PS. Just like to say again (I know I have before), this is such a wonderful channel and a delight to watch and listen. It needs to be said occasionally.
A cross-over notch with a time interval of 1 µs is impossible to hear. John, I think, the combination of this high slew rate and high gain output transistors will work very good. I would also suggest to apply a Boucherot network at the output in order to stabilize the negative feedback.
I think JAT should put the 0.1 uF in series with the 10 ohm resistor from output to ground. I've seen this called the Zobel Network. It might take some of the high frequencies out of the crossover notch. If the output had a choke paralleled by a 10 ohm resistor between output and speaker, it might help further in removing some higher frequencies from the notch.
Hi I love the simplicity of your explanations I don't know much amplifiers you were talking about the cross over not being able to be biased can you use for second opamp operate the negative part of the signal
It would be interesting to learn if the cross-over notch was due to the limit of the op amp's slew rate, or if the output transistor was slow to turn on after having been so heavily reversed biased.
_1) Thanks for video John. _2) The strange sound (at PlbackPos 5min12sec) reminded me of a video from xrayTonyB 's channel, he had very similar challenge. -->Though he discovered that if he unplugged his PowerLine-cord from his oscilloscope, then his slow-oscillation phenomena stopped. (..Even when oscilloscope turnedOff/Standby, the phenomena still existed, so he 'pulled-the-plugg' and things became Ok.) //Thanks again for sharing videos.
I'd like to see how this simple circuit would ultimately perform with various additions (up to of a pair of diodes)... Time to get the breadboard out. Very inspiring..And of course then the second thing that springs to mind is FETs.. Do you prefer bipolars over FETS?
A power MOSFET takes several volts to turn on, which would make a much larger 'dead zone' where the zero bias causes a large amount of crossover distortion. Plus the output stage would lose several volts of swing because of this extra voltage needed.
@@iceberg789 this is, of course, true. And if you look behind the curton of the opamp. It draws 10 mA of idle current, the most of it is the bias current of the output stages. I guess there is at least 4 mA of bias current through the output stages. It means you are in class a mode until 4 mA of output current are reached. The maximum output current is 26 mA, so you have around 20% class a operation on a well fittet load resistance. If you operate the op amp on a load resistance of 3.3 kOhm at +/- 15 V it will always operate in class A mode. This and its 140 dB voltage gain is also reason of its THD+N of 0.00003%.
perhaps would be possible to connect the two OPs sections of LM4562 in parallel in order to reduce the 150-Ohm resistor to 82-Ohm. I espect in this way to reduce the crossover notch
John, best wishes for 2024. I look forward to more of your informative examples. I'd like to ask 1/4 watt rated resistor is OK or is a higher rating recommended?
I really appreciated your TDA7498E test (i have the non E version of it) I’d really really appreciate if you could test the XH-M252 dual mono config TDA8954TH board (the one with 8 power capacitors and blue heasink) I have it and i’d like to see a proper test of it... it requires a dual psu, i wanted to remove the full bridge rectifiers to power it directly with 2 36/48v 350w smps... please let me know if you could do it... thx a LOT
Do a Fourier transform and find the high frequency element causing the notch and use the appropriate low pass filter to smooth out the notch Or This just occurred to me … don’t know if it will work! Find the offending frequency using spectrum analyser and utilise the feedback opamp as a selen-key lowpass filter ?
Isn't this how a Quad amp works?? I wonder how about using Germanium output to lower the crossover dead zone for shits and giggles. Now that I think about it....does germanium have enough gain to make this design work? It seems to me that this amp is simpler then the EZ amp project your working on?
I can't hear that little notch so I can sure use that simple thing for a simple 'in-the-shed-amplifier'. Gonna use it for hardcore so that little thing won't matter. A long time ago I was into audio amplifiers but lost that when I got into digital electronics and as I'm not a real music lover and I sure can live without music but your vids sure made me wanna do something like this again, not a perfect amplifier, a thing like this one but boost it to lets say 50W so I can really feel those fast paced music in my guts 😂 I do though always through my years have wanted to build a class-d amplifier but haven't got into it and as it's more advanced than a class-a, I never bothered, but I don't want a 10 pounds heat sink 😋
Hello. If i put a pair of drivers in darlington mode, and an higher voltage op amp like opa2604, is there a chance to increase the power up to 30w in to 4 ohms load?
Hi John. A simple modification can reduce the deadzone by half, while still leaving the amplifier in class B. I tried to post a link but it keeps deleting my comments :)
Hi! Thanks for the experiment. How about adding a current source to transform it into a class XD amp? Please check “en.m.wikipedia.org/wiki/Class-XD_Amplifier”. I am quite curious… I’ll try it by simulation…
PS. Just like to say again (I know I have before), this is such a wonderful channel and a delight to watch and listen. It needs to be said occasionally.
No - it needs to be said very often!
@@johncoops6897 😸
Thanks
Big Big thanks!
That class-B doesnt sound all that bad, I honestly didn't expect that! Nice video John 👍🏼
A cross-over notch with a time interval of 1 µs is impossible to hear. John, I think, the combination of this high slew rate and high gain output transistors will work very good. I would also suggest to apply a Boucherot network at the output in order to stabilize the negative feedback.
No thermal comp needed, love it!
I think JAT should put the 0.1 uF in series with the 10 ohm resistor from output to ground. I've seen this called the Zobel Network. It might take some of the high frequencies out of the crossover notch. If the output had a choke paralleled by a 10 ohm resistor between output and speaker, it might help further in removing some higher frequencies from the notch.
