| |||||||||||||||||||||||||||
| Moderators: cmoy | Welcome. Please log in. | ||||||||||||||||||||||||||
Registration is required to post a new topic or a reply. 
| |||||||||||||||||||||||||||
gewa  HeadWizer Joined: Sep. 22, 2006 |  Message [#100] posted on: 10-20-2006 07:11 PM CST (US).   |  |  tomb, digi01 Should look something like this. No PS decoupling cap's yet. Regards Attachment: C2181.jpg |
tomb Headphone Council Joined: Mar. 1, 2006 | Message [#101] posted on: 10-20-2006 08:01 PM CST (US). | | Looks great to me. I guess we need Digi to tell us what's possible, and how much, then pick which ones to be made? It would be nice if the board mfr charges by the sq.in. and allows some saw cuts. We might get 'em all that way. |
| searleb HeadWizer Joined: Jan. 2, 2005 | Message [#102] posted on: 10-20-2006 10:20 PM CST (US). | | Making cuts (as long as they're straight) is pretty easy with a dremel tool, as with Tangent's MINT amp. We could get single, larger boards with several of these (and several options) printed and users would cut what they'd want. [Edited by searleb on 10-21-2006 at 03:46 AM.] |
GregVDS HeadWizer Joined: Oct. 25, 2005 | Message [#103] posted on: 10-21-2006 02:43 AM CST (US). | |
Gewa, I wonder, did you just arrived at the same result as me, or did you reproduced my design in your software? I'm practically sure this kind of design can be achieved also for the Intersil 5002 or something like that. for this one, it could be useful to have Epsilon12 board in the output circuit, in order to have headphones plugged before turning on the amp, but still really connect left and right signals after a delay for the tube to heat and stabilize. I'm working on a reduced epsilon12 board which do this delay plugging. I think from my point of view the double single opamp is more interesting, if you already want to go this way, why stack this board, and then a browndog, just longer signal paths. And if I insist, the smallest, the best, for the same signal path length reason, and for boxing convenience. The idea to buy a big board to be sawed with different buffer designs is interesting, at least this could satisfy everyone. I don't know what's possible, Digi01, it's up to you all the best, GregVDS |
| gewa HeadWizer Joined: Sep. 22, 2006 | Message [#104] posted on: 10-21-2006 06:57 AM CST (US). | |
Yes and no. I already had designed it partialy but after I saw yours I used it as a guide to tidy up mine. If you look closely you wil see that they are not the same. I left out two connections because they are already on the SOHA board. For the 5002 board: are those decoupling cap's mandatory or not? If yes the board will slightly increase in outer dimensions and a connection have to be made to GND. All my boards have a groundplane BTW, just for this one it has to have a physical connection to the SOHA board. Regards |
| GregVDS HeadWizer Joined: Oct. 25, 2005 | Message [#105] posted on: 10-21-2006 07:22 AM CST (US). | | I absolutely don't know about the Intersil stability, maybe a look at its spec pdf could help. I find the use of this one quite dangerous, for if you want to compare two headphones (and so deplugging - replugging while the amp is working, so shorting output of the chip), you absolutely need output resistors so as to avoid frying the chip, and I'm not inclined to use them. All the best, GregVDS |
| gewa HeadWizer Joined: Sep. 22, 2006 | Message [#106] posted on: 10-21-2006 09:32 AM CST (US). | | Attached a part I took out from the datasheet. If this does the same job to prevent the chip from frying (tomb post#92) than we don't have a resistor in the signal pad. For the more intelligent people here, correct me if I'm wrong. Regards Attachment: C2182.jpg |
