Horns with SS amp instead of Tube (se)!?
- Thread starter byrdparis
- Start date
The big issue you have with SETs is the output transformer. This has been the Achilles heel since their inception. In a nutshell, the bigger you make the amp, the less bandwidth you get. In many ways 7-8 watts is often the practical limit since making power over that usually means that you can't get 'hifi' bandwidth of 20Hz-20KHz. Now if you aren't running feedback, if you wish to avoid phase shift (which is audible in the bass as a lack of impact and audible in the highs as a darkness with diminished soundstage) then you have to have bandwidth well beyond the audio range in both directions.
This is a bit of a trick with single-ended output transformers. In most cases, since the power tube draws its plate current through the primary winding of the OPT (OutPut Transformer), it has to have a cut in its core to reduce saturation. This is not a perfect fix! So it becomes much harder to deal with the additional currents required to make higher levels of power. To get around that, tubes that run at higher voltages are often used; this allows the transformer design to not see an increase in current through the core. There are other solutions- one is to use a coupling capacitor to the primary winding so that DC current is not present in the core. But this method is inefficient- and now you have a capacitance and inductance working together, and they will have a resonance...
So paralleling power DHTs to get more power is problematic. So much depends on design and parts variables that absolute generalities can't be drawn, but it is very safe to say that as more power is demanded of the circuit, the design parameters and compromises close up around the designer!
Why can't people just connect several amplifiers with lower output impedance and gain in series?
We have examples of top quality SETs around 30W, four would achieve 120W.
You can, but actually you put them in parallel. If you use a pair, the 16 ohm taps get used for an 8 ohm speaker. But in doing this I find that the resulting amplifier does not have the delicacy that non-monostrapped amps have. If you are running feedback in the amps though, this scheme works better.
Probably because it is almost impossible to get two exactly similar gain SE amplifiers, it would need extraordinary matching of tubes - a real problem of tube amplifiers without feedback.
It is why I would prefer running two 2 ohm taps in series.
SS sounds shrill and hifi even honky on most horn speakers I tried and own, somehow the JBL M9500s don't show any of those negative qualities and it comes down to preferences, the ss is a little dead sounding compared to tubes on this speaker. All my comparisons where done with Lamm amplifiers, M1.1 for solid state, ML1 & ML2 for tubes. There horn speakers that you can't drive properly with an 18 watt SET and need a more powerful amplifier, either tube or solid state, you have to determine that by yourself for your speakers.
david
JBL also is very careful not to over-damp their woofers. Horns thrive on higher dampen drivers but high feedback with different acoustic characteristics on each side of a driver can cause a conflict that turns sound into small or even shrill.
The shrill is caused by distortion introduced by the feedback. The only way around this is to have a lot of feedback, north of 35dB or more- at that point the amp can then compensate for the distortion caused by the feedback (which tends to be higher ordered harmonics). Most amps simply lack the Gain Bandwidth Product and Phase Margin to add that much feedback!
I'm certainly not certain that I'd be certain that the feedback is certainly causing any certain distortion. The delayed acoustic feedback is an issue for real time electrical feedback.
Nothing is really changed... certain things mix, others don't.
my experience is that feedback is the root of all evil and the cause of antiseptic gladpack sound
That feedback does this has been known for a very long time (see the writings of Norman Crowhurst, so effectively 60 years) so you can be certain now
The problem has been that nearly all amps with feedback made in the last 60-70 years have had insufficient amounts of it. And it has been nearly impossible to build an amp that actually had enough until very recently- with tubes you have too many problems with phase margin, and with transistors simply not enough gain and bandwidth at the same time. However there are now ways around this problem, so it appears that its now possible to add enough feedback that such is able to compensate for the distortion it causes. The minimum amount to do this seems to be around 35dB; when you consider that in a power amp its going to have to have about 25dB of gain beyond that- well that's a lot of gain, and it takes a competent designer to build an amp that works properly (is stable) with that sort of feedback. But the result is that the annoying higher ordered harmonics which makes amps harsh goes away.
Ya... I can't find any real evidence of what you're getting at. All I read is a big mess. You have to make some wonky amps to get more distortion by applying more feedback. Also the measurements don't seem to show that. All I'm reading is bogyman stuff about feedback. Honestly I'm not sure why it's so hard to believe most amps suck as opposed to subjective claims proving objective things.
Yes, most amps suck...the question is why do they suck? The interface between the objective and subjective is where things in audio, beyond enjoying music, get interesting.
This was Putsey’s argument for his class D solution...did those ncore amps sound natural to you? They don’t to me...That feedback does this has been known for a very long time (see the writings of Norman Crowhurst, so effectively 60 years) so you can be certain now
However there are now ways around this problem, so it appears that its now possible to add enough feedback that such is able to compensate for the distortion it causes. The minimum amount to do this seems to be around 35dB; when you consider that in a power amp its going to have to have about 25dB of gain beyond that- well that's a lot of gain, and it takes a competent designer to build an amp that works properly (is stable) with that sort of feedback. But the result is that the annoying higher ordered harmonics which makes amps harsh goes away.
It was his arguement... Nelson Pass espoused the same idea, although he sounded like it was impractical. I don't think I've ever heard an NCore, but I have heard some newer class D amps that are as smooth and detailed as tube amps. And in at least one of those the feedback was about 41dB, so there seems to be some merit to it. In a class D amp, you don't worry so much about the phase margin, because its expected that the amp will oscillate. That oscillation is the switching frequency. At that point, you can add as much feedback as you want although it does get tricky!
But if you can't do that, IMO you are better off with no loop feedback at all.
That is why I recommended Norman Crowhurst.
I am of the opinion that a lot of amps suck. There is something called the 'GedLee Metric' which applies a formula to harmonic content; essentially placing a weighting system on the higher orders. The idea here is that we will never make an amp without distortion, so instead let's make one with the least audible distortion. That was the idea I had with the Atma-Sphere stuff... anyway, Geddes (IIRC) found a modest Pioneer amp that seemed to meet the requirements of the metric. I suspect the design was from the early 1970s, when circuits that were entirely single-ended right to the output transistors were more common. Such a circuit tends to make a fair amount of second order harmonic, which can mask higher orders. I've got an amplifier that is built that way, and as long as you don't push it hard its very smooth and easy to listen to. But it only makes 6 watts. I use it in my bedroom system, but right now I'm designing a little PP tube amp to replace it, using EL95s...
Brad's comment wasn't about smoothness and detail but wether class D amps sound "natural". Computer sources can be very smooth and detailed but they're far from sounding "natural"
david
That feedback does this has been known for a very long time (see the writings of Norman Crowhurst, so effectively 60 years) so you can be certain now
However there are now ways around this problem, so it appears that its now possible to add enough feedback that such is able to compensate for the distortion it causes. The minimum amount to do this seems to be around 35dB; when you consider that in a power amp its going to have to have about 25dB of gain beyond that- well that's a lot of gain, and it takes a competent designer to build an amp that works properly (is stable) with that sort of feedback. But the result is that the annoying higher ordered harmonics which makes amps harsh goes away.
The Bakoon amp I keep promoting here is a solid state zero feedback design.
And for Class D, Bruno Pultzey, positively uses negative feedback. Excellent interview with John Darko where he speaks of amp designers being myopically obsessed with feedback and not paying attention to what he's doing.
Which Class D might those be? I am far from ignorant when it comes to Class D having owned 3 different technologies and heard at length many more but none sounded what would call natural. There was always a dryness that one never hears with live unamplified music...or a good tube amp for that matter.
Similar threads
