| |||||||||||||||||||||||||||
| Moderators: cmoy | Welcome. Please log in. | ||||||||||||||||||||||||||
Registration is required to post a new topic or a reply. 
| |||||||||||||||||||||||||||
Clutz HeadWize Fanatic Joined: Jun. 1, 2002 |  Message [#120] posted on: 11-26-2006 12:39 AM CST (US).   |  |  You guys answered my questions perfectly, thank you. As with a lot of things in hifi audio, there seems to be a lof of dogmatic opinions on the "right' way of doing things, and I've often noticed that when you ask people on one side of the argument about the other - they give an incomplete, or distorted picture. I really appreciate the effort you guys put into this response.  Bottom line is - now I want to build both |
amb Headphone Council Joined: Apr. 1, 2004 | Message [#121] posted on: 11-26-2006 03:41 AM CST (US). | | Btw, regarding our recent modification to use 2SK170/2SJ74 JFETs at the input (as opposed to BJTs); I ran some distortion spectrum simulations earlier on both, and found that the JFET version lower in distortion across the board. I didn't capture the graphs but suffice to say that the change to JFETs does a lot of good things for this amp. |
| Dave_M HeadWizer Joined: Jul. 1, 2006 | Message [#122] posted on: 11-26-2006 09:25 AM CST (US). | | Can the CKIII be improved with global feedback the same way as the b22? If the amp topology will allow it? I also noticed you are sticking with the +/- 15V rails. Is that because of the problems with the servo with my prototype? You could use LM317 type regulator adjusted for +/- 24V (then you can have capacitor on the Adj pin to improve the PSU performance). And use small 15v regs like in TO-92 style for the opamp's supply. It would fit without having to make the board bigger I think. It's not like 15V isn't high enough but it's good to have a a lot of headroom for those dynamic peeks or at least to allow the transistors to work completely in the linear region. |
runeight Headphone Council Joined: Mar. 8, 2002 | Message [#123] posted on: 11-26-2006 12:49 PM CST (US). | | It is possible to add global nfb to the CKIII, but it doesn't really help too much. This is because the amp, being designed to have no nfb, has so little gain. When an amp is designed to use lots of nfb it also must have alot of open loop gain so that when nfb is applied the gain is reduced to something sensible. The b22 has an open loop gain of about 800x which is reduced by the nfb to something less than 10x. As amb points out, opamps have gains in the millions and take lots of nfb. The CKIII only has a gain of 9x. I've tried adding about 6db of nfb which gets the gain to 4x, but the improvement is not worth whatever negative effects might come from adding nfb to an amp not design to have it. Talked this over with amb a few days ago and we agreed that application of nfb to this amp tends to defeat the original idea behind the design. So we're leaving it out. Dave_M, I don't think that this amp is unstable at 24V. I think there must have been a problem with your particular situation that we weren't able to figure out at the time. The amp doesn't really need 24V to run well. All this does is add more power dissipation and bigger capacitors and a bigger transformer. If you're using 32R headphones and the amp can get to 14V, that's 435mA into your headphones. That's quite alot of power. If the amp can get to 23V that's 720mA. I'm pretty sure that you won't get there with most headphones. The amp is otherwise running fairly linearly. I've reset some of the bias currents to do this and the CFP output stage runs much more linearly than the original darlington. So, I guess I don't see the need. But, I have been wrong before. |
| Dave_M HeadWizer Joined: Jul. 1, 2006 | Message [#124] posted on: 11-26-2006 02:56 PM CST (US). | | ok point taken. I was just trying to extract the most performance from the amp. I doubt it would make much difference anyway. But I think with 15v supplies the amp would be able to put out 9-10V RMS without clipping. Senn HD650 are rated at 103dB SPL at 1V RMS so the point is moot anyway. But then you might say why not make the supplies +/- 5v or less because that looks to be more than enough even for high impedance headphones. Then you can use smaller transformer and reduce power consumption. I just think you need a good dollop of over engineering to make high end audio gear |
