This may be old news to many of you but Elektor published a new amp design in the September edition of their magazine:
Now in Hearing Range: Elektor Q-Watt Audio Power Amplifier - ELEKTOR.com | Electronics: Microcontrollers Embedded Audio Digital Analogue Test Measurement
Just an FYI, Gang.
Now in Hearing Range: Elektor Q-Watt Audio Power Amplifier - ELEKTOR.com | Electronics: Microcontrollers Embedded Audio Digital Analogue Test Measurement
Just an FYI, Gang.
Member
Joined 2009
Paid Member
Poorly thought out implementation of a EF2 mje15032/33 - MG6330/9410 OPS.
Highly doubtful of the inflated power ratings with those outputs , as well.
4R operation would also be "borderline" with the MJE's -VERY low Hfe MG's.
National's datasheet even recommends "additional driver stages" (-a bloody triple)
and multiple output pairs. The datasheet even details the low current gain (<50) issue.
Yet another example of poor design , like the many DIYA amps that "pile on"
the 5-10 pairs of outputs onto 1 poor little pair of mje15032/33's and expect
reliable 4R operation.
OS
Highly doubtful of the inflated power ratings with those outputs , as well.
4R operation would also be "borderline" with the MJE's -VERY low Hfe MG's.
National's datasheet even recommends "additional driver stages" (-a bloody triple)
and multiple output pairs. The datasheet even details the low current gain (<50) issue.
Yet another example of poor design , like the many DIYA amps that "pile on"
the 5-10 pairs of outputs onto 1 poor little pair of mje15032/33's and expect
reliable 4R operation.
OS
Are you sure ? "Best case" spec's... but not "real world".
(attachment 1) .. is semilabs -4V ! MG9410 Hfe plot.
(attachment 2) .. is ON semi's MJL21193 -20V Vce vs. gain plot.
I could only guess what the plot would be like at 60V -Vce.
At High Vce + high current, neither would have that "best case" gain.
In fact , I would guess the MG's , having a larger die than the ON's,
WOULD have better Hfe at a given current.
60V @ 10A, both devices would need a more robust driver.
The Japanese can use small drivers in their consumer electronics by
using those high gain Sankens ... the 2SA1492 has a "Y" grade of 200Hfe.
Even with these special devices , most of their amps are EF3's - VAS running
at 2-3ma ... that's why they last 30 years
.I'm having a hard time choosing between the 2 above devices - are the MG's
really worth 10USD each ?
OS
Attachments
140 HFE is my measurement done at 100ma for 4Are you sure ? "Best case" spec's... but not "real world".
(attachment 1) .. is semilabs -4V ! MG9410 Hfe plot.
(attachment 2) .. is ON semi's MJL21193 -20V Vce vs. gain plot.
I could only guess what the plot would be like at 60V -Vce.
At High Vce + high current, neither would have that "best case" gain.
In fact , I would guess the MG's , having a larger die than the ON's,
WOULD have better Hfe at a given current.
60V @ 10A, both devices would need a more robust driver.
The Japanese can use small drivers in their consumer electronics by
using those high gain Sankens ... the 2SA1492 has a "Y" grade of 200Hfe.
Even with these special devices , most of their amps are EF3's - VAS running
at 2-3ma ... that's why they last 30 years
I'm having a hard time choosing between the 2 above devices - are the MG's
really worth 10USD each ?
OS
MG devices of the same lot. Mine cost me 3.70$ can / each at Newark.
They have sustained HFE for high current.
However, they are still not installed in an amp yet so I ca not comment about the sound and the claim of Semelab that they are better than the Onsemi...
Fab
Dont worry too much, they are far better than those onsemis, BTW hfe increases with incease in voltage so comparing one at 4v to one at 20v is like comparing apples to olives.
The semelab devices are also known as 2sa2223, they make them for sanken. They use a improved method that results in improved heat transfer effeciency and this allows these devices to have the SOA capabilty of bigger packaged devices without the penalty of higher paracitic capacitances and thus low speed. They right up the table of the best outputs if not the best available BJTs for audio.
Thanks for these comments I will try to find a good amp design for them.
By the way I meant 100ma for all 4 devices and they were measured at 15vdc.
Fab
Elektor Power Amp
Compared to the current tendencies, the darlingtons in the output stage are a bit puzzling, just like is the sub-optimal implementation of the temperature compensation, drivers and power transistors, all simply mounted side by side on the same heatsink.
Can we now avoid a current protection for the power transistors ?
Is the large capacitive local decoupling really a good idea ?
.
Compared to the current tendencies, the darlingtons in the output stage are a bit puzzling, just like is the sub-optimal implementation of the temperature compensation, drivers and power transistors, all simply mounted side by side on the same heatsink.
Can we now avoid a current protection for the power transistors ?
Is the large capacitive local decoupling really a good idea ?
.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.