| tomb Headphone Council Joined: Mar. 1, 2006 | Message [#107] posted on: 10-21-2006 11:48 AM CST (US). | | Yep, my 5002 datasheet shows the same thing. Problem is, I don't think this is conducive to stacking. My only suggestion to using the output resistors was that despite reports to the contrary, I can't tell the difference in sound on various Milletts with a minimal output resistor (10ohms) or none at all. Seeing as how the resistor pads are already there on the basic SOHA board, this might be the easiest way to go. There is a additional problem identified by Amb in a parallel conversation over on DIYForums.org. If both these things are true, then the only option for parallel Intersils is a separate board mounting, with resistor circuits inbetween. Amb is stating that in a general case, current-sharing resistors are used inbetween paralleled output devices to prevent one output device from becoming the load of the other. You see this all the time in transistor circuits, where the output is taken from the junction of two resistors between two transistors especially MOSFET's. I'm confused, though - because I don't understand how the BUF634 includes this internally (as Amb states). There is a "diode clamp" combined with an input resistor, but this is for the basic short-circuit protection. There seems to be nothing that would form a resistance between the outputs of stacked BUF634's. There are resistors between output trannies, but none on the trannies behind those, while the reverse is true on the Intersil's. Wouldn't the un-resistored trannies be equally exposed to the "OUT" or "Vo" current in both cases? Following the history of year's past posts, this assumption was believed when the BUF634 first came out: that jumper resistors were required. Then it wasn't. Besides the aforementioned bandwidth resistors (not related), there are caps and resistors shown in recommended circuits between the BUF and OPA2134 - and recommended for many other opamps. Yet, Runeight recommended neither and Dbel has confirmed that it works. I definitely stacked the Intersil's on one of my Milletts, with great sound and no bad results (so far). Instead of all us being lucky, maybe it's the unity gain that allows us to get away with all of this. It's not like I'm any kind of expert, though - maybe one of you other guys can give us some insight: P.S. In any case, the prudent thing would be to prototype each proposed circuit before we ran an entire production run. (Dbel's may already count.) [Edited by tomb on 10-21-2006 at 12:41 PM.] |
| dBel84 HeadWize Fanatic Joined: May 5, 2006 | Message [#108] posted on: 10-22-2006 11:38 PM CST (US). | | hmmmm, in a better place now than I have been this weekend. I thought I had blown my power amp thanx to a 10V dc offset at switch on that I had never looked for with the buff634 in place. I did all the initial testing with the 2134 in place and it appeared stable enough and I ALWAYS power up the SOHA before my amp ( the SOHA has been in it's rightfull place as a buffer between my cd and amp for a while ) thank the forefathers for input protection fuses!! What then started off as a quick addition of bandwidth limiting resistors ended in smoking my soha - it was well past the witching hour and i had wanted to test it quickly after adding (much needed) heatsinks and resistors - I inadvertently shorted the ac line(15VAC) in to ground and actually blew a section of trace off the pcb... a first for me. ( see picture attached ) Today was trouble shooting and I have just restored the SOHA after replacing the rectifier and using a variac to power up and measure ... sigh of relief  . Just a caution to those adding buffers etc - watch for the dc offest - it is 10V at power up and 1V when powered down ....I need an epsilon. The Vregs also need heatsinking , they were getting too hot to touch with the resistors in place. well that's all folk , am off to bed. dBedit - pictures giving me grief and not in the mood - I will try another time if anyone wants a laugh [Edited by dBel84 on 10-23-2006 at 11:06 AM.] Attachment: C2185.jpg |
| GregVDS HeadWizer Joined: Oct. 25, 2005 | Message [#109] posted on: 10-23-2006 09:19 AM CST (US). | | Not a lot of movement here since Friday, Digi01, are you still there? All the best, GregVDS |
| dBel84 HeadWize Fanatic Joined: May 5, 2006 | Message [#110] posted on: 10-23-2006 11:05 AM CST (US). | | I stumbled across this on an archived diyaudio forum re protecting the intersils against short circuit As per data sheets of HA-3-5002 page 4, under heading SHORT CIRCUIT PROTECTION, it is mentioned how to limit the output current of HA-5002, i have used the same technique to limit the output current of HA-5002 and its working fine for me.
Attachment: C2184.jpg |
| gewa HeadWizer Joined: Sep. 22, 2006 | Message [#111] posted on: 10-23-2006 01:47 PM CST (US). | | Well if it worked for dBel84 the board could look something like this.  Two small SMD(1206) resistors per chip soldered on the bottom.
