Hi all,
I have been listening to my leach amps for some time, and it seemed to me that there was something wrong with the sound. Hard to pinpoint, so I took it to my schools lab.
I did THD measurements, and found that there is about 0.1 % THD at 130 W into 8 ohms from 1kHz to 5 kHz . Haven’t made a TIM measurement yet.
Where is this caused? I tried to adjust the Qpoint in the output transistors, but it did not seem to make any difference.
I have used the following for my amps.
Input stage BJT’s:
BC546B / BC556B
VAS BJT’s:
BD139/140
Drivers:
MJE15032/33
Output:
Mjl1302A/3281A
Resistors are Standard 5% ones., except for the ones used in the emitters of the output, they are wirewound.
All caps are foil, except C4+5, they are standard 2.2µF ones. I know they are too small, but are they important to the sound? (haven’t got any others to replace them at the moment). C1, C7, C9, C10, C11 are all ceramic.
Thanks in advance
\Jens
I have been listening to my leach amps for some time, and it seemed to me that there was something wrong with the sound. Hard to pinpoint, so I took it to my schools lab.
I did THD measurements, and found that there is about 0.1 % THD at 130 W into 8 ohms from 1kHz to 5 kHz . Haven’t made a TIM measurement yet.
Where is this caused? I tried to adjust the Qpoint in the output transistors, but it did not seem to make any difference.
I have used the following for my amps.
Input stage BJT’s:
BC546B / BC556B
VAS BJT’s:
BD139/140
Drivers:
MJE15032/33
Output:
Mjl1302A/3281A
Resistors are Standard 5% ones., except for the ones used in the emitters of the output, they are wirewound.
All caps are foil, except C4+5, they are standard 2.2µF ones. I know they are too small, but are they important to the sound? (haven’t got any others to replace them at the moment). C1, C7, C9, C10, C11 are all ceramic.
Thanks in advance
\Jens
Jens:
Without getting into too much detail, Prof. Leach seems to have spent considerable effort over the years optimizing this amp with "proven technology" (in other words, "well known" or "old" 🙂 ) transistors. In this case, being old has one benefit that the device characteristics are fairly well known. I can't speak about the BCxxx and BDxxx input and driver devices, but the MJL/2SC3281-MJL/2SA1302 outputs have rather radical spec differences from the Motorola bipolars the amp was initially designed with. See Randy Slone's 1st book on amplifiers for some interesting comments about output stage frequency response due to using these high fT, extended beta devices. I think you would do well to spend some time reworking (reoptimizing) the bias points, frequency compensation, etc. for the amp as you have built it. I'm assuming, of course, that all the other "little" things that impact performance have been addressed (for example; good grounding scheme and general lead dress, clean rails, etc....)
On the other hand, are you sure that "wrong" sound you are hearing may simply mean that you're not used to listening to a low feedback design? 😉
Did you make other changes?
mlloyd1
Without getting into too much detail, Prof. Leach seems to have spent considerable effort over the years optimizing this amp with "proven technology" (in other words, "well known" or "old" 🙂 ) transistors. In this case, being old has one benefit that the device characteristics are fairly well known. I can't speak about the BCxxx and BDxxx input and driver devices, but the MJL/2SC3281-MJL/2SA1302 outputs have rather radical spec differences from the Motorola bipolars the amp was initially designed with. See Randy Slone's 1st book on amplifiers for some interesting comments about output stage frequency response due to using these high fT, extended beta devices. I think you would do well to spend some time reworking (reoptimizing) the bias points, frequency compensation, etc. for the amp as you have built it. I'm assuming, of course, that all the other "little" things that impact performance have been addressed (for example; good grounding scheme and general lead dress, clean rails, etc....)
On the other hand, are you sure that "wrong" sound you are hearing may simply mean that you're not used to listening to a low feedback design? 😉
Did you make other changes?
mlloyd1
JensRasmussen said:
Lodspeakers
They must be the dominating factor
in putting a sound caracter onto the music.
And also to some degree the match between
amplifier/loudspeaker.
See this interesting post:
http://www.diyaudio.com/forums/showthread.php?s=&postid=80556#post80556
I wish we had speakers with 0.01% dist.
Then they could be used for evaluate amplifiers
with 0.1% dist.
Say you want to to messure microvolts.
Would you use a volt meter with in-accuracy of millivolts?
I would not!
They must be the dominating factor
in putting a sound caracter onto the music.
And also to some degree the match between
amplifier/loudspeaker.
See this interesting post:
http://www.diyaudio.com/forums/showthread.php?s=&postid=80556#post80556
I wish we had speakers with 0.01% dist.
Then they could be used for evaluate amplifiers
with 0.1% dist.
Say you want to to messure microvolts.
Would you use a volt meter with in-accuracy of millivolts?
I would not!
Jens,
Can you try to describe in what way it doesn't sound right? Does this happen in both left and right channels or just one? Has it always sounded like this or is this a recent problem?
Can you try to describe in what way it doesn't sound right? Does this happen in both left and right channels or just one? Has it always sounded like this or is this a recent problem?