Zobel and boucherot networks, and a set of 0.33 ohm emitter resistors nights help. Cool circuit.
Great breadboard project to play with....cheers.
Polka party with John!
Continue making informative content! Thank you!
Hi I love the simplicity of your explanations I don't know much amplifiers you were talking about the cross over not being able to be biased can you use for second opamp operate the negative part of the signal
It would be interesting to learn if the cross-over notch was due to the limit of the op amp's slew rate, or if the output transistor was slow to turn on after having been so heavily reversed biased.
thanks that looks like a nice amp
_1) Thanks for video John. _2) The strange sound (at PlbackPos 5min12sec) reminded me of a video from xrayTonyB 's channel, he had very similar challenge. -->Though he discovered that if he unplugged his PowerLine-cord from his oscilloscope, then his slow-oscillation phenomena stopped. (..Even when oscilloscope turnedOff/Standby, the phenomena still existed, so he 'pulled-the-plugg' and things became Ok.) //Thanks again for sharing videos.
Could you add a current gain stage in the middle? Just a single BJT in class A? It might help push enough current for the crossover distortion.
Stick the spare opamp in parallel.
Nice piano music
I'd like to see how this simple circuit would ultimately perform with various additions (up to of a pair of diodes)... Time to get the breadboard out. Very inspiring..And of course then the second thing that springs to mind is FETs..
Do you prefer bipolars over FETS?
A power MOSFET takes several volts to turn on, which would make a much larger 'dead zone' where the zero bias causes a large amount of crossover distortion.
Plus the output stage would lose several volts of swing because of this extra voltage needed.
@@acmefixer1 I was not suggesting a direct replacement, they need a 100ma to bias, not a problem.
@@martinda7446 Biasing this circuit would defeat the entire purpose..The goal was to see what you could do with class B, not AB
@@elpechos Yeah. I started wandering...
This is probably why Class-B isn't used much; the crossover distortion issue.
well, this isn't the traditional class b, its a modified active class b.
@@iceberg789 this is, of course, true. And if you look behind the curton of the opamp. It draws 10 mA of idle current, the most of it is the bias current of the output stages. I guess there is at least 4 mA of bias current through the output stages. It means you are in class a mode until 4 mA of output current are reached.
The maximum output current is 26 mA, so you have around 20% class a operation on a well fittet load resistance. If you operate the op amp on a load resistance of 3.3 kOhm at +/- 15 V it will always operate in class A mode. This and its 140 dB voltage gain is also reason of its THD+N of 0.00003%.
perhaps would be possible to connect the two OPs sections of LM4562 in parallel in order to reduce the 150-Ohm resistor to 82-Ohm. I espect in this way to reduce the crossover notch
John, best wishes for 2024. I look forward to more of your informative examples. I'd like to ask 1/4 watt rated resistor is OK or is a higher rating recommended?
I was wondering about using MOSFETS instead
So what would cause a hum without any input signal? I have changed out a few caps that i thought were suspect.
I really appreciated your TDA7498E test (i have the non E version of it)
I’d really really appreciate if you could test the XH-M252 dual mono config TDA8954TH board (the one with 8 power capacitors and blue heasink)
I have it and i’d like to see a proper test of it... it requires a dual psu, i wanted to remove the full bridge rectifiers to power it directly with 2 36/48v 350w smps... please let me know if you could do it... thx a LOT
Do a Fourier transform and find the high frequency element causing the notch and use the appropriate low pass filter to smooth out the notch
Or
This just occurred to me … don’t know if it will work!
Find the offending frequency using spectrum analyser and utilise the feedback opamp as a selen-key lowpass filter ?
Hahah I will try this on my own
Interesting 🤓
Does any one know where to buy the 4562 op amps?
Mouser, farnell, loads of other places. Search 'LM4562'. Great op amp.
Isn't this how a Quad amp works?? I wonder how about using Germanium output to lower the crossover dead zone for shits and giggles. Now that I think about it....does germanium have enough gain to make this design work? It seems to me that this amp is simpler then the EZ amp project your working on?
2022 slotted screw still exsist😯
I can't hear that little notch so I can sure use that simple thing for a simple 'in-the-shed-amplifier'. Gonna use it for hardcore so that little thing won't matter.
A long time ago I was into audio amplifiers but lost that when I got into digital electronics and as I'm not a real music lover and I sure can live without music but your vids sure made me wanna do something like this again, not a perfect amplifier, a thing like this one but boost it to lets say 50W so I can really feel those fast paced music in my guts 😂
I do though always through my years have wanted to build a class-d amplifier but haven't got into it and as it's more advanced than a class-a,
I never bothered, but I don't want a 10 pounds heat sink 😋
Can be used battery operated
Lm741 ?
Hello. If i put a pair of drivers in darlington mode, and an higher voltage op amp like opa2604, is there a chance to increase the power up to 30w in to 4 ohms load?
Darlington requires a higher base voltage to turn them on. I'd try a CFP output stage instead.
yeeeeeeeeee new video :D
Is that Floppa?
@@JohnAudioTech yup :D
Circuit please
Hi John. A simple modification can reduce the deadzone by half, while still leaving the amplifier in class B. I tried to post a link but it keeps deleting my comments :)
''Pure Class B' Ha ha.
Hi! Thanks for the experiment. How about adding a current source to transform it into a class XD amp? Please check “en.m.wikipedia.org/wiki/Class-XD_Amplifier”. I am quite curious… I’ll try it by simulation…