Why have I never heard of the third option? A true switching amplifier...
The Mag Amp—World?s First Switching Amp | Stereophile.com
The Mag Amp—World?s First Switching Amp | Stereophile.com
There is so much ******** in the post you linked to it's hard to decide where to begin:
Magamps are not switching amplifiers. AC amplifiers where envelope is recovered (think of AM detection).
Magamps were not designed in 1960s. Try decades earlier. Perhaps this particular one was which seems rather weird, considering that far superior alternatives were already avaliable at that time.
The price they charge is downright silly and I cannot possibly imagine magamp displaying superior performance in terms of sound quality than either of the "first two options", as you implicitly dubbed them, simply due to nature of magamp's operation.
Why haven't you ever heard about them ? I guess you don't have much interest in the obsolete technology, just like most people. Then again, who better to answer this question than yourself ?
Magamps are not switching amplifiers. AC amplifiers where envelope is recovered (think of AM detection).
Magamps were not designed in 1960s. Try decades earlier. Perhaps this particular one was which seems rather weird, considering that far superior alternatives were already avaliable at that time.
The price they charge is downright silly and I cannot possibly imagine magamp displaying superior performance in terms of sound quality than either of the "first two options", as you implicitly dubbed them, simply due to nature of magamp's operation.
Why haven't you ever heard about them ? I guess you don't have much interest in the obsolete technology, just like most people. Then again, who better to answer this question than yourself ?
Search around for "Carver magnetic amplifiers" he had them commercially avail. not very long after his "auto-correlator" sound processors and pre-amps were "the buzz" (in the general shake the purist wannabe audiophiles tree".
Indeed the concepts and practical applications are rooted in military electronics of at least the 1940's, and very probably somewhat earlier.
Magnetic "amplifiers" were around way back, used mainly for motor speed control, power supply regulators and light dimmers. Slow, non-linear, heavy, noisy, but reasonably efficient at low carrier (60 Hz) frequencies. Not really an amplifier, but an AC modulator with diode envelope detection. Transformer saturation via a DC control winding is used to control inductive AC current passage thru it. Looking at the Lundahl patent, I see that two are used in P-P to improve distortion.
I see many problems with using mag amps for audio. A high frequency carrier would have to be generated, so probably plenty of power transistors and maybe an IC in there. The high frequency carrier (similar to class D amplifiers) would lead to very hot magnetics, since they have to run into saturation at the HF. This is a definite NO-NO for swithcing amps. This would be very inefficient, likely fans are needed to cool the box. The output has to be diode demodulated and then Low Pass filtered to remove the HF carrier. EMI considerations, regulatory approval/license issues. Almost certainly limited audio HF frequency response due to the magnetics HF carrier limitations.
The LP output filter makes for a high Z/poor damping factor output, just like for class D amps, so feedback around the whole thing is needed then. That drives the HF carrier freq. requirement up even higher, leading to near melt down of the magnetics. A vacuum tube front end is needed to provide the power to run the two xfmr control windings and provide excess gain for the feedback. Thats probably close to a VT power amp or two in there already. The distortion figures are likely quite poor due to limited feedback bandwidth available. I don't see any specs posted for the amp for distortion, power consumption, frequency response, or damping factor. At that price though, any buyers are probably a little light headed anyway.
I see many problems with using mag amps for audio. A high frequency carrier would have to be generated, so probably plenty of power transistors and maybe an IC in there. The high frequency carrier (similar to class D amplifiers) would lead to very hot magnetics, since they have to run into saturation at the HF. This is a definite NO-NO for swithcing amps. This would be very inefficient, likely fans are needed to cool the box. The output has to be diode demodulated and then Low Pass filtered to remove the HF carrier. EMI considerations, regulatory approval/license issues. Almost certainly limited audio HF frequency response due to the magnetics HF carrier limitations.
The LP output filter makes for a high Z/poor damping factor output, just like for class D amps, so feedback around the whole thing is needed then. That drives the HF carrier freq. requirement up even higher, leading to near melt down of the magnetics. A vacuum tube front end is needed to provide the power to run the two xfmr control windings and provide excess gain for the feedback. Thats probably close to a VT power amp or two in there already. The distortion figures are likely quite poor due to limited feedback bandwidth available. I don't see any specs posted for the amp for distortion, power consumption, frequency response, or damping factor. At that price though, any buyers are probably a little light headed anyway.
