| |||||||||||||||||||||||||||
| Moderators: cmoy | Welcome. Please log in. | ||||||||||||||||||||||||||
Registration is required to post a new topic or a reply. 
| |||||||||||||||||||||||||||
DIY Workshop » discrete buffer: jisbos (air wired) Steinchen  HeadWizer Joined: Oct. 12, 2005 |  Message [#1] posted on: 01-14-2007 12:08 PM CST (US).   |  |  this is a follow up from the interest check at the Marketplace jisbos board size: 43mm x 35mm or 1.7" x 1.4" - no smd parts, all parts are through hole schematic board layout: board legend:
Attachment: C2515.jpg,C2516.jpg,C2517.jpg |
| dBel84 HeadWize Fanatic Joined: May 5, 2006 | Message [#2] posted on: 01-14-2007 05:01 PM CST (US). | | Hi Steinchen, I was just curious about the offset drift on the jisbos. Have you found that you need to adjust the zero periodically or is it relatively stable? (if I got it right , this is the buffer you have on your SOHA) Any reason not to add a servo like the DB?..dB |
| Steinchen HeadWizer Joined: Oct. 12, 2005 | Message [#3] posted on: 01-14-2007 05:25 PM CST (US). | | I haven't encountered offset drift with this buffer, the offset stayed well below 1mV over hours. We could control the offset by an opamp like suggested by runeight: What's the opinion of the to-be builders ? Should we change the offset control ? [Edited by Steinchen on 01-14-2007 at 05:25 PM.] Attachment: C2522.gif |
| dBel84 HeadWize Fanatic Joined: May 5, 2006 | Message [#4] posted on: 01-14-2007 05:55 PM CST (US). | | It was a comment from PRR that made me think that offset might be an issue as the tube ages / tubes are rolled etc. I guess the tube bias will need to be adjusted in any case but it would be one less thing to have to think about. Servos can be a pain but Runeight generally gets it right the first time  ..dB |
raromachine HeadWize Fanatic Joined: Apr. 15, 2005 | Message [#5] posted on: 01-14-2007 08:26 PM CST (US). | | Is it feasible to line up the input and output and v -/+ points so one of those pin-header style plugs can be used? Sorta like: x x x x x i o g v v I'm happy to play with layouts if they can be opened with the free ver' of Eagle |
| Steinchen HeadWizer Joined: Oct. 12, 2005 | Message [#6] posted on: 01-15-2007 03:14 AM CST (US). | |
there is a coupling cap between tube and buffer with the SOHA, tube rolling / aging does not affect dc offset, offset depends on the jfets only.
that's possible, but the board would become larger. I understand that it would be nice to have input/ground, output/ground and V+/ground/V- headers but I don't understand why we should line them up all together. I'll try to add Molex headers tonight. [Edited by Steinchen on 01-15-2007 at 03:14 AM.] |
| dBel84 HeadWize Fanatic Joined: May 5, 2006 | Message [#7] posted on: 01-15-2007 06:55 AM CST (US). | | yes, I had overlooked the coupling cap. thanx..dB |
| Steinchen HeadWizer Joined: Oct. 12, 2005 | Message [#8] posted on: 01-17-2007 05:09 PM CST (US). | | changed and added pads for 2-pin and 3-pin Molex connectors: layout legend
layout: legend: Attachment: C2538.jpg,C2539.jpg,C2540.jpg,C2541.jpg |
| raromachine HeadWize Fanatic Joined: Apr. 15, 2005 | Message [#9] posted on: 01-17-2007 05:53 PM CST (US). | | They look good - the servo'd one is especially good for folk like me  |
PRR Headphone Council Joined: Mar. 18, 2002 | Message [#10] posted on: 01-19-2007 08:16 PM CST (US). | | In C2515.jpg, Q2 is drawn "wrong", source and drain interchanged. Both Sources should point to the output (via the offset trim network). In fact it will work fine this way. In a JFET, the difference between Source and Drain is zero to none. A few RF JFETs have a hair less gate capacitance if you use the designated Drain; this is moot in audio. The "problem" is just that someone who does not know that JFETs are 99.9% bidirectional may try to understant Q2 as a "drain follower" and wonder "WTF???". The board layout seems nominally correct. [Edited by PRR on 01-21-2007 at 01:23 AM.] |
| Steinchen HeadWizer Joined: Oct. 12, 2005 | Message [#11] posted on: 01-20-2007 04:13 AM CST (US). | | oops, thanks for checking that reminded me I had to rework the package of Q1, I used 2n3819 for the first layout which has a reversed pinout to 2sk170 attached updated schematic and layout schematic: board layout: legend print: [Edited by Steinchen on 01-20-2007 at 04:14 AM.] Attachment: C2566.jpg,C2567.jpg,C2568.jpg |
| PRR Headphone Council Joined: Mar. 18, 2002 | Message [#12] posted on: 01-21-2007 02:33 AM CST (US). | | > I haven't encountered offset drift with this buffer, the offset stayed well below 1mV over hours. I suspect that in domestic (not "industrial" or "military") use, the offset will be stable once trimmed. > We could control the offset by an opamp like suggested by runeight: Maybe the cold has frozen my brain, but.... isn't the TL081 being asked to follow the full AC signal, and drive the FET Sources? And isn't the '081 a low-speed low-current device? It is, in effect, "in the audio path". Its normal faults are inverted (when the '081 is too weak or too slow, the FET/BJT amp will slew harder and faster trying to bang the '081 into compliance); that's still a fault. Plus the '081 noise is injected into the input at unity gain. That may not be a problem at headphone level, but why pay a lot for big lo-noise discrete JFETs and flood them with 10X the noise from the dinky dirty JFETs on the '081? Personally, I think the '074 series is OK for light audio duty. But if I'm right, then why do we need these exotic JFETs and elaborations? Hang a BJT pair off the ''074 and be done with it. I may not be right, the '074/'081 chips may be