Tascam UH-7000 power supply noise

[mansr]

The Tascam UH-7000 is in many ways a respectable device. There is, however, a minor problem with power supply noise being picked up by the ADC.

 

Recording a bit of silence at 192 kHz, and this is spectrum obtained:

Ugh! That's not what I paid for! The noise being mostly confined to the left channel isn't much consolation.

 

This is where we void the warranty.

 

Opening the case is easy, a simple matter of removing the silver-coloured side plates, removing two more screws, and finally sliding the top cover off. The power supply is a separate PCB connected to the main board with a short, detachable wire, making it easy to remove.

 

At this point, rummaging through the crate of spare power supplies should turn up a few suitable replacement candidates. The original outputs 11.2 V, so 12 V ought to be fine.

 

First up, a small SMPS model FSP036-RAB, for some reason with labelled "shannon" with a yellow sticker:

At least it's different. Still not pretty though.

 

Moving on, another SMPS from the same manufacturer, model FSP040-DGAA1, yields this spectrum:

Well, that explains why it has the words "Sparkle Power" on the spec label.

 

A few variable voltage units, one of them a wall-wart, from Maplin are worth trying:

Much better. Just a lone spike in each case, and all of them above 50 kHz, outside the range of a 96 kHz recording. Still mildly annoying.

 

Maybe a Seasonic SSA-0601S-1 is what we're looking for:

Now we're talking. Although there is a very small spike at 51 kHz, this is really nothing to be concerned about. Looks like we have a winner.

 

Finally, while both impractical and a little dangerous without a proper enclosure, connecting the original power supply from outside the main case proves interesting:

Apparently the output of this power supply is quite clean, and the original problem was caused by radiated noise.

 

All the recordings were done with the pre-amp gain set to minimum.

[sandyk]

A nice piece of detective work !

[ginetto61]

The Tascam UH-7000 is in many ways a respectable device. There is, however, a minor problem with power supply noise being picked up by the ADC....

 

 

Hi ! where i can vote for the best thread of the year ?

Ok i am biased because i am a very recent owner of this nice but with a little defect unit from Tascam

But i have to say sincerely that this is indeed a very valuable thread.

I am everything but an expert but the outcome is spectacular.

The idea would be to extract the power supply and place a dc socket in place of the mains receptacle.

Maybe with some capacitors inside the Tascam to buffer a little ?

Thanks a lot again ! very kind of you.

Kindest regards, gino

 

P.S. however this little black thing is very very clean ... moreover with the right linear psu i think that it would be possible to keep the noise down also after 60kHz ?

[mansr]

P.S. however this little black thing is very very clean ... moreover with the right linear psu i think that it would be possible to keep the noise down also after 60kHz ?

 

The noise above 60 kHz is from the sigma-delta modulator. There is no way to get rid of it.

[ginetto61]

The noise above 60 kHz is from the sigma-delta modulator. There is no way to get rid of it.

 

Hi ! thanks a lot again. My ignorance shows always.

However is also interesting to see that some not expensive smps have very low noise.

SMPS are very very handy. Reliable, cheap ... but usually it is very difficult to understand if one model is a noisy one or what.

I would really like to mod this unit placing a nice dc socket in the back panel (i like the one on the right much better but i cannot find it around).

 

 

and then buy the one with which you measured the lower noise.

I will try to do that if i understand how are the connections.

Thanks a lot again for the very interesting testing, gino

 

P.S. some idea for the new regulator/power supply

 

TPS7A4701 Ultra Low Noise Adjustable DC 20V 32V 1A Power Supply | eBay

 

Finished 65VA Ultra Low Noise Linear Power Supply 5V 9V 12V 15V 18V Etc | eBay

[mansr]

Hi ! thanks a lot again. My ignorance shows always.

However is also interesting to see that some not expensive smps have very low noise.

SMPS are very very handy. Reliable, cheap ... but usually it is very difficult to understand if one model is a noisy one or what.

I would really like to mod this unit placing a nice dc socket in the back panel

and then buy the one with which you measured the lower noise.

I will try to do that if i understand how are the connections.

 

The power connection is simple. The red wire and the adjacent one are positive voltage, the other two ground. Because the original power supply provides 11.2 V, and for some protection, I added a diode in series with the 12 V power supplies. It's probably safe to connect 12 V directly, but I didn't want to risk letting out the smoke.

 

My plan now is to add a barrel jack where the AC input used to be and put the original power supply in a plastic case.

 

Don't put too much into the specific PSU models I mentioned. Those were just what I found in my crate (you don't have one?) with suitable specs.

[ginetto61]

The power connection is simple.

The red wire and the adjacent one are positive voltage, the other two ground.

Because the original power supply provides 11.2 V, and for some protection, I added a diode in series with the 12 V power supplies. It's probably safe to connect 12 V directly, but I didn't want to risk letting out the smoke.

 

Hi again. Thanks a lot for the very helpful advice.

I will measure with the tester. But clearly a 12VDC unit would be easier to find around. It is a very standard value.

The difference with 11.2V looks minimal to me. But i am worried now.

 

My plan now is to add a barrel jack where the AC input used to be and put the original power supply in a plastic case.

 

Please. If you will decide to proceed please kindly take some pictures of the mods. I am a dog at electronics but i like the idea of the external power supply a lot indeed. And your very interesting measurements show that some benefits can be obtained. Thanks a lot !

 

Don't put too much into the specific PSU models I mentioned. Those were just what I found in my crate (you don't have one?) with suitable specs.

 

I have many 12V smps not used at hand but i do not know nothing about their quality (i.e. noise).

One day i would like to be able to carry out some measurements like the wonderful yours. I love measurements and lab reports.

However the one on ebay i know that is very good indeed from what i have read. Very very good.

It has a very low noise on a very wide Hz range.

It has a regulator made with discrete components (i.e. not single chip) and a very nice R-core mains transformers.

I think that it has also a very handy trimmer to vary a little the Vout.

Maybe a little big and heavy ... but very nice quality.

Please if you decide for the mod the pictures ... i will try to do the same.

Thanks a lot again, gino

[mansr]

Hi again. Thanks a lot for the very helpful advice.

I will measure with the tester. But clearly a 12VDC unit would be easier to find around. It is a very standard value.

The difference with 11.2V looks minimal to me. But i am worried now.

 

A series diode will bring the voltage down to 11.0-11.3 V depending on the exact type. Mine didn't blow up with this configuration, so if you do the same, you shouldn't need to worry.

 

Please. If you will decide to proceed please kindly take some pictures of the mods. I am a dog at electronics but i like the idea of the external power supply a lot indeed. And your very interesting measurements show that some benefits can be obtained. Thanks a lot !

 

I have many 12V smps not used at hand but i do not know nothing about their quality (i.e. noise).

One day i would like to be able to carry out some measurements like the wonderful yours. I love measurements and lab reports.

 

In case it wasn't clear, the graphs are simply FFT plots of silence (nothing connected) recorded with the different power supplies.

[ginetto61]

A series diode will bring the voltage down to 11.0-11.3 V depending on the exact type.

Mine didn't blow up with this configuration, so if you do the same, you shouldn't need to worry.

 

Thanks again. Please excuse me but i have really no background in electronics.

I can only ready a datasheet or make some very easy soldering. But in general i like external power supplies.

There is one more box but you very well showed the possible benefits.

However the idea is to use the original given that is indeed pretty silent. Just a nice box to contain it.

There are many aluminum boxes on ebay and some are really fantastic.

 

In case it wasn't clear, the graphs are simply FFT plots of silence (nothing connected) recorded with the different power supplies

 

one small step for you is one giant leap for ginokind ...

however, how did you get those graphs ? do you have a scope ? they are excellent and they go very deep down with dB.

Thanks again, gino

[mansr]

however, how did you get those graphs ? do you have a scope ? they are excellent and they go very deep down with dB.

 

I didn't measure the power supplies directly, only the effect they have on the UH-7000. To do this, I recorded a few seconds of silence to a WAV file using Audacity, then used Octave (a free Matlab clone) to calculate the FFT and plot the graphs. There are many other programs that will do this kind of analysis too.

[ginetto61]

I didn't measure the power supplies directly, only the effect they have on the UH-7000.

 

Perfect approach. This is what really matters in the end isn't it ?

 

To do this, I recorded a few seconds of silence to a WAV file using Audacity,

 

i know Audacity i have it installed on a pc.

 

then used Octave (a free Matlab clone) to calculate the FFT and plot the graphs. There are many other programs that will do this kind of analysis too

Thanks a lot for the explanation. I will try in the next weekend. I am quite curious.

However all considered for me the critical point will be the connections. Therefore i do not want to stress you too much, but if one day you will decide for the mod please take some pictures of the critical passages.

Thanks a lot again, gino

[ginetto61]

I didn't measure the power supplies directly, only the effect they have on the UH-7000. To do this, I recorded a few seconds of silence to a WAV file using Audacity, then used Octave (a free Matlab clone) to calculate the FFT and plot the graphs. There are many other programs that will do this kind of analysis too.

 

Hi ! i have downloaded both Audacity and Octave.

1) in which format i have to record the silence ? i mean bit/sample rate

2) once that i have the wav how can i get the FFT and graphs with Octave ? is it difficult ?

Thanks a lot again, gino

[ginetto61]

Hi ! these are some pictures of the smps

 

In order to improve the shielding of the smps i was thinking to stick to the existing partition separating smps and circuits board some adhesive copper foil or other suitable material (advice much welcome and appreciated).

The 3M material is expensive. But i heard that it works quite well in adsorbing RFI/EMI (?) ...

Thanks a lot, gino

[ginetto61]

Hi !

just to say that also my unit shows some spikes above the noise floor.

Graph obtained with Arta software, mic input open and gain to zero.

I am trying to do something with ferrite beads on dc cable connection between the smps and the circuits board before going with an external dc power supply like the OP recommends.

The very low noise floor allows for some expectations of very good sound.

Also quality of conversion is said to be something.

 

Thanks to the OP and kind regards, gino

 

[ginetto61]

Hi !

