Anyone built Sijosae's class AB amp ?

Discussion in 'DIY (Do-It-Yourself) Discussions' started by ozshadow, Jun 1, 2006.

  1. Seaside
    Quote:

    Originally Posted by .:tuxianer:. /img/forum/go_quote.gif
    Ok. I think I will take the standard parts. Is ther any final schematic? And wich version sounds better? I or II?

    And what's this: IMG




    The red circled section of that image is TREAD like power supply. I think this is thedoctor's board, and the section is optional. He did it that way because that is simply better in terms of stability.

    I think you should start with standard parts and standard layout just as the way sijosae did. It is not like I do not want you to be creative, but I think his layout is very carefully designed one, vey clever layout that did not take up whole lot of space. He is not only an expert at electonics but also at miniaturation too.

    Which is better I or II... that is good question.
    They are basically similiar things with different idea.
    I do not think one is better than the other though, I found I like "I" more than "II".

    Just for your Information, one comment on the power section of his layout, as you can see, there are 4 transitors. 1 pair of transitor close to the power source are... not really working like transistors, since their leads are connected. They are working just like diodes. So, you can replace it with diodes, 1N917 or 1N4148 type. In fact I highly recommend you do that, because in that way you don't have to worry about things like TR hfe matching. Sijosae himself once stated that he originally designed it with diodes but used TRs as alternative because he did not have diodes at hand when he made his.
     
  2. cetoole
    Quote:

    Originally Posted by Seaside /img/forum/go_quote.gif
    Just for your Information, one comment on the power section of his layout, as you can see, there are 4 transitors. 1 pair of transitor close to the power source are... not really working like transistors, since their leads are connected. They are working just like diodes. So, you can replace it with diodes, 1N917 or 1N4148 type. In fact I highly recommend you do that, because in that way you don't have to worry about things like TR hfe matching. Sijosae himself once stated that he originally designed it with diodes but used TRs as alternative because he did not have diodes at hand when he made his.



    I dont see why the beta of the transistor plays a factor when it is diode connected. When used like this, the transistor IS a diode, and one which better matches the characteristics of the output transistors; a benefit in my mind.
     
  3. Seaside
    You're right cetoole. My point is to avoid tedious job like TR matching. You may able to find schematic by sijosae, that uses diodes in place of those transtors. Any of both will work fine, becasue they basically are the same thing. I happened to use diodes because I had them in my drawer.
     
  4. .:tuxianer:.
    Ok but can anyone post the latest layout? And what function have the Diamond puferr and where should i put th buffer? The Railsplitter makes a virtual Ground right?


    edit: ok the buffer strengthens the signal again! Is this right? What of the shown Buffer layouts is the best?
     
  5. FallenAngel Contributor
    Quote:

    Originally Posted by thedoctor /img/forum/go_quote.gif
    i'm not exactly educated in Eagle, but here's a sketch. trying to learn how to use it to draw schemtics, but free time is running low.
    [​IMG][​IMG]




    Thread revival from quite a bit ago, but I think I'm allowed for one reason... I built this amp, and in a Hammond 1455C802@ [​IMG]

    I threw in a nicer PSU, similar to PPAS:
    TLE2426
    2x 470uF V+ to V-
    100uF & 1uF V+ to ground
    100uF & 1uF V- to ground

    No rail isolating JFETs but there is a trickle charger for the 9V battery hard-wired to the DC jack; as you'll see in the photos, there was just absolutely no room.

    After some testing, I settled on OPA2107 on Left/Right and AD744 compensation pin output on Ground because I really like the sound of OPA2107 and the current draw is nice and low. Full amp current draw is 22mA, so the 270mAh AccuPower 9V lasts over 12 hours. Now what kind of diamond buffered portable amp can say that! [​IMG]

    Yes, it sounds very good, I'll just say that. I don't have many portable amps (only PPAS) and I don't trust to compare something I built and worked on hard to other amps on pure memory so I'll have to get somebody else to compare for me.

    Without further ado, pics!
    [​IMG][​IMG][​IMG][​IMG]
     
  6. TzeYang
    Nice!

    The only issue i have is the OPAMPs. Both chips don't seem to love low voltages.
     
  7. FallenAngel Contributor
    While both chips perform better at higher voltages, according to the datasheet and while listening, they run perfectly well off 9V.

    I unfortunately burned the AD8066 I had in there before and didn't like the AD823 that much.

    The opamps driving buffers does make it a lot easier on them instead of driving headphones directly so they should never clip. [​IMG]
     
  8. Spacehead
    I am sorry for digging up old thread. But today I made this amp, but with active ground. I think Sijosae's discrete rail splitter sounds really bad. It just ruins good stereo image and bass.
    I did left and right channels exactly following Sijosae's picture. It took some hours, I dunno what took so long, but maybe it is just so precise work. I made sure that I didn't get any cold joints.
    [​IMG]
    The ground is very very simple dual op amp buffered TLE2426. The dual op amp is fed TLE2426 output and bridged with 5 ohm resistor for double current, and unity voltage gain. It is stable and high current TLE2082AI, for now.
    [​IMG]
    Voltage to the 4700 uF capacitor comes from a regulated (adjustable LM7805) 15V supply.

