HeadWize: DIY Workshop > β22: A discrete, cascoded, fully complementary, pure class A amplifier (part 3)

Moderators: cmoy

Welcome. Please log in.

Registration is required to post a new topic or a reply.
User action bar

Forum:

-

DIY Workshop » β22: A discrete, cascoded, fully complementary, pure class A amplifier (part 3)
Page 1 2 3 4 5 6 7 8 9 10 11 12 13 14 > >> | Next Page | Prev Topic | Next Topic

Author Post

amb

Headphone Council

Joined: Apr. 1, 2004
Locale: Sunnyvale, CA. USA
Total Posts: 4408

Message [#160] posted on: 12-18-2008 09:40 PM CST (US).

Hi aos, yes, you can use 1.5uF for those, as long as they fit properly. The capacitance is not critical.

aos

HeadWizer

Joined: Jun. 15, 2008
Locale: Vancouver, Canada
Total Posts: 128

Message [#161] posted on: 12-18-2008 09:57 PM CST (US).

Thanks for confirming it, amb. How about C1? The compensation table lists C2-C5, omitting C1. Does this mean that C1 should be 33pF no matter what the gain setting is, or is this a typo and it should be the same value as C2-C5?

amb

Headphone Council

Joined: Apr. 1, 2004
Locale: Sunnyvale, CA. USA
Total Posts: 4408

Message [#162] posted on: 12-18-2008 10:06 PM CST (US).

C1 should be 33pF in all configurations. The design calls for changes in compensation cap values wrt. gain only for C2-C5.

aos

HeadWizer

Joined: Jun. 15, 2008
Locale: Vancouver, Canada
Total Posts: 128

Message [#163] posted on: 12-18-2008 10:16 PM CST (US).

Thanks. You might consider revising the parts list to separate C1 from C2-C5 in its own row to make it completely clear.

FYI for anyone considering using J201 for current sources, you will need at least 150 for 4 boards. And even then count on at least one board having some CRD's on the opposite end of the allowable range. I ordered 100, then another 50, then ended up using only 24 FETs and 8 CRD's I had spare in order to get a better match. Not sure if you end up saving any money.

housing

HeadWizer

Joined: Dec. 16, 2006
Locale: N/A
Total Posts: 163

Message [#164] posted on: 12-19-2008 06:07 AM CST (US).

[Quoting aos]

FYI for anyone considering using J201 for current sources, you will need at least 150 for 4 boards. And even then count on at least one board having some CRD's on the opposite end of the allowable range. I ordered 100, then another 50, then ended up using only 24 FETs and 8 CRD's I had spare in order to get a better match. Not sure if you end up saving any money.

I got 36 pieces of 0.56ma plus or miunus 5% J201 from my 200 pieces. I think it would be much cheaper to use J201 than 1N5291 in this configuration. From Mouser :

32 pieces of 1N5291 = US$57.28
200 pieces of J201 = US$$25.80

J201 40v fet Idss can be matched with a tolerance of 5%
1N5291 100v = 10%
J503 50v = 20%

May be I have better luck this time. <big grin>

[Edited by housing on 12-19-2008 at 06:08 AM.]

amb

Headphone Council

Joined: Apr. 1, 2004
Locale: Sunnyvale, CA. USA
Total Posts: 4408

Message [#165] posted on: 12-19-2008 06:43 AM CST (US).

J201 Idss varies wildly from device to device, and you do really need a large number to get reasonable matches. Perhaps they wind up costing less than 1N5291, but you will have to spend a lot of time and effort to match them. My experience with 1N5291 is that, despite the 10% tolerance spec, most of them are fairly accurate, and you don't have to do any matching at all. Using J201s might save you some money, but it will only be a small fraction of the total cost of the amp, and the hassle factor is high, so it's of questionable worth.

aos

HeadWizer

Joined: Jun. 15, 2008
Locale: Vancouver, Canada
Total Posts: 128

Message [#166] posted on: 12-19-2008 12:56 PM CST (US).

