ProSoundWeb Community

Please login or register.

Login with username, password and session length
Advanced search  
Pages: 1 2 [3] 4 5 ... 7   Go Down

Author Topic: adavantage?  (Read 17625 times)

Nathan Short

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 624
    • http://www.myspace.com/nathanshort
Re: adavantage?
« Reply #20 on: February 20, 2010, 03:09:44 pm »

Damnit, I had a very well worded and long winded response, then my damn computer glitched.

So here to sum up again.  

Noise generator->midas Venice console->amp->load->2 types of multimeter B&K and Extech= Sencore Audio consultant to measure THD%

I actually love these amps, but have noticed in the past, that my subs on a few installs in which the system was properly limited, had the symptoms of 2 subs on channel 1 of a 3602  would be blown out of the basket. Not literally.  Surround partially ripped, spider partially torn, leads ripped off the input, slightly rubbed coil, but not burnt. Basically very fast Xmech damage.  The other subs would be fine on the other channel, and in most cases still work to this day.  I am always delighted to go do a service call 4 years later to find everything working fine even when they have DJ's 5 nights a week.  I try to make my systems very Idiot resistant.  

So, the real question is this.  Is this an isolated batch?  An individual component tolerance issue?  Has anyone else noticed this behavior?

Go ahead and try it.  But I will make damn sure that from now on, that little red clip light on channel 1  never lights up EVER.


Logged
AIS Chicago--System Designer

Freelance FOH/Monitor/System Tech/Stage Manager/
Stagehand Lead/Production Coordinator/Sound Enthusiast/Speaker Design

WWW.MYSPACE.COM/NATHANSHORT
WWW.Facebook.com/NathanShort

Nathan Short

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 624
    • http://www.myspace.com/nathanshort
Re: adavantage?
« Reply #21 on: February 20, 2010, 03:16:01 pm »

My purpose in widening the scope of my testing with tone into near limit on the amps is the growing use by us in electronic music.

Dubstep is changing the way we need to measure amplifiers as many of the bass lines are long droning synth lines.  

This will reveal a weak sub amp very quickly, and many amp manufacturers are stepping up the robustness of the amplifiers.

I would hate to be doing a show where the DJ or Electronic musician played a high note, or used a "sweep" effect while performing, and 1- I nuke my drivers on Ch 1, 2 - I blow half my PD.  Who knows?  This is why I try to abuse the amps any way I can think of before I let the professional Idiots beat them up in my installs.
Logged
AIS Chicago--System Designer

Freelance FOH/Monitor/System Tech/Stage Manager/
Stagehand Lead/Production Coordinator/Sound Enthusiast/Speaker Design

WWW.MYSPACE.COM/NATHANSHORT
WWW.Facebook.com/NathanShort

Art Welter

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 4874
Re: adavantage?
« Reply #22 on: February 20, 2010, 03:26:36 pm »

Nathan,

You wrote previously:

“The most interesting thing was finding a flaw in the PLX3602 that at 30 Hz and 15K channel 2 would behave normally and channel 1 would try and put out 81v!!! Danger Danger! and the amp would go into overdrive and try and pull 33amps from the wall.”

Did you mean 30 Hz and 15K simultaneously, or either tone separately caused the runaway?
Logged

Nathan Short

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 624
    • http://www.myspace.com/nathanshort
Re: adavantage?
« Reply #23 on: February 20, 2010, 03:29:02 pm »

Detailed notes about voltage and current draw over time, were taken at Pink 6db crest, 30Hz, 50Hz, 1k, 15k, Dance Music, Dance Music into hard limit.  Separate.
Logged
AIS Chicago--System Designer

Freelance FOH/Monitor/System Tech/Stage Manager/
Stagehand Lead/Production Coordinator/Sound Enthusiast/Speaker Design

WWW.MYSPACE.COM/NATHANSHORT
WWW.Facebook.com/NathanShort

John Roberts {JR}

  • Newbie
  • *
  • Offline Offline
  • Posts: 0
Re: adavantage?
« Reply #24 on: February 20, 2010, 04:34:10 pm »

Nathan Short wrote on Sat, 20 February 2010 14:16

My purpose in widening the scope of my testing with tone into near limit on the amps is the growing use by us in electronic music.

Dubstep is changing the way we need to measure amplifiers as many of the bass lines are long droning synth lines.  

This will reveal a weak sub amp very quickly, and many amp manufacturers are stepping up the robustness of the amplifiers.

I would hate to be doing a show where the DJ or Electronic musician played a high note, or used a "sweep" effect while performing, and 1- I nuke my drivers on Ch 1, 2 - I blow half my PD.  Who knows?  This is why I try to abuse the amps any way I can think of before I let the professional Idiots beat them up in my installs.


