Strange noises coming from a Tascam Interface

[rob aylestone]

When I upgrade, I sort of move the old gear about. In the office, working happily is my old Tascam US-1641 interface. I was doing some recording yesterday, and in Cubase, looking at the waveforms something looked odd. I was trying to record with the mics a long way away - not music, I'm testing for some sound effects recording, so the waveform levels on the screen were very low. I knew I'd have to normalise, and wanted to know if the noise would be acceptable. I had on a pair of Beyer DT100 headphones to try to keep the room noise out, and with an omni mic, the computer fans were there, and switching to cardioid, wanted sounds was OK and the fans went down quite a bit, as expected. While I was recording I noticed the occasional waveform appear when I heard nothing? I normalised it and there was a clearly obvious sound source. I could not hear this thump. I used Cubase's spectrum feature which draws a graph, and there was a peak - a very sharp one - just below 10K. Bunging this into Audition, as I like the spectrum display there shows a continual feint line at just below 10K and quite a bit right at the very bottom. A direct test into Audition reveals that this signal at 10K occasionally gets louder. Not by much, and with the room noise and stuff, It's not a 'tone', it's hidden by the background noise, but it is there. It is not the mic - unplugging makes no difference and it appears generated on the tascam output on all channels.

Anybody ever heard/seen this before. A phantom tone generated by the interface. Look at this image - you can see the silent parts of the recording, but that tone is there all the time?

 

[TalismanRich]

I've never seen that in anything except maybe a 60 cycle hum. A while back I did a direct comparison of my Tascam 16x08, my AW1600 and my R24. I had some residual noise from the mics picking up frequencies around 2-300 and a few further up. By 5K, everything was clean. The thing that intrigued me was that the Tascam noise floor from about 5K to 20K was at-90dB, the AW1600 was about -75dB and the R-24 was about -80. Somewhere I have the graphs, snapped with BlueCat's analyzer. If I find them, I'll stick it up here.

 

[rob aylestone]

I tried drivers, and even found a firmware update, but whatever I record has that strange 10K line. It's quiet enough that I don't hear it, but every now and then, it fades in and out to a louder level. Really strange.

 

[TalismanRich]

It looks like it's about 96 or 9700Hz. I can't think of anything that is a multiple of that frequency that might be generating a sub-harmonic or an intermodulation product around that frequency. Definitely weird.

You mentioned a "thump" as well that you couldn't hear. Any idea what that is from?

 

[Gtoboy]

I tried drivers, and even found a firmware update, but whatever I record has that strange 10K line. It's quiet enough that I don't hear it, but every now and then, it fades in and out to a louder level. Really strange.

Filter cap ghosty?

 

[rob aylestone]

Hmmm not thought along those lines, but filtering makes sense - but why that frequency? Maths wise, 44.1 or 48K don't seem linked?

 

[ecc83]

I am reminded of the ~6kHz "whistle" we got with early 16 bit converters used in USB mixers (and don't just blame Behrrys for once! My Mk1 ZED10 does it!) But that tone is related to Windows sounds, cut the input gain to five %, up the input drive and the tone dives into the ~-85dBFS noise floor.

Two other sources come to mind? The DC-DC converters used to derive +&- supply rails and the 48V phantom power. Switching spook juice off might not in fact fix things as the converter might still be running and just its output killed? Removing the 48V load might change the tone's frequency mind?

Thing is though, when you read of the fantastically wide dynamic ranges of the really top grade interfaces, north of 120dB, you have to realize that those figures are achieved only by THE most scrupulous attention to keeping audio and digital ground paths from interacting. Great value though the Tascams are they are not Prism or Grace products!

Can I guess Rob that the tone/noise is at better than -80dB FS? If so I guess it is in specc' for the weighted noise figure?

Tascam US1641

Put simply, if you want to send lots of audio inputs into your computer simultaneously, but don't want to pay too much for the privilege, the US1641 could be exactly what you're looking for...

www.soundonsound.com

Aha! ^ See the box re "artifacts" with Nvidia USB 2.0 controllers. Also, phantom power is not switched globally but in groups of 4 so I doubt the DC-DC converter can be switched off.

Dave.

 

[rob aylestone]

I'm trying to work out when I swapped it out of my audio studio to the video studio - I'll have to download something old, created on it and see if that line is evident.

 

[jamesperrett]

I've got a fluorescent magnifying lamp that puts out lots of noise around 25kHz. Most switch mode power supplies work at 100kHz upwards although there are a few that work much lower. I also have an LCD screen that sometimes puts out little chirps as the picture changes.

 

[ecc83]

I've got a fluorescent magnifying lamp that puts out lots of noise around 25kHz. Most switch mode power supplies work at 100kHz upwards although there are a few that work much lower. I also have an LCD screen that sometimes puts out little chirps as the picture changes.

Could be due to a 'heterodyne' problem? When the Blackstar HT pedals had to change from a 16V AC supply to a 22V DC one they had a problem, how to derive a -15V supply? I suggested to Bruce, chief designer that they used a small dc-dc converter. "Might beat with the 300V boost circuit" he told me. That 22-300V supply used the UC3843A chip and a MOSFET and definitely ran at higher than 100kHz so another converter running anywhere near that frequency could have been trouble.
I shan't tell how he solved the problem because it was pretty clever and I don't want to make Behringer's work any easier!

Dave.