Last edited:
Actually I think they go way further back, to early 1900s, when Alexandersson alternator was invented. It was used in conjunction with mag-amps for AM audio broadcasting.
ok, it may move up my to-do list after seeing this
Homemade Magnetic Audio Amplifier.
YouTube - Homemade magnetic amplifiers made from common materials.
Homemade Magnetic Audio Amplifier.
YouTube - Homemade magnetic amplifiers made from common materials.
Hmm, Steiner's circuit is using pulsating DC in the Xfmr cores, which would be more efficient than the full swing AC that appears to be used in Lundahl's patent (#3,131,360 figure 2). (lower hysteresis losses with SE flux changes) The patent mentions saturable reactors and a power gain of 200. Steiner is getting a power gain of 2000 with the DC pulses. Maybe some hope for a practical design that way. But the carrier will have to be a lot higher freq. than 35 KHz for HiFi. (more bandwidth needed to linearize the gain with feedback)
Last edited:
I heard about a "White Cube" magnetic amp that was used in Moscow institute of acoustic research; yesterday a topic was started on Russian forum. It is scary synchronicity how the same thoughts are coming in different heads around the globe almost simultaneously. That amp had something like hundred kilohertz of carrier frequency.
I still have a Carver M-400 "Magnetic Field Amplifier" and a Phase Linear 4000 Auto-Correlating Preaplifier. I got talked into buying them when we had a "high end audio club" at work. These components were LOUD and I guess many people associated LOUD with good sound. I listened to this stuff for several years and accepted it as "good sound". A friend gave me an old Scott tube amp that didn't work. After fixing it, the Carver stuff went into the warehouse where it still sits.
The "magnetic field" amp is just marketing B.S. The amp is a solid state design with 3 sets of output devices running from 3 sets of supply rails. Different sets are engaged at different power levels to improve efficiency and reduce dissipation. This is a good thing because there are no heat sinks.
The power supply (the magnetic field device) is a crude and early switching power supply running in the 2 to 60 Hz range. It used a small power transformer that would melt quickly if plugged continuously into the wall outlet. There is a triac between the transformer and the power cord. The triac is pulsed on at a low rate (about 2 Hz) when the amp is idled, and the the pulse rate increases as the power is cranked up. The triac goes into continuous conduction if both channels are driven to full power. The pulsing of the power supply can be heard and felt if you place your hand on the amp. It took me a lot of tinkering to keep the pulsing power line currents out of the turntable wiring.
The entire amp is packaged in a cube that is about 8 inches on a side. The manual advises that the amp should not be tested with sine waves, used in sound reinforcement, commercial applications, or used as a guitar amp. This amp and many of Carvers amps absolutely rely on the fact that the average power level produced by this 400 watt amp is never above 20 watts or so for more than about a minute.
Of course I wired mine into bridged mono mode, connected up two 4 X 12 cabinets and connected my guitar to it. It was #*$!ing loud, but it would shut itself off before blowing up. You could crank it quite loud for a while before triggering the protection stuff though.
True "switching amps" require a SWITCH device. It can be a tube, transistor, triac, SCR, thyratron, or other type of switch but it must be capable of switching at a rate at of least twice the highest audio frequency being amplified. For good fidelity 10X is the given norm.
Magnetic amps have indeed been around for a long time. The typical application was for industrial control and lamp dimming. Google "saturable core reactor" for some history. I once maintained some high power (200 watt) laser equipment that used a saturable core reactor in the primary circuit of a 25,000 volt 1 amp transformer. A small DC current was used to regulate the HV to 30 KV.
The reactor (a choke) is in series with the primary of the main power transformer. Its inductance limits the current to a low value. A small DC current is applied to a control winding in the reactor that causes partial core saturation reducing the inductance. A large DC current causes complete core saturation reducing the inductance to near zero.This enabled a single 2N3055 transistor and a small PC board to regulate the HV.
