A few days ago I recieved my Tascam UH-7000 unit for measuring audio signals from DAC’s and recording my daughters voice through microphone and mixing them into small songs. This unit has proven to have a very good and deep noise floor. That is exactly what I wanted, as I want to be able to measure good DAC’s and audible differences with just changing cables. My project will expand to several years of testing, so I needed something I could play around with. Images from this unit told me its something I can play with!
Upon ordering the Tascam unit, I knew there was an issue with electrostatic noise above 16 bit 44/48 khz recordings. I read other blogs, and I found that another blogger detached the different cables from the cabinet of the unit, and especially the jack plug solved the issue. Before ordering the unit, I thought that disconnecting the jack plug is something I can live with. So I ordered on ebay, and a few days later it arrived at and time to start testing the unit.
The first thing I wanted to test was the noise floor and find a good measurement area where I keep settings at the same. I wanted to keep the pre-volume as low as possible but the normal volume control at 0db. My first spectrogram was therefore pre-volume AND the main volume set to as low desibel as possible, so the noise floor went too deep. The main reason for this, is that at this point I did not measure anything from any DAC’s and I just had the electrostatic noise as a reference.
The first spectrogram of the noise floor, with both volume controls set to one step above mute:
This was recorded at 24 bit 192khz. The top spectrogram is left channel, the bottom is right channel. The red line is the average top, the blue line is the current sample of the spectrogram. What is interesting with this is the fact that the average top noise above 21khz is not higher than the actual line. That means the noise is constant and not moving. That points out ground issue with the unit.
I live in in North Europe and we use 230 volt 50hz. The measurement I saw with another blogger was the US 110 volt 60 hz power. Before I connected the unit to my circuits, I checked that the power adapter inside the cabinet can deal with 230v/60hz. It can, but it is certified for US only. Since I planned to void the warranty anyway, I didnt really care much about what it was certified for. Inside the unit its identical as the ones shipped to Europe.
The first thing I did after I thought about ground issue, was to look at my wall socket. Is it grounded ? No its not. So I started looking for grounded sockets in my apartment. I found them in my kitchen and in my bathroom. The unit got moved to kitchen and measured the same spectrogram at 2 different wall sockets. I also know that some wall sockets are NOT grounded even if they have the socket type. So I moved to Bathroom. That socket HAVE to be grounded, due to humified air from bathroom. I got the same spectrogram.
So after my Tascam unit and my laptop had a nice tour of my apartment, even to the toilet, its time to open this unit up:
4 screws on the silver plate on each side and 2 screws on the side of the unit behind the silver plates is all that takes to open the unit up. I took a look at the screws and there is NO lock-tight on the screws, and I did not hear any click sound un-screwing the screws. So if I were to return the unit, they can not claim the lock tight has been voided. Good news.
I immediately found the Jackplug wire towards the main PCB, and disconnected it. I noticed the wire is 1 mm next to the power plug, so this might be the reason for some of the noise. I put the lid back on and measured again. The noise at 21khz, the first round circle went away. Nice I thought. what about the rest ? So I started to look around the circuits near the power plug and Line 1 volume knob is sitting next to it. Since I use Line1 for one of the channels, I disconnected the signal cable for the volume know for Line1 and Line2. They are basically next to each other on the PCB.
Here is a picture of the 3 plugs disconnected, Line1/2 the 2 circles on top, and jack-plug at the bottom.
Then I put the lid back on, and started to measure again.
Before I show you the measurement, I have to be honest and say, that the recording of this specific measurement were done with the pre- volume in Tascam software set to 0 db, and the main volume within the software set to 0db. That means that the ground floor can be compared to eg. a DAC’s ground floor with DAC’s digital volume set to full, as it should be set to when measuring. Also I did test with my Hegel HD25 dac, and my Oppo HA-1 dac before this recording were taken. I wanted to record once I reached a conclusion!
As you can see, there is a really nice ground floor without any wires to Line 1 and 2. There is a small peak at 70khz, but it is free-running as it is not near the red line, hence not the ground issue I saw earlier.
I am not worried about a small peak at 70khz anyway as most of my measurement will be done with 16bit 44/48 khz recordings and 24bit 96 khz recordings. Only at 24 bit I can go this high in hertz, and its NOT audible what so ever. I would be more worried in that case of the free-running of the bass at 10-100 hz. But this is just ground floor noice. There is NO tone sent to the unit, and no cables attached to the XLR3 inputs on the backside of the unit.
My conclusion is that the ground floor noise with 0 decibel pre-volume and 0 decibel for main volume inside the software with all EQ effects turned off, this ground floor noice spectrogram is good enough to let me start measure audio effects such as jitter tests, different DAC’s and also why I can hear audible difference in the soundstage when my Oppo HA-1 is just turned on ?
A great deal can be made purchasing this product on Amazon:
At the time of writing this, I already did my first tests of Oppo HA-1 and I have seen measurements both at ground floor and J-test that indicate that the unit HAS to warm up half hour before using it, but more about that in next article!
The next article will be about my first measurement of the Oppo HA-1.