A review with some measured data and some modifications to further better this amazing mid-power Pro Amp
I’ve been always fascinated by this line of Pro power amplifiers of the famous brand of Meridian, Mississippi, for several reasons, among which I count
Modular construction, there is almost one board for each section: input & and gain, pre/driver & protection, output stage, power supply.
“All Transistor” construction, excluding the input like in many Hi-Fi amps, which gives this series a superior sound quality; this is almost an Hi-Fi like amp, and the version with the output binding posts is considered such this way. Moreover there is only 1 ( I say 1) polypropylene capacitor in series with the signal… (I’ll show You later)
It uses and advanced (= well engineered ) Modulated Class H output stage, which exhibits no glitches at high frequency during rail switching, also without using RC networks across the switching diode; I simulated the whole circuit with LTSpice and there is always a clean wave and a perfect switch also at 20Khz.
Triple Emitter Follower output configuration with a total of 12 output transistors for each channel; this translates into better behavior on lower loads and less distortion, and more reliable operating conditions due to a great number of devices compared to other brands for the same rating, which use typically 8.
So I decided to search for a second hand unit on the market looking over the Italian “Mercatino Musicale” and found the one I currently own, bought for 350€ from guy who was using it to drive an Electric Bass passive box with only one channel; I was able to do a first functional test with the owner at his home and the amp working ok. At first inspection the amplifier was in very good conditions, both aesthetics and electronics: it was already clean enough and with no signs of any kind of electronic damage.
This is the first pic I took: as I told before the modularity of this amp is clearly visible, with the input board on the back of the amp, the power supply in the center just in front of the transformer and the driver board mounted on top of the power board on each channel. This translates into having more cabling compared to other brands, and also models of the same manufacturer, but this is the price that should be paid for any modular construction; I must admit I saw several “Hi-End” commercial products of K$ with more flying cables than the CA6. Ooohhhh there is a “hole” in the list of the transistors, with an already screwed drill! Yes, the CA6 and the bigger CA9 share almost all, from the supply board to the output stage, with the CA9 having an additional pair of transistor in the lower rail (4cp vs 3cp)
One of the things that immediately appears from this photo is that the toroidal transformer is surrounded with a GOSS band, so a clear intention of keeping the radiated magnetic field to the minimum to favor the noise behavior of this amps and finally (again) the overall sound quality.
Another important thing is the revision marking on the back of the amplifier, which tells REV05, that seems to be the latest, and the one with the same transistor package on the whole output stage; as You can see on the net there are other revision where the transistors in the HI Rail have different, smaller, package.
Here are other pictures, taken after having disassembled the various modules for better cleaning of them.
The power supply unit
Even if the brand name is not reported the symbol clearly tells that they are Nippon (or United) Chemicon; once again we have room for additional 4 capacitors, used in the CA9. Mmmhhh …. interesting!
Detail of the output and driver board with a snapshot of the huge output inductor, built with a copper wire of an unusual (big) diameter, also for Hi-Fi amps
Let’s now discover some numbers of this very interesting amplifier; I connected my usual test bench and the amplifier was driven with a 100Hz sin wave, 3s On and 15s Off
The CA6 clipped at around 440W/8Ohm, 630W/4Ohm and topped 1500W/4Ohm in bridge mode configuration, so 750W/2Ohm stereo; compared to the CC4000 the first clipping on the scope is visible before the associated LED turns on, which happens very few watts later, when the distortion is still very low. I ‘d like to remember that the “clipping” is considered such when the overall distortion reaches 1%, and the clip LED of the CA6 seem to follow rule if You look at the waveform when it is on. Finally the sin wave at 15Khz and 20Khz with 2,83V RMS output is very clean suggesting a proper polarization for less high frequency distortion.
The amplifier met its specifications also connected to a standard wall outlet like mine, and it is a very good sounding and reliable unit, not a monster of power but not always you need tons of Watts! Given this I’m already searching for a CA2 (or CA4) unit to drive my mid-high horns, in order to replace my CPX900.
After the test I continued simulating the circuit with LTSpice, in order to fully understand its behavior; I started looking at the emitter resistors of the lower rail transistors and found them to be 0,33Ohm, with that of the higher rail to be 0,22Ohm. I fully understand that with higher values we have better current sharing among the devices we have, but in several constructions 0,22Ohm are already good enough. So in the simulation I set the rail voltages to get 750W/2Ohm with visible clipping, then I replaced the 0,33Ohm resistors with 0,22Ohm and the circuit showed the same behavior at 890W/2Ohm. Wow, simulations are almost perfect but for sure I’ll have some gain in doing this mod, and also some kind of improvement in the dumping factor; so I replaced the original emitter resistors with 0,22Ohm units, and used low inductance ones, and did it without removing the power board from the heatsink. After this I set the quiescent current to be the same as before, turning the only trimmer present in the circuit and put the amplifier again under test.
Basically 8Ohm power didn’t changed, but I was expecting this, 4Ohm increased to 650W and the bridged 4Ohm power was 1670W ; 835W/2Ohm stereo was not exactly the simulated but is however something better than 750W. So this confirmed that either the current limit was acting too early or the emitter resistors were causing too much power loss at lower loads.
Finally my curiosity was so high that I put in the circuit and additional transistor pair to simulate the CA9 output, and LTSpice said 950W/2Ohm with the same conditions as above…
Stay tuned because my next steps are to add an additional pair in the output, because all the components surrounding them are already in place, and to fill the 4 places in the supply board with additional capacitors.
Update on 1 July 2013
Last week I bought 4 x 15000uF Nippon capacitors of the KMH series (105°) from a German supplier and filled the holes on the supply board, so that now the total capacitance is 120000uF instead of 60000uF.
I didn’t checked the new power ratings after the mod, I’ll do during the next days; however I don’t expect I great increase of the RMS values, but for sure there is more storage for transient power