Technical Assistance/Advice Thread

[mwl168]

Do I understand correctly that most of these issues can be effectively resolved or drastically minimized  if we don't expect to have the music played real time as do the LP playback? 

What I mean is that if the digital signal transmitted from the play back device is buffered at the DAC end where measures can be employed to "re-align" and cleanse the data before feeding the conversion circuits, similar to how digital data is transmitted between computers, then the effect of digital cable can be eliminated as long as they are properly spec'ed? 

Or am I over-simplifying things?   

[sbelyo]

I can tell you that I used a $5 usb cable from Amazon in my setup for a year or so.  I figured there was enough isolation and reclocking in the Amanero/Chronos/Rhea stack that a better cable wouldn't matter.  I was able to get a hold of an Audioquest Cinnamon USB cable for $35 out of pocket so I thought let me try it.  I'll tell you I tried to remain as objective as humanly possible but I heard a positive difference with the better quality cable.

[vivolo]

Does anyone have any tips for building an enclosure?

I need to make one for a CFA. How do I go about mounting the transistors to the heat sinks?

[n_maher]

Start with some commonly available instructions, like the excellent stuff at AMB's website.  Like this - https://www.amb.org/audio/beta22/, among others.

[mwl168]

12 hours ago, vivolo said:

 How do I go about mounting the transistors to the heat sinks?

You really have two options when it comes to mounting resistors to heat sinks:

1.  mount the transistors directly to the heat sink which requires you to drill and tap the heatsink

2.  mount the transistors to an aluminum angle then mount the angle to the heatsink

Of the two options, I would opt for option 2 when possible.

Drilling and tapping the heat sink is tricky and unforgiving. If you make a mistake like broken the tapper while tapping the heat sink, you may be forced to abandon that heat sink altogether.

Drilling and tapping (optional) 3mm- thick aluminum angle (which I typically use) is much simpler and forgiving. Aluminum angle is also easier and cheaper to replace.

I speak from personal experience. 

 

[Pars]

You still have to drill and tap the heat sink (preferably blind holes) to mount the angle to it. The mount directly to the sink method would work well for the least thermal resistance IF you are good at drilling fairly precisely AND you aren’t prone to breaking taps. Something like the Shars drill guide helps as well (though I haven’t used mine yet.)


Sent from my iPhone using Tapatalk

[sbelyo]

For lower current and heat I use aluminum cases and mount the transistors to the case bottom.  Then secured with a 4-40 hex head screw and locknut

For higher current and heat I use the angle iron method mounted to the case bottom then mount the heatsink to the angle iron.

When tapping the heatsinks it's best to use cutting fluid.  WD-40 works fine if you don't have cutting fluid.  Take care to back the tap out a half turn or so for each two full turns.  This will break off the metal pieces that bind taps when you're cutting the threads. 

[mwl168]

59 minutes ago, Pars said:

You still have to drill and tap the heat sink (preferably blind holes) to mount the angle to it. 

Good point.

I should have mentioned that I mount the aluminum angle to the heat sink using M5 screws, which, in my experience, on top of having fewer holes to work with, is much easier to tap and far less likely to break the tapper. 

I drill the mounting halls on the angle slightly larger than the M5 screw diameter to provide some margin of error so I don't need to be as precise when I drill and tap the heat sink.. 

[jose]

2 hours ago, mwl168 said:

2.  mount the transistors to an aluminum angle then mount the angle to the heatsink

 

For convenience, I also believe that it is the best option in most cases.

If I have to do few holes, I don't mind practicing them directly on the heat sink, but here you have to be much more careful… although after having done those of the CFAe I can already do anything.

 

[muskyhuntr]

I like to start my blind holes on a heat sink (or any thing else) with a thru hole tap and then finish with a closed end tap.  this

allows the threads to be better formed deeper in the hole.

I use a drill press set to the proper depth to avoid drilling all the way through but a hand drill with tape wound around the drill bit

at the proper depth works also, just a little less precise.

Me

 

[Helium]

Check this video by Haas:

You may want a special tapping drill to make thread go deeper in the limited depth of blind hole.

[jose]

I found this guy's channel. I did not know where to share it so I have put it in this thread in case someone is useful.

It has enough videos on how to build planar speakers.

I think it's quite interesting.

 

 

 

 

[Kerry]

A few months ago a broke a #4-40 tap in a heat sink.  After a lot of cursing and futile attempts to get it out, I just left it for another day.

I felt compelled a few days ago to go after it again.  This time I came across an article on the use of alum powder to dissolve the ferrous metal (tap) from the aluminum without affecting the aluminum.  Long story short, I got some potassium alum powder, set up a double boiler with a Pyrex container and six hours later and a little baby sitting, the tap was gone and I was able to save the threads and all.  Kind of amazing.

You can see the small bubbles coming from where the tap was.  Note that I don't remember any chemistry, so I'm hoping this was safe.

