Hifisonix kx-Amplifier

www.hifisonix.com
Joined 2003
Paid Member
Vunce, ideally you want to terminate at the 0V on the amplifier module. It was for reasons of space and compatibility with the existing sx-Amp builders that I terminated it back on the PSU.

However, if you are building a new PSU from scratch, as long as the connection between the Zobel connector on the amp board and the PSU is short, you will be good to go. By short I mean c. 6"-8" and tightly twisted with the 0V, + and - supply wires to keep the loop inductance down.
 
www.hifisonix.com
Joined 2003
Paid Member
Hello Asch,
for the input wiring and the bias switch, its standard 2.54mm pitch. But for low level stuff I always solder directly to the board.

For the power and speaker connections, I use 6.3mm push on tabs - there a link to suitable type numbers in the build doc.

Remember you will need the mating push on receptacles as well. You can solder your wire into the receptacle, but a better way is to use a crimp tool - they are about $15-$20 and usually available at most hardware stores in the electrical section.
 
Quick question.
Have the SX amp. I just couldn´t be happier. Its awesome!
Thinking of upgrading to the KX. (Just have to finish the PeeCeeBee v4h) I have a very big case with big heatsinks. (4U full length modushop). It also probably would not hurt with a little extra power.
Are there any other downsides than excessive heat when juicing up the PSU to +-35Vdc (Class AAB) which you write about? Sincerely Magnus
 
Last edited:
www.hifisonix.com
Joined 2003
Paid Member
Hello bambadoo, glad you are still happy with your sx-Amp.

If you run on +-35 volt rails I recommend class AAB at a total of ~600mA per amp. My rails are +-28V and I can switch between full class A or class AAB. The kx-Amp still sounds very clean and warm on AAB (I am using KEF LS50 and B&W 703 - so not the horns like you have).
 
KSA1381

Hello.
I am also hopping on this bandwagon with KX owners, but seems that i have bad start.
Can't find any KSA1381 on web. Is there any complementary pair that I can put in there?

Nearest mouser stock is on late september for 900 ordered items, next stock on november...
I want to make this amp, as it's the only silicon one, that can possibly postpone Tubelab SE2 project.
 
Actually I've found 12 in stock in local electronics store. Time to order cookies from jims...
Shall i go for Ripple Eater or pcb in kit is enough?

My trafo will be toroidal 2x25V 6A, so i suppose it's enough. Has anyone calculated total power usage in kWh? My wife keeps the bills and complains about.
 
Hello.
What would it take to run the KX in class A mode from 35V DC rails ?

Is it just the issue of bettering the 0,3 K/W of the heat-sink, providing yet something bigger in terms of cooling, or are there other, electronics related restrictions? Would the output, driver or other transistors exceed their SOA in such a scenario?
What would happen?

Any chance of alternate resistor values to "accomodate" for Class A at 35 V rails?
 
Any chance of alternate resistor values to "accomodate" for Class A at 35 V rails?
The reason / origin of my earlier question is that 24V AC toroidal transformers are avialable and abundant. I happen to have such in the drawer somewhere.
Any chance of re-using them AND achieving Class A ?
My reasoning is as follows:
sqrt(2) * 24VAC = 1,414 * 24 = 33.93 V DC.
Now, I suppose that I can deduct two forward voltages of the diodes in the rectifier bridge, working under heavy current load conditions:
33,93V - 2V = 31,93V DC (after filtering).
I would also assume that there will be a bit of 'sag' due to the load current? Say, 1V of sag? Or maybe more?
That gives probably something around 30 .. 31V DC on the rail, as measured under real operations conditions.
Such a voltage is 2 .. 3 Volts higher than the upper limit for class A, as specified: 28V.
Any chance of "streching" the 28V as specified, up to 31V AND running class A ?
Even if a bit hotter and on a bigger heat sink?
Any ideas how to "loose" the 3V, to hammer down the rail voltage to 28V ?
Would / could an off-board, external, ripple eater be used as a means to successfully get rid of the unwanted, superfluous 3V?


Not to mention a totally ... brutal idea ... of simply using more than one rectifier bridge, so as to loose not two, but multiple Vf's on multiple bridges (... yeah, stupid idea, but asking pro-forma anyway).
 
Last edited:
Is there any reason, why the KX could not be used in bridge mode?
Idea as follows: purchase TWO sets of the PCB kits, start up all four amplifier boards, and then, after checking that all boards work ok, pair them up, connect the speaker in bridge mode, symmetricaly, between the hot speaker outputs of two boards per channel ?
There would be a need for some inverter at the input, or inputs taken from a balanced XLR connection.
Is this possible to do, or is there something in the electronics of the KX schematic, that would render such attempt impossible?
 
Last edited:
How do I wire up / connect the Zobel networks, when pairs of KX boards are to be used in Bridge mode?
Simply treating each board independently, with Zobel connection from each of the speaker outputs on each of the boards to the ground connectors of associated PSU's, or is there a better way?
Question rather crucial, considering that the Zobels are critical and can not be omitted.
Since in bridged mode, there will be two PSU boards, ... are there any critical issues to look out for when wiring up / connecting the two PSU boards with each other?
I would imagine that the ground connection between the two PSU boards would need to be shared/joined. What about the impact of this joining of the PSU's upon the rest of the wiring scheme? Anything to worry about?
 
Last edited: