90db 4 ohm with tubes?

Assuming you are referring to the current Diamond DMD version, the impedance curve shown in the review Samhar referred to indicates an impedance rise to around 28 ohms in the upper mid-range, some smaller impedance peaks in the deep bass region, and, according to the text of the review, 4 ohm impedance minima with "overall measured impedance [of] 6.5 ohms, so this is nominally a 6 ohm loudspeaker."

The rise to around 28 ohms in the upper mid-range will result with your amplifier in somewhat greater emphasis of those frequencies than would occur with nearly all solid state amplifiers, and also with some tube amplifiers (those having lower output impedance than I believe yours does). There will also be a bit more emphasis of frequencies in the vicinity of the two impedance peaks in the bass region.

I don't know whether or not with those particular speakers and your particular amplifier that would be good, bad, or something that comes down to a matter of taste and room-matching. But it's something to be aware of.

Regards,
-- Al

As a follow-up to my and Marakanetz' earlier posts, I found the following statements in John Atkinson's report on his measurements of the V12, which appears to be generally similar to the V12R aside from differences in the input section:

... allowing for some variation in the calculated output impedance with the level and load impedance used, we're looking at source impedances of 5.5 ohms (8 ohm tap) and 2.8 ohms (4 ohm tap) in triode mode. While these are both high, they are exceeded by the ultralinear figures: 12 and 6 ohms, respectively! Fortunately, the impedances don't change much across the audioband, but, as figs.1-3 reveal, there will be a large and audible change in frequency response depending on which loudspeaker is used with the Cary and which transformer tap and mode of operation are used....

... I was concerned by the highish levels of HF intermodulation and the very high source impedances, particularly in ultralinear mode. Unless you really need the extra 3dB of dynamic range you get from ultralinear operation compared with triode, and the extra 3dB you also get from the amplifier's 8 ohm tap compared with the 4 ohm tap, my advice is to use the CAD-280SA in triode mode with the speakers hooked up to its 4 ohm outputs.

Caution would seem to be in order here. I would not advise pairing such an amplifier with a speaker having such wide variation of impedance as a function of frequency, without first carefully researching the results others have obtained using those speakers with the same or similar amplifiers, AND carefully auditioning the speaker with that specific amplifier.

Regards,
-- Al

Dr. M, the concern is NOT the speaker's 4 ohm or whatever nominal impedance. The concern is with the wide VARIATION of that impedance as a function of frequency (from about 4 ohms to about 28 ohms according to the measurements of the particular speaker that were cited earlier, with the 28 ohm peak occurring in a critical part of the spectrum), and the INTERACTION of that VARIATION with the amplifier's high output impedance.

Yes, if the sonic results turn out to be poor "it won't be the amp's fault." It won't be the speaker's fault either. It will be the result of a mismatch.

Regards,
-- Al

Bruce, thanks very much for your comments. Charles and Hifitime as well. Here are my comments on some of what has been said:

06-10-13: Bifwynne
The ability of a tube amp to drive a speaker with wide impedance fluctuations with tight output regulation is a function of the amp's output impedance rather than the muscle of the power or output trannies.

Yes. Of course, if the impedance of the speaker reaches low values at some frequencies muscle, good trannies, and good power supplies are ALSO likely to be necessary.

06-10-13: Bifwynne
I think Al might concur with the surmise that if an amp's output impedance is higher, say 5.5 ohms, its output regulation would be much higher in dbs as a function of the speaker's impedance variations.

Yes, absolutely.

06-10-13: Bifwynne
It is my anecdotal understanding that there are a few speakers that have relatively flat impedance curves that would make them tube friendly.

Certainly. My Daedalus speakers are one example. I believe that the Coincident speakers, one of which Charles uses, are another. I'm sure many other examples could be cited.

I would add that widely varying impedance does not NECESSARILY mean that the speaker won't work well with an amp having high output impedance. For example, electrostatic speakers commonly have very high impedances at low frequencies, which descend to very low values in the upper treble region. Check out this curve for the classic Quad ESL, which, like many electrostatics, works well with tubes. Or, among dynamic speakers, check out this curve for the Harbeth M40.1, which a lot of people use with tubes with excellent results.

