HeadWize: DIY Workshop > M³ headphone amplifier (part 2)

Moderators: cmoy

Welcome. Please log in.

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

Forum:

-

DIY Workshop » M³ headphone amplifier (part 2)
Page 1 2 3 4 5 6 7 8 9 10 11 12 13 14 > >> | Next Page | Prev Topic | Next Topic

Author Post

headphone1235

Member

Joined: Nov. 7, 2008
Locale: N/A
Total Posts: 12

Message [#140] posted on: 11-07-2008 07:46 PM CST (US).

[Quoting amb]
headphone1235, for the 2W 1.0 ohm resistors you can use Mouser 282-1.0-RC. For the 2N3904 and 2N3906 I recommend using the same manufacturer for both, so 512-2N3904TFR and 512-2N3906TFR would work.
The AD8610 needs a SOIC-8 to DIP-8 adapter, so you need to order one from Cimarron Technology (part number 970601). They used to sell AD8610s pre-mounted on the adapters, but looks like it's been discontinued, but it probably wouldn't hurt to call and inquire. What you listed in your spreadsheet as the adapter is actually the opamp socket, which you also need.
For the LED, how about Mouser 78-TLHB4400 (blue diffused 3mm)?
You did not list anything for the power supply or the casing and hardware.

Thanks for the quick reply, I've updated the spreadsheet to your recommendations.

For the power supply I'll use the σ11, Im a bit confused in which transformer I'd have to buy. The mains I have is at 240V and I'm looking for the right transformer, I've got a website that sells a selection of transformers: http://www.altro...ea=prod&grp=354

Which one do you recommend for my kit?

amb

Headphone Council

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

Message [#141] posted on: 11-08-2008 01:38 AM CST (US).

headphone1235, the M4912A would work well (for σ11 configured to output 24VDC).

headphone1235

Member

Joined: Nov. 7, 2008
Locale: N/A
Total Posts: 12

Message [#142] posted on: 11-08-2008 01:55 AM CST (US).

[Quoting amb]
headphone1235, the M4912A would work well (for σ11 configured to output 24VDC).

Thanks for all your help amb, In a couple of days I'll start ordering.

I've also uploaded the finished revision for my build. Later I'll start dealing with the power supply.

[Edited by headphone1235 on 11-08-2008 at 02:49 AM.]

Attachment: C4397.xls

headphone1235

Member

Joined: Nov. 7, 2008
Locale: N/A
Total Posts: 12

Message [#143] posted on: 11-18-2008 01:30 AM CST (US).

I've received my delivery today, both pcb boards are pretty good quality and the delivery was very quick.

soundwitty

Member

Joined: Oct. 28, 2007
Locale: London, UK
Total Posts: 16

Message [#144] posted on: 11-19-2008 03:12 PM CST (US).

It is few months I've finished my M3 and I still have issues with hissing and distorted HF when listening at two third of the volume's knob way. This is more evident with highly dynamic programs.

I use a sigma11 at 32V and AD834 chips and the quiescent voltage bios is at 90mV at the moment. Bass boost pot used for gain adjustment.

Raising the bios moves the distortion threshold further up but melts my mosfet heatsinks. At the moment the only solution looks to increase the heat transfer.

So I've a couple of questions:

- For the sake of linearity is it better to lower the psu voltage and increase the bios one or vice versa? Will 27V at 120mV be more linear than 32V at 90mV with the same chips?

- Is the solution using smaller heatsinks and the horizontal aluminum bar under the top cover going to be more efficient in dissipating the heat than 1.5" heatsinks and holes?

Thanks.

amb

Headphone Council

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

Message [#145] posted on: 11-19-2008 03:17 PM CST (US).

soundwitty, AD834 is not an opamp. What do you actually have in there? I suspect you have an oscillation problem.

[Edited by amb on 11-19-2008 at 03:18 PM.]

soundwitty

Member

Joined: Oct. 28, 2007
Locale: London, UK
Total Posts: 16

Message [#146] posted on: 11-20-2008 03:52 PM CST (US).

Hello Amb,

sorry it was ad843 of course. And exactly this is what I've got:
http://uk.farnel...jsp?SKU=1438585

Thanks.

amb

Headphone Council

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

Message [#147] posted on: 11-20-2008 04:14 PM CST (US).

OK, AD843 is a "supported" opamp for the M³. Did you use the default values of C1L, C1R and C1G? Also, what headphones are you using?

soundwitty

Member

Joined: Oct. 28, 2007
Locale: London, UK
Total Posts: 16

Message [#148] posted on: 11-20-2008 05:24 PM CST (US).

Yep. 33pf. My can is an HD880.

Cheers.

soundwitty

Member

Joined: Oct. 28, 2007
Locale: London, UK
Total Posts: 16

Message [#149] posted on: 11-20-2008 05:25 PM CST (US).

Sorry again. The headphone is DT880.

amb

Headphone Council

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

Message [#150] posted on: 11-20-2008 09:52 PM CST (US).

OK, there are different versions of DT880, and even though I have not heard of anyone having trouble with them on the M³, maybe you could try adding a Zobel network to the output of the amp to see if this problem would be resolved. For details about the Zobel, see the β22 website, under "Wiring & ground".

soundwitty

Member

Joined: Oct. 28, 2007
Locale: London, UK
Total Posts: 16

Message [#151] posted on: 11-22-2008 04:34 PM CST (US).

