i'm on a roll... the kgsshv

[ujamerstand]

The new boards are looking good. You could even use the unused ccs holes as mounting holes to the angled brackets. Quite nice so far.

[livewire]

What (if any) is the sonic difference between the ixys and a1968 current sources?

[DouglasQuaid]

If a group buy happens, I'm in. I may be new, but I can get a few vouches.

Edited April 9, 2011 by DouglasQuaid

[El_Doug]

Seems there is enough interest for me to at least ask if Justin will let us use his heatsinks. If not, the hifi2000's don't look too bad, I'll try and see if we can fit inside of them (without going for the 4U version, which is just too big for my tastes)

[MASantos]

Though the onboard heatsink version will be a lot easier and cheaper to build.

x2

If the performance doesn't change with the internal heatsinks it will be much easier to build, not to mention much cheaper!

[monsieurguzel]

I vote that if people are going to go through all the trouble and cost of building a KGSSHV with externally mounted heatsink fins, that the amp should run at a higher power than the previous incarnation. Otherwise deepak has a point that the other version is a lot easier to build with little/no audible difference.

[deepak]

This is what KG said earlier in the thread wrt to the externally mounted heatsink version when I inquired.

It does require a lot more skill and tools. I'm perfectly happy with the onboard heatsink version after he posted that.

Original design can run on +/-600 volts and absolutely requires T2 style heatsinking.

Which puts the chassis at about triple the cost of the rest of the parts combined. Maybe more.

I don't think that is what people want. This is supposed to be an everyman kind of thing.

If you want nuts, go and build a T2. Or you have to be someone like luvdunhill or n_maher with

the equipment at home to do the chassis. Or even kerry which manages to do that kind of stuff

on a balcony overlooking central park. Even i don't do chassis like that at home and i have

all the right machine tools. Getting the heatsink interface right pretty much requires a NC machine.

Or a whole bunch of patience.

I'm working on yet a third power supply design. Shunt regulator. Sure, why not. Except my

version does not blow up without a load, it uses the ixys parts as the current source, and i measure

the current thru the shunt, and if its too much, i shut down the amount the current source supplies.

Sort of the best of both worlds.

On another note if there is an interest in group buy for the semis I can spearhead that in another month when I have a lot more free time. Though I see some of them are already ~ $8 a piece at BDent with no price breaks at 50.

[kevin gilmore]

justin pointed out that the transistors are at the top of the heatsink, but you

also could just turn the board upside down. I don't have any decent thermal

analysis software, but if you figure on about 18 watts of heat, the temperature

differential (as measured on my T2) from the bottom of the heatsink to the

top of the heatsink is only about .3 degrees C. (and the T2 is more power)

Nothing to worry about.

But still, i'm going to try and calculate by hand what the differential would

be on a 3U size heatsink. I've seen ayre's and firstwatts with the transistors

at the top... krell with top and bottom, B24 with top and bottom...

I'm making all of the board files available. If people want to make multiple

board runs of the multiple different versions, that is certainly OK.

I of course will have at least one of each. But i can do my own chassis so

for me its easy.

What we need is some company like front panel express that can take standard

heatsinks and drill and tap holes...

Edited April 10, 2011 by kevin gilmore

[digger945]

justin pointed out that the transistors are at the top of the heatsink, but you

also could just turn the board upside down. I don't have any decent thermal

analysis software, but if you figure on about 18 watts of heat, the temperature

differential (as measured on my T2) from the bottom of the heatsink to the

top of the heatsink is only about .3 degrees C. (and the T2 is more power)

Nothing to worry about.

But still, i'm going to try and calculate by hand what the differential would

be on a 3U size heatsink. I've seen ayre's and firstwatts with the transistors

at the top... krell with top and bottom, B24 with top and bottom...

I'm making all of the board files available. If people want to make multiple

board runs of the multiple different versions, that is certainly OK.

I of course will have at least one of each. But i can do my own chassis so

for me its easy.

What we need is some company like front panel express that can take standard

heatsinks and drill and tap holes...