I have cut the power cable between power supply and the circuits main board and used instead a refurbished DELL 12VDC smps to power the board (cost 10 euro).

The bumps in the noise floor have gone away !!!

So i guess the issue is in the original smps in the unit injecting noise in the circuits.

It is not a bad shielding issue (i tried moving cables away from the smps and nothing changed).

One solution could be to place a dc socket on the rear panel and use a decent 12VDC power supply, linear or smps.

I am quite sure that just a little troubleshooting on the original smps could have solved the issue in a much more elegant way.

I do not have a scope but i would like to check if some additional bypassing cap could trim those noise bumps.

I will try something in the next days. But i am shooting blind ... i do not even know a suitable value for this output cap. Maybe 10 uF could work ?

Thanks a lot again, gino

[manpowre]

I wish there was something on the original SMPS that could get modded to get rid of these artifacts.

[ginetto61]

I wish there was something on the original SMPS that could get modded to get rid of these artifacts.

 

Hi ! yes you are right.

But troubleshooting a smps i guess it is a very challenging task in terms of knowledge and equipment.

And there is also high voltage.

Much easier to replace it with something more quiet.

I see that some parts look missing and replaced by jumper ?

But without a schematic i have no idea really

I would like to take the opportunity to thank you for your initial testing.

Very very important in order to define the issue.

Kind regards, gino

[mansr]

Hi ! yes you are right.

But troubleshooting a smps i guess it is a very challenging task in terms of knowledge and equipment.

And there is also high voltage.

Much easier to replace it with something more quiet.

I see that some parts look missing and replaced by jumper ?

But without a schematic i have no idea really

I would like to take the opportunity to thank you for your initial testing.

Very very important in order to define the issue.

Kind regards, gino

 

There are many possible explanations for the "missing" parts. It could be that the same PCB is used for several models, e.g. different voltages. It's also common to design PCBs to allow alternative components to be used in case of supply problems.

[ginetto61]

There are many possible explanations for the "missing" parts.

It could be that the same PCB is used for several models, e.g. different voltages.

It's also common to design PCBs to allow alternative components to be used in case of supply problems

 

Yes you are right. Still they could have done better with very little effort.

However you said at the beginning

 

Finally, while both impractical and a little dangerous without a proper enclosure, connecting the original power supply from outside the main case proves interesting:

[ATTACH=CONFIG]26006[/ATTACH]

Apparently the output of this power supply is quite clean, and the original problem was caused by radiated noise

 

So you have tested the stock one outside the case and did not find any noise ?

Could it be a bad grounding issue ?

[mansr]

Yes you are right. Still they could have done better with very little effort.

However you said at the beginning

 

 

 

So you have tested the stock one outside the case and did not find any noise ?

Could it be a bad grounding issue ?

I've been meaning to do some more testing to verify that. I'll post the results here when I do.

[jfaok]

Hi I am new to the forum and have an interest in this device purely to make "needle drops". Could the addition of a Jitterbug help reduce the noise?

[mansr]

Hi I am new to the forum and have an interest in this device purely to make "needle drops". Could the addition of a Jitterbug help reduce the noise?

 

No. The Jitterbug is meant to reduce noise carried over the USB cable. This noise comes from the SMPS internal to the UH-7000, and no amount of external gadgetry will help with that. As for USB-related noise, this unit doesn't seem particularly sensitive to that.

[mansr]

I've been meaning to do some more testing to verify that. I'll post the results here when I do.

 

In doing a little more testing with the original SMPS in an external enclosure, I found that at least in some configurations, there is still some residual noise getting through. Maybe I was simply lucky in that first test. A clip-on ferrite core on the power wire helped a bit, and attaching a "power filter" (a simple circuit with a couple of capacitors and inductors) I happened to find in a box (no idea where it came from) got rid of the remainder.

[junkyardawg1]

I went through the same thing. That is how I found this page. Everyone raves on how quiet this thing is. I saw the same noise profile in iZotope. When I put the supply on an oscilloscope I was able to read the noise. The first thing I am going to try will be putting the supply in a faraday cage and checking it again. Great detective work on your part!

[ginetto61]

In doing a little more testing with the original SMPS in an external enclosure, I found that at least in some configurations, there is still some residual noise getting through.

Maybe I was simply lucky in that first test.

A clip-on ferrite core on the power wire helped a bit, and attaching a "power filter" (a simple circuit with a couple of capacitors and inductors) I happened to find in a box (no idea where it came from) got rid of the remainder

 

Hi ! thanks a lot again for the very helpful advice.

I have still to try the original psu in the out of the box mode.

I am on holiday now but i will try to do it as soon as i will get back home.

However i can already confirm that almost any other 12V psu i tried did not show this kind of noise.

This applies to both smps and linear ones. Even some 10 USD model !!!

The result noise floor has been always very low in level.

Now i can also carry out some thd+noise tests with a loop-back arrangement.

I will post the result with the original psu connected outside the box ... a little dangerous but worthwhile i guess.

As i said before i see some parts missing and replaced by jumpers. Capacitors mainly.

Thanks again, gino