Jens,
To me it seems that C4 and C5 are used only smooth the bias diode reference voltage and thus they should not be too critical for SIZE but 2u2 might be a little small for the 100 uF requested. Try a larger one as soon as possible. Who knows it might affect the bias voltage which is not stable and that might very well make the amp sound not optimal.
/UrSv
To me it seems that C4 and C5 are used only smooth the bias diode reference voltage and thus they should not be too critical for SIZE but 2u2 might be a little small for the 100 uF requested. Try a larger one as soon as possible. Who knows it might affect the bias voltage which is not stable and that might very well make the amp sound not optimal.
/UrSv
The sound of things
Well, the sound is.....
Blury is the first word that comes to me, for lack of better.
It's the same for both channels. Thats ok, since they are twins.
Not sure what else to say, It's just not clean at all.... 🙁
Hope you have som idears...I'll change the two caps in the imput (C4 + C5) as soon as possible.
Thanks for any idears
\Jens
Well, the sound is.....
Blury is the first word that comes to me, for lack of better.
It's the same for both channels. Thats ok, since they are twins.
Not sure what else to say, It's just not clean at all.... 🙁
Hope you have som idears...I'll change the two caps in the imput (C4 + C5) as soon as possible.
Thanks for any idears
\Jens
Ceramic Capacitors.
I wouldn't use ceramic in an audio amplifier.
Not if I could avoid it.
They are noisy.
Try some plastic film. Polypropylene?
I wouldn't use ceramic in an audio amplifier.
Not if I could avoid it.
They are noisy.
Try some plastic film. Polypropylene?
JensRasmussen said:
Jens, when you look at the THD waveform, how does it look? Can you get a pic for us, or even a (good-looking) hand drawing?
Jan Didden
THD
Hi Jan,
I don't have the patterns, the instrument used just have an analog meter that reads out the THD.
I'll do an FFT on the thing when I have the time (school is busy at the moment, so there is not much time for DIY stuff)
What shall I look for even or odd harmonics? both?
Thanks for your time
\Jens
Hi Jan,
I don't have the patterns, the instrument used just have an analog meter that reads out the THD.
I'll do an FFT on the thing when I have the time (school is busy at the moment, so there is not much time for DIY stuff)
What shall I look for even or odd harmonics? both?
Thanks for your time
\Jens
Member
Joined 2002
ive had this problems with amps too it was my blown transofrmer i found this buy using my scope and looking at the wave on the psu see if there is any ac in the dc line and also if it is actualy the same wave on both windings. also cheak your sorce and wires.. might be your source creting the distortion and the amps are just amplifieing it.
Jason
Jason
Re: THD
Jens, depending on the THD meter there will be an output connector (I would hope!) before you go into the meter rectifier circuits that gives out the THD signal for a scope. What you should look for is the THD waveform relative to the fundamental. That will let you see if the THD is 2nd or 3rd harmonic, whether it has just noise and/or hum (it may not be THD at all, you know! These meters are very indiscriminating).
It will also let you see if you have xover distortion (should be 0-crossing related).
What signal generator did you use? Grounding arrangement? What type of THD analyzer?
If you can just get the waveform and the fundamental on the scope, put a transparent sheet over it and just draw the picture. I know, this is a quite revolutionary technique, but someone must be the first to do it. 😉 .
I know, questions, questions....
Jan Didden
There is nothing so practical as a really good theory - Ludwig Boltzmann
JensRasmussen said:
Jens, depending on the THD meter there will be an output connector (I would hope!) before you go into the meter rectifier circuits that gives out the THD signal for a scope. What you should look for is the THD waveform relative to the fundamental. That will let you see if the THD is 2nd or 3rd harmonic, whether it has just noise and/or hum (it may not be THD at all, you know! These meters are very indiscriminating).
It will also let you see if you have xover distortion (should be 0-crossing related).
What signal generator did you use? Grounding arrangement? What type of THD analyzer?
If you can just get the waveform and the fundamental on the scope, put a transparent sheet over it and just draw the picture. I know, this is a quite revolutionary technique, but someone must be the first to do it. 😉 .
I know, questions, questions....
Jan Didden
There is nothing so practical as a really good theory - Ludwig Boltzmann
Jens,
I have never heard the Leach amp. So I'm considering whether this blurriness is inherent to the design or whether your implementation is not optimal. Since both channels behave the same it is likely that the problem, if there is a problem, is in your power connections or ground connections.
In my opinion you are unlikely to find the problem by making THD measurements unless it is really severe. Better to use an oscilloscope and signal generator to see how the amp reproduces sinewaves and squarewaves into various loads. This usually shows up most things.
But before you do this it is very important to check the ground connections. Can you post a diagram showing how you connected all the grounds and earths? This is a common cause of blurred sound and amplifier instability and it is really easy to make a mistake here. For example the zobel network needs to be connected at the pcb and have a completely separate earth line to the psu star point from all other grounds.
Are you using two separate psus for each channel? If so, it really matters how you connect their grounds together. Please show how you have wired the supplies.
Have you made sure the grounds of the input connectors are insulated from the chassis? How do these ground get connected back to the psu star point?
Enough questions for now!
