Aria Line-Stage

Much research since last post. Space limitations for this mighty atom mean the tone control won't fit so we'll have to do that as an accessory. The Alps RK271 motorized pot takes up too much room. We have the option of a smaller Alps motorized pot if the European supplier would only provide a sample (we asked for one and await its arrival). The minimum order quantity is 400 pieces which is an expensive way to find out if it's any good. If it leaves more room then we'll look into doing the see-saw bass/treble control a few of you asked for.

This is hopefully the last lash-up before we commit the design to a PCB...



It uses a variation of JL Hood's class-A transistor triple more akin to the output stage of the great NE5534 classic audio op-amp, and with strong similarities to the gain-clone output stage. Unlike op-amps which deliver class-B current, this discrete BJT design runs with 10mA quiescent current per stage and is therefore able to drive its full output into a 1k Ohm load without leaving the class-A bias envelope. It has near 90 degree phase margin and a massive gain margin for stability under all expected loads. An earlier test stage measured 0.015% THD+N (distortion) and better than -85dB noise (CCIR quasi peak 22Kz - 22kHz ... A wtd. it will be nearer -95dB).

There are two stages per channel. They are inverting stages with the first providing all the gain and the anti-phase side of the balanced output, while the second stage re-inverts the signal to give the in-phase side of the balanced output and also the single ended output.

There is also a stereo balance control.

Input selection has not yet been prototyped but should be a matter of course being the passive section.

The space to the left will shortly be occupied by the remote control receiver/decoder board.


Edited by Graham Slee - 11 Oct 2011 at 11:04pm

Cyreg,

Is this UL? Yes! It is the very epitome of ultra-linear. It is just that the technique is supplied in a different configuration, but the technique still pays off.

Let's restate what IMO ultra-linear is: 1] linear input (or as close as possible); 2] wide bandwidth open loop gain; 3] the smallest amount of negative feedback to result in acceptable distortion and noise - in our case we consider that to be less than 0.05%THD and for a line stage preamp better than -80dB noise; 4] similar distortion at all frequencies - usually THD grows with increasing frequency; 5] class-A wherever possible (class AB acceptable for a power amp provided the bias is sufficient to null out to the lowest THD reading); 6] Good phase and gain margins for high frequency stability...

This is achieved in all our ultra-linear adaptations.

Sound? Try cheating a Harbeth M20. Since these things have run in they tell me immediately if I got anything wrong. More so than anything else I have used to monitor with. Near-field they are so unforgiving. The lash-up pictured above is being thrashed to reveal all and will be suitably tweaked as and where necessary over the next few weeks. At first switch on, and as is usual for us, it lacked much. But after 48 hours it is starting to do its (dare I say) magic.

We developed the Proprius to be as transparent as possible because we felt we would exhaust ourselves and use up valuable time (and lots of cash) trying other power amplifiers. We have not yet committed the Proprius to final design because we have detected a SQ issue through the M20 we'd not picked out on other speakers.

You can be assured that with our many years of experience that none of these products skew the sound toward one signature or another and as such will work independently in other systems but will contribute to each other when used together.

As I grow older I take much more time over the finest details - I never want our designs to be the subject of blame. They must impress in every area for a very long lifetime and must be ambassadors of correctness.

And yes, they must sound tube like as the very best of tube designs of which I believe there are very few on the market today. I would say Adrian's aging tubed Dansette record player could give many tube designs of today the run around. The best sounding tube designs I ever heard were used in theatre and cinema.

Latest update:

Remote Control working!

The remote control is a proprietary in-house design. The simplest implementation I've been able to think up, it runs in "hardware" rather than "software" simply because the learning curve in software is too much for my "grey cells" right now. I guess the functions could be put on a programmable IC (PIC) but that would only save one IC and whilst ever there's the appropriate chips on sale, I'd rather put my time toward SQ than showing off.

The designs lend themselves to considerable miniaturisation so don't worry about the remote being the size of a brick. It will be remote control size!



The development board now separated into transmit and receive sections. Transmit board on left. Receiver board on right.



And another photo showing the transmitter battery holder (4 x AAA cells).

So how well does it work? I placed the receiver on a shelf in the back room, and from the filing cabinet 6 metres away no hesitation. Aiming the transmitter 90 degrees off axis... still worked. Aiming at ceiling... still worked. Aiming in the opposite direction... still worked. Aiming at the floor under my desk... didn't work, but that was asking for a miracle anyway. To find out the actual range would mean testing it outdoors but as it's raining....

With just about everything tested and working it's PCB and case design time... stay tuned!