Yea that's what I was talking about earlier. Mounting the board on a parallel plane with the HS face. The devices could be mounted with the legs bent back a bit like Ti did on his amp, if the brackets get in the way. It would always be nicer though if there was a hole in the pcb to stick an allen wrench into for removal and troubleshooting purposes, unless the board is mounted on high enough standoffs that it wouldn't matter. Parallel plane mounted boards leave lots of room for PS and other stuff.

As for heatsink machining, as long as the pcb mounting holes were spot on, you could be off quite a bit for the device mounting holes and it wouldn't matter much at all.

Edited April 10, 2011 by digger945

[kevin gilmore]

There are holes in the board to stick a screwdriver in. i may have to make them bigger.

I figure the size of a hex head #4-40 should do it. About .2 inch.

look here...

http://www.amb.org/audio/beta22/

off board heatsinks, bottom mount.

edit: like this

http://gilmore.chem.northwestern.edu/kgsshvampv7.jpg

now with .2 inch holes for hexhead screws.

Edited April 10, 2011 by kevin gilmore

[El_Doug]

4-40 should be more than enough for anyone to mount the transistors

is there a sufficiently common type of mounting hardware that we could use for calculating standoff height? several of the early builders are using different types, which I assume do not necessarily have the same thickness :/

[kevin gilmore]

For the ixys part, the 4171g is required which is .07 thick

The to220's are .1780 thick. So the absolute minimum standoff

height should be .250 inch. But i would go .375 inch.

edit: they lie.

the .07 is actually .0780

which would make the minimum spacer distance .2560

So the .375 is exactly the board to heatsink spacer needed.

Edited April 10, 2011 by kevin gilmore

[digger945]

Can we see an updated schematic Kevin? I'm missing something here. Is the IXYS part an alternate current source? I don't see the +500v for the source pin.

edit ... maybe I'm looking at the symbol upside down

nevermind, I see now that that whole part of my schematic is X'ed out with a highlighter. I couldn't see it.

edit #3: How can you make it any more easier to build than this?

Edited April 10, 2011 by digger945

[DouglasQuaid]

Can we see an updated schematic Kevin? I'm missing something here. Is the IXYS part an alternate current source? I don't see the +500v for the source pin.

edit ... maybe I'm looking at the symbol upside down

nevermind, I see now that that whole part of my schematic is X'ed out with a highlighter. I couldn't see it.

edit #3: How can you make it any more easier to build than this?

The only way i can see to make it easier would be some ridiculous 900v servo to avoid needing the bias and offset pots, but that's asking for way too much.

If you want it easier to build, go with the boards with the individual stand-up heatsinks, and stick with the 1968.

As for the schematic, it appears the only changes since kgsshvproduction.pdf was posted are a few compensation capacitors.

Edited April 10, 2011 by DouglasQuaid

[livewire]

The comp caps exist on the last design. KG didnt put them on the schematic.

They do show on the BOM and the board overlay jpeg drawing.

FWIW if the latest outboard heatsink design sees the light of day,

sign me up for a board set and some sexy heatsinks!

[digger945]

If you have room, this might be another way to secure the IXYS part using a thermal pad with no mounting holes. This man'f is in Chicago, and has another similar, high temp device with a fulcrum opposite the device side.

[spritzer]

Yes, but the top and bottom bracket on the heatsink might get in the way of the transistors.

Need much better real picture.

I think my F5 is in a 3U box from Modushop...

[livewire]

What (if any) is the sonic difference between the ixys and the 2SA1968 current sources?

Any other benefits of the ixys over the other? (besides price)

[ujamerstand]

I remember KG said something about 2sa1968 sounds sweeter. But you can run the ixys part at a higher voltage, resulting in a larger voltage swing; and it is not being discontinued like the 2sa1968.

[kevin gilmore]

I'll answer the sonic differences in about another month. I have one of each built up.

Really the differences are very minor.

For those smart people like me who have accumulated a lifetime supply of 2sa1968's its

no big deal.

The little to220 clip is probably going to push the size out just enough that it won't work.