BAM
I have never heard the Leach amp. So I'm considering whether this blurriness is inherent to the design or whether your implementation is not optimal. Since both channels behave the same it is likely that the problem, if there is a problem, is in your power connections or ground connections.
In my opinion you are unlikely to find the problem by making THD measurements unless it is really severe. Better to use an oscilloscope and signal generator to see how the amp reproduces sinewaves and squarewaves into various loads. This usually shows up most things.
But before you do this it is very important to check the ground connections. Can you post a diagram showing how you connected all the grounds and earths? This is a common cause of blurred sound and amplifier instability and it is really easy to make a mistake here. For example the zobel network needs to be connected at the pcb and have a completely separate earth line to the psu star point from all other grounds.
Are you using two separate psus for each channel? If so, it really matters how you connect their grounds together. Please show how you have wired the supplies.
Have you made sure the grounds of the input connectors are insulated from the chassis? How do these ground get connected back to the psu star point?
Enough questions for now!
BAM
JensRasmussen said:
For my new leach amp, it seems to work alright, but I am getting distortion in the higher frequencies, much more then the regular leach amp. It looks like crossover distortion on a scope.
I measured my first leach amp several months ago:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=3057&highlight=thd
It was perfectly low throughout the entire frequency band. The new layout of mine has problems with higher frequencies. The only difference in the two, besides the new pcbs is the output devices. I am using the MJL1302A/MJL3281A devices. If I get time, I will swap them for MJL21193/4 devices and try it again. I just have been sidetracked.
I really haven't spent much time trying to find out the cause of the problem with my amplifier yet, as I have been sidetracked with other projects and work.
--
Brian
THD pattern
Hi
Made an xy plot of the output from the THD meter, it's like a semi circle, but a littel flat on the top.
I'll borrow a storage scope this weekend, and post pictures...
thanks
\Jens
Hi
Made an xy plot of the output from the THD meter, it's like a semi circle, but a littel flat on the top.
I'll borrow a storage scope this weekend, and post pictures...
thanks
\Jens
I would check all bias-currents first of all.
Then I would check hum&noise levels, as they are probably included in your measurements.
🙂
Also, try to measure not only total harmoniv distortion, but have a glance at each harmonic. The distortion-waveform will tell You a lot. Zero-crossing-distortion does in generall sound very bad 😡 😡 😡
best regards\Morello😎
Then I would check hum&noise levels, as they are probably included in your measurements.
🙂
Also, try to measure not only total harmoniv distortion, but have a glance at each harmonic. The distortion-waveform will tell You a lot. Zero-crossing-distortion does in generall sound very bad 😡 😡 😡
best regards\Morello😎
Found a problem maby
Hi.
Today I started invetigatin the amp, to see if I could find the problem causing the high THD.
I found that changing the railvoltage down made the THD fall also.
I started wondering about the spec. of the transistors used.
In the VAS stage i had used BD139/40 rated at 80 V CE, could it be that they are not rated high enough for the application?
And that too high a voltage CE could cause them to be very unlinear?
I have changed them in into MJE340/350, and will measure the THD as soon as possible.
Any idears are welcome with other transistors suited for the VAS
Thanks
\Jens
Hi.
Today I started invetigatin the amp, to see if I could find the problem causing the high THD.
I found that changing the railvoltage down made the THD fall also.
I started wondering about the spec. of the transistors used.
In the VAS stage i had used BD139/40 rated at 80 V CE, could it be that they are not rated high enough for the application?
And that too high a voltage CE could cause them to be very unlinear?
I have changed them in into MJE340/350, and will measure the THD as soon as possible.
Any idears are welcome with other transistors suited for the VAS
Thanks
\Jens
One step forward....
Just measured the THD with the new transistors. (MJE340/50)
At 1 kHz it's now 0.03% which is alot better than the 0.1% I measured with the BD139/40 in the VAS.
I have one question. How do one subtract the distortion from the generator itself? at 1 kHz the generator produces about 0.02% THD acording to my meter. What do you do to get the actual performance of the amp?
Conclusion. Don't use BD139/40 in applications where the CE voltage may exceed the 80 V specified 😉
Thanks for any help
\Jens
Just measured the THD with the new transistors. (MJE340/50)
At 1 kHz it's now 0.03% which is alot better than the 0.1% I measured with the BD139/40 in the VAS.
I have one question. How do one subtract the distortion from the generator itself? at 1 kHz the generator produces about 0.02% THD acording to my meter. What do you do to get the actual performance of the amp?
Conclusion. Don't use BD139/40 in applications where the CE voltage may exceed the 80 V specified 😉
Thanks for any help
\Jens
Without being able to compare the waveform of the
distortion of generator and amplifier, it's pretty iffy.
On the other hand, it's quite easy to rig an op amp and a
light bulb and get .00X%.
distortion of generator and amplifier, it's pretty iffy.
On the other hand, it's quite easy to rig an op amp and a
light bulb and get .00X%.
Op amp and a light bulb
I guess it's for an osccilator?
Could you draw a schmatic please?
Thanks
\Jens
Nelson Pass said:
I guess it's for an osccilator?
Could you draw a schmatic please?
Thanks
\Jens
- Status
- Not open for further replies.