| runeight Headphone Council Joined: Mar. 8, 2002 | Message [#125] posted on: 11-26-2006 11:03 PM CST (US). | | yes, that's a good point too. So, we're basically making a trade at 15V. The op points at 5V would not be good. The op points at 15V are good. The op points at 24V might be a little better. But the trade is probably not worth the extra effort and parts changes. You are right that the amp will make about 9V before clipping. Is that enough? I think that it is, but I am happy to hear other opinions. I don't think that amb is going to mess with the boards any further (although I cannot speak for him), but now is the time to talk this over and not after the boards are ordered from the pcb supplier. One other thing is that dBel84 has the 15V amp. Maybe he can supply further thoughts on its performance. |
| awpagan HeadWizer Joined: Jan. 6, 2005 | Message [#126] posted on: 11-26-2006 11:04 PM CST (US). | | Just a note for the lm317 case http://www.acous...reg_notes1.html I'll see if i can find other other site's comparing noise between 78/79's and 317/337's but then? ....noise levels |
| dBel84 HeadWize Fanatic Joined: May 5, 2006 | Message [#127] posted on: 11-27-2006 09:29 PM CST (US). | | I am not sure that I can add anything more about the performance of this amp. I once again look forward to relaxing at the end of the day listening to music, if that is a tribute. The beauty about the CK2 / CKIII is that it doesn't add anything to the music. It has taken me a while to come to this realisation and it only really dawned on me after I had been listening to a PIMETA and SOHA before listening to the CKIII. The other amps have a distinct signature, not a bad thing and both enjoyable to listen to, but... the music just sounds right with the CKIII. I honestly don't see the need to change anything on this amp eventhough I realise that the final schematic differs somewhat from the version I built. Perhaps one day I will find someone who can do some measurements, not that it will change my opinion; and that is all I have to offer , my opinion ..dB |
| awpagan HeadWizer Joined: Jan. 6, 2005 | Message [#128] posted on: 11-28-2006 12:07 AM CST (US). | |
Then don't change this amp. The sign of a good system(or amp/preamp/source etc) is you don't hear it, just the music. Now the next step is to improve on it allan |
| amb Headphone Council Joined: Apr. 1, 2004 | Message [#129] posted on: 11-29-2006 07:18 PM CST (US). | | I think I've run enough simulations now, and am happy with it. At the moment I am busy creating new websites for the β22 and σ22 projects, but as soon as I get done with that, I will prepare and order boards. dbel84 has already built a preliminary prototype, and even though it differs a bit from what we have now, I am going to take a leap of faith with the simulation results, skip the prototyping phase and go straight to production boards. Since this will be an easy amp to build, I will build one and test it before releasing the boards. My best guess now is that boards should be available by the end of the year. The response in the marketplace interest check thread has been most impressive and I will keep it open. |
| amb Headphone Council Joined: Apr. 1, 2004 | Message [#130] posted on: 12-01-2006 04:04 PM CST (US). | | A hint: Regarding the Amveco TE62033 transforer, just a few days ago Digi-key was out of stock on this item and now they have some. There are 95 in stock as I write this, so if you're planning to build this amp soon, you might want to get it while they last. Who knows how long the lead time will be the next time they run out?  |
alfie Member Joined: Jul. 12, 2006 | Message [#131] posted on: 12-02-2006 05:23 AM CST (US). | | MUSQ: I can only find BD139s and BD140s in TO-126 package. Is it fine? Otherwise, what other transistors could I use? |
| amb Headphone Council Joined: Apr. 1, 2004 | Message [#132] posted on: 12-02-2006 05:38 AM CST (US). | |
TO-126 is effectively the same as TO-225AA. So they should work fine. Btw, what brand of these transistors are you looking at? |
