kevin gilmore

Birgir is going to make prototype boards soon

The LF is a rip on a early 1980's design with the jfets replaced by tubes, and vertical mosfet outputs replaced with standard international rectifier parts (now vishay). Not bad, not great either. Power supply limits current delivery especially on inefficent headphones like he6's. A single gain stage, no feedback design like some of the pass labs amps, or the circlotron it ain't. A S-30 blows the crap out of a LF for about the same money, or less used (if you can find one) and drives he6's or LCDX with true authority. And rise and fall times more than 50 times faster than the LF.

At least for lcd2's i have not found a tube amp that makes them sound right. Neither OTL or transformer amps sound right. Well except for an atmasphere.

Thanks for trying. Was worth a shot. When is a standard not a standard?? Pars you are up... But really, shouldn't the board house do the panelizing for you??? (without charging you a setup fee?)

all the files are there and seem to work, the extra file can be ignored because the tool sizes are in the echelon files ??

try this, but the sizes are in the one extra file http://gilmore.chem.northwestern.edu/doug.zip

Newest version of the blackhole amp is 3.45 x 3.67 inches http://gilmore.chem....u/blackhole.zip not done in eagle however.

Very Nice

Fully loaded, on the 20v lines about 500ma On the 15v lines about 250 ma 12v line peaks of 30 ma 5v line I gotta look up

Have not even thought about the power supply yet. In the simplest version the smallest power one tripple works great. If you go with the outboard amplifier modules then you need something that does +/-20, then you can regulate that down to +/-15 with simple 3 terminal regulators, and +12 to drive the relays.

I modified the kgsshv power supply and batepower boards for the next board run. Bias is now a 3 pin connector, TST which is the 580v direct, 580V which is after the 5M resistor and the ADJ which is the center of a resistor divider for any other voltage. If you have a 10 meg input resistance dvm, then the correct voltage measured after the 5M resistor would be 386V.

Each balanced board should pull 600ma on each of the power supply lines.

courtesy the stax mafia http://gilmore.chem....sr1_sm_sony.pdf

This is what we do

can't seem to find the schematic for free, but a few places are selling it for $5 to $10. If you get desperate, buy it and post... If its actually a bad solder joint to the big black cap, pulling up on the relay end of it may temporarily cure the problem, in which case soldering just that pin should do it.

next time it cuts out, tap the relay first. If you can be certain that it is the relay and can't find a replacement, you can always short across the relay. On second thought it does look like a bad solder joint on C334. third thought bb204 is a grounding bar likely, so the relay may short the audio to ground, so shorting across the relay would actually make the sound disappear. They sure don't build them like this anymore. Going to try and find a schematic.

you are doing a single ended input right?? unbalanced to balanced is never absolutely perfect and the input tube matching side to side is the most important.

although i think that this is a bit expensive compared to everything else, it would be a complete plug and play stack. http://www.liquidware.com/shop/show/TSL/TouchShield+Slide plus and arduino uno and the ethernet board. as long as there are enough pins left over to drive the i2s bus and a rotary encoder with buttons, should be ok.

Anything is ok by me. Just has to write to the pcf8574 chips and write to a color display. We already have some of the code for arudino for this,it's minimal. I don't know if the linux for the raspberry pi does real time or not, an interrupt in the middle of programming the volume control could lead to interesting noises. The video output or the hdmi is probably not suitable for our uses unless someone knows of a really small display that takes composite or hdmi. Touch screen also a real sweet idea for local control.

disconnect all wires to the transformer and measure the transformer for shorts between the windings.

They should be 4700uf/25v caps, and the unregulated voltage should be 21 volts.

It should be clear by now that all of us are nuts. Some of us are VERY nuts. At least a couple of the members of the stax mafia are criminally insane. My dream is to make the all dht direct coupled triode thing in a machined non magnetic damascus steel chassis. I'm trying to find a suitable piece of lapis now.

I certainly have not built one yet. Don't have enough time.

I'm almost done. Going to order different opamps for the servo's today. And i've redone the output using the audio quality opamp for that purpose and now a different opamp for the servo. Currently the input stage dc offset is 200 microvolts which is perfectly fine, but i'll get it down more. Without the servo its actually 4mv which is known to most people that use ths part, but evidently not others. The output stage with the audio opamp as the servo comes in at about 5mv, and while fine for headphones, driving monster power amps is not a great idea. I got to listen to it over the weekend, and compared it to a number of the most expensive preamps out there, and it holds its own just fine. Kerry is testing the multiamp outboard amplifier. Sure seems that for control the way to go is an arduino with the ethernet board, a ethernet to wireless converter, and then your favorite ipad like device talking to the preamp via a webpage. And then an oled display on the preamp to show levels. The biggest mess up with this was the fact that the discontinued omron relays had a different pinout than the panasonic relays i'm now using.

You want +/-500 volt power supplys, you have to give a little. Both the 10m90s and the 2sa1968 are much more thermally stable with respect to current, but are only 900 volt parts. The 10m90s is designed to be a sharp cutoff device and 50ppm. The ixtp01n100d is not designed this way. Stacked current sources don't perform as well in a standard configuration and have to be bootstrapped to perform at low Thd levels.