missed it. The video does not measure the effects of supply ripple noise on the output of the amplifier. It does on the other hand demonstrate what the cap does before the amp. Here, there is no argument about physics or theory since the test can be easily replicated by any car audio shop. That, by the way, is what I wanted to accomplish. I wished to democratize knowledge in a way that runs counter to the normal engineering tendency to make knowledge exclusive.
In one of your responses to JoyLove you ask how is it that the Cap seems to have done nothing to the transient measurement on your light bulbs. Well, you yourself said it, there is no AC component. Outside of the almost instantaneous "on" portion, which probably happens at a higher frequency than your meter can measure, there is no AC ripple to eliminate. Caps do nothing with pure DC and have a very hard time with very low frequencies (assuming that any battery power drop over time could be construed as being part of a very long cycle). So, I would anticipate for the Cap to do nothing to your test results; which coincidently is what I believe JoyLove is stating too.
I chose a class D amp, as JoyLove stated, because they switch in sympathy with the input (typically audible music). Class AB amps switch at a higher frequency and thus would not be directly measured by an RTA, which again is what shops have. Since I wanted to show people how to measure noise in the car's power grid using common shop supplies, a Class D amp allowed me to show the effects of a noisy source across the audible spectrum.
But the argument was not about whether these amps contain or need caps internally. The argument was purely on their effect on the supply to the whole system. While Amps may have voltage regulation or large caps and inductors in them, many other components used in the same signal chain won't. That is the problem!
Next, many may ask why I did not give specs about the magnitude of the noise in a more engineer-approved form, well because I referenced the whole exercise to the alternator noise. If you look at the video, I even placed a large arrow to show the amplitude and frequency of it within the screen. I used alternator noise because "that" is what shops understand. Shops know what pre-amps or head-units are susceptible to alternator noise and by how much. Now, add the noise from many loaded class D amps in the system.
Finally, I used an amp without a load to eliminate the chance someone may misunderstand the measurements as being "after" the amp. I wanted to show that, since no speakers were present, the audible noise must be coming from some other place. The other place being the car's power system. Now that once you place a load on the amp (connect a woofer), the noise gets bigger and then more stiffening (not less) is needed. Again, the stiffening is needed to prevent the resulting noise from affecting components down or upstream.
I would like to thank JayLove for clarifying the many parts I missed in the video and for you to add value to this great forum.
Alberto Lopez
Former UK champion and forever indebted to the awesome people in the UK who adopted me for a few years while away from the US.
Alberto...you are an utter Diamond Geezer and just for the record folks... Alberto was, with Paul Richardson, part of the team that had the UK's ONLY EVER best-sounding winning install, that also was the UK's loudest at the time at 157dB SPL!!

Only Mark Fukuda in the USA has ever done this!

You are a STAR! Proud to count you as a friend, Sir!
:-))))