With a purely resistive load, the frequency response variations drop to +/-5dB in the audio range. A trifle. That would probably help the WATTs that the reviewer (and John) use. Of course, so would $1 worth of passive parts, but that's not as fun of a story, and most passive parts won't give you the piles of higher order harmonics.
I did see an amp once that probably was worth what the Wavac folks are charging. Performance was considerably better, not outstanding but closer to acceptable, but the "coolness" factor was far, far higher. Built by one of the diyAudio regulars, Vinylsavor.
I did see an amp once that probably was worth what the Wavac folks are charging. Performance was considerably better, not outstanding but closer to acceptable, but the "coolness" factor was far, far higher. Built by one of the diyAudio regulars, Vinylsavor.
It is VERY SIMPLE. Just put a 4 to 5 ohm resistor in SERIES with the output you YOUR AMP, and you will have the same effect. It is called DAMPING FACTOR. It is so simple that it has alluded many here, including me, UNTIL I looked at the review. It is not magic, or careless design, it is an artifact of operating open loop. That is all there is to it.
No, it will not have the same effect unless the crappy amp's source impedance is constant with frequency and level. And, of course, unless you add all of the distortion. Sorry, John, you can't put a 5R on the output of a JC-1 and raise the price by a factor of 100. To quote Max Bialystock, "Nahhh, too good." 
As SY said - it's not just output resistance - because there is same unlinearity into the pure resistive load. I bet on some kind of CL coupling.
I presume that the effect of an "idiosyncratic" damping factor is not just to change the system's frequency response..?
I'm imagining loss of control of the speaker cone having interesting effects on complex signals and transients - effects that are not possible to quantify from testing with steady sine waves.
I'm imagining loss of control of the speaker cone having interesting effects on complex signals and transients - effects that are not possible to quantify from testing with steady sine waves.
Possible, but I think that you can parse what the amp is doing with a lot of sine wave measurements. If you measured the output impedance as a function of frequency, then repeated the measurement at 5 or 10 different power levels, you could then design an output network that simulated the dynamic variation that the bad amp gives, and add that dynamic distortion to your JC-1. You still won't have the overlay of other distortions (intermod and harmonic), but you'd at least get the dynamic variation in frequency response. It would be easier to use impulse testing or MLS, but you'd still need to repeat it at multiple power levels. Because this amp is much worse than the ones Carver was playing with (conrad johnson), a simple resistor on the output of a competent amp won't simulate the transfer function.
John may "give up," but if he puts his mind to it, he can stop designing the neutral, transparent amps for which he is famous and start making inefficient effects boxes. I have great confidence in his ability.
Just to be clear, it my understanding from the test article (as SY mentions) that the frequency response curves are done with:
a speaker simulator load, then with 4 different purely resistive loads.
I was assuming the speaker simulator load was done to try and give some rationale for the intentional bump which shows up during the resistive scans, (since the peaks seem to line up for some reason) but that's a big assumption.
For the scans with purely resistive loads I don't see how to get any frequency dependency with a purely resistive output impedance, simply forming a resistive divider.
Hope this helps
-Antonio
a speaker simulator load, then with 4 different purely resistive loads.
I was assuming the speaker simulator load was done to try and give some rationale for the intentional bump which shows up during the resistive scans, (since the peaks seem to line up for some reason) but that's a big assumption.
For the scans with purely resistive loads I don't see how to get any frequency dependency with a purely resistive output impedance, simply forming a resistive divider.
Hope this helps
-Antonio
How do you know that? Have you measured them both with an interestingnessmometer?
Disclaimer: I'm drunk now. Well, duh - why else would I read the crud in this thread? Is anybody here sober?
Funny, isn't it, that audioholics care about performance in direct correlation with the achievable accuracy and measurability of what they're measuring. DACs and ADCs are really very accurate, and conveniently measurable, so imperfections at -100dB are claimed to cause physical pain to the listener. Amplifiers are pretty good but not quite so predictable into real world loads, so 0.01% THD conveniently measured across a resistor rather than an inconveniently loud speaker load is regarded as the nausea-inducing benchmark. Subtleties of speaker performance are not really measurable, and while speakers reputedly introduce basic distortion of 0.1%-5% THD and big wobbles in the frequency response, these and more subtle characteristics are only measurable in an expensive anechoic chamber with an expensive microphone (which itself has unmeasurably subtle characteristics). Lo and behold, audioholics seem very laissez faire about speaker measurements, claiming to pick out tiny differences between amps but not really caring whether the speaker THD is 1% and frequency response flat +/- 6dB. And then there's the room characteristics: here, the science goes no further than recommending shuffling the speakers about until it sounds 'about right'.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.