Is my Rotel RC-1590 preamp THIS bad or am I measuring wrong?

[LandscapeJohn]

I have a Rotel RC-1590 preamp and have noticed that when it's not muted there is an audible hiss from my tweeters. This prompted me to hook up my QuantAsylum 402 analyzer and measure the SNR to see what was up. The spec on the Rotel web page lists the 'IHF "A" Weighted' SNR for digital inputs at 108 dB. I haven't been able to measure a SNR above 90 dB using any setting I've tried so far.

Since I was primarily listening using the Bluetooth input my signal chain for testing was iPhone -> Bluetooth -> RC-1590 -> Balanced output -> QA-402. Playing a 1kHz test tone on the iPhone at max volume wouldn't yield more than around 70 dB SNR:

I switched to one of the optical inputs using a signal chain of iPhone -> AirPlay -> AppleTV -> Toslink -> RC-1590 -> Balanced Output -> QA-402. Playing the same 1kHz test tone on the iPhone at max volume yielded a much better SNR of around 87 dB but still nowhere close to the spec of 108 dB provided by Rotel.

The last screenshot shows a measurement with no signal playing and after the amplifier has gone into the standby "mute" feature after 10 or so seconds of no signal. This seems to me to rule out that the noise is being picked up someone after the Rotel in the measurement chain and also coincides with the hissing I hear from the tweeters that goes away once the mute relay clicks.

I've attached the settings file I used with the QA-402 but the highlights are:

RC-1590 Volume Setting: 65 (Around 66 the THD would start climbing (presumably overloading the input of the QA-402?) and much higher would result in the QA-402 trying to engage the input attenuator)
Sampling rate: 192k
Input Attenuator: 0
FFT: 256k
Averaging: 5
Weighting: A
Window: Hann

 

[staticV3]

Bluetooth and Airplay can both introduce their own noise and distortion into the measurement.

Rotel specify 112dB(A) SNR with line in and 108dB(A) with S/PDIF and USB in, so I would try that.

If you do, make sure you're either using an ultra clean DAC for line in, or you're sending the test tone via USB or S/PDIF bit-perfect without any resampling.

 

[sarumbear]

Bluetooth and Airplay can both introduce their own noise and distortion into the measurement.

How can Airplay, which is a lossless 16-but data transmission protocol can introduce noise?

 

[staticV3]

How can Airplay, which is a lossless 16-but data transmission protocol can introduce noise?

By downsampling to 16bit, Airplay (1) will limit measured SNR to at most 96dB.

If you're trying to confirm a DUT's SNR of 112dB(A), ~110dB(no-wt), you can't do it via Airplay 1. You need to send 24bit audio to the DAC, or not have a DAC in the measurement chain in the first place.

 

[VintageFlanker]

but still nowhere close to the spec of 108 dB provided by Rotel.

108dB SNR can't be achieved with a 16b/44.1kHz 1kHz sine wave. Plus, your QA-402 adds noise by its own. You're measuring Airplay and Bluetooth integrations, not the RC-1590 overall D/A performance.

As @staticV3 said, try 24bits signal through S/PDIF or USB.

 

[sarumbear]

By downsampling to 16bit, Airplay (1) will limit measured SNR to at most 96dB.

That is the limit of 16-bit, which is not an AirPlay limit. Besides, that is still much higher than what the OP is measuring. I do not see the point you are making.

 

[staticV3]

That is the limit of 16-bit, which is not an AirPlay limit.

Airplay is limited to 16bit:


If you don't believe me, maybe you'll believe Amir's words

 

[sarumbear]

Airplay is limited to 16bit:
View attachment 248828

Duh! That is what I said!

 

[staticV3]

Duh! That is what I said!

Huh?! You said that 16bit is not an Airplay limit!

That is the limit of 16-bit, which is not an AirPlay limit.

Evidently, 16bit is an Airplay limit.

 

[sarumbear]

Huh?! You said that 16bit is not an Airplay limit!

16-bit is not a limit specifically for Airplay. The 96dB limit you said is a limit for any 16-bit system, as well as for all CD systems and until hi-res arrived for every streaming service.

Evidently, 16bit is an Airplay limit.

Nope. 16-bit is a limitation of Airplay. There is a big difference between the declarations.

Anyway, I understood what you said.

 

[staticV3]

....

however you want to put it, neither AAC nor Airplay are valid interfaces for measuring the Rotel's SNR.
No matter what OP measures using those, the numbers will (justifiably) be put into question, unless the aim is to measure the SNR of Airplay+Rotel, which in this case, it isn't.

 

[Ageve]

SNR (A-wtd): 90 dB
Frequency response: 20 Hz - 100 kHz +/- 0.05 dB
THD: 0.0002 - 0.0005 % (20 Hz - 20 kHz, 0dBV)

It looks like Hifi News messed the numbers up though. It's impossible for analog SNR to be 90dB, if it's better than the digital (DAC), since they write:

The digital ... THD is slightly higher than the analogue in/out at 0.0011 - 0.0035%, the A-wtd S/N only 'average' at 103.5 dB and jitter just a little untidy at 380 psec.

