I used a calibrated microphone and professional 1/6th octave RTA, so I don't really care about your opinions about a minimum standard.
Ron’s New System
- Thread starter Ron Resnick
- Start date
Nobody said it couldn't be done, but bass isn't the top discipline of a set amp. It can be quite enjoyable if the bass is more powerful and slower; it depends on the speaker. If you do an impulse measurement, it looks disastrous if you demand power. As long as you don't add more than 3% thd, it's hardly possible because the bass already has plenty of it.
As long as the impedance is relatively straight and not too low-impedance, it will work. If the speaker's impedance curve looks like a bumpy track, that's not a good idea.
As long as the impedance is relatively straight and not too low-impedance, it will work. If the speaker's impedance curve looks like a bumpy track, that's not a good idea.
Depends completely on the quality and size of the output transformer how much distortion is being generated in the bass. Many lower performance SETs have a problem but the best ones do not.
Then supply the relevant details, including loudness levels and distortion of the measurements.
No, it mainly depends on the speaker's impedance curve. For example, if you have an impedance peak in the bass at 50 Hz (10 ohms), at 80 Hz (4.5 ohms), 50 Hz will be significantly louder. If you also have room mode there, it will sound badly unbalanced. Add a little feedback to the amp, and you won't hear it. This is independent of how good your transformer is.
P.S
Feedback is not bad, it depends on the dose
Last edited:
Your example won’t make a big difference to an amp with <1 ohm impedance, like a properly designed SET will have. I have never had issues with my frequency response. Feedback is like lead or mercury, no safe amount…No, it mainly depends on the speaker's impedance curve. For example, if you have an impedance peak in the bass at 50 Hz (10 ohms), at 80 Hz (4.5 ohms), 50 Hz will be significantly louder. If you also have room mode there, it will sound badly unbalanced. Add a little feedback to the amp, and you won't hear it. This is independent of how good your transformer is.
P.S
Feedback is not bad, it depends on the dose
Last edited:
Ron, I presume your friend was auditioning different speakers and amplifiers and DACs because he is considering changes to his system. I may be wrong.
If I am correct, do you plan on making any comments about these various visits or were you simply announcing on your system thread that you’ve heard a lot of different gear at a friends house?
If I am correct, do you plan on making any comments about these various visits or were you simply announcing on your system thread that you’ve heard a lot of different gear at a friends house?
The black line shows a loudspeaker dummy load; in reality, it can look much worse. That's why you hear statements like a set amp sounds so lively because it simply plays louder, for example, graphics at 1-2 kHz because of the impedance peak.
Exsample lamm amp
It gets much worse in the bass range, where, for example, the port tuning frequency and driver (fs) of the bass reflex speaker are located.
Exsample
Last edited:
Those frequency response fluctuations are for a more standard type of speaker with complex crossover. Also, notice the +-2 dB range, far less than most speakers in-room response anyway. Besides, you have switched the topic from non-linear distortion from transformer saturation to linear distortion of frequency response...why? It is also clear that the Lamm does just fine down to about 10Hz and so should perform fine in the bass.
Well, once we put numbers things become more interesting. High-efficiency speakers have typically high impedance and should be used with the 8 or 16 ohm taps that have tipically higher than 1 ohm impedance - even Lamm's ML2.1 that get low output impedance using " a little overall negative feedback" have 1.2 and 2 ohm respectively. And avoiding feedback would increase output impedance.
It seems the great majority of SETs are not so well designed ...
BTW, the nice Stereophile frequency response down to 10 Hz shown by Stereophile are taken at 1W at 8 ohm 2,83V. Are they assuming SET owners have 110 dB/W speakers?
As this is Ron's system thread, has he not mentioned many times that he loves the sound of Lamm SETs? I assume he would share such an opinion because he has surely heard the Lamm SETs in various systems in which they were properly matched to the speakers. I heard the ML2 matched with the wrong speakers, my Magico Q3, and yet, and yet, I heard a magic in the midrange that convinced me to replace those speakers with more sensitive ones, 105 dB, 16 0hm. A very good match.
And why did you use the ML3 with the 16 ohm tap as your example? Who uses the 16 ohm tap? If you look at the rest of the measurements, the 8 ohm tap has <2dB swings and the 4 ohm tap has < 1 dB swings. Not sure why you cherry picked to try to take the worst case. If you had tougher speakers, then the 4 ohm tap would probably be the best choice (although in practice I have found the 8 ohm tap always seems to sound better) and then you have rather inconsequential swings.
BTW, if you look at the ML2.2 measurements, you will see that the swings are even smaller than the ML3, although the ML2.2 has a couple of dB negative feedback to improve that and the measurements shown in STereophile for the ML3 are without feedback.
It was just an example to show what happens with a non-feedback set amp when the impedance fluctuates. A simple trick on the speaker results in better bass and significantly less strain. This gives you more dynamic headroom and a significantly higher achievable volume. The OPT saturates later. Do you want to know the trick or are you looking for a confrontation? The fact that the lamm amp is pure coincidence—the first graphic I found when searching for images on Google to illustrate the problem.
