Messages

[Same as above but with expanded frequency scale]

Same as above but with expanded frequency scale

[Same as above but with expanded frequency scale]

Same as above but with expanded frequency scale

[Same as above but with expanded frequency scale]

Same as above but with expanded frequency scale

[Introduction:]

Introduction:

In this exercise we continue on with the jitter tests using the J-test methodology outlined in our original jitter test thread but this time we want to see how the jitter performance varies as we change the DAC Digital PLL bandwidth. The ES9038PRO DAC has 15 different settings for Digital PLL bandwidth but obviously the tighter the bandwidth the better the jitter performance but this comes at the expense of the DAC being able to lock onto an asynchronous signal with inherent jitter so there is a trade-off between bandwidth and jitter tolerance.

Test Methodology:

There are a variety ways of measuring jitter but the single single RMS or peak jitter metric is not very revealing since it does not reveal the nature of the jitter, so a more detailed analysis such as the J-test method is used to test the susceptibility and tolerance of the device under test. The J-test methodology is described on the ASR forum J-test methodology 1 and J-test methodology 2

Test Results

• DPLL BW Register = 1
• DPLL BW Register = 2
• DPLL BW Register = 3
• DPLL BW Register = 4
• DPLL BW Register = 5 (default)
• DPLL BW Register = 6
• DPLL BW Register = 7
• DPLL BW Register = 8
• DPLL BW Register = 9
• DPLL BW Register = 10
• DPLL BW Register = 11
• DPLL BW Register = 12
• DPLL BW Register = 13
• DPLL BW Register = 14
• DPLL BW Register = 15

Conclusion:

Below the DPLL BW setting of 5 there doesn't seem to be any more improvement in jitter to be had so it seems like this setting is the best compromise between jitter performance and PLL bandwidth.

[Test Methodology:]

Test Methodology:

In this test we excite the DAC with a 1KHz square wave at -3dBFS and observe the step response on a scope for each of the seven different reconstruction filters available in the DAC settings menu.

Test Results:

• Fast roll-off, linear phase filter
• Slow roll-off, linear phase filter
• Fast roll-off, minimum phase filter
• Slow roll-off, minimum phase filter
• Apodizing, fast roll-off, linear phase filter
• Hybrid, fast roll-off, minimum phase filter
• Brickwall filter

Conclusion:

[These preamps operate in perfect sync with each other and the master preamp controls the volume, balance and muting on all of the other slave preamps in perfect unison.]

We had a recent enquiry by someone who wanted us to build a preamp with 16 outputs ! But what if someone asked us to build a 24 or 32 or even a 48 channel or more hi-end stereo pre-amp and our preamp only had 8 outputs? Fear not with the external Link-Port on both the UP2 and UPP you can chain as many preamps together as you need to get as many outputs as you need with no loss of information or performance ! Why would you need such high channel count ? One application comes to mind and that is a phased array loudspeaker or wave shaped loudspeaker with many independent driver channels such as the B&O kii audio etc or even a multiple active speaker that are all controlled by the same pre-amp.

By itself the Ultimate Preamp 2/+ has 8 outputs plus an optional digital output to drive an external stereo DAC bringing the total output count to 10 channels which is still far short of 48 channels. However, the main DSP board used in the preamp has provision for connecting multiple DSP boards as a master and many slave arrangement in the one chassis using the on-board Link-Port connector.

However, from a practical point of view this would require a redesign of the enclosure and power supply to handle multiple boards and output connectors which is just not practical at this point in time. We just needed to extend this facility outside the chassis so that we could link multiple preamps in 8-output clusters and they all needed to sync perfectly together when changing things like the volume, balance, mute etc because you don’t want to have to adjust each preamp’s volume control every time you need to adjust the volume.

Because this was a planned feature for the UP2 and UPP we needed to make sure it worked perfectly before we started shipping the preamps so whilst we still have all of the preamps in our lab undergoing testing before we ship them what better time to test out this feature which is why we have spent the last week updating the firmware and hardware to make sure it all works as advertised. As an example we took 6 preamps and chained them all together in a one-master five-slave system to give a total channel count of 48 channels !. We also loaded the simple Audioweaver pass through example which feeds a stereo signal directly to all 8 channels but this could be any design of your choosing. It also helps us test the HDMI I2S input and the Link-Port HDMI output ports on the preamp at the same time 😉

The bottom preamp is assigned the master preamp and the top five preamps are assigned slave preamps connected in daisy chain fashion using HDMI cables.

These preamps operate in perfect sync with each other and the master preamp controls the volume, balance and muting on all of the other slave preamps in perfect unison.



48 channel VU meter operating in real time !



Muting works across all preamps as well.



A 48 channel Preamp playing vinyl

While I am intrigued by the Ultimate Preamplifier, at least for me, some of the use-cases for it are unclear.
For example - what is the easiest way to create a crossover using the UP? I know DEQX has a simple process that is using a mic and does its magic automatically.
How easy is it to create a correct, good sounding crossover? I am looking to replace my NHT XDA amp with the Ultimate Preamp + multi-channel amp.

Well - nothing like reading the manual 
The manual directs you to download sample 2,3 or 4 way crossover templates from the download section. In other words, it is next to trivial to implement a xover with Audio Weaver + Ultimate Preamp.

Most people will use the available templates from our downloads page https://analog-precision.com/downloads/  and modify them for their own use case.

DEQX is truly a lightweight DSP platform compared to the UP whiuch uses a 4th generation SHARC DSP chip. By comparison DEQX uses a first generation SHARC DSP from Analog Devices and needs two of them just to process at 96kHz ! Their hardware design hasn't changed in decades and is hamstrung by legacy code. See here for a direct comparison with the UP https://analog-precision.com/home/pcg/

regards
David

In this test we increased the analyzer bandwidth to 80 kHz and increased the sampling rate from 48kHz to 192kHz to view the ultrasonic spectrum above the audio bandwidth just to see if there was any surprises. As can be seen from the spectrum below it is fairly clean with some artifacts centered at 55kHz at extremely low level of around -120dB which could be some slight peaking in the IV circuit so nothing to be worried about because it is at such low level and outside the audible bandwidth. Running a simulator on the implimentation of the IV converter suggest no peaking at all so something else at play because the actual peaking is dependent on audio level. Further thoughts are that the peaking may be some beating with the on-board switching regulator but at 120dB down we have done a pretty good job at suppressing it. The broad peaking indicates some spread spectrum which is typical of a switching regulator that continuously regulates itself by varying it's pulse width.