high pitch frequency on motu 896mk3 hybrid inputs

[wylie1]

Thanx for the info ME.
I'm surprised they wouldn't have replied or contacted you directly.I would think MOTU would have taken this seriously

[Spencil_Splodge]

Hi,

I have received a reply on my Motu TechLink after forwarding the initial details you posted here, Midnight_Engineer (I posted them a link to your post also). They have replied to this and are, in their words, "looking into it very seriously". That was what they told me back on the 15th Feb.

I might poke them again (I get replies fairly quickly now as I've been talking with them a lot regarding this issue) before the end of the week to see if there has been any progress.

Once again, thanks for your hard work.

Best Regards,

Spence.

[Klaus]

I called MOTU on Feb 8th and talked to a service rep and he said he would forward the message that I had solved my noise issue. To date, I have received no call back. Also, I posted a techlink on Feb 9th on MOTU’s site which remains "unread" to this day, Feb 27th. I said I would wait two weeks in a prior PM, so off we go….

Me, and a great friend, modified my unit on Feb 7th. I've been running it off and on for hours at a time over the past several weeks with no problems. The mod appears to work. I’m passing this on just in case I happen to pass on.

BACKGROUND: The Power Supply Board (PSB) is distinct from Main Board (MB). Utilizing a low impedance Shure Sm7b placed in an anechoic chamber, approx 60mV of 1500-1600 Hz AC shows up on several of the dual header jumpers between PSB to MB, those specifically labeled 3.3, 3.8 and 5. All of these pins have actual DC slightly higher than labeled. I assume this provides headroom for downstream regulation on the MB. The PSB corruption is easily audible at high gain thru the headphone jack and is recorded.

My power supply board is labeled "PS-3501" rev4, layout dated as 2006. The PSB is a pulse-width modulated switcher with error feedback wherein the dc output is continuously compared to a reference and pulse width is adjusted accordingly. In this case, the PSB utilizes an optical coupler, precision reference, and amplifier in one handy chip.

http://www.fairchildsemi.com/ds/FO/FOD2741A.pdf" onclick="window.open(this.href);return false;" onclick="window.open(this.href);return false;

The 896mk3 PSB closely resembles the "typical application" in the data sheet. Be cautious studying this datasheet as my cohort noticed that Figures 7 and 8 have the opamp pins incorrectly labeled. I could easily examine the PSB feedback oscillation with an ac-coupled scope probe on U1 pins 6, 7, and 8. Pins 6 & 7 exhibit a sinusoidal waveform and Pin 8 is asymmetrically compressed. Since we only want pure dc out of the PSB, this feedback signal needs to be clean and, ideally, it should only waiver as the MB load rapidly varies. From my experience, the layout of the PSB should have the compensation network much closer to the chip. Component changes or a new layout would be a “better” fix but at least there is an “easy” fix , we need to simply add a part.

SOLUTION: A very similar chip is shown in

http://www.scantec.de/uploads/media/Fly ... ration.pdf" onclick="window.open(this.href);return false;" onclick="window.open(this.href);return false;

Specifically, Figure 3 “Control Loop Section at the Output” provides for an additional capacitor. The cap connects between pins 6 & 7. We added a X7R 4700pF, 10%, 50 Volt capacitor to PSB U1. We used a small surface mount part but you can use anything non-polarized that will fit nicely - a 1206, 0805 or even a thru-hole part. Adding this cap lowers the oscillation amplitude on pins 6, 7, & 8 and moves the dominant frequency to well over 300kHz which is well above Nyquist frequency for the highest onboard sample rate. Better yet, the oscillation is NOT AT ALL EVIDENT on the PSB to MB jumpers due to the existing output filtering and so it can only be found on the feedback network after the mod.

PHYSICAL IMAGE: http://s1095.photobucket.com/albums/i46 ... gineering/" onclick="window.open(this.href);return false;" onclick="window.open(this.href);return false;

Regards,
ME

Thank you very much ME

Now they could sell for less and we fix it with a cap !

Best

Klaus

[Spencil_Splodge]

Hi all,

I have some fantastic news!

I have been in touch with someone from Motu's QA department and they have found the cause of the issue to be related to manufacturing tolerances. Basically, not all units have the issue but a fair few will. They have now implemented a change to the units (probably along the lines of Midnight_Engineer's suggestion!) so all units coming out from manufacturing from this point will be good regardless of the tolerance problems - I guess the tolerance issue is related to the Switching Power Supply.

I will eventually have a replacement unit that has the issue resolved and I will post again with what I expect to be gloriously tone-free news.

My sincere thanks to Midnight_Engineer for your efforts and helping to push this issue through (and of course Klaus for your initial test results!) and to Motu too for working hard on this, implementing a solution and offering me fantastic support. I doubt all companies would be quite so open with an issue like this and turnaround a fix and associated support in the time frame Motu have. I know not everyone whom posts here has something good to say about Motu support, but for me I've had a really good experience with it.

