kevin gilmore

here you go its better than a lot of other eddie current stuff I have seen people that think that shunt regulated power supplies are better than really good linear supplies are wrong the input stage is very familiar really weak output stage with virtually no heatsinking painted really cheap steel chassis there are a number of ksa5 clone boards on ebay, some are mine and are obvious, but no one putting my board inside a chassis at the moment. if you have to buy one now on ebay, this one is complete, the power supply is not as strong as the original, but its cheap http://www.ebay.com/itm/Headphone-amplifier-Imitation-KRELL-ksa5-line-CPI-full-DC-amp-machine-ClassA-amp-/112050882595?hash=item1a16c0c823:g:61kAAOSwu4BVkJvH this one has my board but a different power supply http://www.ebay.com/itm/Finished-Kevin-Gilmore-KG-version-KSA5-headphone-amp-preamp-amplifier-HL-230-/142031703076?hash=item2111bfcc24:g:6GMAAOSwM4xXaLnE this one is my power supply with a different heatsink http://www.ebay.com/itm/Assembeled-Power-supply-board-for-KG-version-KSA5-headphone-amp-PSU-/322047929078?hash=item4afb8d6af6:g:-~UAAOSwoudW8MxP so you could buy the right board, and the right power supply, put it in your own chassis with a transformer for about $300 total...

capacitors floating in air filament transformers also floating in air with chipamp drive circuits not really a headphone amp is it.

singlepower made a sds xlr gold plated that was over $30k. inside it looked like all the other ones. craig published one picture of a obviously hand built electra that looked decent except for the el-crapo connectors. I have been told production models are the same garbage as always. companies that refuse to show whats inside are obviously not proud of their work. buy a Chinese made ksa5 klone.

get me one for a couple of days, then everyone will know. lately things getting high praise by many over there turn into massive pieces of poorly build crap.

with 2 x p-channel fets as the inputs, yes

updated board, one of the fets was labeled wrong

ether e not shown with ll2 or ll2t? these would not be the first stats to require 2500vppss

maybe they need 800v bias. otherwise the efficiency, or lack of it is going to require a cheap chip amp feeding transformers. Not exactly high end.

How d'agostino ish i would do the whole thing in copper, but that is stupid expensive. make the holes in the copper a bit big to take care of the thermal expansion differences

how thick is the mounting bracket?

The black color is supposed to make it harder to reverse engineer. Turns out it makes it easier with the appropriate lighting and photoshop. definitely the black color makes repairs harder, but its really the size of the circuit lands, the 1 oz copper and the diameter of the pads that makes it worse.

http://www.us.tdk-lambda.com/ftp/Specs/km.pdf medical grade, switches at 132 kHz. definitely can fit everything inside the box. This would be the most expensive part at $80 even cheaper if you go for the 15 watt version. definitely everything fits in the box, and no phase splitter or other crap because its a true differential balanced amplifier. the quad alpha pot would be the 2nd most expensive part at $30 so say $200 in parts plus about $25 for the box. and it won't blow up even into a 4 ohm load. or a dead short.

My point was not that made in china was crappy, it was that made in china is priced such that made in usa cannot compete. Audio-GD for example makes ultra high caliber equipment, and at a price that i cannot even buy the parts for. Ksa5 boards now on ebay that are my layout stuffed with parts and assembled and sell for a fair bit less than i can buy the parts for.

Headroom designs were ahead of their time and infinitely better quality i still think I can build something much better and still get it all inside the same size box

so here is what I think is going to happen. up to 1000 of these things in the field. as already reported elsewhere, plug anything into both the se and balanced output and the source resistors and possibly the fets fry. may be ok for the first 500 which have a lifetime transferable warranty, but once any of the second batch show up used and someone does the thing some damage and its not covered under warranty, or the warranty repair is Mikhail class, people are going to get mad. The pictures posted show that cavalli cannot do decent repairs to his own stuff. when this thing shows up on ebay as made in china kits for say $150, and are a better and more repairable product, the used price is going to fall thru the floor making even more people upset. having my own web server ( ) I know how many times the schematic has been looked at, and where the hits are coming from. More hits from china than I ever thought possible.

mosfet current sources are depletion mode. made a very different way from enhancement mode mosfets.

on srm717/727 only I think stax used this relay with a detect circuit that opened the relay if either of the + outputs exceeded + or -400V or the difference between the + and - power supply exceeded +/-30V. which is why that fucking idiot over there who thinks he was gods gift to modifying stax amplifiers had all sorts of trouble. on the T2 pretty sure it was a delay circuit only. will look later.

absolutely we need to cryo all parts.

actually it was the GG hard to keep track of who is building what.

what is likely the final version posted including board file. actually performs better with bipolar outputs.

you have to love open architecture

The depleted uranium and the beryllium are enough thank you A tungsten knob would be nice. Almost as hard as the uranium

great build not sure where to put this (because i'm lazy) but congo5 figured out what was wrong with his grounded grid dht amp. hint: 4.7M resistors replacing the 5.1k output resistors do all sorts of bad things to the sound. (unless you are a severe bass head) anyone know somebody that will make me a couple of metal knobs? I will supply the material. warning: toxic, and slightly radioactive

so it helps to have it on the bench where you can monitor whats going on while you listen to it. unloaded balanced (which would be a 804 ohm load due to the feedback resistors), 40 volts peak to peak. and its a hard clip, but thats ok. Now add a 50 ohm balanced load and the output voltage dips to under 20 volts peak to peak balanced. Turn on the other amplifier and do the same, and with both running, 17 volts peak to peak. At which point everything, the output transistors, and everything else is getting pretty warm. A number of the surface mount transistors are running pretty close to their power limits. The why is probably obvious, the power supply is sagging. Down to less than +/-12V and the rails ride the music (or test signals) Plus you loose almost 7 volts due to the Vgs of the output fets. The liquid fire had a pretty crappy underpowered power supply too. Which is why even on hd800 you can hear compression in the music. And audeze... ouch. And he6 well don't even bother. The singlepower SquarewaveXL did a much better job of sounding better (till it burned up) A decent amplifier can actually be built in a box this size. But its not easy. And every single input/output connector just takes away from valuable board space Also makes you wonder how much better a liquid gold would sound with a real power supply. I should be able to take the goldreference power supply and get it to mount in the same holes. edit: i should mention that tests were done at 115 VAC. The switcher is likely to produce more power when run on 220V

Singlepower had a 5 year warranty too. Now nothing currently made is anywhere near that bad, if cavalli is experiencing a return rate for repair of more than 2% he is in a world of hurt.