It was called 'Sigma Drive' and used on several amps, including the KA-800 and KA-1000, and other TOTL models.
Basically it's a remote load sense, using a 4-wire system. Two wires are the normal output, which are internally connected to the sensing terminals via small value resistors (on the order of 10 ohms). This takes care of default feedback from the amp terminals in case a normal 2-wire connection is used, or the return wires get disconnected. The return wires connect to the other end of the 10-ohm resistors where the actual feedback terminals are. The low cable resistance swamps out the default return path through the resistors, so the actual voltage at the load is what the feedback sees.
Biggest practical problem with these amps were users erroneously connecting the speaker to the return terminals, thus feeding the speaker through the relatively low power resistors, resulting in them burning out and the feedback loop getting disconnected, at which point the output would get stuck to one of the supply rails and trip the protection, which would remain engaged and the amp would have to be serviced. They fixed that in later version with anti-parallel diodes, which then generated further service requests when people got horribly distorted sound because of the wrong connection. Finally, Kenwood sold their own speaker cables but of course that did not sit well with the prospective buyers, so the concept fell out of favor with the next model.
The only point I wanted to make is that speaker cable interactions are often too simplified to be realistic and measured reality provides a different picture. A remote feedback loop might indeed provide enough value to be worth playing.Back to the 1980s and the Kenwood Basic 1 amp with "Sigma Drive" and using the speaker leads for feedback.
By connecting to the Sigma Drive feedback point (and disconnecting the feedback from the amp's speaker terminal), the amp is reading the voltage at the speaker terminal at the far end of the speaker lead (right at the speaker + terminal). When I did the math, it turned out (under the impedance assumptions I used) that there is a clear enhancement in the feedback that tracks the speaker's varying motions and erroneous motions. Do the math yourself.
As I modelled it, the signal from the amp first goes through the internal impedance (about .5 ohms), the speaker supply lead (.3, say), the speaker (I modelled 8 to 40 ohms), and the return lead. With normal feedback, your are reading the voltage from the far end of the internal impedance. With Sigma Drive, you are reading the voltage from the far end of the internal impedance plus the supply lead. Yes, you are picking up more speaker variation (and feeding it back to the amp) simply by moving the feedback pickup point to the far end of the supply lead.
With Sigma Drive, the feedback fraction goes from .91 to .98. With regular feedback, the fraction only goes from .95 to .99. In other words, the amp is more responsive to the speaker's back EMF (motional feedback) and so corrects for the exaggerated (and erroneous) cone motion.
Does this make sense?
Footnote: shouldn't be hard to configure almost any amp this way by changing the amp feedback from the speaker terminal right on the amp to the remote-sensed terminal right on the speaker.
Second footnote: I bought a Basic M1. Quite amazing checking amps today with a computer spectrum analyzer - garbage in the signal shows up like lights on a Christmas tree. Anybody know of free osciloscope/spectrum analyzer software for the current Mac OS?
By connecting to the Sigma Drive feedback point (and disconnecting the feedback from the amp's speaker terminal), the amp is reading the voltage at the speaker terminal at the far end of the speaker lead (right at the speaker + terminal). When I did the math, it turned out (under the impedance assumptions I used) that there is a clear enhancement in the feedback that tracks the speaker's varying motions and erroneous motions. Do the math yourself.
As I modelled it, the signal from the amp first goes through the internal impedance (about .5 ohms), the speaker supply lead (.3, say), the speaker (I modelled 8 to 40 ohms), and the return lead. With normal feedback, your are reading the voltage from the far end of the internal impedance. With Sigma Drive, you are reading the voltage from the far end of the internal impedance plus the supply lead. Yes, you are picking up more speaker variation (and feeding it back to the amp) simply by moving the feedback pickup point to the far end of the supply lead.
With Sigma Drive, the feedback fraction goes from .91 to .98. With regular feedback, the fraction only goes from .95 to .99. In other words, the amp is more responsive to the speaker's back EMF (motional feedback) and so corrects for the exaggerated (and erroneous) cone motion.
Does this make sense?
Footnote: shouldn't be hard to configure almost any amp this way by changing the amp feedback from the speaker terminal right on the amp to the remote-sensed terminal right on the speaker.
Second footnote: I bought a Basic M1. Quite amazing checking amps today with a computer spectrum analyzer - garbage in the signal shows up like lights on a Christmas tree. Anybody know of free osciloscope/spectrum analyzer software for the current Mac OS?
Last edited:
i run DL4YHF's Spectrum Lab (a windows program) on linux using WINE. WINE can also be installed on Mac OS-X. the link on how to do this is here:
MacOSX - The Official Wine Wiki
MacOSX - The Official Wine Wiki
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.