junk, and if so.... Don't let the "junk" even see the audio! And you sure don't want this junk output shoved in under your input stage. The job of the '081 is DC. Clobber the AC before it gets to the '081. Don't drop the '081 output right to your input. What's the worst the DC offset can be, 100mV? What can the '081 swing, 10V? Knock-down the '081 output by a factor of 100 before bringing to the good amp's input. Then the self-noise of the '081 is much lower than fat-JFET noise. Any stray garbage the '081 makes is at least cut-down very significantly before it gets to the good input. Here's a model. I've reduced the JFET/BJT forward amp to an ideal opamp plus an arbitrary 69mV input error (as if the N-FET had 200mV Ggs and the P-FET had 269mV Ggs). I've modeled the '081 as an ideal opamp, with 10^5 gain to show the residual. The working gain is (10^5)/101 from the 5K/50 divider, so good-amp error is reduced about 1000:1. I did ignore '081 offset error.... this will add directly to output offset error. 99% of current BiFet chips will be well under 2mV error, and 2mV at the headphone jack is nothing. Note that the '081 has to swing ~7V to fight the 69mV error through the 5K:50 divider, but it has nothing else to do. Oh, and the R-C values shown give -0.03dB at 20Hz, -1dB at 3.1Hz. Attachment: C2571.gif |
amb Headphone Council Joined: Apr. 1, 2004 | Message [#13] posted on: 01-21-2007 03:24 AM CST (US). | | Yes, PRR is right. Note that in his drawing there is an additional capacitor to ground at the servo opamp's + input. Note also that the CK2III has the same cap. This filters out the AC signal component from the servo and lets it handle only the DC offset. However you cannot simply insert the modified servo in place of the original DC offset adjustment trimpot because we need 100% AC NFB to keep the buffer at unity gain. Thus PRR's drawing has R1 which provides the path for AC to get around the servo, and at the same time R4 and R1 forms a voltage divider to reduce the servo's output swing. All that said, I think for its intended use, the DC servo adds unnecessary complexity here. The original trimpot solution is just fine and saves quite a few parts. |
| Steinchen HeadWizer Joined: Oct. 12, 2005 | Message [#14] posted on: 01-23-2007 07:23 PM CST (US). | |
even worse, such offsets could occur just due to an unbalanced 7815/7915 psu, on top of that you get the offset from the buffer itself. We'd have to take offsets of several hundred mV into consideration. I already changed the layout to corporate your servo suggestion and I really like it's sophisticated design. What makes me hesitate is the fact that this is a significant change that has to be prototyped before proceeding and may imply intricacies. I'm totally swamped with projects and work to do, hence I'd like to take the easy way, keep the circuit simple and go back to the trimpot version which does it's job in my SOHA just fine. PRR, thank you very much for your input and suggestions. I was able to cram a small cap across the vbe multiplier into the existing layout so board size stays the same. Furthermore I fixed the wrong pinouts of the jfet packages. Moreover changed the lead spacing of the electrolytics from 5 to 3.5mm schematic: board layout: board legend: [Edited by Steinchen on 01-23-2007 at 07:25 PM.] Attachment: C2577.jpg,C2578.jpg,C2579.jpg |
| Steinchen HeadWizer Joined: Oct. 12, 2005 | Message [#15] posted on: 02-05-2007 07:57 AM CST (US). | | I ordered proto PCBs, ETA is Friday 16.02.2007 |
FallenAngel Member Joined: Sep. 27, 2006 | Message [#16] posted on: 02-06-2007 09:25 AM CST (US). | | Hey Steinchen, I'm about to finish this buffer on breadboard and I'm wandering on a couple of things. 1) With this buffer having 1K inline on the input, then 1M to ground, do we still keep the 1M to ground on the input of the opamp (R5/R15 in the original SOHA)? 2) Are the 100uF to ground electrolytics really necessary, I'm using 100uF after the regulators already anyway. Thanks! |
| Steinchen HeadWizer Joined: Oct. 12, 2005 | Message [#17] posted on: 02-06-2007 10:57 AM CST (US). | |
the 1M to gound on the buffer board is for versatility with other amps, with the SOHA one of the 1M resistors should be omitted
those elcos might be beneficial with long wires from the psu to the buffer board, with short wires of 3 inches or so they can be omitted |
| amb Headphone Council Joined: Apr. 1, 2004 | Message [#18] posted on: 02-06-2007 02:27 PM CST (US). | | I don't know if it's obvious, but this buffer is also suitable as a standalone unity-gain headphone amp for low-Z headphones. Just add a volume pot and power supply.  |
| Ryssen Member Joined: Feb. 16, 2007 | Message [#19] posted on: 02-18-2007 06:02 PM CST (US). | | Well,if I missed the interesst check can I still be interessted? |
| Steinchen HeadWizer Joined: Oct. 12, 2005 | Message [#20] posted on: 02-18-2007 06:38 PM CST (US). | | received the proto pcbs:  almost completed the build, expect my report the next days
[Edited by Steinchen on 02-18-2007 at 06:39 PM.] Attachment: C2661.jpg |
| Steinchen HeadWizer Joined: Oct. 12, 2005 | Message [#21] posted on: 02-25-2007 07:57 AM CST (US). | | proto finished, working fine and sounds excellent. I thought about bridging the 51 Ohms resistors and changing the offset trimpot from 100 to 200 Ohms but the present circuit easily adjusts more than 3mA IDSS difference at the input JFETs. I'm just going to give a little more space to the molex connectors, everything else is just fine.
[Edited by Steinchen on 02-25-2007 at 07:58 AM.] Attachment: C2680.jpg |
| 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.