Go ahead Gary. I can always rename the thread if needed. I would love to know the differences
.
OK.... you asked for it.... this is a long post.
I think a bit of a background first - about 5 years ago, I was looking for a replacement Class D amplifier module for the servo-bass section of my speakers. We've been using Class D modules since 1991 because of high-power with high-efficiency (low heat) and very high damping: all necessary because of the brute-force nature of servo-controlled bass, and the fact that the amplifier would be built into a loudspeaker.
With impending ROHS, and stricter CE requirements, the module we used needed replacement. We tried making the ICEpower 500A modules work - primarily because they were the "high-end" flavor of the day. The ICE was a drop-in (almost) replacement for our existing Class D module - so my team happily proceeded along those lines.
Then, my designer-in-chief retired, and in the next three months, I learned far, far more about the craziness in this "audiophile" world than I ever wanted to or ever imagined.
The first thing that caught my eye was this:
The frequency response wasn't bad, but the phase went from +20deg to zero in the frequency range (10Hz to 140Hz) that we were going to be using it in. Isn't that going to play havoc with the servo-feedback, I asked? "Well, we'll just slow down the servo so that it doesn't matter. Nobody's going to notice," I was told.
Being the curious audiophile PITA I was, I started asking questions of the engineering team. Why the ICE module in particular? What others have you investigated? What are the advantages? What are the disadvantages? And no, just because Jeff Rowland uses it in an amplifier that sells for more than our loudspeakers wasn't a good enough reason!!
By then, we had already invested in tooling, parts, etc. but I wasn't willing to proceed on the technology direction for possibly the next 10 years without at least
some due diligence. Other technologies considered then were a module from Spectron and Nuforce. However, at that time I did not consider Nuforce a mature technology - they updated their modules almost every 3 months when they first launched. Spectron did not have a module, but the technology to design a Class D amp for us.
So, it ended up with the ICEPower again, and the UcD - which was very new at that time. So the comparison:
1)
Full bridge (ICE) vs Half bridge (UcD). The advantage of a full bridge design is that power supply bus pumping was far less than with the half bridge design, and this was a very important consideration for a subwoofer amplifier. Since the built-in subwoofer amplifiers are single-channel, we couldn't take advantage of the reverse-phase pair (as I mentioned in post #5 on this thread). Taking care of bus pumping in the power supply for the UcD would make it more complicated (+1 to ICE).
Since the full bridge design only requires a single-ended power supply (0V, +75V) the power supply requires half the parts of the half bridge design. The parts for the half bridge power supply (-40V, 0V, +40V) are cheaper and smaller than for a 75V power supply but not half as cheap and half as small (+1 to ICE).
With a full bridge design, BOTH speaker output terminals are "hot" and with DC level equal to half the power supply voltage - this means that both speaker output terminals float at +37.5V, and shorting either of them to ground which could happen in failure modes is disastrous. (+1 to UcD)
2)
Specifications. The ICE power module is 500W into 4ohms, the UcD 400W. For marketing purposes, 500W has a great advantage over 400W. In the real world, it made no difference. Otherwise, specifications are close - except for the phase response. (+1 ICE)
3)
Implementation issues. The ICE module has no protection of any sort - over current or thermal. If a woofer voice-coil melts (for example) and shorts to the basket there could be dangerous consequences if overcurrent protection isn't implemented. The UcD module has over current clipping, current limiting, short circuit and thermal shutdown. (+1 UcD)
The ICE module requires additional -12V.0V,+12V well-regulated power supply for the input stage. This is built-in on the UcD (but they provide instructions on how to add on your own for better sound). That Hypex has this built in gives (+1 UcD), that they care enough to provide instructions to improving gave my audiophile heart a nice warm feeling.
4)
Listening. At the end of all that, it was tied. So, the question was - can the phase shift of 20deg be heard in the bass? We built two subs using the existing servo-control and preamp from the original amp. Listened to the three and did some ABX - old against ICE, old against UcD and ICE against UcD. It was difficult to tell the difference between the old amp against the ICE module - so no improvement there, but there was a distinct rhythmic improvement in the UcD module. The bass seemed snappier and tighter, and PRaT was better.
So, with the UcD module the trade-off was slightly better sound and easier implementation with all the on-board safety against slightly lower power and more complexity in the power supply. With the better phase response, we also tighten up the servo control for even better bass rhythm and pace. The side-effects of this was better ambiance and a wider soundstage.
While all this was going on, I was learning about loudspeaker design - and realizing that soldering crossover components on while music was playing isn't a very good idea because one slip and you blow up the amplifier. Since the UcD had such good output protection, I built an amplifier for development purposes, and it sounded fab! Not as good as my reference amplifier at that time, but far safer as a short circuit merely made it burp instead of burn.
Building a similar amp for full-range duty with the ICE module, the upper midrange and high frequencies were far grittier. Violins had an unnatural "glare" and it made female singers sound unforgiving and hard. Now, this is not saying that the ICE module can be made to sound good - this is merely implementing the UcD with the available power supply and transformer that they sell for DIYers and the ICE module with the power supply as recommended in their application notes.
With the full-range amplifiers, the voltage in both speaker outputs made a much greater difference to the sound of loudspeaker cables. I found the difference between cables to be far more with the full-bridge amplifier than with the half-bridge amplifier. I can only speculate to the cause, but with the speaker cables 37V above ground at all times, it might have something to do with the capacitance between the cables and the carpet.
