Topic: TASCAM DR-680 portable 8-track #6

[florian.ardelean]

I did some testing tonight with phantom power noise and difference between hi-lo gain.

Test setup was this:
mic -> Radial ProMS2 "direct 1" -> DR680 channel 1 +P48.
Radial ProMS2 "direct 2" -> DR680 channel 4.
Channel 1+4 = left and right of stereo mix. Direct 2 on the ProMS2 is transformer isolated from the direct 1 that carries the phantom power.

The idea was that if there is a difference between a channel with and without phantom power in terms of noise this test ought to show it. If I am making a mistake in my reasoning, feel free to let me know.

If I understand correctly, the small capacitors in the stock unit lead to a raise in the self-noise level of the microphone. So your method would not work, given the fact that the problem is not a ground loop hum, but one of mic self-noise - the signal from the mic coming out of your splitter should be identical.

A proper test would involve an external phantom power adapter, and using only one input channel of the DR680 with a condenser mic with P48 turned on VS P48 OFF with the mic powered by the external phantom adapter box (such as the millenium PP2B)

Users on a french forum tested the DR680 when it came out and found that the noise floor is slightly different on each channel on both mic and line.

The HI-GAIN- LOW-GAIN noise difference is not a problem, it's a feature actually - works as a PAD - helps you NOT to overload the inputs with high-level SPL. I'm pretty sure it's the same in the MKii - so there was nothing to fix there!

[lukpac]

The HI-GAIN- LOW-GAIN noise difference is not a problem, it's a feature actually - works as a PAD - helps you NOT to overload the inputs with high-level SPL. I'm pretty sure it's the same in the MKii - so there was nothing to fix there!

But the gain control is in the analog domain, isn't it? Why would a pad even be necessary most of the time?

And wouldn't it make more sense if it was actually called a pad, like on mics and other interfaces?

[Gutbucket]

Input gain is via a digitally controlled analog PGA, AFAIK.  I'm pretty sure the switch changes the gross range of gain-adjustment of the PGA, and is a function built into it rather than being implemented prior to it.  If implemented in it, it's shifting the gain range up and down and is not a pad, which would instead be attenuating the signal prior to it reaching the PGA.

[lukpac]

Input gain is via a digitally controlled analog PGA, AFAIK.  I'm pretty sure the switch changes the gross range of gain-adjustment of the PGA, and is a function built into it rather than being implemented prior to it.  If implemented in it, it's shifting the gain range up and down and is not a pad, which would instead be attenuating the signal prior to it reaching the PGA.

That would make more sense, but at the same time, if that was the case, why would the noise level differ at the same overall gain depending on where the switch was set?

[Gutbucket]

It may do it by changing the gain of a separate gain stage within the PGA but before the more granular trim, or in a completely separate stage prior to the PGA.  For best noise performance, most of the gain should be made in the first stage. The gain of the second amplifies the noise of the first. 

Wait, I'm forgetting about the HDA discrete stuff, that would be the first stage.  It probably has two gain levels, line/mic switched, and after that, the fine trim adjustment is made via the digitally controlled PGA.  So if the PGA is having to make more of the gain (input LOW, trim cranked) input noise increases.  With input HIGH, the quieter HDA stage is doing a greater share of the work, so input noise is lower. 

Improved first-stage HDA performance over the first gen machine would translate as better noise performance either way, but you still want to keep the gain of the first stage HIGH and the second stage lower, unless you'd clip the first stage, in which you'd set it to LOW, in which case the signal will be hot enough that input noise isn't an issue.

Start HIGH, then go LOW if necessary to avoid overloading the input.

[connloyalist]

A proper test would involve an external phantom power adapter, and using only one input channel of the DR680 with a condenser mic with P48 turned on VS P48 OFF with the mic powered by the external phantom adapter box (such as the millenium PP2B)

In other words, have the mic go into a DR680 channel with phantom turned on, but that phantom power isn't actually powering the mic. That would require the P48 channel on the DR680 to be attached to the transformer isolated side of the mic splitter to avoid the risk of the mic receiving double P48.

The non-P48 channel of the DR680 can be hooked to the non-isolated side of the mic splitter.

Or am I understanding you incorrectly?

Regards, Christine

[lukpac]

A proper test would involve an external phantom power adapter, and using only one input channel of the DR680 with a condenser mic with P48 turned on VS P48 OFF with the mic powered by the external phantom adapter box (such as the millenium PP2B)

In other words, have the mic go into a DR680 channel with phantom turned on, but that phantom power isn't actually powering the mic. That would require the P48 channel on the DR680 to be attached to the transformer isolated side of the mic splitter to avoid the risk of the mic receiving double P48.

