Hypex based amplifier review


The Merrill Veritas mono blocs get a pretty good review here:


1000 hours to break in…

I very nearly bought a pair of these before going with the ARC ref 150.

Still hoping to hear some good news from Paul on the PWA.


I’ve been so involved with the NPC launch I haven’t had time to work on the PWA. Finally got the newsletter out yesterday and today I’ll be able to jump back in and design the new input stage.

We haven’t given up on the Hypex - it’s still a viable candidate for the back end - but we’ll see how it fairs with the new front end once I get it figured out.

The topology will be extremely low feedback double differential complementary symmetry gain stage.


Of course! An ELFDDCSGS is just the ticket.

By the way Paul, I particularly enjoyed the newsletter this time around; very open and inviting.

Of course! An ELFDDCSGS is just the ticket.


That's why Paul is at the helm...


@elk Thanks Elk!

For those of you interested in what this all means I’ll try and give a short description. Our standard output or input stage is one we’ve used for many years that’s a rather simple discrete differential op amp - composed of an input differential pair, feeding a current sourced gain stage that then feeds a complementary output pair. We then use negative feedback to achieve the gain we want. By running this stage at 60 volts we get good linearity and low distortion.

One of the problems we’ve recently discovered has to do with the openness of this circuit - or the lack thereof - and I believe it is due to the amount of negative feedback we’re using. Because we use a current source on the collector of the gain stage, it appears as an infinitely big resistor. Since the simple voltage gain of a transistor is the ratio of its emitter resistor divided into its collector resistor, that’s a lot of gain - and limited only by the beta of the transistor.

This may sound complicated but it isn’t. Just imagine a single transistor and it has an input (its base) and two other leads (a collector and an emitter). If you want gain out of a transistor you get it from the collector - making sure the collector resistor is bigger than the emitter resistor. So if you have a collector resistor that’s 10K and an emitter resistor that’s 1K then you have a gain of 10 (20dB). But, if you use a current source instead of the collector resistor, that current source looks like a HUGE resistor so the gain is also HUGE. This makes for a lot of open loop gain and we feed that back to keep the stage linear and low distortion.

We use a current source because it is a more linear approach to running the transistor - it’s also called a constant current source - because that’s what it does - maintains a constant current through the stage regardless of what the transistor wants or is being told to do.

So the first thing I did when i got back from the first amp shootout at Arnies’ was to figure out how to reduce the open loop gain without removing the constant current source. I did that using an old trick Dr. Bob Odell taught us years ago - placing a resistor at the junction of the collectors (of the gain stage and current source) to ground. If done properly you keep the benefits of the current source and now the gain is determined by the ratio of the gain stage emitter divided into this new resistor to ground. Bingo - lower open loop gain.

I got a magnitude more open sound from this arrangement, but it wasn’t enough. Now I want to eliminate all the negative feedback, but to do that I need a new circuit topology which supports low or no feedback - and that’s a full complementary design - equal halves of the circuit make it linear without using feedback - but also doubles the number of parts.



Very, very interesting!

I would never have realized the impact of a current source on the collector of the gain stage. It is embarrassingly apparent - after you explain it. :slight_smile:


We get an education from Paul and also the various members here. That is one of the reasons that I like the PS Audio forums so much. That and the friendly crowd.



@elk The first time I realized this was back when we made the 200CX power amplifier - it didn’t have a current source but it did have another collector of the opposite sex transistor - just picture the full complementary symmetry amplifier topology - you have two collectors tied together and it’s the same sort of thing.

Of course most designers are delighted to get that much OLG - they strive for it - but then that’s the difference in design philosophy.


@wingsounds13 Thanks JP. It’s fun and we have a wonderful group here. Thanks for all you do.