| dBel84 HeadWize Fanatic Joined: May 5, 2006 | Message [#133] posted on: 12-02-2006 06:05 AM CST (US). | | I don't think I have ever mentioned this but I used MJE340 / MJE350 instead of BD139/140 for the O/P transistors. This was after a coment in the original thread that referred to the MJ series as having better sonic characteristics over the BD?..dB |
| alfie Member Joined: Jul. 12, 2006 | Message [#134] posted on: 12-02-2006 06:29 AM CST (US). | |
I was looking at ON Semiconductor as in http://tinyurl.com/y5hozn |
| Dave_M HeadWizer Joined: Jul. 1, 2006 | Message [#135] posted on: 12-02-2006 06:59 AM CST (US). | |
Yeah, I made that comment. I haven't compared them myself, but the general opinion over at diyaudio.com is that BD139/140 are not bad but they are just general purpose types that you would put in a power supply. The MJE340/350 are nicer and intended for audio use. I think mine are TO-126. I have them mounted to TO-220 heatsinks. Power dissipation isn't high enough for the smaller form factor to be a problem. [Edited by Dave_M on 12-02-2006 at 07:00 AM.] |
| amb Headphone Council Joined: Apr. 1, 2004 | Message [#136] posted on: 12-02-2006 07:04 AM CST (US). | |
The datasheet you linked to shows it as TO-225. Btw, that's an old version of the datasheet. The modern ones are here: The Fairchild, STmicroelectronics, Philips and Siemens versions of these all lists them as TO-126 (=SOT32). As you can see, they are all just about identical. |
| amb Headphone Council Joined: Apr. 1, 2004 | Message [#137] posted on: 12-02-2006 07:21 AM CST (US). | |
I don't really buy that. The only objective things to consider are their specs, and the BD139/140 are just fine. A couple of important spec that I look for that isn't in any of their datasheets is the Cob and Ft, but I found that the Philips datasheet for the BD139/140 with an Ft rating -- 190MHz, which is quite good for a medium power transistor. The MJ340/350 is rated at only 0.5A whereas the BD139/140 is 1.5A, but the MJ340/350 has much higher Vce and Vcb ratings which are not important in this amp (our rail voltages are far lower than the limits of any of them). The Hfe for the -16 suffix version of the BD139/140 is in a somewhat higher span than the MJ340/350, which is a good thing. The Hfe vs. Ic graphs are comparable. While 0.5A is still a fair margin higher than the output current that this amp will ever encounter, I think the BD139/140 would have a better chance at surviving an unintended momentary output short to ground than the MJ340/350. I would really like to know the Ft and Cob specs of the MJ340/350... Another popular alternative are the MJ243/253. These are rated at 4A, but the Ft is lower at 40MHz. |
| alfie Member Joined: Jul. 12, 2006 | Message [#138] posted on: 12-02-2006 07:30 AM CST (US). | | And the final question is... How do you see this amp in a balanced configuration? |
Steinchen HeadWizer Joined: Oct. 12, 2005 | Message [#139] posted on: 12-02-2006 09:03 AM CST (US). | |
the datasheet of MJE340/350 states them for "general purpose applications" and the Hfe plot does not look very linear, therefore I'd be surprised if they performed stellar. I tried a dozen of different output transistors while developing the dDB and the sonical differences have been rather subtle (among those who worked fine). From my experience I can recommend: Please don't take this list as an all time top ten or so, in the end performance depend on various dependencies like operating points, bias currents, rail voltages, ... A transistor that performs well in a +-60V power amp may be inferior in a +-15V headamp. All the curious among you are invited to join me in the transistor shoot-out once the CK-III pcbs are available |
| dBel84 HeadWize Fanatic Joined: May 5, 2006 | Message [#140] posted on: 12-02-2006 11:26 AM CST (US). | | That would be interesting indeed Steinchen. This would apply to all CFP designs which I can see me using again. ( and DaveM - sorry for not referencing you, I remembered that it was you after I had posted)..dB |
| Page 1 2 3 4 5 6 7 8 9 10 11 12 13 14 > >> | Next Page | Prev Topic | Next Topic | ||||
 (remove _nospam_ ) |
© Chu Moy, 2001.