diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Solid State (http://www.diyaudio.com/forums/solid-state/)
-   -   1200w+ Amp Shootout March 27-28 (http://www.diyaudio.com/forums/solid-state/27835-1200w-amp-shootout-march-27-28-a.html)

Bink 9th February 2004 12:10 AM

1200w+ Amp Shootout March 27-28
 
Hello, All -

I'm organizing an Amp Shootout for serious touring amplifiers -- at least 600w per channel into a pair of 4 ohm loads or 1200w into 8 ohms bridged. This event is March 27-28 in Emeryville, California (on the eastern edge of the San Francisco Bay) and will take place at a small art college called Expression Center for New Media.

My questions have to do with testing and analysis, not amp design. I'm sure the folks here are better skilled at analysis than I am! As a freelance live audio mixer who's a complete analysis noob I need all the help I can get.

At the Amp Shootout I intend to monitor AC current consumption and thus calculate amplifier efficiency but the standard clamp ammeter is too slow to respond to brief surges. And I mean BRIEF -- at least one of the amplifiers taking part is able to draw 125 amps (QSC PL9.0) at full power but will self-protect if this condition continues for longer than a few milliseconds. What kind of ammeter samples quickly enough and accurately graphs current surges over time? PC control of such an ammeter would be ideal.

Another focus of mine is to examine the changing AC harmonics which result from a large amplifier pulling very hard on the AC supply. Will a cheap Mouser step-down transformer from 120vac to, say, 12vac help me in fabbing a project box designed to turn AC line voltage into audio line level signal for real-time analysis? Or will such a cheap solution alter my frequency response so much that it won't be useful? I want to end up with a representation of AC line voltage but have it isolated and padded down to about a half a volt so that my IBM laptop can accept the signal into its minijack line level input. I have an FFT analysis program on the laptop which looks at frequency changes (SmaartLive) so I'm good to go once I get AC turned into audio.

FYI, the output of the amplifier under test is going to be analyzed by one of the best rigs in the world: the Audio Precision 2722. Testing trial procedure is still being developed but the first Shootout day I intend to test amplifiers under heavy load and under low-voltage conditions. On Day Two randomized double-blind istening tests will be conducted with audience responses tabulated. Reference speakers are the SPL TD-1 (multi-entrant horn-loaded mid-high) and the SPL B-DEAP32 (V-drive folded horn/boundary dependent subwoofer) which are reknowned in the concert audio field for their clarity, linearity and lack of distortion artifacts. The 200-person listening room (Meyer Sound Performance Space) has 600A of 3-phase power and boasts damped acoustics suitable for recording studio work. All DIY-ers who can attend are invited. RSVP binkster@binkster.net.

Load resistors: After researching the subject (and perusing the posts here on DIY) I am forced to concede that expensive low-inductance resistors will be necessary. If the testing were more casual I could just use a bunch of 4500w 12.8 ohm 240vac water heater elements paralleled into ~4 ohm and ~2 ohm loads. Water cooled, of course! :cool: However, the Audio Precision 2722 will be faced with inductive back-EMF which will mess with the results if I use heater elements as loads. Instead, what I intend to do is have low-inductance resistors made for me by Milwaukee Resistors. Their most appropriate model is 2kW, 20 ohms, 36 uH -- it costs $85. Ouch. I would need twenty-four of these to make a pair of 1.67 ohm loads but the side benefit is that the loads would each dissipate 24kw. A cheaper solution would be to use just 12 of these resistors, 6 per channel, and when it comes time to load the amplifier down below safe impedance to examine output under protection mode then I could bridge it and use 6 resistors at 3.33 ohms which amounts to 1.67 ohms per channel. Why so low? A few amplifiers I'm looking at use a protection scheme which kicks in below 1.8 ohms. Too many end-users operating professional concert systems are hooking up stereo 2 ohm loads for me NOT to examine low-impedance performance. A 2 ohm subwoofer load will dip below 1.8 ohms many, many times during the show.

AC for International users: Since I'm faced with a 120vac 3-phase power distro that offers 208v between any two hot legs, how can I achieve valid results for international users who will be operating at 230 or 240v? I need a high-power transformer that can get 240v from my 120v source. 100A capacity would do the trick.

