The Multi Amp aka Dynalo Mk2

[astro]

I tried aluminum heat sink before. It does have wider fin gap, but less fin count and much shorter fin height.

The final stable temp is higher than the one you see in the pic. At last,  copper has higher thermal conductivity.

I have to say sorry that currently I have no capability to handle a group buy, but I would like to provide the drawing and seller link if you want.

 

[Pars]

I’d be interested in that as well Astro (dwg and seller info).



Sent from my iPhone using Tapatalk

[Pars]

Question: if someone wanted to use this both with normal sources, and with a source with digital volume control, is turning the pot all the way up when your source has volume equivalent to not having the pot in the amp? Or is it better to short the pot out (wiper to input)?
It would seem that the wiper to input resistance would be 0 ohms or close with the pot all the way. Of course real world doesn’t always match theory


Sent from my iPhone using Tapatalk

[DrizztDo]

I wonder who that someone might be...

[Pars]

On 1/17/2017 at 6:12 AM, A_Zero said:

Also some interesting results. The THAT340 was then replaced by hand-matched ksc1815/ksa992, beta deviation within 0.1% - Huge improvement in transparency and details!

 

Interesting, and sorry to chop out your casing pics

I was looking at the KSC/KSA, then saw reference to the KSA1015 being the actual complement to the KSC1815. Then it dawned on me... are these the same as the Toshiba 2SA1015/2SC1815 that Kevin used to use in the Dynalo? If so, I've got some of these and may try them. The KSC1815 appears to be problematic to get in the same grade as the 992 which would make matching nearly impossible.

[Pars]

Also, regarding regulators, it would seem that the Amb-style externals with heatsink (U5, U7) would have a thermal advantage over on-board mounting (U2, U3) of these?

Or is this an incorrect assumption for those of you who have built one or more of these?

[johnwmclean]

I have built only with the AMB regs with heatsinks, I closely watched temperatures and in comparison to Astro I’m running slightly cooler... and that’s without any heatsinking attached to the transistors. 

I’m think directly soldering the regs to the board creates a thermal bottleneck.... ymmv

 

 

Edited December 14, 2017 by johnwmclean

[jose]

Can I change that340 or 2sk170/2sj74 for ksa992/ksc1845 or ksc1815/ksa1015?

 

I need to do an order on Mouser, so I can add this fets on my order.

 

 

[MASantos]

On ‎06‎-‎12‎-‎2017 at 2:57 PM, astro said:

I tried aluminum heat sink before. It does have wider fin gap, but less fin count and much shorter fin height.

The final stable temp is higher than the one you see in the pic. At last,  copper has higher thermal conductivity.

I have to say sorry that currently I have no capability to handle a group buy, but I would like to provide the drawing and seller link if you want.

 

Astro could you share the seller info and your case drawings please? A few of us might be interested in a group buy if the price is good.

[Pars]

[whisper] Ditch Hakko... buy Metcal...[/whisper]

Had the wrong post linked in my previous reply. This person sounds like they know what they are gliding with packages like the regulators use...


Sent from my iPhone using Tapatalk

[A_Zero]

11 hours ago, Pars said:

Interesting, and sorry to chop out your casing pics

I was looking at the KSC/KSA, then saw reference to the KSA1015 being the actual complement to the KSC1815. Then it dawned on me... are these the same as the Toshiba 2SA1015/2SC1815 that Kevin used to use in the Dynalo? If so, I've got some of these and may try them. The KSC1815 appears to be problematic to get in the same grade as the 992 which would make matching nearly impossible.

Hi Pars,

it's okay Now the casing was changed to an extruded aluminium product from taobao, the soundstage become somewhat more stable and background is also darker than before...

Sorry for the typo, it's actually KSC1845 and KSA992... Only the same polarity pairs are matched for beta. I also tried to match pnp with pnp, but quite ironically, the beta deviation within the same batch was too small so that a match between different batches (eg. PNP with NPN) was made impossible...Even if I buy 1845 and 992 in same beta-grade, their beta grouping just avoided each other perfectly....

[ang728]

actually I tried 992/1815 from project spare parts .With some simple matching you can still get low offset results

sounding is also not bad tho

will keep trying other combos  

Edited December 14, 2017 by ang728

[astro]

18 hours ago, MASantos said:

Astro could you share the seller info and your case drawings please? A few of us might be interested in a group buy if the price is good.

Sure, left your email in message box. I'll send it to you.

@Pars Hi, I left a message, but have no response. Do you still need them?

[astro]

19 hours ago, johnwmclean said:

I have built only with the AMB regs with heatsinks, I closely watched temperatures and in comparison to Astro I’m running slightly cooler... and that’s without any heatsinking attached to the transistors. 

I’m think directly soldering the regs to the board creates a thermal bottleneck.... ymmv

 

 

Agree. The final resistor value chosen is 220 ohm and get 18ma bias current.

The final sink temp on the transistors is 74c~76c

 The regs are pretty hot, especially the negative one.(63c sink temp on positive/71c sink temp on negative)

The top surface case temp is 38c. The ambient temp inside the case is 52c.

 Anyway still like this power solution.

 

[Pars]

Astro sent the .dwg file for the case to me, seeing if I can post it here... it looks like it worked. I don't have autocad or anything else on my Mac to open it (suggestions welcome).

Dynalo.dwg.zip

[Pars]

I have built only with the AMB regs with heatsinks, I closely watched temperatures and in comparison to Astro I’m running slightly cooler... and that’s without any heatsinking attached to the transistors. 
I’m think directly soldering the regs to the board creates a thermal bottleneck.... ymmv
 
 

It would be interesting to see builds from a voltage/biasing R value/heat perspective including whether using onboard vs. external regs.


Sent from my iPhone using Tapatalk

[gepardcv]

That file imports cleanly into Fusion360. Grab it from the Mac App Store and register for a free "hobby" license. Its all-"cloud"-all-the-time data management system leaves much to be desired, but in brief, you need to (1) create a project in the data panel, and (2) upload the .dwg into the project.

Post-processing the design into something you can manufacture is the usual can of worms. I recommend at least splitting each panel (currently all in the same file) into separate part files. Fusion360 can export STLs if that helps. If you have (access to) a CNC mill with Fusion360 drivers, it should "just work."

[Pars]

Thanks! I actually have Fusion360 on my Mac but haven’t spent the time to learn how to use it yet. Any recs on tutorials for that would be helpful. I’ve never used any CAD type programs... closest would be things like Eagle, etc.


Sent from my iPhone using Tapatalk

[Pars]

Took a look at the dwg file tonight using Fusion 360 and also the SolidWorks eDrawings app. One thing I noticed is the corner blocks don't seem to show drilled/tapped holes on them. When I look at in the SolidWorks app, there is some Chinese writing (I presume), which of course I can't read. Are these instructions to whoever is making the parts for you?

Here's a screenshot from the Solidworks app:

 

[jose]

My dynalo

[Pars]

Very nice Jose!

I take it that is input switching back by the jacks? Is the other PSU back by the transformer for a momentary power switch, or is that just for the input switching? It looks like some thought went into this layout and good job on the wire routing, etc.

[jose]

Thanks Pars!

Yes I use an input selector SE and Bal with a little Psu. I like use this detail on my amps because I have unbal and bal sources.

[ang728]

Very Nice

[jose]

Thanks guys!

[cspirou]

Okay, I think I fixed it. I wont know for sure until I test it later tonight.

Just one question. The pins on the bottom of the negative regulator are bridged and I was having a hard time removing the excess solder. Then I noticed that all those pins were connected to the ground plane anyway. Does it matter if they are bridged?