Quote:
Resulting in distortion figures 10x better and independent to load - indicates that a lot of measurements before were wrong - including Krell Evolution 600??

I don't believe so. There was only one connector affected.

John Atkinson
Editor, Stereophile

John, thanks for doing all that detective work on the DM38 Follow-Up. It is a lesson for all of us.

I enjoyed the Follow-Up, but I wish you could have published your usual 19+20 kHz IM spectra findings. I suppose they fell victim to the limited space allocated to a Follow-Up. Did you do the twin-tone measurements, and, if so, can you tell us how many dB down were the spectral components at 18 kHz (3rd order) and 17 kHz (5th order)?

Also, although your elimination of the banana plugs in the load arrangement made a very big improvement in the 1 kHz THD, it looks from your Figure 5 that the third-order distortion is real distortion from the DM38, at the 0.00072% value. This is down 103 dB. Although this is a good number, it is still way above the -120 dB or more that Halcro specifies at 1 kHz.

Thanks again,
Bob

Quote:
John, thanks for doing all that detective work on the DM38 Follow-Up. It is a lesson for all of us.

Thank you Bob. It was just one of the 8 or so dual-banana-fitted adaptor cables I use, but of course, that was the one that was used to measure both the dm88 and the active channel of the dm38. (For the spectral analysis and clipping tests, I only monitor one channel, even if both are being driven.)

Quote:
I enjoyed the Follow-Up, but I wish you could have published your usual 19+20 kHz IM spectra findings. I suppose they fell victim to the limited space allocated to a Follow-Up. Did you do the twin-tone measurements, and, if so, can you tell us how many dB down were the spectral components at 18 kHz (3rd order) and 17 kHz (5th order)?

To my regret, I did not do this test. I wanted to use two analog signal generators and a passive summing network (rather than the 24-bit digital signal generator and the very-low-distortion DAC I usually use) to ensure that all that appeared in the Halcro's output was due its behavior. I just ran out of time, I am afraid.

Quote:
Also, although your elimination of the banana plugs in the load arrangement made a very big improvement in the 1 kHz THD, it looks from your Figure 5 that the third-order distortion is real distortion from the DM38, at the 0.00072% value. This is down 103 dB. Although this is a good number, it is still way above the -120 dB or more that Halcro specifies at 1 kHz.

Yes. In the Web reprint -- www.stereophile.com/solidpoweramps/1004halcro/index5.html -- I could include the graphs that had to omitted from the paper book for space reasons. If you compare fig.8 in the Web reprint (paper fig.5) with fig.4, which shows the harmonic spuriae of my Audio Precision generator, plotted to the same scale, each with 32 readings averaged to reduce the contribution of noise, this does suggest that the Halcro is still introducing some 3rd harmonic (although it might be thought academic, or at least only of interest to designers such as yourself, given its low level).

Thanks for your thoughts, Bob.

John Atkinson
Editor, Stereophile

John, as always, thanks for your detailed response. You've got the patience of Job, both with equipment and people.

Here's a thought. Let's see if we can rustle up a Halco at HE 2007 and measure it in real time at the show with the special equipment I have. With my twin-tone generator and Distortion Magnifier, I can see down to below -120 dB on the twin tone test.

Best regrads,
Bob