Cheers,

Spence.

[Klaus]

Hi all,

I have some fantastic news!

I have been in touch with someone from Motu's QA department and they have found the cause of the issue to be related to manufacturing tolerances. Basically, not all units have the issue but a fair few will. They have now implemented a change to the units (probably along the lines of Midnight_Engineer's suggestion!) so all units coming out from manufacturing from this point will be good regardless of the tolerance problems - I guess the tolerance issue is related to the Switching Power Supply.

I will eventually have a replacement unit that has the issue resolved and I will post again with what I expect to be gloriously tone-free news.

My sincere thanks to Midnight_Engineer for your efforts and helping to push this issue through (and of course Klaus for your initial test results!) and to Motu too for working hard on this, implementing a solution and offering me fantastic support. I doubt all companies would be quite so open with an issue like this and turnaround a fix and associated support in the time frame Motu have. I know not everyone whom posts here has something good to say about Motu support, but for me I've had a really good experience with it.

Cheers,

Spence.

I'm a little skeptical that tolerances are the cause, but may Motu convince me...
The previous 896mk3 ( firewire only ) has it too, just 8 dB less...
I have 2 of them...

Klaus

[Spencil_Splodge]

Hi Klaus,

Interestingly, Motu QA were not able to reproduce the issue on the 896mk3 Firewire units that they have but they were able to reproduce the 896mk3 Hybrid issues on several units (but not all). Widening their investigation of the Hybrids showed that there were manufacturing tolerance issues (as they reported to me at least) and I guess that the 896mk3 Firewire is not manufactured anymore but the issue is probably likely to be the same albeit not on the units they have available for testing on - I understand that this itself might cause some skepticism

I'm just glad that we know what it is, Motu openly say there is an issue on some units and that a fix is in the pipeline...

Thanks,

Spence.

[Klaus]

Re: some have the issue ( hybrid, or firewire - only )...
I tested two 896mk3, and the distributor some
I tested one 896mk3H, and the distributor some

All showed the same issue :

http://www.redmountain.ch/896mk3-896mk3 ... 1k65hz.pdf" onclick="window.open(this.href);return false;

Let's say we tested seven, chances are every eighth is ok...

Point.

If I were Motu, I wouldn't even talk about this, just fix and make my customers happy !
At least that's what counts here...
Customer relation is one thing, but support efficiency can degrade that as well...

Klaus

[Spencil_Splodge]

Klaus, totally agree - if they fix it, then that's the main thing...

Of course, as you say, the support that comes with this kind of thing is also important. I've had the support from Motu so far and hopefully will have a good unit at some point next week which I'll post my test results for. Just to say it again - this forum has helped so much, as have you and midnight_engineer. Hopefully now, all 896's moving forward will be rid of this issue.

Thanks,

Spence.

[Spencil_Splodge]

Just to let you know - I've received a replacement unit which, after some initial testing, appears to be rid of the tone/noise issue! I've tested a couple of mics, a variety of preamp levels with my headphone volume set loud and I've not been able to reproduce the issue. I've also checked the FFT display and cannot see any abnormal peaks - just white noise. Yay!

I'll do some more extensive tests this weekend when I'll actually be recording several tracks at once from multiple mics/sources. Should be fairly conclusive.

Thanks,

Spence.

[chakatrain]

I've been following this thread with interest, as I'm also getting high frequency noise on my 896mk3H.

Great to hear you got a new device and that it's quiet. I think I need to tell MOTU about my issues and see about getting them resolved.

[EnjoyRC]

Unfortunately, I'm joining the group that have experienced the high pitched whine when plugging a microphone into the analog microphone input. I've tried multiple microphones and cables. Even used a microphone that had a switch on it. I heard the high pitched sound even with the switch turned off.

[EnjoyRC]

Just noticed that a 1/4" TRS cable does the same thing. Plugged it into my Virus TI, Kronos, and other keyboards and it also creates the high pitched whine. Even with the keyboard turned off.

[Spencil_Splodge]

Sorry to hear you are in the same boat that as I was in until recently.

I now have a new unit and after testing for a period of time, all is well. My old unit would have the tone on anything that was plugged into an analog input - there is no exception. You can even make the tone by engaging the PAD with nothing plugged in.

I suggest you contact Motu tech support and confirm with them the issue you are having.

Best Regards,

Spence.

[chakatrain]

Hi all,

I fixed the high frequency issue on my 896mk3H, but I'm afraid it may not help the original poster's issue as I had thought.

My issue turned out to be significant hiss on the Main Outs, not on the inputs. I did a factory reset of the settings and voila!, the output is now pretty much dead quiet. I'm happy as can be, now.

[hk_lydian]

I have been followed the thread and sincere appreciate to all of you sharing the experience about the 896mk3. I am still interested in buying a 896mk3, anyone bought that unit recently who can tell the high pitch issue have already been solved by motu or not?