Anyway, the noise level is low (should not be audible), and the distortion is very low.

https://www.bwgroup.ch/wp-content/uploads/2016/11/Hi-Fi-News-11-2015-Test-Rotel-RC-1590-RB-1590.pdf

 

[LandscapeJohn]

A lot of great information here, thanks to everyone that's replied. The short answer is that I evidently WAS measuring incorrectly but it took me a while to figure it out. From my first post I noted:

RC-1590 Volume Setting: 65 (Around 66 the THD would start climbing (presumably overloading the input of the QA-402?) and much higher would result in the QA-402 trying to engage the input attenuator)

It turns out that engaging the attenuator and to a lesser extent allowing distortion to increase slightly yielded results much closer to what would be expected based on the Rotel's specification.

To confirm that I was on the right track and set a baseline I hooked the QA-402 up using RCA cables to the CD input and took a measurement. I messed with various settings to get the highest possible SNR. I lowered the sampling rate to 48kHz so any higher harmonics wouldn't be contributing to the measurement then increased the QA-402 output by 1 dB at a time until distortion started to rise and then backed it down by 1 dB for an overall input level of 11.00 dBV. Note that this corresponds to approximately 3.5 volts while Rotel specifies the input sensitivity as 150mv with an overload of 4 volts. From there I started increasing the volume of the RC-1590 and setting the QA-402 attenuator higher to avoid overloading. I kept doing this until I got the highest SNR without dramatically increasing distortion.

This yielded the following measurement of 104 dBA SNR with less than 0.003% THD:

Continuing to increase the volume resulted in THD climbing to just under 0.3% but allowed the SNR to come up to almost 107 dBA.

With the baseline analog measurement completed and showing promise I switched back to the Airplay via Toslink setup which yielded over 95 dBA SNR. I feel like this is probably close enough to the theoretical 16-bit limit that you all have been talking about so I feel like there's a good chance that the preamp is fine in this regard.

If you remember, the thing that got me going down this rabbit hole in the first place was the hissing I hear from my tweeters when music isn't playing. At this point I believe that is probably down to the gain setting on the power amps. I get inconsistent channel matching when setting the gains to anything less than max but that means that most of the time I'm nowhere near a high enough volume on the preamp to get the higher SNRs that I've seen with the volume cranked up.

Using the optical input with no signal and the volume set to 1 shows that there's around 32 uV of noise that is subsequently getting amplified to audible levels by the power amps:

I'm thinking of permanently settings a lower gain on the power amps by replacing the troublesome gain pots with precision resistors so that I will have a stable setting somewhere less than at max gain and then using the upper ranges of the RC-1590.

Thanks again for everyone that helped me get to the bottom of what was going on!

 

[The Guy who cares]

I have a Rotel RC-1590 preamp and have noticed that when it's not muted there is an audible hiss from my tweeters. This prompted me to hook up my QuantAsylum 402 analyzer and measure the SNR to see what was up. The spec on the Rotel web page lists the 'IHF "A" Weighted' SNR for digital inputs at 108 dB. I haven't been able to measure a SNR above 90 dB using any setting I've tried so far.

Since I was primarily listening using the Bluetooth input my signal chain for testing was iPhone -> Bluetooth -> RC-1590 -> Balanced output -> QA-402. Playing a 1kHz test tone on the iPhone at max volume wouldn't yield more than around 70 dB SNR:

View attachment 248673

I switched to one of the optical inputs using a signal chain of iPhone -> AirPlay -> AppleTV -> Toslink -> RC-1590 -> Balanced Output -> QA-402. Playing the same 1kHz test tone on the iPhone at max volume yielded a much better SNR of around 87 dB but still nowhere close to the spec of 108 dB provided by Rotel.

View attachment 248674

The last screenshot shows a measurement with no signal playing and after the amplifier has gone into the standby "mute" feature after 10 or so seconds of no signal. This seems to me to rule out that the noise is being picked up someone after the Rotel in the measurement chain and also coincides with the hissing I hear from the tweeters that goes away once the mute relay clicks.

View attachment 248675

I've attached the settings file I used with the QA-402 but the highlights are:

RC-1590 Volume Setting: 65 (Around 66 the THD would start climbing (presumably overloading the input of the QA-402?) and much higher would result in the QA-402 trying to engage the input attenuator)
Sampling rate: 192k
Input Attenuator: 0
FFT: 256k
Averaging: 5
Weighting: A
Window: Hann

I have two RC-1590MKII's and I have no such issues. Your problem could be your patch cords. If RCA Patch cords are too long and near other electronic cables it can cause problems. Try moving your cables away from other cables. If this doesn't work use RCA Coaxial cables instead

 

[restorer-john]

I'm thinking of permanently settings a lower gain on the power amps by replacing the troublesome gain pots with precision resistors so that I will have a stable setting somewhere less than at max gain and then using the upper ranges of the RC-1590.

What you need to do is measure the residual noise of the preamplfier with shorted inputs at various volume setting on the preamp. It is unlikely to be a constant 32uV across the range to maximum volume. Plenty of preamps go from noisy, to less noisy, to dead quiet at full volume. The worst-case volume position is quite likely to be in the middle of the range you are currently using- hence the audible hiss.

Balancing your power amp gain pots and preamp volume position along with source output level is vital if you want the lowest residual noise at your typical listeing levels.

Also, what is the sensitivity of the inputs you are using compared to the output level of the connected source?

Edit: I see it is, as I expected, 150mV. For S/N measurements, you short the input, measure the residual in V (uV), reference that to the full measured output voltage when driven with the specified 150mV. 1khz is not sufficient nor is it representative. Rotel specs are always pretty much spot on or a little conservative.