Low voltage AC: What's the best way for me to conduct low-AC voltage tests? Yer standard variac can only handle 2kw. In catalogs I've seen bigger ones that can handle 50A but since I have a few amplifiers that can pull tons of power at low voltage I need a 200A solution. What device do I need in order to bring 120vac down to 90v? I'm guessing there's some kind of 200A transformer that has stepped output voltage levels available via taps. I would be happy enough with 10v steps but 5v steps would be better.

Wow. That's a ton of questions. I know I'm pushing the envelope here but if there's a reader who is familiar with industrial power then I'd appreciate any help you can offer.

-Bink
http://www.binkster.net/index.shtml

Eva 9th February 2004 04:36 AM

Current consumption measurements and analysis are easy to perform with a current transformer

The purpose of the transformer is to scale down the instanteneous mains current and get it as a voltage waveform with galvanic isolation so you can perform FFT, THD and other measurements on it

In a transformer with two windings, current in each winding is inversely proportional to its number of turns provided that magnetizing current is low enough to be negligible

To get negligible magnetizing current, transformer must be operated well below its rated volts-turn [half or less]

So all you'll need is a small 50Hz toroid and a load resistor of the right value, as an example I'll take some 120V to 12V 50VA lighting transformer

First, you have to find the number of turns for each winding by adding an additional coil of known number of turns to the transformer, connecting the primary to a known 50-60Hz AC voltage source and measuring voltage ratio between source and the additional winding

Lets say it has 800T primary and 80T secondary

Then, pass one of mains wires throug the toroid [one or two turns] and select the maximum current you want to measure, for example 100Arms sinweave

Calculate the maximum scaled rms current, in this case [assuming 2 sense turns] : 100A * 2 / 800 = 0,25 Arms

Calculate a load resistor for the 800 turn winding so the transformer is operated at a low volts per turn :

With 0.25Arms we have to get 60Vrms or less in the 120Vrms winding so resistor value would be about :

50/0.25 = 200 ohms [ie : 100+100 ohms in series]

Resistor should be rated at 0.25^2 * 200 = about 12.5 watts or more [10W + 10W should run not too hot]

Well, with 1A instantaneous mains current we should get 1 * 2 / 800 = 0.0025A in the 800T winding, that should cause 0.5V voltage drop in the 200 Ohm load, so we have the mains current waveform sampled at 0.5V / A

In the 80T winding, voltage is 10 times lower so we have 0.05V / A

So 125A instantaneous current will be read as 5V peak in the 80T winding, ready to be connected to the analyzer or the oscilloscope

I prefer to use the primary for load resistor and the secondary for measuremens due to less leakage inductance error

There are other alternatives for current measurements, like hall-effect sensors, but current transformers are easier and cheaper to use for mains current

Bink 9th February 2004 04:51 AM

Toroidal xfrmr current measurement
 
Cool, Eva, you give very good hints and suggestions.

You say I'll get less inductive leakage if I use the primary for load resistor and the secondary for measurement -- good point. This helps me get the best frequency response.

Can you tell me the benefit of having two sense turns of the mains through the toroid rather than one?

Thanks -

-Bink
www.binkster.net

Eva 9th February 2004 07:47 AM

More turns mean better coupling [less leakage inductance] as long as they are spread over the entire toroid

Also, you may use two or three turns as a trick to help getting the desirded current scaling as long as transformer VA rating is not exceeded [when currents get too high you have to take into account winding resistance as part of the load resistance since it's no longer negligible]

Anyway, the dominant leakage inductance will be the one found in mains line and power transformers so the setup I described usually has far better frequency response than te mains line itself

Most current consumption of transformer + rectifier supplies is concentrated at 3*F, 5*F and 7*F, falling very quickly for higher armonics

Anyway, current consumption on the field with live music is not so high. A year or two ago, I measured current consumption on a live performance using 4x QSC EX4000 loaded with 4+4 ohms [16x 18" 8 ohm drivers horn loaded] and 4x QSC EX1600 loaded with 8 + 8 ohms [16x 10" 16 ohm drivers horn loaded plus 16x 1" exit 16 ohm compression drivers]

With amplifiers clipping a little, peak current was no more than 40A for the entire system and average current was no more than 10A [using 230V line since I live in Europe]


All times are GMT. The time now is 02:27 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2