Seems like a more sensible use for the magnetic amplifier would be to use it to make a tracking B+ or +/- Drain voltage rails for a normal linear device amplifier. Then the AC frequency could be relaxed some from 100 or 200 KHz as mentioned to maybe 40 KHz. Linearity demands then relaxed on the mag. amp. too. Efficiency of the Linear amp increased then. Hard to say whether the overall efficiency would be increased though (versus fixed rails and no mag amp). Let's see, I guess the mag amp dissipates more power in the magnetics at low power output, so idle power would be increased.
Last edited:
YouTube - Saturable Reactor
BTW, mag amps DO switch. A square permalloy core will generate harmonics into the MHz (i.e., sub-us switching speeds). The on/off ratio is quite reasonable, considering a permeability of ca. 100k abruptly changing to about 1. Don't know if anything is intrinsically faster (square ferrites possibly??).
I haven't yet made a proper magnetic switching amplifier with fundamental voltage gain; the simplest circuit is actually a voltage follower, which obviously makes problems for producing voltage gain.
Tim
BTW, mag amps DO switch. A square permalloy core will generate harmonics into the MHz (i.e., sub-us switching speeds). The on/off ratio is quite reasonable, considering a permeability of ca. 100k abruptly changing to about 1. Don't know if anything is intrinsically faster (square ferrites possibly??).
I haven't yet made a proper magnetic switching amplifier with fundamental voltage gain; the simplest circuit is actually a voltage follower, which obviously makes problems for producing voltage gain.
Tim
Been there done that, except the normal linear device was a vacuum tube cathode follower, and the magnetic amplifier was a dsPIC chip controlled agile SMPS tracking the audio signal. The efficiency improvement allowed for nearly doubling the output power.
I have the service manual for both the Carver amp and the Phase Linear preamp around here somewhere. The manual explained the "crest fator" and peak to average ratio of most music. This means that a 400 watt amp pushed to clipping on peaks will be producing between 4 and 40 watts of AVERAGE power, depending on the type of music being played, and how much clipping is tolerated. Carver exploited this fact to build 400 watt amps with 40 watts worth of heat sinks and 40 watt power supplies augmented with big fat caps to handle the peak demands.
What does this mean to us? It means that we can design a P-P class AB tube amp that exhibits some tube glow when cranked to full power. We can even bend a few specs since the full power peaks are few. You should be able to play it to the edge of clipping without overstressing any component
I just noticed that the Steiner Mag amp only puts out modulated DC for the speaker. To get AC one would need two of these. But they cannot be just parallel connected at the speaker due to the diodes shorting everything out. So this would need a P-P audio output Xfmr just like the Lundahl patent (figure 2) to combine the two polarities. May require two -separate- primary windings too (not the usual center tapped OT primary).
I'm starting to wonder just how many xfmrs are used in the Lundahl design. I thought I read on that website that it had two mag amp xfmrs and an audio OT per channel. Hard to tell from the patent diagram, but if it has 4 mag amp xfmrs and a audio OT (per channel), it would be running with DC pulses in the xfmrs also, like the Steiner setup.
-----------------------------------------
"(square ferrites possibly??)"
Yes, ferrite mag amp xfmrs would be mandatory at 200 KHz carrier freq. But it seems the Lundahl is using silicon steel, so it must be running at lower frequency. Not likely very "HiFi" if it has no feedback.
I'm starting to wonder just how many xfmrs are used in the Lundahl design. I thought I read on that website that it had two mag amp xfmrs and an audio OT per channel. Hard to tell from the patent diagram, but if it has 4 mag amp xfmrs and a audio OT (per channel), it would be running with DC pulses in the xfmrs also, like the Steiner setup.
-----------------------------------------
"(square ferrites possibly??)"
Yes, ferrite mag amp xfmrs would be mandatory at 200 KHz carrier freq. But it seems the Lundahl is using silicon steel, so it must be running at lower frequency. Not likely very "HiFi" if it has no feedback.
Last edited:
With the rotten CD mastering these days, it wouldn't work. Maybe 9dB average below peak, lot's of clipping. But that's for another thread!
Pretty much my experience, too. I was originally fascinated with the auto-correlation noise thing as it was actually derived from Bob's NASA imaging research.
Saturable core reactor motor control centers were long "the Cats pajamas".
Thank you for the memories, loud and very flat (as in no inner detail)!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.