While I agree the use of 1N5291 is the most practical and hassle-free solution, $60 US is easily 15-20% of the cost of the amplifier (ignoring exotic case costs) which isn't a small portion by any means. It really stands out in the bill of materials.

amb

Headphone Council

Joined: Apr. 1, 2004
Locale: Sunnyvale, CA. USA
Total Posts: 4408

Message [#167] posted on: 12-19-2008 07:55 PM CST (US).

Then your β22 only cost $300-400? I think most people spend way more than that.

aos

HeadWizer

Joined: Jun. 15, 2008
Locale: Vancouver, Canada
Total Posts: 128

Message [#168] posted on: 12-19-2008 08:33 PM CST (US).

I think it's $400-500 without case (and PS, likely) for 4 boards, but I haven't exactly been adding up numbers (I might later, but I'm afraid). If you buy boutique jacks, switches, knobs and attenuators (as well as resistors and capacitors) then I'm sure it can easily go much higher. But speaking strictly about components to stuff boards, I think CRDs are a considerable expense. Although discussing it at this point is academic exercise at best.

I am considering getting Galaxy Maggiorato from HiFi200. The case GX383 is 80 mm high, 330 wide and 230 deep. It should be able to host 4 boards if I'm correct. What? Is it possible it's only 28 Euro?

Mister X

HeadWize Fanatic

Joined: Jul. 23, 2003
Locale: Kalamazoo, Michigan
Total Posts: 606

Message [#169] posted on: 12-19-2008 08:51 PM CST (US).

[Quote]
Is it possible it's only 28 Euro?

Yes but the shipping is going to kill that. <frown>

[Edited by Mister X on 12-19-2008 at 08:57 PM.]

jazzist

Member

Joined: May 9, 2008
Locale: N/A
Total Posts: 30

Message [#170] posted on: 12-20-2008 03:33 AM CST (US).

I've been really impressed with Hi-Fi 2000, shipping reasonable within Europe, answering emails quickly, and providing me with the drawings I wanted. They also sent additional items quickly when some parts arrived damaged.

miles hand

Member

Joined: Jun. 28, 2007
Locale: uk
Total Posts: 65

Message [#171] posted on: 12-20-2008 04:53 AM CST (US).

The internal measurement of that Galaxy case is 310mm. The sides of the case are 10mm thick extrusions. According to my crappy CAD drawing four boards fit with 5mm to spare as long as they're butted up to each other.

I found the Galaxy cases accurate and easy to work with.

aos

HeadWizer

Joined: Jun. 15, 2008
Locale: Vancouver, Canada
Total Posts: 128

Message [#172] posted on: 12-20-2008 10:57 AM CST (US).

Thanks miles hand. I was hoping the case is 330mm internally as 310 will be cutting it close. Do you know how deep the case is, relative to their nominal measurement? I.e. the one I want is 230mm deep nominally. I need to place some other stuff in it at the front.

Ferrari

HeadWizer

Joined: Apr. 29, 2006
Locale: Netherlands
Total Posts: 224

Message [#173] posted on: 12-20-2008 06:02 PM CST (US).

The Galaxy Max GX383 is 227mm depth (230mm nominal).

miles hand

Member

Joined: Jun. 28, 2007
Locale: uk
Total Posts: 65

Message [#174] posted on: 12-20-2008 07:36 PM CST (US).

Here's the drawing again with an internal front to back of 227mm (cheers, Ferrari).

Just for fun I've added a four-deck Goldpoint attenuator (two extra decks roughly stuck on to give the right front to back dimension), an Alps RK27112A, some Neutrik phonos and a no-name headphone jack, all to scale. The front view shows a Beta22 board on 5MM and 10MM standoffs with 2 inch heatsinks. The positions of the capacitors on the board are approximate, but the heights are accurate. The rectangular shape is an Alps pot and the oval is a Goldpoint, both mounted in the centre of the front panel. With the boards in this position there is 26mm of space from the edge of the board to the back face of the front. If you move the boards right back to just in front of the phono sockets, you get 33mm clear at the front.

