Done.
Not sure if R17 in my schematic is even needed.
Sim works better without it.
If using 'normal' biasing with one transistor, R8 removal also slightly improves
sim results.
I just realized that relatively early clipping in my tests (at 62 Vpp) might be due to
my lab PSU running out of steam (amps).. Will re-test with real (linear) PSU later on.
Not sure if R17 in my schematic is even needed.
Sim works better without it.
If using 'normal' biasing with one transistor, R8 removal also slightly improves
sim results.
I just realized that relatively early clipping in my tests (at 62 Vpp) might be due to
my lab PSU running out of steam (amps).. Will re-test with real (linear) PSU later on.
Last edited:
Also, since I have PCBs left, I tried a simplified hexFET version of this amp.
'Normal' biasing has been used, with Si transistors with low beta
(ZTX, but also two MJE 340 can be used). Attached schematic is 'as-built'.
Even though it's much simplier (no drivers), it simulates in spice better than BJT version.
I quickly built it last night, using junk parts salvaged from the past, failed projects, and it seems to work just fine.
All test with 40V rails, 8 Ohm resistive load, and idle current set to approx 50mA for each of the output devices.
It's actually on this FET version I discovered that D3 slows rise time, so it's left as optional in the schematic.
'Normal' biasing has been used, with Si transistors with low beta
(ZTX, but also two MJE 340 can be used). Attached schematic is 'as-built'.
Even though it's much simplier (no drivers), it simulates in spice better than BJT version.
I quickly built it last night, using junk parts salvaged from the past, failed projects, and it seems to work just fine.
All test with 40V rails, 8 Ohm resistive load, and idle current set to approx 50mA for each of the output devices.
It's actually on this FET version I discovered that D3 slows rise time, so it's left as optional in the schematic.
Attachments
Last edited:
While at it, I also tried 2 other kinds of JFET op-amps (vs. TL071 tried first) :
1) LF356
2) LT1056
All look the same, including slew rate (so the op-amp is NOT the limiting factor)
With LT1056, I see some small oscillations in the bottom edge of the sinus,
but only when Vpp is approaching over 60V.
1) LF356
2) LT1056
All look the same, including slew rate (so the op-amp is NOT the limiting factor)
With LT1056, I see some small oscillations in the bottom edge of the sinus,
but only when Vpp is approaching over 60V.
LT spice
I tried the other op-amps in the real build, not in the sim. I tried what I had
I work at MS-12.
BJT (original) version of the amp is mounted on the chassis, connected to PSU (big box below, with VU meters)
and playing music.
So far so good. Music sounds great, chassis is cool,
driver's temperature is fine (53 C), so these small heatsinks are adequate.
DC offset is minimal (2-3 mV), amp is perfectly stable.
Still running with 100mA idle current for output devices.
As far as Alu chassis goes, this is the smallest one I built so far (15 x 14 cm), with only one fan (Noctua).
From past experiences it should be enough.
and playing music.
So far so good. Music sounds great, chassis is cool,
driver's temperature is fine (53 C), so these small heatsinks are adequate.
DC offset is minimal (2-3 mV), amp is perfectly stable.
Still running with 100mA idle current for output devices.
As far as Alu chassis goes, this is the smallest one I built so far (15 x 14 cm), with only one fan (Noctua).
From past experiences it should be enough.
Attachments
Last edited:
