For small signal work yes.
As Jan has pointed out, slew rate and rise time are not the same. For an opamp operating in the small signal region, you can expect rise times of less than 100 ns - i.e. damn fast and more than you will need for audio.
Further, as Scott has noted, opamps with 100's V/us are readily available now with very good distortion and noise performance.
For small signal work (line amps, RIAA etc) you have no problem in these modern times. Stride forth, avail yourself of the opportunity to appply some leading edge technology and use with confidence!
Just follow the data sheet guidelines for layout, decoupling etc.
As mentioned a few pages back, here is a link to my class A headphone amp. This uses a similar control technique for the class A bias being discussed here. I have a subsequent version using two transistors to control the bias.
With this version, using just an EF, distortion is sub 1ppm into 32 Ohms. Certainly with an EF2, higher supply rails to support more power, this kind of performance would be maintained.
I am using a simpler scheme in my Symphony preamp. The current source loads either side of the bias controller make a big difference. If you use resistors instead, the distortion is about 5x higher.
http://www.diyaudio.com/forums/head...-1-class-headphone-amplifier.html#post3988037
With this version, using just an EF, distortion is sub 1ppm into 32 Ohms. Certainly with an EF2, higher supply rails to support more power, this kind of performance would be maintained.
I am using a simpler scheme in my Symphony preamp. The current source loads either side of the bias controller make a big difference. If you use resistors instead, the distortion is about 5x higher.
http://www.diyaudio.com/forums/head...-1-class-headphone-amplifier.html#post3988037
On the whole 20Hz period is a total of 140000 samples (2,8MHz / 20Hz = 140000).
On the whole 20kHz period is a total of 140 samples (2,8MHz / 20kHz = 140). 35 samples on the rising edge of the signal. What would be the quality at full amplitude?
It depends on the filter being used, how much voltage for the filter allows much noise leaves (sample) voltage.
SACD converter can afford to 20kHz full level, but in the real sound of such a value is not present.
@PMA,
Those little caps are Bruno's solution for not needing an output Zobel/Thiele network, I never gave them much thought as he seemed pretty sure of himself that these would be the ticket.
Regardless, thanks for posting the loop gain plot, I was meaning to give that another try tonight as my day was jam packed full of other stuff and I simply didn't have time.
Those little caps are Bruno's solution for not needing an output Zobel/Thiele network, I never gave them much thought as he seemed pretty sure of himself that these would be the ticket.
Regardless, thanks for posting the loop gain plot, I was meaning to give that another try tonight as my day was jam packed full of other stuff and I simply didn't have time.
Let's forget for a moment that these caps are connected to the OPS emitters and pretend they're connected directly to the output. In this case they are in parallel with the feedback resistors (47k) ; since they're 2.2pF, they'll provide some phase lead around 1MHz, which fits very well with the open loop gain/phase plot on the previous page.
Now, they are connected to the emitters and not the output, and that's why the zobel seems to be unnecessary.
Now, they are connected to the emitters and not the output, and that's why the zobel seems to be unnecessary.
I was doing a manual filter on stacks of AES papers and came across this one --
Concept described in AES 1683 (D-4) preprint? George Dodson, JBL.
View attachment WW Class B.pdf
THx-RNMarsh
Concept described in AES 1683 (D-4) preprint? George Dodson, JBL.
View attachment WW Class B.pdf
THx-RNMarsh
FWIW what is called the Allison connection was published by Barry Blesser in 1969 in the JSSC.
Thx- I'll take a look. I believe I read it already as JC and I flew to NY on a great airplane deal and were at this AES presentation. During the Q and A afterwards, I said to JC that this had already been shown.
[The commercial JBL power amp used the diode across the Re of the OPS.]
AES paper shows a smoother transistion from class A to AB.
[presented at 67th convention. 1980 Oct31/Nov 3 New York]
THx-RNMarsh
Last edited:
Sorry, submitted in 1969 the actual publication was June issue 1970. He shows the basic connection using feedback to eliminate the Vbe error of the output device, but does not go into a complementary version for use as an output stage. The obviousness or non-obviousness of that is in the eye and imagination of the beholder.
EDIT - I forgot what an interesting guy he was...http://www.blesser.net/downloads/eContact Loud Music.pdf
Last edited:
Nice paper.... also saw his web site.... good references there but hasnt been updated since 2010. where is he today? [I fear the worse].
Mainly, I am wondering if this concept can be used here for the ExtremA to keep the distortion down when leaving class A.
THx-RNMarsh
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.