I hope that gives an idea of what room you have to play with.

aos

HeadWizer

Joined: Jun. 15, 2008
Locale: Vancouver, Canada
Total Posts: 128

Message [#175] posted on: 12-20-2008 10:22 PM CST (US).

Thanks a lot for the info and drawings, Ferrari and miles hand.

kschong

Member

Joined: Oct. 16, 2008
Locale: N/A
Total Posts: 13

Message [#176] posted on: 12-21-2008 11:14 PM CST (US).

Hi Amb, last time I had a problem with noise and you told me to case up the amp first. Okay I have did that and the problem is solved.

Now I have another problem... <frown> I'm using a 3 board layout with the s22 power supply.

Every time after about 5 mins of usage, when I turn the pot to a specific 9 o'clock position, there's a sharp continous humming sound through the headphones. Below that position, the humming goes away and the volume diminishes quickly. At the minimum position, the humming sound comes on again.

I also realise that there's a low humming sound coming from the board, literally.

Other than this issues, the amp seems to function normally.

Anything I can do to fix the problem?

amb

Headphone Council

Joined: Apr. 1, 2004
Locale: Sunnyvale, CA. USA
Total Posts: 4408

Message [#177] posted on: 12-21-2008 11:29 PM CST (US).

[Quoting kschong]
Every time after about 5 mins of usage, when I turn the pot to a specific 9 o'clock position, there's a sharp continous humming sound through the headphones. Below that position, the humming goes away and the volume diminishes quickly. At the minimum position, the humming sound comes on again.

Is this a two-chassis or single-chassis build? What volume pot are you using? How is it wired up? How did you do the grounding? What headphones? Maybe a clear photo would help. Also, do you have a zobel network connected to the outputs? Sometimes a high frequency oscillation can have sub-harmonics that would sound like hum.

[Quote]
I also realise that there's a low humming sound coming from the board, literally.

On which board? Can you identify which exact part(s) that's making the noise?

jazzist

Member

Joined: May 9, 2008
Locale: N/A
Total Posts: 30

Message [#178] posted on: 12-23-2008 07:06 AM CST (US).

Work in progress:

Machined brass, gold plated, 3A rated, Samtec SIL Header/Socket:

Thank you AMB:

Fixing details:

Overall (1 more to come, for other 3 B22 boards, is essentially a mirror of this, hasn't left Altium yet <wink> ):

[Edited by jazzist on 12-23-2008 at 10:57 AM.]

masantos

HeadWizer

Joined: Dec. 13, 2005
Locale: N/A
Total Posts: 103

Message [#179] posted on: 12-23-2008 08:58 AM CST (US).

That's an impressive PCB jazzist!!! <shocked>

It looks like you took amb's backplane board one step further and added the volume pot, jacks, and so on to the board.

Is it your design? Could you share more info abou this? A pic from the top and circuit description? <cool>

Thanks in advance!

jazzist

Member

Joined: May 9, 2008
Locale: N/A
Total Posts: 30

Message [#180] posted on: 12-23-2008 09:19 PM CST (US).

Sure, will post info soon.

Is now 03:15AM... debugging the above pictured and my first B22 board made so far, the VAS stage was not behaving at all.. it does help if C3 is connected accross the pads designated by the II marking on the silkscreen <crying> ...big d'oh!, what a goof. Thanks AMB for your voltages PDF on your website, else i'd never have spotted it <smile>

Now is all good, 4.5-4.6 R9-R12 dialled in, 0.18mV DC offset (no adjustment!).

Very relieved now <smile>

[Edited by jazzist on 12-23-2008 at 09:40 PM.]

Page 1 2 3 4 5 6 7 8 9 10 11 12 13 14 > >> \| Next Page \| Prev Topic \| Next Topic

Contact Webmaster - HeadWize welcomes comments and suggestions!

(remove _nospam_ )