[livewire]

Will be showing the KGSSHV at a local HF meet next Saturday.

We have reserved a meeting room at a hotel in downtown San Diego with a private patio overlooking the beach.

A sponsor will be picking up the hotel tab, forty or so expected to be in attendance.

I'm wiring in a 50K DACT that I recently picked up from Justin.

Curious to see the improvement over the existing quad RK27.

Also gonna do some surgery on the bias circuits.

Just picked up a Koss ESP-950 and want to do 610 volt bias

instead of the existing low voltage Stax bias which I will prolly never use.

Finished "big box" pics to follow.

The theme will be "Back To The Future".

Ha-ha, it kind of looks like a piece of lab equipment designed by a mad professor.

[tims]

Congrats on getting it up and running - any thoughts on how it compares to other amps you may have tried?

Looking forward to how it goes down at the meet on Saturday; I'm sure it will be of interest to many.

Tim

Will be showing the KGSSHV at a local HF meet next Saturday.

We have reserved a meeting room at a hotel in downtown San Diego with a private patio overlooking the beach.

A sponsor will be picking up the hotel tab, forty or so expected to be in attendance.

I'm wiring in a 50K DACT that I recently picked up from Justin.

Curious to see the improvement over the existing quad RK27.

Also gonna do some surgery on the bias circuits.

Just picked up a Koss ESP-950 and want to do 610 volt bias

instead of the existing low voltage Stax bias which I will prolly never use.

Finished "big box" pics to follow.

The theme will be "Back To The Future".

Ha-ha, it kind of looks like a piece of lab equipment designed by a mad professor.

[livewire]

Thanks, I'm a stat n00b so I dont have much in the way of comparisons to other stat amps.

All I've heard is what I own. The Stax SR-202 & SRM-252II amp, also the Koss ESP-950 system.

Compared to any dynamic headphone setup in general, stats rule. No real comparison IMHO.

Here again, I'm a "mid-fi" guy so take what i say with a grain (or truckload) of salt,

I have not heard any of the real high end dynamic stuff. (but want to, someday)

Actually the KGSSHV's performance is a bit of a let down for me due to the fact it is

a mismatch with the Stax SR-202s. It tends to overdrive them at higher volume levels.(@SPL>95dB)

Distortion sets in and no decent improvement in bass response.

The Stax Basic headset works great with the puny SRM-252II amp it came with, a match made in heaven.

The sound is as good as it is with the KGSSHV but limited to a SPL of ~95dB.

So I guess the Basic Stax setup is matched well, just not head shakers or bass thumpers.

The Koss ESP-950 on the other hand.....

Wow! These cans can handle all the power the KGSSHV throws at them.

For a stat there is gobs of butt stompin bass and they can produce insane amounts of volume without distorting.

They sound fine with the little plastic box amp they came with,

but with these cans the KGSSHV makes a marked difference, although on the extreme end of things.

I dont know if I can (or really want to) take those SPL levels for very long.

The overall sound of the KGSSHV is very neutral, no coloration,

no perceived noise at all at any volume setting.

It suits me fine, just what I was looking for in a DIY stat amp project.

I have found that with the quad DACT attenuator I just installed,

the highs are sharper. There was a slight "upper veil" with the RK27 pot.

The log taper is different to what I am used to, there is more finite adjustment to the middle range.

Just as well, I dont need to have my head blasted off with the volume maxing at 12 o'clock

therefore having the upper half of the range unusable.

Edited April 21, 2011 by livewire

[livewire]

One thing I forgot to mention is that the KGSSHV having balanced inputs makes a big difference

in the quality of sound texture and minute detail retreival compared to the single ended stock amps.

I guess also having a good balanced source helps here too.

[kevin gilmore]

The design of the amp, and especially the power supply is for the

most neutral sound, and absolutely lowest noise. It should come

as no surprise that this amplifier is capable of driving many

of the older headphones to levels that exceed their capabilities.

This and the T2 are really for the top of the line stax headphones,

especially the new generation.

95db spl is already plenty loud.