I only could find polyester and carbon at the street shop.
A bit better but the issue is still there. It is like a strong ringing in the high part of MF. A strong female acute, not necessarily high in frequency but powerful and rich in harmonics, will make it very clear and annoying.

Cheers,
A

amb

Headphone Council

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

Message [#152] posted on: 11-22-2008 10:49 PM CST (US).

soundwitty, since the Zobel does seem to have an effect, then perhaps you need somewhat more aggressive part values. Try 10 ohms + 0.1uF instead of 22 ohm + 0.047uF.

If you have other headphones to try, it would also be useful to know if the issue is confined to just the DT880s you have.

Dougigs

HeadWize Fanatic

Joined: Sep. 21, 2005
Locale: London England
Total Posts: 372

Message [#153] posted on: 11-23-2008 05:35 AM CST (US).

Also, Soundwitty, make sure it isn't the source. If you're listening to recordings from before the 1990s (or pop recordings from more recently) it could be that you're simply hearing recording imperfections that weren't audible before you had an amp of this quality. Also, if your source is less than great, it could be that - - try a good-quality digital recording from the line out of a decent CD player or DAC and see if it sounds different.

soundwitty

Member

Joined: Oct. 28, 2007
Locale: London, UK
Total Posts: 16

Message [#154] posted on: 11-23-2008 06:17 AM CST (US).

Hello,

thanks to all for your help.

I will try to augment the zobel effect.
About my source it is a modified cd0 with every single lytic changed to fresh Rubycon or Silmic, all opamps changed to opa607 and a Tent's clock. At the moment the phone output in the cdp sounds better (less distorted and more musical) than the M3 which is doesn't seem possible.
The same cdp connected to my LUX valve amplifier (300B SE driven by EF86 with a cathode follower interstage) sounds pretty good. Sure a better I/V+output stage for the cdp will improve things but I think it is fair to say it is a reasonably good source.
All cds I'm using for tests are recend editions. The cd where the defect comes immediately out is Hesperion XXI's "Invocation a la nuit". The first track is a good example.

Also, the heatsink are running very hot. I can't touch them for longer than 10 seconds. Is that normal?

Cheers,
Antonio

soundwitty

Member

Joined: Oct. 28, 2007
Locale: London, UK
Total Posts: 16

Message [#155] posted on: 11-23-2008 06:20 AM CST (US).

Again my kepyboard skill doesn't excell. The cdp is a CD80 and the OPamps in it are OPA627s.

Cheers,
Antonio

amb

Headphone Council

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

Message [#156] posted on: 11-23-2008 07:21 AM CST (US).

[Quoting soundwitty]
Also, the heatsink are running very hot. I can't touch them for longer than 10 seconds. Is that normal?

It depends on how much quiescent current you're running through the MOSFETs, how high you supply voltage is, how big the heatsinks are, and how well is the ventilation (if cased). At the default of 80mA and 24V supply, the heatsinks should get hot, but not burning hot.

soundwitty

Member

Joined: Oct. 28, 2007
Locale: London, UK
Total Posts: 16

Message [#157] posted on: 11-23-2008 07:52 AM CST (US).

I've a sigma 11 runing at 32.5V on a separate case. The heatsinks are the 1.5" ones, the case is the Hammond standard one in aluminum and I didn't solve yet the problem of ventilation as the top cover is still without holes and I operate the amp with the top open.
I've an aluminum bar I can use to implement Morsel's idea of transferring the heat to the top cover without making holes but I will have to change the heatsinks to the one 1" ones or they won't fit.
The quiescent current is set for 90mV.

Thanks.
A

soundwitty

Member

Joined: Oct. 28, 2007
Locale: London, UK
Total Posts: 16

Message [#158] posted on: 11-24-2008 07:54 PM CST (US).

Ok.
I plugged my HD280 pro I use at the office. Of course it takes much more volume to get the defect to the point is not bearable.
The doubled zobel didn't have any effect.
My original question is still pending. Being this SS rather than tubes wouldn't be better to lower the psu voltage and increase the bias current?
Also are 1" heat sinks with the aluminum bar on top joint to the box cover being efficient with this level of power?

Thanks.
Regards.

amb

Headphone Council

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

Message [#159] posted on: 11-24-2008 11:05 PM CST (US).

soundwitty, even at the default bias current (or lower) you should not be experiencing the problems you describe, so I think that there is something else wrong. If increasing the bias alleviates the problem, then it's masking the symptom, not fixing the root cause.

Unfortunately, to really see whether there is a problem would require an oscilloscope. That would be very telling.

Just for additional data points. Do you have any other "approved" opamps to try? Also, check all of your wiring to make sure they are correct. In particular, the output ground (OG) isn't shorted to input ground (IG), and the power supply V- isn't shorted to either, etc.

[Edited by amb on 11-24-2008 at 11:10 PM.]

soundwitty

Member

Joined: Oct. 28, 2007
Locale: London, UK
Total Posts: 16

Message [#160] posted on: 11-25-2008 04:53 PM CST (US).

Tanks Amb.

I may try to build a MilletMax as I've the board and the most difficult to get bits.

In the main time I will recheck the whole board for cold joints and short. There is no shirtcut between ig and og and betwen v- and og even if the multimeter start low in resistance to raise slowly as it would go through a capacitor.

I've an old low frequency oscilloscope, cumbersome and probably compeletely out of tune. Not sure this will be helpful though. I know that I will have to get a computer based oscilloscope one day or another.

Thanks for your support.

Cheers.

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_ )