Jump to content

MDS-JB980 aout clips at -6dB or higher at default 0dB aout setting

Rate this topic


Perus

Recommended Posts

Hi!

Just noticed while listening to music on the JB980 via aout and the phones jack that with the default output (setup/aout) level setting at 0dB the outputs clips for signals higher than -6dB
(Also noticed this while injecting a -0.1dB 315Hz via the optical input and verified with a scope on the aout)

Is that really normal? I guess since its a consumer device and recordings reaching 0dB levels with the aout set at 0dB should result in clipped output signal - why have aout at 0dB chosen as default then?
My tape decks certainly doesn't behave this way...

Anyone else noticed this?

Going to compare with the JE470 later but since it doesn't have the setup/aout level settings I hope its going to be fine.

/Per

Link to comment
Share on other sites

I'd agree that does sound weird and I don't really have any other ideas at the moment. Clipping in the analogue world (on a well designed circuit) is usually due to inadequate power supply headroom (power supply rails too low). I'll need to take a look at the schematics.

Do you think this is a new phenomenon or that you've only just noticed it?

Does the 980 support SF (Scale Factor)? Probably not relevant as I read the above that you used "monitor" mode in your testing and I don't think SF would affect this.

Link to comment
Share on other sites

This is the analogue I/O section if you wanted to check the +/-6V rails from IC141 and IC241. Unlikely to be wrong I know, but as you appear adept enough with you test-tone stuff I'm thinking you're up to this check! This looks like fixed gain buffer stages so any attenuation control will be in the digital domain prior to the DAC.

(As an aside, for anyone who is interested in electronics, there are some nice simple op-amp configurations here that one could analyse and research - unit gain non-inverting buffers/followers (around IC101), differential to single ended buffers with stability compensation (around IC111) and active low-pass filter around IC112.)

 

image.thumb.png.a4d4e70248404571e3feb1c3b8202adb.png

Link to comment
Share on other sites

Another aside - I notice some schematics note to use slightly different capacitor values in the circuits like that above, noted "UK" and "Except UK". Such capacitor changes would slightly alter the analogue sound as the filter characteristics are slightly modified. I wonder if this aligns with units with the "UK Tuned" badges?

(Note not the case in the 980 schematic above on C143/C243, they are just reservoir/decoupling caps on the op-amp power supplies. That won't change the analogue sound. No idea why they would spec the increase to a 25V cap rather than 10V on a 6V power supply, other than for long-term reliability.)

I just noticed this looking at the analogue output schematics for a Sony CDP-XE530 CD player. Example below, C807/C907 are 1.5nF in most models but 1.2nF in the UK model:

image.png.7c60f2724a3e96137c2332af5b0b7edd.png image.png.13739d79e80a1ffbf133155323442d07.png

Note: JB930 is claimed to have a "UK Tuned" version but the schematics don't claim any cap value changes for the UK model. Possible the "UK Tuned" variant came out later and the SM schematics don't reflect that. Sometimes there is a SM supplement (like I have for the E10).

Mods - I'd like to delete the following image, but seem unable to...

image.png

Link to comment
Share on other sites

The 980 does support SFE, but you should not have to use it for something like this. By my observation, not normal at all...I have never had anything like that happen on the 930, 940, D400, et al. At least I learned that the 980 has the same ATRAC chip as the D400. Would like to know if you get the same result with the 470.

Link to comment
Share on other sites

Hi!

Ok, the JE470 behaves in the same manner, as soon as the recorded signal rises above -7dB the analog output signal is clipped.

For the JB980 this is fixable either via using the digital output or decreasing the analog output level via setup/aout temporarily since its not possible to save.
This "phenomena" happens both while listening to Minidiscs and monitoring through the input, digital or analog doesn't matter.

I really don't understand why higher (>-7dB) signals recorded on the media is distorted in this way - especially since the manual states that the recommended recording level is one step below "over" which is close to 0dB. I've seen other recommendations for a maxium recording level of -12dB for some decks but not these decks.

On the JB980 this little issue affects the headphone output as well - this is how I first started to suspect something was wrong.

Any more suggestions on this? I'm certainly going to measure the supply rails for the OP-amps later as suggested (thanks!) and also trace where the signal clipping occurs.

Btw, this is the "UK tuned" version of the JB980

/Per

Link to comment
Share on other sites

Could be off the mark here, so feel free to respond!