The non-P48 channel of the DR680 can be hooked to the non-isolated side of the mic splitter.

Or am I understanding you incorrectly?

Regards, Christine

If I'm understanding correctly, you wouldn't be able to use a splitter, you'd either have to do two separate tests with the same mic, or use two mics of the same model, one powered by an external source and the other powered by phantom from the 680.

[EarlyMorningRain]

If I'm understanding correctly, you'd either have to do two separate tests with the same mic, or use two mics of the same model, one powered by an external source and the other powered by phantom from the 680.   

This is what I am envisioning one would want to do/acccomplish

[florian.ardelean]

In other words, have the mic go into a DR680 channel with phantom turned on, but that phantom power isn't actually powering the mic. That would require the P48 channel on the DR680 to be attached to the transformer isolated side of the mic splitter to avoid the risk of the mic receiving double P48.

The non-P48 channel of the DR680 can be hooked to the non-isolated side of the mic splitter.

Sorry if I was a bit vague last night.

What I meant was, I don't think the mic splitter would be useful in this case. OFC it's a great tool to compare mic pres, but in this instance, a simpler approach is needed, because we need to test the influence of the phantom-48 on the mic's own noise floor. So we're not actually testing for the differences in mic pres, but the quality of the electrical current provided to the mic.

Hence, using a mic splitter would yield almost identical results on different channels of the DR680, since the signal it's splitting is already "polluted" by the +48V the DR680 is feeding it.

So a proper test would include just a condenser mic, the DR680, and an external phantom power supply.
You would make two consecutive recordings:

First, mic -> DR680 P48 ON
Second, mic-> phantom power box -> DR680 P48 OFF.

There is no case I can think of, even including the mic splitter, which could accomodate a single recording of two different channels which would show the difference in this case, but then again I'm not familiar with mic splitters.

As far as I know there are many sources which influence the noise in a recording:

- mic self-noise (except dynamic mics, but it's a different topic),
- the p48 supply,
- the mic pre EIN,
- the a/d converter "resolution" or EIN
- electromagnetic interference (this is why expensve recorders have all-metal chassis - it creates a faraday cage
- the quality of the power feeding the recorder (esp. in case of AC power - the AC outlet never gives a perfect sine wave, it may be distorted, there may be many harmonics and parasitic noise, some people even hear radio...)



Here are some links:

http://www.beat-net.blogspot.tw/2014/08/dr-680-phantom-power-noise.html

http://outrecording.com/tascam-dr-680-noise-test/


This is only ever a problem with low-SPL recordings. It would likely never influence anything going over 70 dBFS (normal speech), and in NO CASE would it matter for taping amped shows.

It might make a difference in recording quiet acoustic performances in quiet places though..

[lukpac]

http://www.beat-net.blogspot.tw/2014/08/dr-680-phantom-power-noise.html

http://outrecording.com/tascam-dr-680-noise-test/

Now...both of those links seem to suggest the issue does *not* have to do with mic self-noise. For instance this quote from the second link:

"Testing phantom power noise was as simple as turning up the gain and flicking it on and off. Of course with no mic plugged in. You can hear a stark difference when turned on and off at a HIGH GAIN."

So now I'm confused.

[florian.ardelean]

Now...both of those links seem to suggest the issue does *not* have to do with mic self-noise. For instance this quote from the second link:

"Testing phantom power noise was as simple as turning up the gain and flicking it on and off. Of course with no mic plugged in. You can hear a stark difference when turned on and off at a HIGH GAIN."

So now I'm confused.

Technically speaking, it is distinct from the mic self-noise, but it influences it. And there's no way to tell it apart unless you have an external +48V supply. I was using the term as an example to why using a splitter wouldn't work in this particular case (testing mic ->splitter -> dr680 two channels -one with +48 on, the other with +48off)

However testing the noise level of an input with nothing attached to it doesn't work - the noise will be very high as it's not terminated, not balanced and picks up em interference. The author of the second link does say that his test "is hardly scientific or conclusive" though

Turning phantom power on with nothing attached to the input should produce some more noise on any machine as long as there is no termination. An impedance termination is needed, and for +48V, a resistance as well. The noise, even if you hear it with no mic attached, should balance itself out when the mic is connected (all the blah-blah of the balanced connection, 90 degree delayed phase of the signal on one of the wires, noise being the same on both wires thus being phased out etc.)