As Ralph (Atmasphere) has said, and this also addresses the point in Hifitime's first post above, it depends on the intentions of the designer, and tonal balance problems usually result when the amp and speaker are not of the same paradigm.

And in the case of a speaker having an impedance that is near 4 ohms across much of the spectrum, but rises to 28 ohms in the upper mid-range, it would seem to be a good bet that it was not designed to sound its best with an amplifier whose output impedance is even higher than that of most tube amps.

Regards,
-- Al

Hi Bruce,

That all sounds generally correct, but I'll add some qualifications:

1)Re the last paragraph, yes a 1.2 db fluctuation is pretty minor in comparison to room effects, and in your case I suspect is nothing to worry about. More generally, however, and particularly where the fluctuation may be significantly greater (as it would be in the OP's proposed configuration), I would keep in mind that a given variation in measured frequency response at the listening position, caused by room effects, may be subjectively much less objectionable than an identical frequency response variation caused by impedance incompatibilities in the system. At least in the situation where the in-room response is measured in the traditional simplistic ways that do not take arrival times into account.

The reason for that is the ear's ability to discriminate between "first arrival" sounds and reflected sounds that arrive subsequently. See this Wikipedia writeup on the Haas and Precedence Effects.

2)Re the ARC spec you quoted for "output regulation," it may help to clarify matters if I describe how I would calculate the approximate variation in db that would correspond to a damping factor of 8, and a load variation between 8 ohms and an open circuit (i.e., infinity ohms). On the other hand it may just confuse matters further, but I'll do it anyway :-)

(a)First, see this Wikipedia writeup on the Voltage Divider Effect.

(b)In the first figure, consider Vin to be the voltage the amp is "trying" to put out at some instant of time (i.e., the voltage it would output under open circuit/no load conditions (hypothetically speaking, as of course a tube amp that has an output transformer should not be run unloaded)).

(c)To keep things simple(!), let's assume that the speaker impedance and the output impedance of the amp are purely resistive.

(d)Think of Z1 as the output impedance of the amp, and Z2 as the impedance of the speaker.

(e)For an 8 ohm tap, or for an amplifier having no output taps, damping factor is normally defined as output impedance divided into 8 ohms. So the damping factor of 8 for the VS-115's 8 ohm tap means that the output impedance is about 8/8 = 1 ohm (much lower than for the OP's amp, and in fact lower than for the majority of tube amps).

(f)Referring to the Wikipedia page on the voltage divider effect, under no load conditions Z2 is infinite, no current will flow through Z1 (because a complete circuit is not present if Z2 is infinite), therefore the voltage drop across Z1 will be zero, and therefore Vout = Vin.

(g)If Z2 were 8 ohms, based on the voltage divider effect Vout = (Vin) x (8/(8+1)) = 0.888Vin

(h)Voltage ratios are converted to db as 20 times the logarithm of the voltage ratio.

20log(0.888) = -1.02 db.

So the output variation from an 8 ohm load to an open circuit, assuming a damping factor of 8 and based on the oversimplified assumption that the speaker impedance is purely resistive, would be 1.02 db, a little less than the stated figure.

That value would be somewhat worse, of course, from 4 ohms to an open circuit, if the 8 ohm tap were used. Using the 4 ohm tap on the other hand, which probably has an output impedance that is around half the output impedance of the 8 ohm tap, would of course reduce the variation significantly.

Finally, although you most likely have this in mind, it should be noted that a variation of 1.2 db is only half as much as +/- 1.2 db.

Best regards,
-- Al

Hi Ralph,

The manual for the V12, which appears to be very similar to the OP's V12R except for differences in the input section, indicates that it is a zero feedback design. (The manual for the V12R is silent on that question, though). Also, JA's output impedance measurements of the V12, and his measurements and statements about the sensitivity of its frequency response to load impedance, seem very consistent with the use of little or no feedback.

Best regards,
-- Al