[TrevorNetwork]

Does anyone have or know where to get the angled aluminum?

Kevin:

Will the power supply you're designing for the Dynahi be superior or more economical than the original reference supplies? When do you expect it to be finished?

Sorry for all the questions!

 

[hadron]

A warning about the HS-TWS, turns out the web page dimensions have a typo, the fin height is 29/32" not 2-9/32". My rough calculations give about 2.5 °C/W, should still be plenty for 15W, so I think I'll keep them without complaint.

I snapped a photo to give a sense of the size, the heat sink is just over 4-1/2" wide.

For thermal compound, ordered some paraffin based phase-change interface material, kinda pricey at $17, but I can't stand using that other goopy stuff either. Digi-Key part HS243-ND.

For the PS, I've made a few further refinements (adding pre-regs, a couple more caps, and opamp protection diodes) to Kevin's tracking version of the Didden-Jung circuit, from my previous post on the subject. Should have all the parts soon if I can remember to order the transformer tomorrow

I can draw up a new schematic when I get the chance if anyone is interested (PM me rather than cluttering up this thread).

 

[pooooo]

The transistor looks like this.

As you can see, electrical isolation between transistors and heatsinks are not necessary. However, some soft material between them might improve the thermal coupling in case the surface of the heatsink is not that smooth.

I think I'll stick with Dr. Gilmore's opamp-based power supply. I will power-up the 317/337 with darlington power transistors so that they won't get fried.

 

[hadron]

pooooo, why did you include the darlingtons? A LT1085/1185 reg pair will do 3A each. [Edit] and should have better ripple rejection and lower noise, important since the final stage opamps are supplied from this.

 

[pooooo]

Quote:

Originally Posted by hadron pooooo, why did you include the darlingtons? A LT1085/1185 reg pair will do 3A each. [Edit] and should have better ripple rejection and lower noise, important since the final stage opamps are supplied from this.

I found all the parts I need in my drawer while I was trying to build this, so some of the part values are a little bit different from Dr. Gilmore's original schematic. That's also a reason why the current through each transistor became 150ma instead of Kevin's 100ma. What makes things worse is that when I built this monster amp, I was thinking about using it also as a low power amplifier for my speaker. So I used 6 pairs of transistors per channel in the output stage instead of two, and my power transformer is a 300VA one. Well, different goal from most of you.

 

[sft]

I suggest that you can build a balaced bridge dynahi to drive your speaker for more power and control.

From My "Gilmore experience", balanced bridge version sounds great !

 

[pooooo]

Quote:

Originally Posted by sft I suggest that you can build a balaced bridge dynahi to drive your speaker for more power and control. From My "Gilmore experience", balanced bridge version sounds great !

Yeah, that will be great. Thanks for you advice,
However, this is only for my bookshelf speakers in my bedroom, so I don't think I'm going to need that much power. And I don't have a pre-amp with balanced output. This is also a problem.
The speakers of my main system is not a suitable load for dynahi, I guess. The impedance is less than 1ohm.

 

[sft]

Quote:

Originally Posted by pooooo How are you guys going to deal with the power supply?

Simple Dynahi PSU(output 30V) or Pre-regulator part(output 39V) using LT1085, 1033

 

[sft]

Quote:

Originally Posted by kevin gilmore ... opa541's work as the whole thing. 2 of them plus a 5 volt reference. just like original power supply. way busy, hope to get to this by the end of 2004.

According to Dr.Gilmore's thought, I've built this and it works.

Space is ready for two large heatsinks.

Some close-up photos:

OPA541

two large heatsinks:

Current limit resistor:FUKUSHIMA's 0.22R/5W

 

[chillysalsa]

I think I went overboard.. I ordered 2 of these GFC Tectrol 28V, 4A power supplies. They hum/vibrate a bit, so I think I will make a remote / separate enclosure for these puppies.

I'm hoping it's worth it, looking at their specs:

Is that any good??? Seems better than the PowerOne / Condor supplies that were suggested...

 

[Edwood]

Quote:

Originally Posted by TrevorNetwork Does anyone have or know where to get the angled aluminum? Kevin: Will the power supply you're designing for the Dynahi be superior or more economical than the original reference supplies? When do you expect it to be finished? Sorry for all the questions!

www.onlinemetals.com ?

-Ed

 

[hadron]

My (hopefully) final draft of the Gilmore-Didden-Jung tracking regulator. R15 & C13 may not be necessary, I included them for equal opamp input impedance and symmetry. Shouldn't be too difficult to convert back to a more standard (non-tracking) Jung super-regulator, if that is your preference. Also hoping the D44H11 Hfe is a bit more than the 20 min spec, otherwise might need to reduce the R9 value.

The transformer is not stocked by anyone I've found, put an order in to Parts Connexion but it will be at least a week...

 

[steel]

Quote:

Originally Posted by sft According to Dr.Gilmore's thought, I've built this and it works.

It's great
I'll to try it...

where to buy the Pre-regulator's PCB ?

 

[sft]

Sorry, I've forgot where to buy.

I found them in my drawer, then cut and collect.
Three terminal regulator kit is common and easily to get.

 

[kevin gilmore]

Here is how i make the angle brackets for the surplus sales heatsinks.
I realize most of you will not have this equipment, but for future reference
this is how i do it.

(Regular size pictures from my old camera, refuse to leave good camera
at work)

1) cut the stock aluminum pieces to 5 x 4.4 inch with one of these
http://gilmore.chem.northwestern.edu/heatsink3.jpg
I use 3/4 hard 1.8 thick aluminum sheet

2) Layout and centerpunch the holes
http://gilmore.chem.northwestern.edu/heatsink7.jpg

3) Punch the holes
http://gilmore.chem.northwestern.edu/heatsink2.jpg

4) Drill the heatsink holes with a #28 drill and tap with 8-32
http://gilmore.chem.northwestern.edu/heatsink5.jpg
large cup of coffee absolutely required, the resultant shaking
keeps the tap from binding.

Result
http://gilmore.chem.northwestern.edu/heatsink1.jpg
http://gilmore.chem.northwestern.edu/heatsink6.jpg

Total construction time about 45 minutes.
(includes making a pot of coffee)