After some small modifications this little amp started to work. Here is bc639/640 drivers and bd135/136 output transistors. C1 is not needed and C2&C3 is now 47pF. Also added one diode 1N4148 between R3 and R12. And there is 1500uF10V capacitor parallel with them, without this it can oscillate at some levels.
1kHz square wave is clean, no overshoot. 10kHz sq wave is rounded (no picture).
Thanks for all, don't know how it sounds, only measurements now.
1kHz square wave is clean, no overshoot. 10kHz sq wave is rounded (no picture).
Thanks for all, don't know how it sounds, only measurements now.
Attachments
The LED reference CCS is not a great choice because LEDs are a ~2V reference which subtracts from the available voltage swing. A Transistor based CCS provides about 2-0.65~= 1.35V more voltage swing, which means higher clipping power.
Lower voltage/ low power amps usually use a bootstrap or a CFP with gain in order to get closer to a rail-rail output. One of the automotive chip-amps (I don't remember the #) features a CPF with gain in order to get the most out of 14.3V and maybe chip-amps "own" this range of power amps.
Lower voltage/ low power amps usually use a bootstrap or a CFP with gain in order to get closer to a rail-rail output. One of the automotive chip-amps (I don't remember the #) features a CPF with gain in order to get the most out of 14.3V and maybe chip-amps "own" this range of power amps.
Yes, LED based CCS "steals" some voltage, but have good temperature stability if it is important. I have adjusted "middle point" voltage at 5.75Volts, which gives symmetrical clipping and 1.5W output with no load. Power supply is 12Volt like original one.
This can be easily modified to use 19Volts laptop power supply, like original Nelson Pass ACA v1.0 which can give little bit more power, say around 4 watt (?)
But in any case this sounds good and is simple and cheap design.
Can be easily prototyped using blanco PCB board, look at picture 😎
This can be easily modified to use 19Volts laptop power supply, like original Nelson Pass ACA v1.0 which can give little bit more power, say around 4 watt (?)
But in any case this sounds good and is simple and cheap design.
Can be easily prototyped using blanco PCB board, look at picture 😎
Maybe use one IR led around 1.1 volt instead of red led around 1.6 volt, this gives little bit more margin ?
Congratulations! There’s nothing like your first successful amplifier design!
This is a great place to get your design scrutinized by people who actually know what they’re talking about. I just had one of my designs thoroughly picked apart and built back up again and I learned more than I could ever have imagined!
Keep designing and sharing!
This is a great place to get your design scrutinized by people who actually know what they’re talking about. I just had one of my designs thoroughly picked apart and built back up again and I learned more than I could ever have imagined!
Keep designing and sharing!
Thanks, maybe you noticed that original design is from Wattnik (another member from Finland), with help from lineup and other members.
I just modified it a little bit.
This bc639/640 have Ft=100MHz and bd139/140 also higher than tip41/42 pair.
Maybe current source instead of LED CCS like steveu suggest and Vbe multiplier also.
I have done many different amplifiers before, but this is something special, like small SE tube desing without any output transformer.
I just modified it a little bit.
This bc639/640 have Ft=100MHz and bd139/140 also higher than tip41/42 pair.
Maybe current source instead of LED CCS like steveu suggest and Vbe multiplier also.
I have done many different amplifiers before, but this is something special, like small SE tube desing without any output transformer.
One more modified design using IR led instead of red led. Not rail-to-rail design but little bit better.
Modified prototype here
2kHz square wave (no load) is little bit rounded, but no overshoot, 1kHz square is more clean. Timediv=0.1mS/div voltsdiv=1V/div
Time to listen how it sounds, but should be same as before.
Modified prototype here
2kHz square wave (no load) is little bit rounded, but no overshoot, 1kHz square is more clean. Timediv=0.1mS/div voltsdiv=1V/div
Time to listen how it sounds, but should be same as before.
Many years ago, I threw together a small amplifier in this range but, given the limited voltage and power, I used a single EF output stage with bootstrap and a LTP input. I remember experimenting with the LTP CCS and favored the sound of a JFET. The outputs were TIP TO220s with little heat sinks, but it made enough sound for my landlord to complain.
Hell, a clock radio can go loud enough to make a landlord complain.
If an LTP is used, you can double bootstrap (total current thru input stage needs to be constant for it to work). That and EF1 output stage using low quasi saturation outputs like D44/5H give you about as much as you can get from a limited supply. Distortion is always higher than circuits like these, though.
If an LTP is used, you can double bootstrap (total current thru input stage needs to be constant for it to work). That and EF1 output stage using low quasi saturation outputs like D44/5H give you about as much as you can get from a limited supply. Distortion is always higher than circuits like these, though.
Yeah. A double boot strapped EF1 would be hard pressed to get to .04%, even driven from an op amp. 2 stages of current gain goes a long way.
My little desktop stereo amp is tubes. Why? Because I can. No other technical or audiophile idiocy reason. Nice “little” sealed box speakers the same size as my KRK powered monitors. It’s as small as I GO.
My little desktop stereo amp is tubes. Why? Because I can. No other technical or audiophile idiocy reason. Nice “little” sealed box speakers the same size as my KRK powered monitors. It’s as small as I GO.