There is nothing wrong with sine wave (tone) testing on amps, loudspeakers not so much.

I'm sure Bob Lee will sort this out but 81V sounds a little hi for a single channel of 3602 under load, unless square waved.

Excessive current at HF can be caused by a condition called mutual conduction, when both the pull up power devices and pull down power devices are pulling both directions at the same time. This was an issue with early mosfet power amps (due to poor driver stage design) but not common in modern power amps. If such an amp doesn't trip mains breakers it could easily release device smoke, since the power devices are working up a good sweat playing tug of war with each other.

I doubt this is happening inside a properly functioning PLX, just saying what the symptom sounds like.

JR
Logged
 https://www.resotune.com/


Tune it, or don't play it...
-----

Bob Lee (QSC)

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1788
Re: adavantage?
« Reply #25 on: February 22, 2010, 02:18:30 pm »

Sorry Nathan, but when I read the thread it seemed that your "test" was (and is) not described in any detail.

Do you mean the gain pots on the RMX2450? There are no pots on the inputs. The gain pots are a weak spot on the RMX amps; it's probably good that they don't also have input pots.

Your description of "overdrive" is still sketchy. Can you describe your test setup, what you were measuring, etc.? "Jumped to 81V" and "drew 33 A" doesn't tell us much. I can make a perfectly good PLX3602 put out 81 V and draw 33 A too. Can you describe what the amp did versus what it was supposed to do?
Logged
Bob Lee
Applications Engineer, Tech Services Group QSC Audio
Secretary, Audio Engineering Society
www.linkedin.com/in/qscbob

"If it sounds good, it is good."
 -Duke Ellington

Nathan Short

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 624
    • http://www.myspace.com/nathanshort
Re: adavantage?
« Reply #26 on: February 22, 2010, 02:38:12 pm »

Never mind Bob,

I don't have time to argue semantics over my quickly worded posts.

Everyone knows that the potentiometers on the RMX series blow, are cheap, degrade quickly, and otherwise ruin a perfectly good budget amp line.  I will never sell them for an install, but make plenty of money upgrading and fixing them then reselling to my friends.

It is wonderful To see that the PL340 and 380 series have wonderful current management and do what they are supposed to do.

It is very odd that I can easily, get one channel of a 3602 to put out 81 volts while the other measures perfectly,  and that the amp even tries to pull 33amps from the wall.  

This is my only question.  How and why is the 3602 able to do this.  Seems pretty dangerous, and a complete failure of the "clip limiter" circuit in the amp. Whatever that might be.
Logged
AIS Chicago--System Designer

Freelance FOH/Monitor/System Tech/Stage Manager/
Stagehand Lead/Production Coordinator/Sound Enthusiast/Speaker Design

WWW.MYSPACE.COM/NATHANSHORT
WWW.Facebook.com/NathanShort

Mac Kerr

  • SR Forums
  • Hero Member
  • *
  • Offline Offline
  • Posts: 10223
Re: advantage?
« Reply #27 on: February 22, 2010, 02:55:28 pm »

Nathan Short wrote on Mon, 22 February 2010 14:38

Never mind Bob,

I don't have time to argue semantics over my quickly worded posts.

Everyone knows that the potentiometers on the RMX series blow, are cheap, degrade quickly, and otherwise ruin a perfectly good budget amp line.  I will never sell them for an install, but make plenty of money upgrading and fixing them then reselling to my friends.

It is wonderful To see that the PL340 and 380 series have wonderful current management and do what they are supposed to do.

It is very odd that I can easily, get one channel of a 3602 to put out 81 volts while the other measures perfectly,  and that the amp even tries to pull 33amps from the wall.  

This is my only question.  How and why is the 3602 able to do this.  Seems pretty dangerous, and a complete failure of the "clip limiter" circuit in the amp. Whatever that might be.


This is a pretty cheap shot that basically puts your credibility at zero. Bob asked a couple of questions about your testing methods that you have never explained.

Bob is not the only one who can't figure out what you are talking about.

Mac
Logged

Nathan Short

  • Hero Member
  • *****
  • Offline Offline
  • Posts: 624
    • http://www.myspace.com/nathanshort
Re: advantage?
« Reply #28 on: February 22, 2010, 03:31:08 pm »

I don't see how it is a cheap shot, just the truth.  

The quick and easy method to reproduce my findings.

