JBL JRX518S -- help me measure output without blowing things up

[xtian]

Gentleman says his JBL JRX518S sounds distorted, but assures me it's not the 18" driver. He gave me the complete backplate amp chassis, but did not bring the cabinet and speaker. Inside, everything passes visual inspection. I powered it up, noting it draws ~20 watts at idle, and played music into my guitar speaker cabinet--sounds fine at low volume, up until everything in my office starts rattling. So my next step is to play a 100Hz sine wave into my 4R7/1000w resistive load.

But! I'm wary of scoping differential amps, because my scope is earthed, and thus the probe's ground. In the JRX518S schematic (snip below, see page two of PDF for output schematic), it looks like the output is "center referenced" to ground. I'm wary of scoping that.

How can I proceed safely?


Full schematic:
JBL_PRX518S.pdf

[Enzo]

You don't connect the scope GROUND probe to either side of a bridged output, both sides are hot. But both sides are just amps, they just happen to work together. So ground your scope to ground and look at the two sides individually

[xtian]

Thanks, Enzo!

[Enzo]

That is what I do in service. If as a final check you really want to see the whole thing, use two channels of scope. One to either side, then set the scope to ADD so they both display together

[xtian]

This isn't working like I expect. First, I measured for DC on the output, and found only 0.020vDC. I am feeding a 120Hz sine wave at 1.0v p-p into the input jack, and turning up the Level control. I have the output clipped to my resistive load, and one 'scope probe clipped to one leg of the load. The output is very small, only reaching 3 or 4v p-p, but it show lots of high frequency hash. You can see the 120Hz sine in the first photo, second photo is zoomed into the details.

What am I doing wrong?

[xtian]

I'm using AC coupled mode. That's correct, isn't it?

[Jon Snell]

"but it show lots of high frequency hash." It is an FM amplifier, running at around 500kHZ. There will be full square wave rail to rail on the output before the common mode choke and a lot less HF after the choke. That is your HF hash.
Being class D (or almost), it will either work or it won't.
I would look before the MosFet drivers or even reduced rail voltages perhaps.

[Jazz P Bass]

If you simply want to measure the output power, ditch the scope and use a DVM.

[xtian]

OK, here are my results. First, I notice that at idle, the amp oscillates at 7.3Hz (yes, subsonic) at 1.6vRMS. Not an issue, I guess.

This time I chose a 60Hz sine wave. I had to dial it up past 2.1v p-p at the input before seeing any distortion, then back to 2.1v where the output wave is clean. Although I was able to use the Math menu on my Hantek DSO5072P to visualize CH1+CH2 (had to invert CH2), the Measure function does not show the results for the combined probes. But both probes show equal looking waves, so back to using just CH1: I got 42v p-p max output before clipping. Strangely, the "Cyclic RMS" readout on the Measure function reads 14.4vRMS. I say strange, because I calculate RMS as 42 X 0.7071 = 29.7vRMS. Which one is right?

Extrapolating to both rails, can we say 2 X 42v p-p = 84v p-p, and 84 X 0.7071 = 59.4vRMS? Into 4.7R load, that would be 750 watts, which doesn't seem likely.

As an additional data point, I used my Fluke AND Meterman trueRMS multimeters; both read 24.4vRMS. That's only 127 watts.

Which numbers to believe?

[Helmholtz]

If you simply want to measure the output power, ditch the scope and use a DVM.

Only a scope can tell if the output is clean, distorted or full of HF.

BTW, I don't think that much switching frequency in the ouput signal is normal or even allowable by EMI regulation.

[Helmholtz]

I got 42v p-p max output before clipping. Strangely, the "Cyclic RMS" readout on the Measure function reads 14.4vRMS. I say strange, because I calculate RMS as 42 X 0.7071 = 29.7vRMS. Which one is right?

Extrapolating to both rails, can we say 2 X 42v p-p = 84v p-p, and 84 X 0.7071 = 59.4vRMS? Into 4.7R load, that would be 750 watts, which doesn't seem likely.

As an additional data point, I used my Fluke AND Meterman trueRMS multimeters; both read 24.4vRMS. That's only 127 watts.

Which numbers to believe?

RMS voltage is peak voltage times 0.707. Peak-to-peak voltage is twice the peak voltage for a sine.
So RMS = Vp-p/2 x 0.707.
But the conversion only works for perfect sine signals without any HF hash, which will increase Vp-p but not signal RMS.

I think both the scope and the RMS meters measure correctly. Difference is that the wideband scope includes the HF while the DMMs don't register the HF.
For signal power calculation use the RMS meter readings between the ends of the load.

[xtian]

For signal power calculation use the RMS meter readings between the ends of the load.

And that shows 127 watts. Not great. But this is confirmed observationally, by the fact that, after running for a couple minutes, my 1000w resistor was barely warm to the touch.

[J M Fahey]

And that shows 127 watts. Not great. But this is confirmed observationally, by the fact that, after running for a couple minutes, my 1000w resistor was barely warm to the touch.

Some JBL powered cabinets, this one included, use 1 ohm voice coils.
I already knew it but for confirmation searched and found:

Brand New Genuine JBL 352938-009X 18″ Woofer Model 2044E used in the PRX518S Subwoofer. 1 ohm.

Millennials are sissies, they were born in the 5V and even 3.3V Logic and Arduino World and are scared to Death by High Voltages! (anything above 12V, the highest found inside a Computer) so they make 800W RMS amps fed from 32V supplies. 36V tops.

Which FORCES them to use Class D only (no big deal, they were born with that) , bridged outputs only and as-low-as-needed impedance voice coils.

That same plate amp will provide 4x127=508W RMS into 1 ohm and if you allow for 10% distortion (which is clipping territory), some 600W RMS

EDIT:curious,searched around.
Not very encouraging reliability reports:

Great sound - not so sure about the quality
By Curt from Cedar Rapids, IA on December 1, 2009 ... We purchased tow of these for my sons band about a year ago and have had two issues so far. First was a strong burned circuit card smell- sent it in with no fault found. Next, one unit simply stopped powering up, sending in under warranty now. They have not seen much in the way of live shows to date.

Questionable reliability of the amp module
By Maung Chit Khaing from California on August 29, 2009 .... Crown amp module in one of my 518 subs went dead after the second gig.

Not Happy!
By Kevin McLean from Eureka, CA USA on December 4, 2009 .... We've lost 3 in about 4 months. They just stopped powering up. Something seems amiss.

sweetwater is awesome
By Billy Neal from charleston, ar on November 16, 2009 ...... we have a small church and our system is never pushed very hard, ..... first subs crown amp bit the dust after 2 days. second one quit after about 4 months.

[Helmholtz]

As the 127W output was calculated from 24.4Vrms into a 4.7R load, the theoretical output into 1R is 595W.

[gui_tarzan]

Millennials are sissies, they were born in the 5V and even 3.3V Logic and Arduino World and are scared to Death by High Voltages!

LOL! Truth!