MOTUnation.com

Hi ,

Recently I purchased a new MOTU 8Pre using it as a ADAT converter.

Since the first power up I was observing the follow trouble:

A non linear potentiometer gain action; the most gain changes are at the final potentiometer position, so is quiet difficult optimize a correct gain setup.

Also, on that pot. working zone, I listen typical noise (dirt and/or oxide inside potentiometer) exit on audio monitor system.

So , just to fix that problem by myself , I open the 8Pre cabinet, just to clean with a spray product, all potentiometer.

Looking the preamp board, I seen a empty place with no capacitors installed and a small printed trace jumper was in, instead capacitor. Near every soldering area a “+” sign indicate clearly a couple of polarized capacitor back to back was predicted.

My first thought was: Somebody has forgot to install the low frequency limitation capacitors.
But the printed circuit trace indicate a modification then it isn’t a casual omission.

This means that, with a 56 db (630) gain, some mV of DC offset at input, pass to the next preamp section with a bigger value reducing also the overhead of preamplifier.

So, I know enough well this circuitry, if no capacitors are installed, also a large DC component modulated by normal irregularity of potentiometers pass thru the following section causing the noise similar to old oxided potentiometers I heard.
.
Action done:

I cutout the jumper trace , and installed, on PCB board n. 8 no polarized 1000uF capacitors; one for each channel.

After this modification, preamp work very well, and potentiometer noise disappeared.

I don’t know why MOTU decided this modification to the original project. This is very detrimental respect the device quality. I hope this is not a tentative to produce saving. Because this is the worse way to do that.

Cheers,

Gianluca

Hmm. I think it is brave of you to make that mod without a schematic in hand. I understand your thinking on this, but caution you that this was probably not an omission on MOTU's part. I'm sure if that section of the circuit was DC coupled it was done that way for a reason. (perhaps there's a DC offset adjust somewhere) Did you actually measure a DC offset on the pot, or was this just an assumption? I'd really want to see a schematic before making a mod like that, but if you're right about there being DC on that pot, it's just poor design.

Phil

Ok Phil,

I have to explain something. I know very well this type of preamplifier and before doing any modification I followed carefully the printed circuits to have a confirmation of schematic. I have no doubts the drawing is very typical similar to that employed in many mixer consoles with the exception that at the end of analog path there’s an A to D converter with differential input. The problem is not the offset in itself. This have effect only in the headroom of the preamplifier. The problem is the gain of a DC component because this stigmatize the irregularity of potentiometer track making very noisy every little irregularity. And very audible with high gain. Think that in an audio preamplifier of this type normally the DC gain is fixed to 1 for every position of the trim potentiometer where the AC gain follow the potentiometer position. In this case the DC and the AC
gain swing from 2.3 = 7.3db to 660 = 56db. Last but not less important every thing is confirmed by the excellent final result. After my modification all is working in a perfect way.

Hmm. Very odd that they would do this. I don't know what to say. Even a gain of 660 is not a problem if the DC is 0V (0 X 660 = 0). It comes back to the same question - is there actually a DC component? A simple measurement would confirm this, yes? Perhaps it's worth contacting MOTU. Very odd indeed.

Phil

No now there is not a DC component at the preamplifier output. There was a DC component extremely variable from channel to channel before the modification. Consider that any channel have a couple of op-amp in a differential configuration. If them mounted a 4580 for this function for every channel (as I suspect, and this is not clearly visible because the component is below a rectifier bridge employed as over voltage protection) this double op-amp have a typical offset voltage of 0.5 mV ranging to a maximum of 3mV, at 660 of gain we have a DC output component ranging fro 0 to 1.98 V. In fact I measured some channel to 0.450V and some other very near to 0V. And the channel with the greater DC component was the one with the maximum of noise moving the potentiometer. No I think the situation is very clear. I sent a e-mail with the same text I posted also to Mr. Cooper in MOTU to signal this issue but till now I have received any answer.

Just out of curiosity, is the range of gain now operating in a more linear fashion? I own two MOTU 896 mk1's, and have worked on many other pieces of motu gear, all of which exhibit this extreme non-linearity. There is a good 12db jump in gain around 4 o'clock to 4:30... and a boost in noise.

Thanks for the detailed post.

No, unfortunately this modification have not effect on pot linearity. Simply avoid the noise caused by the wiper sliding on the imperfect potentiometer resistive surface. The non linearity you mentioned is caused by the type of circuit design. This type of preamplifier design is in it quite good but have some problem from point of view of components needed. The gain potentiometer is one of this critical component because every type of commercial potentiometer exhibit a resistance / angle characteristic not suitable for this purpose. The ideal condition will be a customized potentiometer with a special resistance / angle characteristic but this lead to unaffordable cost and the most part of manufacturer accept this compromise. The alternative will be a different type of design but with worst characteristic from point of view of impedance balance. Or better a solution with input coupling balanced transformer but this last will be too mach expensive for a device of this price range.