Take an oscillator, Run it into a midas venice, set the level to 0.
Buss the channel.  Run Masters to 0.  Masters are plugged directly into the amplifier.  Amplifier was tested on 2 dummy loads. I will explain the first.  A 4 element 4x 7.8-8.2 ohm 2000w water heater elements suspended in a bucket of water.  This provides an easy impedance that doesn't seem to drift more than .2ohm when hot.

Test leads from a BnK multimeter were attached to the individual leads one one of the resistors , and at the amp different times.

The amp was plugged into a military grade distro, with a current meter.  The distro was plugged into two seperate 100amp breakers, 1 per circuit of the 2 circuit unit.  The unit has 40amp breakers.  both circuits went back to the same panel ground, and seperate neutral bus.

Channel 1 was found, when starting to engage the clip led by driving the output of the midas past +4, to start putting out a lot more voltage than Channel 2.  Channel 2 behaved normally like almost every other amp at the test.  Channel 1 could be made to put out 81v when the clip light was moderately active.
The current meter on the Distro showed 33amp draw.

What else can I tell you.  A few of my EE friends, as I am not an EE, gave the testing a thumbs up, and a great way to check for thermal limits also.  

The 3602 was found to exibit this behavior at 30Hz and 15kHz, I did not push the test further for fear of ruining the amp.
Logged
AIS Chicago--System Designer

Freelance FOH/Monitor/System Tech/Stage Manager/
Stagehand Lead/Production Coordinator/Sound Enthusiast/Speaker Design

WWW.MYSPACE.COM/NATHANSHORT
WWW.Facebook.com/NathanShort

John Roberts {JR}

  • Newbie
  • *
  • Offline Offline
  • Posts: 0
Re: advantage?
« Reply #29 on: February 22, 2010, 04:10:36 pm »

Nathan Short wrote on Mon, 22 February 2010 14:31

I don't see how it is a cheap shot, just the truth.  

The quick and easy method to reproduce my findings.

Take an oscillator, Run it into a midas venice, set the level to 0.


doesn't matter
Quote:


Buss the channel.  Run Masters to 0.  Masters are plugged directly into the amplifier.


bus.. Console outputs, I presume. Same as level coming from the Oscillator?  IIRC Midas uses peak meters, but I'm not sure how they calibrate for sine waves.  Probably doesn't matter.
Quote:


Amplifier was tested on 2 dummy loads. I will explain the first.  A 4 element 4x 7.8-8.2 ohm 2000w water heater elements suspended in a bucket of water.  This provides an easy impedance that doesn't seem to drift more than .2ohm when hot.


nominal 2 ohm load.. one on each channel I assume.
Quote:


Test leads from a BnK multimeter were attached to the individual leads one one of the resistors , and at the amp different times.


Average AC volts (even if RMS it doesn't tell us the full story since it's clipping).
Quote:


The amp was plugged into a military grade distro, with a current meter.  The distro was plugged into two seperate 100amp breakers, 1 per circuit of the 2 circuit unit.  The unit has 40amp breakers.  both circuits went back to the same panel ground, and seperate neutral bus.


I assume nominal 115-120V mains?
Quote:


Channel 1 was found, when starting to engage the clip led by driving the output of the midas past +4, to start putting out a lot more voltage than Channel 2.  Channel 2 behaved normally like almost every other amp at the test.  Channel 1 could be made to put out 81v when the clip light was moderately active.
The current meter on the Distro showed 33amp draw.


81V/2 ohm is 40A. A perfect amp would require 24A line cord draw for just this one channel, if mains were stiff 120V. Since the amp isn't 100% efficient (even class D amps aren't) the mains draw needs to also supply current to account for losses.

A clip light being "moderately" active on a sine wave means IT'S CLIPPING.  
Quote:


What else can I tell you.  A few of my EE friends, as I am not an EE, gave the testing a thumbs up, and a great way to check for thermal limits also.  

The 3602 was found to exibit this behavior at 30Hz and 15kHz, I did not push the test further for fear of ruining the amp.


My EE friends would ask you for more data.. like current draw and output voltage at 1kHz for the same input voltage.

I wouldn't be surprised to see clipping earlier at 30Hz than 1kHz. More data might help reveal a problem (or not). In general the clip light being on, makes the actual (clean) output voltage rather difficult to know from a clipped waveform measurement.  

If your EE friends approve voltage measurements with clipped waveforms, why stop at only moderate clipping? You could double the power with a full square wave.

JR

PS: Inexpensive pots in a value amp... I'm shocked.  Shocked
Logged
 https://www.resotune.com/


Tune it, or don't play it...
-----
Pages: 1 2 [3] 4 5 ... 7   Go Up
 

Page created in 0.027 seconds with 22 queries.