Indeed, digi01 where are you? Regards Oh yeah, I finaly bit the bullet and ordered some LND150's from Mouser. Seems like that's the only place on earth where you can by them! Attachment: C2193.jpg |
| GregVDS HeadWizer Joined: Oct. 25, 2005 | Message [#112] posted on: 10-23-2006 02:30 PM CST (US). | | It seems this thread begins to be almost belgian I also found them only from Mouser. Gewa, did you knew that the biggest tube stockist is here in Brussels? Look for Halfin, the guy is very helpful. Your last Intersil board is quite nice, but could it be possible to have the opamp deported to the left, not to the right? Damn, my boxing is so small I need a lot to be tweaked. Here is a bottom view of what's planned: and here a top view:
And last, the inside: In green, a Linkwitz crossfeed, its settings are reachable by the bottom, as I realized I didn't changed that much and listen all the time with medium setting. Signal comes very quickly to the board, mini switches help the signal path to be the shortest ever in my amps, then it comes directly to the input pads near the Alps. The mains connector will have a switch dual pole, and fuse in it, hence its size. the transfo will be cased with two alodined pannels to isolate it (dotted traces), but even as it is now (not cased, no shielding device, no shielded signal wires), I hear no hum. You see that the front panel is just next to the limit of the board, so the neutrik locking jack comes far inside the pcb towards the buffer position, hence my will to have something not going this way. The box will be only 120mm broad, 212mm long and only 42mm in height, so the underside position of the two regulators, the flat position of the two small 47µF to the right (not actual in the picture) and 470µF . All caps are Silmic-II, and the three big ones are too tall for the box. C1 and C13 are Black gate, which are just 25mm high, so they enter without problem. R10 is under the pcb too so as not touching the neutrik metallic casing, so are the two adjustable resistors. All the bias setting system is reachable through the bottom, test points and resistors. I planned to decenter C1 enough to have the jack symmetrically placed with the volume knob, but my guessing while soldering was not right, and I will need to shift the jack position 2mm right, pitty! When this one will be finished, I will be very proud (if it happens!) comments most welcome, even bad ones All the best, GregVDS [Edited by GregVDS on 10-23-2006 at 03:06 PM.] |
I don't want to desolder - resolder it, for I had to do it for LND150, and one of the pad went off. I had to retrace with wire, very ugly, but it works flawlessly. When the buffer board size problem will be tackled, there will still be the problem of where to put an Epsilon12, as the dc offset of 10V power up 1V power down is very dangerous I think. There is still some volumetric place near the volume pot, or if I manage to squeeze the design, maybe between the bufferboard and the top. Whatever the place, I will need to redesign and make an Epsilon12 board from scratch, as AMB board is by no means able to enter here, too long AND too broad.
| gewa HeadWizer Joined: Sep. 22, 2006 | Message [#113] posted on: 10-23-2006 03:07 PM CST (US). | | Well, I moved the op-amp to the other side but the PCB still has the same width. I suggest you buy a bigger box Nice drawings by the way. Regards Attachment: C2194.jpg |
| GregVDS HeadWizer Joined: Oct. 25, 2005 | Message [#114] posted on: 10-23-2006 03:26 PM CST (US). | | Wow, that was quick! Thanks, I don't have any box, everything will be machined by FrontPanel Express, and I don't want the box to be bigger than this, even if it forces me to think things twice. I stick to my designs, they are very important to me. Do you found Intersil sources here in Belgium? All the best, GregVDS |
| gewa HeadWizer Joined: Sep. 22, 2006 | Message [#115] posted on: 10-23-2006 03:31 PM CST (US). | |
Belgian way of getting things done.
Haven't looked yet, but Zang said he had a bunch. Would be nice if the little China guy made a post here Regards |
| rreynol Member Joined: Oct. 13, 2005 | Message [#116] posted on: 10-23-2006 08:09 PM CST (US). | | Finally got the last 2 caps I was waiting for. |
| GregVDS HeadWizer Joined: Oct. 25, 2005 | Message [#117] posted on: 10-24-2006 02:35 AM CST (US). | | Nice build rreynol! Were did you find your different caps? What are those gold ones? All the best, GregVDS |
digi01 HeadWize Fanatic Joined: Dec. 3, 2002 | Message [#118] posted on: 10-24-2006 04:06 AM CST (US). | | to gewa,GregVDS, sorry for slow response.because I am not surf net during weekend and monday...there are others hobby I need handle I have read all post and keep go ahead. P.S. I only work on protel99se,any stuffs I would transfer to protel. thank you, |
| digi01 HeadWize Fanatic Joined: Dec. 3, 2002 | Message [#119] posted on: 10-24-2006 04:18 AM CST (US). | |
for buf634 boost configuration,I think unregulated power supply would work fine.just bypassed the VR1,VR2 with a couple of 10 ohm resistors and C7,8,9,10=1000uf/35v. |
| rreynol Member Joined: Oct. 13, 2005 | Message [#120] posted on: 10-24-2006 06:37 AM CST (US). | |
The gold ones are Nichicon Muse FG from Handmade Electronics. The red ones are BG NX from PartsConnexion The rest of the electrolytics are Elna Silmic II from either Handmade or Digikey. The Mundorf films are from Parts Connexion. |
| 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.