Things to do when bored:
Open EDA and just throw parts together, see what comes out.
Turns out my third brain cell lead me to making a Amplifier thats just a bunch of opamps slapped together
on a 10x10cm 4 layer pcb. because why not.
PCBs are ordered, and 2 weeks later they showed up, fully assembled.
Its literally just 40 NE5532 in parallel with a 10ohm balancing resistor. its gain is about 8x, non inverting.
the initial idea was to use opa2891 as they can drive up to 200mA (5532 do just 40mA or so) but it would have gotten too pricey for a experiment...
both channels seperated, so each pcb could be used as a stereo or mono bridged amp, or even a symmetrical mono.
a crude +-18V supply was set up from two caps in series with zener diodes across (after i took that pic). not ideal but works.
both inputs are tied together for a bridged single ended amp. output goes into a 8ohm load.
just one channel:
Hooking up a speaker also sounded good, absolutely no hiss or hum audible!
with 2 channels i got about 15W out of it. not great, not terrible.
i dont trust my signal generator for delivering a pure sine, so i later switched over to my PC, fired up REW and set up my Audio interface (focusrite 4i4 3rd gen)
Keep in mind nothing is calibrated, im just interessted in the difference shown in this graph..
the 8ohm resistor has a 1:10 divier across to not blow up my audio interface... 1:10 is a bit much so i had to use the preamp a little.
White noise, peak hold for a few seconds. its not a sweep. but is flat enough for me
I did a loopback from Foobar2000 into audacity trough this amp, into a 8ohm load trough the 1:10 divider and preamp.
it sounded great. at least when i didnt crank the volume into clipping.
Its getting hot, seen at 80°C at max without any cooling or heatsinking.
15W idle power consumption, and 35W at max.
edit:
Oh and the PCBs can be stacked for more current.. at some point its probably even necessary to use a output transformer to step up the voltage. back to the roots of tube amps, just the other way around
Open EDA and just throw parts together, see what comes out.
Turns out my third brain cell lead me to making a Amplifier thats just a bunch of opamps slapped together
on a 10x10cm 4 layer pcb. because why not.
PCBs are ordered, and 2 weeks later they showed up, fully assembled.
Its literally just 40 NE5532 in parallel with a 10ohm balancing resistor. its gain is about 8x, non inverting.
the initial idea was to use opa2891 as they can drive up to 200mA (5532 do just 40mA or so) but it would have gotten too pricey for a experiment...
both channels seperated, so each pcb could be used as a stereo or mono bridged amp, or even a symmetrical mono.
a crude +-18V supply was set up from two caps in series with zener diodes across (after i took that pic). not ideal but works.
both inputs are tied together for a bridged single ended amp. output goes into a 8ohm load.
just one channel:
Hooking up a speaker also sounded good, absolutely no hiss or hum audible!
with 2 channels i got about 15W out of it. not great, not terrible.
i dont trust my signal generator for delivering a pure sine, so i later switched over to my PC, fired up REW and set up my Audio interface (focusrite 4i4 3rd gen)
Keep in mind nothing is calibrated, im just interessted in the difference shown in this graph..
the 8ohm resistor has a 1:10 divier across to not blow up my audio interface... 1:10 is a bit much so i had to use the preamp a little.
White noise, peak hold for a few seconds. its not a sweep. but is flat enough for me
I did a loopback from Foobar2000 into audacity trough this amp, into a 8ohm load trough the 1:10 divider and preamp.
it sounded great. at least when i didnt crank the volume into clipping.
Its getting hot, seen at 80°C at max without any cooling or heatsinking.
15W idle power consumption, and 35W at max.
edit:
Oh and the PCBs can be stacked for more current.. at some point its probably even necessary to use a output transformer to step up the voltage. back to the roots of tube amps, just the other way around
Added connectors and stacked two ontop for a little stereo amp.
It sounds great. very clear, no noise audible at all. detailed and the imaging is something i immediately noticed.
i could locate every instrument on chris isaaks wicked game live. something i couldnt do on some other amps before.
(powered by a 12V laptop brick with a two pin plug to not have a ground reference, going trough a 12V to 36V boostconverter into my floating ground contraption... works great but limits the power supply selection)
i can crank up the volume knob to the max without it going into clipping on my 4ohm ELAC BS244,
it becomes a little too loud for my taste but thats good having a little headroom. those speakers are not even that efficient at 88 dB / 2.83 V / 1 m.
i added two 2200uF caps in series (back to back to get a bipolar cap at half the capacity) for getting rid of a slight DC offset im having.
this makes the bass roll off a bit early.
two 10.000µF in series would be better (8hz) but i didnt have any in my parts bin.
Input impedance is around 12kOhm, i didnt want to go lower, this would have decreased the offset aswell due to the opamps input offset current i completely missed.
for the next boards i probably add a pre-buffer to it, decrease the balancing resistors from 10 to 1ohms, and the feedback path from high value (100k) resistors to something in the 1..5k range. and maybe use better opamps than 5532's
It sounds great. very clear, no noise audible at all. detailed and the imaging is something i immediately noticed.
i could locate every instrument on chris isaaks wicked game live. something i couldnt do on some other amps before.
(powered by a 12V laptop brick with a two pin plug to not have a ground reference, going trough a 12V to 36V boostconverter into my floating ground contraption... works great but limits the power supply selection)
i can crank up the volume knob to the max without it going into clipping on my 4ohm ELAC BS244,
it becomes a little too loud for my taste but thats good having a little headroom. those speakers are not even that efficient at 88 dB / 2.83 V / 1 m.
i added two 2200uF caps in series (back to back to get a bipolar cap at half the capacity) for getting rid of a slight DC offset im having.
this makes the bass roll off a bit early.
two 10.000µF in series would be better (8hz) but i didnt have any in my parts bin.
Input impedance is around 12kOhm, i didnt want to go lower, this would have decreased the offset aswell due to the opamps input offset current i completely missed.
for the next boards i probably add a pre-buffer to it, decrease the balancing resistors from 10 to 1ohms, and the feedback path from high value (100k) resistors to something in the 1..5k range. and maybe use better opamps than 5532's