20200227_181723000_iOS.MOV

Edited February 28, 2020 by Kerry

[Pars]

 

Edited February 28, 2020 by Pars

[Kerry]

Just a note.   There are a few videos on the internet on this (not many).  I wanted to be clear that I mixed the alum powder with water about 4 tbsp / cup (or as much as can be dissolved).  Some of the videos show mixing it with acid, which is not necessary, so therefore not advisable.

The alum I used was a non-ferrous food grade version, so I could pickle some tomatoes or cucumbers if I'd like with the extra  

Edited February 28, 2020 by Kerry

[mwl168]

How does the pickled tap taste? 

[Kerry]

🤢

[lkong]

Working on build a water cooling Uberamp.

Snatched a cold plate off ebay.

I'm looking for advice to mount the transistors. I dont think I should drill and tap into the cold plate.

So I have 2 option:

1: 1/8 by 1 copper plates drilled to mount on the cold plate and sandwich the transistors in between.

-----copper plate-------

-----transistors----------

-----cold plate-----------

2: mount copper plates directly on the cold plate, drill mounting holes on the copper plates, counterbore the back side (the face contacting the cold plate).

-----transistors----------

-----copper plate-------

-----cold plate-----------

Any suggestions?

Here's a planned layout:

Edited March 2, 2020 by lkong

[MASantos]

I started matching transistors today for my cfa build but quickly started scratching my head.

I'm using Amb.org bjt matching circuit with a 110K resistor  for the MJF1530/1531. I got between 9 and 15 mA for the 1530's and ~30mA range for the 1531's. Shouldn't I get closer readings between N and P devices?

[Pars]

Not necessarily. First, it sounds like you are trying to measure these at too low of current. Look at the datasheet. The min hFe for the MJF15030/15031 is 40mA. They should ideally be matched at about the current they are going to run at, if possible,

[boinger]

On 3/1/2020 at 8:01 PM, lkong said:

Working on build a water cooling Uberamp.

Snatched a cold plate off ebay.

 

I'm looking for advice to mount the transistors. I dont think I should drill and tap into the cold plate.

So I have 2 option:

1: 1/8 by 1 copper plates drilled to mount on the cold plate and sandwich the transistors in between.

-----copper plate-------

-----transistors----------

-----cold plate-----------

2: mount copper plates directly on the cold plate, drill mounting holes on the copper plates, counterbore the back side (the face contacting the cold plate).

-----transistors----------

-----copper plate-------

-----cold plate-----------

Any suggestions?

Here's a planned layout:

 

Drill and tap the transistors into a copper plate and once drilling and tapping is complete.

Use a thermal epoxy to bond the copper plate to the water blocks then you can mount transistors onto the tapped copper plate. 

[Craig Sawyers]

Somewhere I have a CERN report from the 1970 about a 13kA pulsed magnet power supply. That used 7,200 2N3055's as output devices driven  by another 360 of them. All of those were in massive 19% racks mounted on cooling plates with forced water cooling. Might be interesting or relevant.

This was for the 28GeV Synchrotron. Eventually that was used simply as the injector accelerator for the SPS (Super Proton Synchrotron), and of course that has been superceded by the Large Hadron Collider. The current  acceleration scheme is Synchrotron - SPS - LHC to get 13TeV

Found the power supply paper. Attached.

CERN-70-20.pdf

[luvdunhill]

Any recommendations on a multi-zone house setup? I feel like I just want a stack of Raspberry Pi and a Multichannel Amp... I suppose I could just use the on-board audio jack or maybe even a ultra cheap USB stick DAC (suggestions?)

Thoughts?

[Grahame]

Here's one approach

https://support.hifiberry.com/hc/en-us/articles/205699981-How-to-build-a-multiroom-audio-system-based-on-Raspberry-Pi-and-Hifiberry?mobile_site=true

The RasPi's make a commodity streamer / digital  endpoint. Room ( and slimserver ) support multi player / grouping / syncing control.

Centralized (multi channel) Amp (s) and distributed analog signal to passive speaker's ?

Distributed RasPi's to active speaker's?

Distributed RasPi's with (HiFi Berry?) Amp HAT's and passive speaker's?

Plenty of different Audio signal path architectures , all behind the same digital path/control architecture. 

Or mix and match?

Nice problem to have.

 

 

[luvdunhill]

Here's one approach
https://support.hifiberry.com/hc/en-us/articles/205699981-How-to-build-a-multiroom-audio-system-based-on-Raspberry-Pi-and-Hifiberry?mobile_site=true
The RasPi's make a commodity streamer / digital  endpoint. Room ( and slimserver ) support multi player / grouping / syncing control.
Centralized (multi channel) Amp (s) and distributed analog signal to passive speaker's ?
Distributed RasPi's to active speaker's?
Distributed RasPi's with (HiFi Berry?) Amp HAT's and passive speaker's?
Plenty of different Audio signal path architectures , all behind the same digital path/control architecture. 
Or mix and match?
Nice problem to have.
 
 

Yeah that was what I was thinking as well!

I am not sold on the amp solution here, but yeah I have passive speakers to feed with in-room passive / auto-former volume controls that I might as well keep.

I have used a Lepai amp to feed some outdoor speakers via a auto-former control which makes me think these little chip amps won’t be enough... but maybe?