    To the input I added 0.1 uF plastic capacitors


    It sounds excellent! Should be superb for battery operation and portable too!
    I can hear the mixture of Class A and Class B operation. It isn't as smooth as pure Class A amplifier but for the size of it, it still is very good. And it is very accurate with the OPA2107 op amp. I could try other op amps there too. I think the transistors don't have any sound of their own, which of course is a good thing. I am just afraid that they burn up, they run hot.
    I am using 47 ohm output resistors, because I am afraid that possible short circuits when plugging the headphones would burn the transistors. I would just hate to do the amplifier again. Now enjoying this amp. [​IMG]
     
  9. diditmyself
    Quote:

    Originally Posted by Spacehead /img/forum/go_quote.gif
    I think Sijosae's discrete rail splitter sounds really bad.



    Indeed it does. The output impedance is far to high to serve as a ground for low impedance returns like phones. The crosstalk will be bad unless large capacitors from rails to ground are used.
     
  10. Fred_fred2004 Contributor
    I might as well chip in [​IMG]

    I used a TLE2426 rail splitter, and an OPA2227, Sijosae's excellent plans made the layout very easy, and it sounded very respectable, battery life was good as well


    [​IMG]

    [​IMG]

    cheers
    FRED
     
  11. Spacehead
    I checked DC offset and I get plenty of it if I use other op amp than OPA2132. But I don't like OPA2132s sound and wished to use OPA2107. Could I get DC offset because I added capacitors in series to input rca connectors?
    I also built a ground channel which has parallel op amps for higher current capability:
    [​IMG]
    [​IMG]

    The resistor in ground channel are
    op amp bridging 12, 6 and 5 ohm
    feedback 27k / 5.56k to ground
    first op amp to buffer op amps -> 47 ohm

    With OPA2132 I get -30 mV of DC offset, with OPA2107 1100 mV
     
  12. ericj
    Quote:

    Originally Posted by Spacehead /img/forum/go_quote.gif
    I checked DC offset and I get plenty of it if I use other op amp than OPA2132. But I don't like OPA2132s sound and wished to use OPA2107. Could I get DC offset because I added capacitors in series to input rca connectors?
    I also built a ground channel which has parallel op amps for higher current capability:

    The resistor in ground channel are
    op amp bridging 12, 6 and 5 ohm
    feedback 27k / 5.56k to ground
    first op amp to buffer op amps -> 47 ohm

    With OPA2132 I get -30 mV of DC offset, with OPA2107 1100 mV




    Are you getting these offsets with or without the ground channel?

    If with, start by removing the ground channel. Which, fwiw, should be at unity gain.

    Capacitors in-line with the inputs would remove rather than add offset.

    If your opa2107 is showing a DC offset, it's fake. The opa2107 is a precision laser-trimmed jfet-input opamp. That's why it costs so much.
     
  13. Spacehead
    Quote:

    Originally Posted by ericj /img/forum/go_quote.gif
    Are you getting these offsets with or without the ground channel?

    If with, start by removing the ground channel. Which, fwiw, should be at unity gain.

    Capacitors in-line with the inputs would remove rather than add offset.

    If your opa2107 is showing a DC offset, it's fake. The opa2107 is a precision laser-trimmed jfet-input opamp. That's why it costs so much.




    I recall reading something like that if I add capacitors to this amps input I need to add resistors to ground too. This doesn't have input resistors like Cmoy, could that cause op amp to malfunction on certain situations?
    Now I have the ground channel in place. Can I use feedback resistor with unity gain? I am trying to get my ground channel working too, because I need high current capability and I don't have BUF634 and I want to use lot of op amps. At least I want to try if I can make it work.

    I have ordered those OPA2107s from TI as samples. They should be working, yeah, I tried another one and same results.

    I had this working before with another ground channel. But I tried to recreate another with NJM4556S, TLE2426 and Capacitor and it didn't work, I didn't use resistors, and NJM4556S was in unity gain.

    I was pretty disappointed today, but now I am listening to the device with OPA2132 x 4 and it works, has bass too.

    I am doing this because my Sennheiser HD-201s need a LOT of current (24 ohm).
     
  14. Spacehead
    Yeah! I solved it. It was the fact that I didn't have 10k volume pot like Sijosae's drawings suggest! That caused my main op amp to malfunction as it's input bias currents had no place to flow. I added 27K resistors to RCA connector
    [​IMG]
    and adjusted the feedback on ground channel, used 3K feedback resistor and didn't ground it so it is at unity gain. DC Offset is now with OPA2107 very low, 0.6 mV for both channels.
    I power it with 7805 adjusted to +15V
    [​IMG]
     
  15. ericj
    Quote:

    Originally Posted by Spacehead /img/forum/go_quote.gif
    I recall reading something like that if I add capacitors to this amps input I need to add resistors to ground too. This doesn't have input resistors like Cmoy, could that cause op amp to malfunction on certain situations?
    Now I have the ground channel in place. Can I use feedback resistor with unity gain? I am trying to get my ground channel working too, because I need high current capability and I don't have BUF634 and I want to use lot of op amps. At least I want to try if I can make it work.




    Yeah, you need a resistor from +input to ground if you add the caps.

    I still don't understand what you're on about wrt the feedback resistor on the ground channel.
     

Share This Page