Provided pins are used correctly according to NCxxxMP datasheet.
FusionAmp manual doesnt mention anything abt this..
So.. if the correct audio signal and ground pins connected to new XLR inputs and everything else of FusionAmp remain unchanged - is it supposed to work OK? Power on/off, circuit protection functionality etc etc.
Just trying to avoid situation when DSP/DAC part not being able to "See" the input impedance of amp on it's outputs shuts power down or smth like that.
I think that Hypex use the standard amp modules.
By disconnecting original ribbon cables going from the DSP/inputs to amp modules and build ones according to datashetes that fit your needs would let you make all the experimentation you want without major risk
Well, with adequate preamp and following the datasheet I did not suffer from any noise coming from the ACH01 but anyway could you share the ref of your acceleromter ?
A company called Ghent audio do make cable harnesses for these AMPs, nothing that cannot be built by a diy'er, but if you don't fancy building new cables/connectors then they may be an option.
Saying that, I am unsure if they do a cable for the tweeter module so YMMV.
The DSP section is powered from the NC122 or NC252 power supply IIRC. Should just be able to unplug the DSP power supply ribbon cable to turn that off completely.
Wouah, 30 dB feedback on the accelerator loop is that possible ? I'm very curious to see a functional block diagram or some kind of schematic
Well, if you look at the measurements of the Soekris in the previous page you can see that at -6 dB the noise level also shifts down roughly 6 dB. And at -60 dB the noise level is still about 100 dB down. Probably the noise of the shift registers is most dominant, and lower at logical level 0. Volume control is purely digital by the way, it is done in a FPGA and is then put on a 28 bit R2R resistors bank via shift registers.
I wonder indeed what the gain structure on Fusion amps is. I think I heard that the gain is set so that just the full amplifier power can be used. This is overkill if you typically listen to 1 Watt or so, though sometimes more. It may help to take a smaller Fusion amp from that perspective. Though I also wonder whether the preamp gain can be reduced, the one after the DAC, to get less quantisation noise and better SNR etc.
Fedde
That is a tempting proposition! 😛
I very well may come back at this, it is a 120 km drive, not too far...
Fedde
Interesting. So they trade SNR for distortion/non-linearity? 0.01% resistors still only give 14-bit linearity.
Yes and no. Typically the actual resistor precision is better than the rating. But more importantly, the resistors that are turned off (long term) do not contribute to the non linearity anymore. So if you run at 256x lower volume, you still have the same 14-16 bits...
I mostly listen to redbook material, though I have also high def. But I really care about the ability to do digital attenuation as I have bad experience with typical analog attenuators (pots, steppers, relais)... 😱
Fedde
Yes, for stuff like 1% or 0.5% resistors. Not 0.01% that you only get by actually picking and rejecting.
Please explain how that works in your view.
Let's take the trivial example of the smallest non-null signal, with just one bit turned on. At that point there are only two resistors active, forming a voltage divider. The signal value is directly affected by any tolerance.
So you think there is no design margin in 0.01% resistors? It may be the case, but I do not think that all resistors will have exactly that error. I did not measure them, so in the end it is speculation.
The tolerance of the LSB is not so relevant. Assume that indeed the linearity is up to 14 bits and the MSB is active. The two MSB resistors may cause a deviation of a scale 14 bits lower. However assume now that not the MSB but a bit 8 positions lower is active. Then the MSB tolerance is irrelevant, and the tolerance of this bit is still in a scale of 14 bits lower. This effect would be achieved with digital volume control. But you could also say that for recordings with a large dynamic range, the same principal applies because the higher bits only become active in the occasional peak. Also, some resistor tolerances may compensate partly for each other.
I hope we are not getting too much off topic by the way, I hope somebody can shed light on the Fusion gain topic.
Fedde
