I finally got the pamphlet on the original Lundahl Mag Amp to download properly so I could see the simplified schematic. It's only using 30 KHz for the carrier, with no feedback from the output. No low pass filter to remove the carrier from the output either. (at least not in the simplified diagram, I can't imagine it passing the EMI certification without that though) Four separate mag amp xfmrs plus an audio OT xfmr per channel. Another xfmr or two for the 30 KHz carrier distribution. Clearly shows some TO-3 power transistors in the photo, likely for the 30 KHz oscillator. Looks like no regulated power supplies.
The mag amp cores are amorphous metal, high initial permeability, square hysteresis. The low freq. 30 KHz carrier and low loss amorphous alloy cores (and using four of them in DC pulse mode per channel, instead of 2 AC saturable reactors) would explain how they kept power loss under control. 70 Watts line power for two 15 Watt channels output. The square hysteresis core material would also improve the open loop distortion. But most magnetics produce copious 3rd harmonic distortion. The P-P mag amps would cancel even harmonic dist., but not odd harmonic dist. Uses a two stage grounded cathode tube amp for front end gain on each channel. Does not show the core biasing in the simplified schematic, so I have attached the diagram, figure 2, from the Lundahl patent to show details of that. The # 10 winding is where the carrier oscillator power is applied.
No specs were given for distortion, damping factor, or audio frequency response. I doubt seriously if this could be called a HiFi amplifier today with only a 30 KHz carrier and no feedback. This is equivalent to 30 KHz open loop Triac PWM.
-----------------------------------------
A serious attempt at a Mag Amp HiFi amp would require square hysteresis ferrite cores running at about 200 KHz carrier freq. Feedback from the output to fix the damping factor and odd harmonic distortion. Output LP filters. Regulated power supplies.
The mag amp cores are amorphous metal, high initial permeability, square hysteresis. The low freq. 30 KHz carrier and low loss amorphous alloy cores (and using four of them in DC pulse mode per channel, instead of 2 AC saturable reactors) would explain how they kept power loss under control. 70 Watts line power for two 15 Watt channels output. The square hysteresis core material would also improve the open loop distortion. But most magnetics produce copious 3rd harmonic distortion. The P-P mag amps would cancel even harmonic dist., but not odd harmonic dist. Uses a two stage grounded cathode tube amp for front end gain on each channel. Does not show the core biasing in the simplified schematic, so I have attached the diagram, figure 2, from the Lundahl patent to show details of that. The # 10 winding is where the carrier oscillator power is applied.
No specs were given for distortion, damping factor, or audio frequency response. I doubt seriously if this could be called a HiFi amplifier today with only a 30 KHz carrier and no feedback. This is equivalent to 30 KHz open loop Triac PWM.
-----------------------------------------
A serious attempt at a Mag Amp HiFi amp would require square hysteresis ferrite cores running at about 200 KHz carrier freq. Feedback from the output to fix the damping factor and odd harmonic distortion. Output LP filters. Regulated power supplies.
Attachments
Last edited:
Hmmm... I have a pair of LM2575T on my table near the computer.... Worth trying? 
http://www.datasheetcatalog.org/datasheets/480/61690_DS.pdf
http://www.datasheetcatalog.org/datasheets/480/61690_DS.pdf
"Hmmm... I have a pair of LM2575T on my table near the computer.... Worth trying? "
I think you will have to custum order the square hysteresis loop ferrite cores. All the switchmode core materials are linear hysteresis for low loss. Last time anyone did research on square loop ferrites was for core memory! But I think Fair-Rite Products still lists square hysteresis material in their catalog, #83 material.
Last time I ordered custom cores from them (30 years ago) they were quite reasonable actually about pricing for custom orders. The bulk pricing for ferrite stuff was really cheap too. You see those ferrite line cord clip on noise suppressors for like $2 on the surplus sites for example, maybe $.05 to buy them direct from the manufacturer. I think I even have some 83 material noise suppressor beads I custom ordered once. The sales guy thought that would be looney. But the setup charge was cheap, so I custom ordered all kinds of strange stuff to try out.
I was looking at ferrite loaded transmission lines with variable transmission speed. Used 4 hole beads with a DC control "winding" in two holes to control the inductance by orthogonal field of the other two wire transmission line strung through them. Very similar to the magnetic amplifier idea actually, except linear when orthogonal. By the way it worked quite well, I've never seen any product like it. I could load a HF waveform into the line and slow it down to look at on a cheap O'scope or low speed A/D. Of course those applications are silly now with Ghz converters.
-----------------------------------------------
The idea crossing my mind now is whether the mag amp idea can be crossed with the Berning HF ferrite OT in some useful way. Like making tracking B+ HV for the tubes and convert to low impedance for the speaker. If you are using HF and ferrite, might as well do everything there.
The Lundahl amp has still got conventional steel P-P audio OTs.
I think you will have to custum order the square hysteresis loop ferrite cores. All the switchmode core materials are linear hysteresis for low loss. Last time anyone did research on square loop ferrites was for core memory! But I think Fair-Rite Products still lists square hysteresis material in their catalog, #83 material.
Last time I ordered custom cores from them (30 years ago) they were quite reasonable actually about pricing for custom orders. The bulk pricing for ferrite stuff was really cheap too. You see those ferrite line cord clip on noise suppressors for like $2 on the surplus sites for example, maybe $.05 to buy them direct from the manufacturer. I think I even have some 83 material noise suppressor beads I custom ordered once. The sales guy thought that would be looney. But the setup charge was cheap, so I custom ordered all kinds of strange stuff to try out.
I was looking at ferrite loaded transmission lines with variable transmission speed. Used 4 hole beads with a DC control "winding" in two holes to control the inductance by orthogonal field of the other two wire transmission line strung through them. Very similar to the magnetic amplifier idea actually, except linear when orthogonal. By the way it worked quite well, I've never seen any product like it. I could load a HF waveform into the line and slow it down to look at on a cheap O'scope or low speed A/D. Of course those applications are silly now with Ghz converters.
-----------------------------------------------
The idea crossing my mind now is whether the mag amp idea can be crossed with the Berning HF ferrite OT in some useful way. Like making tracking B+ HV for the tubes and convert to low impedance for the speaker. If you are using HF and ferrite, might as well do everything there.
The Lundahl amp has still got conventional steel P-P audio OTs.
Attachments
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.