Analogue output spec is 2Vrms into 50kOhms. This is an ac spec, but the voltage is RMS, so this gives us an "equivalent heating (i.e. power) effect" as dc, so we'll use P= V^2/R, which would give us an output power of (2x2)/50000= 80 microwatts. If we convert this to dBm (decibels referenced 1 milliWatt), this is 10 log(base10) (80uW/1mW) = -11dBm.

So how does the -7dB quoted above relate to the -11dBm output for the analogue outs; does the deck limit the analogue output to -7dB (-7dBm ???) because that is already over-driving the analogue output by 4dBm? Reversing the calculation, the output for -7dBm would be 200 microWatts which is about 3Vrms into 50k. The op-amps seem to have a +/-6V supply so could drive this without clipping in the output driver stage, but this is likely to be overdriving the input stage of whatever the analogue outputs are driving (the amplifier or other connected device).

My analogue electronics is a little fuzzy, so maybe I've made some schoolboy errors here. What does anyone else think?

Link to comment
Share on other sites

Oh that was quick, many years since I jumped between V to dB and dBm that swiftly :)

While investigating this further, I really hope someone else could confirm this as well just in case I've made some kind of mistake here.
I hope to have some time this weekend to provide some oscilloscope traces and spectrum graphs to show the horrible result.

/Per

Link to comment
Share on other sites

Suddenly I had this feeling that what if this is "by design" that a consumer recording shouldn't be as good sounding as pre-recorded media with (possibly) some type of control bit set to change the dynamic range in the playback chain like the "HDCD Peak Extension".

Just a thought...

About SF, the 980 supports it and I don't know when using that if the recording is re-sampled or the range just changed "on the fly" during playback.

/Per 

Link to comment
Share on other sites

I think crippling an own technology and product like that would be more cynical than ever I could achieve!

I’ve got lots of decks and I think they all output at the 0dB setting. Whilst I’m no maestro I think I’ve got quite ‘hi-if’ ears so any horrible sound I’d quite quickly pick up on.

The nature of saturation of any digital signal means that if you saturate the coding you will get an instant and hard clip. Not a soft mushy clip you’d get with a pure analogue device. That’s why with a digital mixer you’d mix to around -18dBFS to give 18dB peaking headroom. With analogue mixing you mix to around 0dB assuming your audio chain had a peak handling of at least +20dB.

In respect of this setting on the MD players, 0dB means ‘unity gain’. Decibel is a relative measure not a physical (measurable) unit. It means nothing on its own, you have to use it with some real world unit. Decibel against a unit, eg dBm or dBV is an absolute unit and means something ‘on its own’.

Consumer line level is considered -10dBV. In your opening statement, is that -6dB actually -6dBV, which if so is possibly over-driving the input.

-10dBV is 0.316V: -10dBV = 20 log(base10) (0.316/1)    [or as I worked out "V". V = 10^(-10/20)*1 = 0.316]

The 980 has a "rated input" of 500mV (into 47k). That translates into -6dBV. However with unity gain on the input stage analogue setting that's probably going to exceed 0dB in the digital domain, which means clipping.

i.e. is your original hypothesis sensible if we're talking the same units?

I'm learning/revising too so I'm not trying to be dismissive with the above, I hope it doesn't come across like that, it's not intended to!

https://www.sweetwater.com/insync/understanding-signal-levels-audio-gear/

Link to comment
Share on other sites

Not sure if any of the below helps?

I recorded a couple of songs from Spotify earlier this week using my MDS-JB980. The recordings are mastered very loud and, with the digital recording level set at default, they show as being pretty much brick walled at 0dB on the meters on the MDS-JB980 and never really fall below -4dB (not a lot of dynamic range...)

Despite this, I cannot hear any audible clipping either through headphones or the analogue outputs.

The MDS-JB980 I did this on is also "UK tuned".

Link to comment
Share on other sites

Ok now I’m officially embarrassed and should have known better than to blame Sony for making flawed equipment ;)

The reason for the clipping in the analog signal chain was due to level limitations in my pc-based measuring solution - the same signal path has been used mostly for tape deck maintenance without any issues but didn’t handle the higher output levels of the minidisc decks very well :(

noticed that while tracing the clipped signal with my oscilloscope later this evening. Sorry to waste anyone’s time and I really hope not to be remembered as the guy who posted some questionable topics as a fresh member ;)

Thanks for your patience,

Per
 

 

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
×
×
  • Create New...