Hey rille I think that I see the problem.

The 10M90S should have better isolation from the heatsink.

I assume that you used the short thermasil plastic washer

and a thin backing pad between the sand and the heatsink.

This may not be enough to prevent a high voltage breakdown

(arc path) between the metal mounting tab and the grounded screw/heatsink.

To insure complete isolation between the two, one needs to use an aluminum oxide ceramic

backing pad with a deeper shoulder plastic washer instead.(along with dielectric heatsink grease)

A discussion of this has been covered in this thread.

(click on my thumbnail pic above to see the enlarged detail of the mounting)

The Mouser part numbers are:

Alumina Pad: 532-4170

Insulating Shoulder Washer: 532-7721-3PPS

Edited April 3, 201115 yr by livewire

In fact any sand with an exposed metal mounting tab

that is used throughout this build should be mounted in this fashion.

If you notice, KG has silkscreened the alumina pad locations on the pc boards.

At the risk of sounding redundant,

In practice any of the transistors or FETs that have exposed metal mounting tabs

which are affixed to any of the heatsinks, mount them this way.

Silkscreened location or not, the alumina pads will fit in there.

(you may need to resolder or replace the transistors to get the correct hole height matchup)

The only exception to this rule are the +/- 15 volt regulators.

Edited April 3, 201115 yr by livewire

I used a fiber-glass reinforced silicone foil which is rated for 4kV.

Mouser is too expensive to order only some small parts. Is this also ok: http://de.rs-online.com/web/search/searchBrowseAction.html?method=getProduct&R=2182609?

The 7721-3PPSG will fit? It has a inner diameter of 2,84 mm for the screw. I still had to tap the heat sinks for M3 because they have 6-32 UNC, which is a little bit thicker than M3 and was too thick for my insulating shoulder washers (these one: http://www.reichelt.de/Waermeleitpaste-Iso-scheiben/IB-2/index.html?;ACTION=3;LA=444;GROUP=C88;GROUPID=3384;ARTICLE=8713, the IB 2).

I know nothing about silicone foil or it's ability to insulate under pressure

where the sharp edge of a transistor mounting tab may cut through.

I've had good luck with the fiberglass reinforced products they sell here

and bad luck with cut-through using the very thin non-reinforced thermasil backing material.

The alumina ceramic product you reference looks to be the right stuff.

The one on the left appears to be correct for use with the TO-220 transistor package.

Generally I do not use metric stuff, so I'll leave it up to you to figure out the fit up

of the products you might consider for use over there in Europe.

I can tell you what I did using American inch standard products.

The 7721-3PPSG did not fit inside the hole of the TO-220 mounting tab.

I used a dremel grinder to enlarge the hole diameter a few thousands of an inch.

Then the plastic shoulder washer slipped right in. The heatsink I used came tapped for a 6-32 screw.

I used a smaller diameter 4-40 screw that fit through the 6-32 heatsink hole without threading into it,

The screw entered from the backside of the heatsink and was nutted from the front side of the transistor.

(with the nut resting on the plastic flanged shoulder washer)

To make the screw fit between the heatsink fins from behind, I parallel ground the screw head flat on two sides.

I assembled it all together in this order:

screw > heatink > thermal coumpound > ceramic pad > more thermal compound > transistor > plastic flange washer > nut.

The 7721-3PPSG washer when used with the alumina ceramic backing plate protrudes

through the transistor tab (and fits) into the hole of the ceramic plate.

This insures that there is no electric arc-through

between the screw threads and the hole in the metal transistor mounting tab.

The normal short shoulder plastic flange washers are not deep enough

to insure that there are no exposed screw threads in this critical area.

Edited April 4, 201115 yr by livewire

As a footnote, I used nylok "aircraft" style nuts.

Both the screws and the nuts are non-magnetic stainless steel.

Hey guys, I've been away for a while and have lost track of the project, any news about boards?

KG is doing some minor revisions to the board design.

A few silkscreen changes and reroute the input fet traces on the amp boards.

don't forget the new hybrid shunt supply!

One thing worth paying attention is that, while their pots are great, they don't have a stable design. AFAIK, they've changed the designs on both types at least 3 times.

The very first ones have PCB pins which would fit for your need but later, they no longer did that way. No idea why.

The SMD one even has more stupid design. The wires are pre-soldered to the pot and you have to make sure about the correct length before ordering. Of course you can change the length but that'd require opening the pot.

The Dale one has very stiff feeling and requires a shitload of force to turn the shaft, I couldn't make any step without using a heavy knob. The SMD, however, feels just like the DACT.

Both are huge, maybe about twice the DACT.

Oh, I forgot about the price but it'd be better to leave it to private.

true, i contacted them and they only make it with wire leads now. oh well. that's really annoying to assemble in production. on the plus side, when the attenuator fails or they change the design again, it doesn't need to fit onto a specific PCB.

it seems like its about the same size as the Alps RK50, maybe a little longer

Edited April 6, 201115 yr by justin

true, i contacted them and they only make it with wire leads now. oh well. that's really annoying to assemble in production. on the plus side, when the attenuator fails or they change the design again, it doesn't need to fit onto a specific PCB.

it seems like its about the same size as the Alps RK50, maybe a little longer

I put it on the T2's pcb to compare with the Rk50 and it's indeed longer.

Been listening to the HGSSHV non-stop.

Sure is different having a fully balanced rig.

Mucho mas detail and headroom to spare.

Now I need some better Stax....

What source are you using?

Lenovo PC Pentium Dual Core 2.5Ghz > Foobar 2000 > FLAC > Neko Audio D100 DAC

A few things I've noted,

When I power it on, the living room lights go dim for a second. Them big caps are thirsty!

Cased up it runs relatively cool. After a few hours, it's just warm to the touch.

Dead silent noise floor. Zero hum or static no matter what the volume setting.

Being a "one box build", I took the extra precaution of ordering a toroid with a "core band".

Minimal shielding around the periphery, it helps cut down on EMI radiation.

I like my new toy. Next up is a better attenuator, although the RK27 dont sound bad.

Eleven or twelve o'clock on the volume control is the sweet spot for me. This thing has mucho grunt to spare.

Edited April 6, 201115 yr by livewire

When I power it on, the living room lights go dim for a second. Them big caps are thirsty!

Here is how to avoid that problem. Turn on the amp remotely with some fancy automation

gidget first.

Then turn on the room lights.

You should see my lights dim when i turn on the krells.

Gidget?

Back in the day she used to turn me on!

Doesnt she do Boniva commercials now?

Gidget?

No, Kevin was referring to the later sequel, Inspector Gidget!

Go Go Gidget Flying Hat!

No, Kevin was referring to the later sequel, Inspector Gidget!

Go Go Gidget Flying Hat!

Now I'm going to have that song stuck in my head for hours.

The greatest American hero has a mother?

Count me in for the next board run. Thanks.

for those of you that want to do a 2 box thing with side heatsinks, i'm almost

done with the amp board for that. Very symetrical.

Oh, that's perfect.

http://gilmore.chem.northwestern.edu/kgsshvvert.jpg

in progress

for those of you that want to do a 2 box thing with side heatsinks, i'm almost

done with the amp board for that. Very symetrical.

EDIT: Upon checking board img vs schematic, I retract my question.

I can only stare at this piece of crap budget monitor in the office for so long before nausea sets in, but haven't found any errors yet. I couldn't read input fet's markings, so I'm not sure there.

Does MMMetals still have the T2DIY heatsink die?