This is what the author of the first link did: "On one leg I (nr. of the XLR input) connected a resistance to ground to simulate mike consumption. On the other leg I connected a capacitor to ground. This forms a RC filter to keep out high frequency signals."

P48 aside, some people recommend using a paper clip to short pins 2 and 3 of the XLR input to test the EIN of the mic pre:
http://www.daking.com/?cat=15

PS. I'm not an an engineer, I learned this all by myself, so if I'm wrong I'd love to be corrected

[connloyalist]

OK, new test.

This is the setup:
KSM141 -> Radial ProMS2
Radial ProMS2: Direct 1 output --> RME Quadmic II
Radial ProMS2: Isolated output --> DR680 (gain: L +15)

In other words: The KSM141 is connected to the input of the mic splitter. The output that does pass P48 is connected to the RME Quadmic II (which I assume to be the most reliable phantom power I own). The isolated out of the mic splitter (which does not pass P48) goes into the DR680. Phantom power on the DR680 is turned off. So the Quadmic is providing the phantom power, the DR680 gets the signal. Theoretically totally separate from any P48.

I took the mono channel of this test and the mono channel of the same test except the DR680 providing phantom power and put them through Adobe Audition CS6. In the attached mp3 the left hand channel is DR680 with phantom power, the right hand channel is DR680 without phantom power (phantom provided by RME Quadmic II). The attached jpg shows the frequency analysis; green is left (DR680 +P48), blue is right (DR680 -P48).

The recording is of my reasonably quiet living room.

The only conclusion I can draw from this specific test is that I cannot prove that anything is wrong with the DR680 (Mk I)'s phantom power. Which doesn't mean there isn't a problem, just that I haven't yet been able to replicate such a result.

Suggestions welcome.

Regards, Christine

[2manyrocks]

You are already well past my level of understanding, but there was a post on GS that I think may be relevant to your question.

https://www.gearslutz.com/board/9675848-post29.html

[florian.ardelean]

This is the setup:
KSM141 -> Radial ProMS2
Radial ProMS2: Direct 1 output --> RME Quadmic II
Radial ProMS2: Isolated output --> DR680 (gain: L +15)

In other words: The KSM141 is connected to the input of the mic splitter. The output that does pass P48 is connected to the RME Quadmic II (which I assume to be the most reliable phantom power I own). The isolated out of the mic splitter (which does not pass P48) goes into the DR680. Phantom power on the DR680 is turned off. So the Quadmic is providing the phantom power, the DR680 gets the signal. Theoretically totally separate from any P48.

Based on the mic's and rated Self-noise, your mics have the following self-noise floor in dBu.

KSM141 Self noise -114 dBA
RODE NT5 self noise -115 dBA. Modified capsule says improved SNR but no specs available...
RODE NT4 self noise -116 dBA
AT4081 -108 dBa
Advanced Audio CM1084 - can't tell (incmplete specs- sensitivity not specified)

TASCAM DR680's mic pre EIN is -128 dB A-weighted, which is a good 12 dB lower than the most silent mic of the bunch, the NT4. So in theory it shouldn't influence any of the mics.

Looking at the noisefloor, the Shure rates even better than specced at -115.5 dB.

It seems like your unit is free of this problem!

Did you buy it new? I wonder what year it was made in, maybe they fixed this silently. It would have been a really easy factory fix - just replace 3 capacitors on the production line

I will soon recieve a DR680 MKi and do the test with an NTG3 as well as an NT2A, with and without external phantom box, can't wait to see the results.


By the way, Christine, didn't you post a comment back in August on outrecording.com's article on dr680 phantom power noise article? Looks like your signature

[connloyalist]

Well, I bought my DR680 (mk I) around March 2014 on eBay from a shop in the UK. I replaced the UK cord for the power supply with one with a European plug (I am in the Netherlands). My impression was that I received a new unit, although I suppose it might have been modified. As far as I can tell from the outside it is entirely standard. Serial number is 0350xxx. But nice to hear that mine doesn't appear to have that problem

Yes, I seem to remember something about posting a comment on the DR680 phantom power noise article.

By the way, my impression is that the senstivity of the CM1084 in "Bright" mode is not very different from the KSM141 or the NT5. It's in the same ballpark at least. In "Vintage" mode it is needs 5 dB more gain. The MJE-384K capsules to the NT5 take about 1dB more gain (that is, if I run the standard NT5's at L+6 then the NT5's with MJE-384K capsules need L+7).

Regards, Christine