(...) The fact that the lamm amp is pure coincidence—the first graphic I found when searching for images on Google to illustrate the problem.
The fact is that comprehensive and reliable data on SET measurements is rare and all the time we are coming to Lamm ML2.X - expensive and rare units that can't be considered the "typical" SET. If we go to other brands, we have values around 2 ~3 ohm or higher, suggesting very low effective damping factors - technically a critical point for bass sound quality. Hard to create general rules in such circumstances.
Anyway, nice to know some SET lovers consider that feedback is acceptable in amplifier design and it is not a damned and cursed technique.
The class Ds have a distortion signature that prevents higher ordered harmonics being unmasked- much like any good quality tube amp, but lower distortion (so more transparent) overall. We have enough customers that have ditched their SETs to know that it works. The goal was to make the amp sound like music which is the same goal we have with our tube amps. Why make something sound like an SET when you can do better??
At some point you might try and see what happens when you filter out some of the bass. The Lamm OPTs are clearly much larger than an 18 Watt amp would otherwise need (compare to the size of output transformers in a Dynaco ST35 for example) so its obvious that Victor was trying to increase low frequency inductance- evidence he knew very well the problem I've been talking about!
But the problem you're up against is not merely that the inductance is much lower at 20Hz than at 100Hz (and thus the load impedance on the power tube is lower as well, causing it to heat up as if it did not run hot enough already); the load line (a theoretical concept to set the tube's operating point) is elliptical at bass frequencies, which causes the amp to make quite a lot of distortion. By filtering out bass the amp will simply do its strengths that much better.
One of the few things upon which we agree, but for entirely different reasons.
But not at full power, and the bass was causing the amp much more distortion, which impedes detail. To do full power your output transformer would be the size of a hotel room refrigerator and would have no HF bandwidth.
Notably, its the bass region where most of the power of musical information lies. There's a lot of lower frequency content that most people don't even know is there, and this is worse if you play LPs due to warp.
The SETs that Ron was using at last report actually runs dual output transformers driven by independent power tubes with their resulting outputs paralleled. That technique helps limit the size of the output transformer while still allowing for greater power. Its clear that the designer is well aware of the bass limitations I've been talking about.
All lower performance SETs have a problem and so do the best ones. You can make outlandish claims like this but they are only that and nothing else; when the rubber meets the road there's no SET ever made that can play 20Hz at full power (clipping). That's hard enough for a regular PP tube amp to do!
@Ron Resnick Have you been able to set up the bass filter for your amps as we discussed?
...if improperly applied. Do it right and it works really well. The trick, in case you're interested, is to do it in the same manner as seen with opamps, which is to say you treat the amplifier as an opamp and apply feedback externally around it. This technique prevents the feedback signal being distorted by non-linearities at the feedback node, since resistors are far more linear than active devices such as the cathode of a tube or base of a transistor.
In the tube world, something called 'Schade feedback' is a similar concept.
On any tube amplifier, the OPT is less efficient with lower impedance taps. You can easily lose an octave of bandwidth between the 4 Ohm and 8 Ohm taps on this account; ask any transformer designer. If the OPT has a 16 Ohm tap and you have a 16 Ohm speaker, since the turns ratio is considerably reduced between input and output, the result is even wider bandwidth, but you need a 16 Ohm speaker to really take advantage of that. That is why 16 Ohm speakers were more common in the 1950s and prior.
@DasguteOhr has made some very insightful comments on this page regarding the load sensitivity issue, literally describing what the issue is when you mix power paradigm and Voltage paradigm technologies (any zero feedback tube amp is power paradigm). The 'standard load impedance' model that Stereophile uses demonstrates amplifier weakness in this regard- the amp reacts more if its output impedance is higher.
The impedance curve of the speaker must be known if you are to get the best out of a zero feedback tube amplifier!!
I would not be surprised if some of the midrange magic with the ML2 on the Magico Q3 was due to the mismatch between amp and speaker, with as a result a frequency curve that was emphasizing certain frequencies in the midrange.
I heard it too when I first heard my corner horns with ML2 amps in Utah. The room was poor due to its modern sheet rock construction and 16" OC studs. Very hollow sounding. I moved the speakers out into the room away from the corners and lost most of the bass. And yet there was that midrange magic. That listening experience got me interested in that specific amp/speaker combination. With the corner reinforcement to the open backed speakers, my room worked fine. The amps simply could not drive the Magicos, but the Lamm M1.1 hybrids sounded superb on the Magicos. It was a great balanced sound, just not quite the same magic as the SET. That is why I went amp first, then speaker.
Ron's poor thread has been completely derailed with this SET and Class D and horn chat. Where is @RonResnick ? Ron, what is the latest with your system?
 | Steve Williams Site Founder | Site Owner | Administrator |  | Ron Resnick Site Owner | Administrator |  | Julian (The Fixer) Website Build | Marketing Managersing |