I can confirm LMac's statement and report other more or less serious shortcomings of this device. And here we are already talking about the improved second version, the first series probably had other problems such as a completely inadequate sensitivity. This problem has only been insufficiently improved in the “improved” version that I own.
But first things first, the problems start with the incorrect printing of the scale. It is clear that this is not an accurate “measuring device”, but the approach should be correct for such a not very cheap device. The current power is displayed in watts and dB. The reference is 0 dB / 100 watts. The power ratio is calculated as 10 * log10 ([current power] / [reference power]).
This means that in addition to the reference value, only -40 dB / 10mW is correct. All other values are printed incorrectly.
-50 dB should be either 1 mW or -49.2 dB,
-30 dB should be either 100 mW or -27 dB,
-20 dB should be either 1 W or -17 dB,
-10 dB should be either 10 W or -7 dB,
+4 dB should be either 250 W or +3 dB,
+6 dB should be either 400 W or +4.8 dB.
However, this device also fails to meet other basic promised features. The missing terminating resistor has already been reported. This is not a problem for digital or conventional transistor amplifiers, but can be fatal for a tube amplifier if you rely on the manufacturer's specifications.
The “measured values” displayed over the frequency are “different” than expected. The display over frequency is not linear with my device. The display value falls at low frequencies and rises at high frequencies.
This is inexplicable, as the greatest power is transmitted at low frequencies. Measurements were taken with a high-quality TRMS bench multimeter up to 100kHz. The outputs were loaded with 8 Ohm resistors.
The display of the channels is also slightly different here, which is not really a problem because, as I said, it is not a “measuring device”, but it is still remarkable. Here are the values I determined; reference 1000Hz. First column is the frequency in Hz, second column right channel in dB, third column left channel in dB.
20 -60 -60
50 -31 -29
100 -16 -15
200 -10 -10
500 -2 -2
1000 0 +0,2
2000 +0,2 +0,8
5000 +2 +2,5
10000 +3 +3,5
20000 +3 +3,5
Without a connected load resistor, the display increases by almost 1 dB. This should not be the case because the connected transistor amplifier has a very low output resistance. The output voltage was nevertheless measured and does not increase. The display should remain the same with or without load.
The transition resistances between the amplifier and speaker terminals are relatively high on my (new) device at 60-80 mOhm per channel. Not dramatic, but it remains to be seen how this will develop over longer periods of operation. 80 mOhm corresponds to 17 mOhm mm²/m after all. With a 4 mm² cable 4.25 mOhm/m this means ~9m of additional cable. The relays used do not appear to be of the highest quality and are probably not gold-plated due to the high contact resistance.
I have not tested them under high load. But it is to be feared that there will be effects at high output powers. After all, with the specified 250W at 4 ohms, around 8 amps flow. The initial posting by binirm shows clearly visible distortions in the 20 watt range and above.
The holes in the connection terminals are only 10 mm deep, but a banana plug has a length of 18 mm. This means that almost half of the plug cannot contribute to the contact and therefore increases the contact resistance if plugs are used.
There is no need to comment on the manual, it is really poor and needs to be comprehensively revised.
It must be stated that quality assurance has not taken place, or not to the extent required. And we are talking here about the improved version that has already been revised. The high-quality feel of the device and its great appearance do not make up for the lack of inner values. A lot of appearance, not much substance.
Naturally, I contacted Fosi support with all these points. They took a week to reply and initially questioned the validity of the points, only to admit a few emails later that the criticisms were correct.
The support team's basic statement that the device is merely decorative in its current state is the bottom of the barrel. You don't know whether to laugh or cry. As “compensation”, $10 was offered - for a new device that cost €170.