MBCV AOut Module Calibration

[jrp]

When calibrating my AOut module for use with my analog VCOs i am having trouble getting the values right. I am measuring against ground of the AOut PCB.

On the site "Music from outer space", that is where i got the vco design from, it is suggested to use a voltage source that is within 2ma tolerance.

 

When trimming to get 2v just right i am reading with my Voltcraft 275:

 

0V (calibration mode off) : 0,001V

1V : 0,992V

2V : 2,000V

4V : 4,02V

8V : 8,04V

 

Is this normal? 20mV is a lot when considering the V/Oct scaling.

 

Maybe my Multimeter is just not accurate enough?

Should i trim so that 1V is just right, or so that 8v is just right?

[TK.]

Maybe my Multimeter is just not accurate enough?

 

Do you mean "Voltcraft 275" or 175?

I'm asking, because when I google for the '275, I only get your posting as the first hit - for the '175 I get the datasheet: http://www.produktinfo.conrad.com/datenblaetter/100000-124999/124457-an-01-ml-VC175_DIGITAL_MULTIMETER_de_en_fr_nl.pdf

 

The datasheet says, that the accuracy is 0.8% reading + 8 steps (which is pretty high).

So, in worst case, instead of 2.000V the actual voltage could be 2.016V or 1.984V +/- 8 steps (so, 2.024 or 1.976)

 

You can compensate this error a bit, if you would measure two or even three calibration points.

Change the gain until 1V *and* 4V are close to the target value

Then measure 2V -> result?

 

If this doesn't help, it might be better to tune the oscillator of your synth with a reference sine tone.

 

Best Regards, Thorsten.

[jrp]

Actually it´s the VC265. I wrote this mistake from memmory... 0,8% accuracy all the same.

 

I posted a similar question on the german ELKO-Forum and got some really helpful replies i would like to share. Maybe you like to quote this in the documentation, calibration of something that should be this accurate is not as straight forward as one might think.

--- BTW it was also mentioned that the Aout module could be improved if the output opamps were part of the feedback loop of the DAC. I didn´t follow this comment because (i´m not sure about it) i think that this is nonsense since the calibration trimmers would then also be in that loop and the output voltage would have to be scaled down in the feedback loop to match the level of the DAC. Just a thought. ---

 

 

With these multimeters some things can be considered to get the most out of them.

 

- Accuracy is the highest just before a measuring range ends. In this case this applies to voltages just below 4v or 40v. Calibrating to 1v or 8v would thereby not be a wise decission since it´s in the lower range of the Multimeter range.

- Below 4v you get 1mv reading, from 4v on you get 10mv as the smallest division.

- There are errors in the measurement that are chaotic and will change over time depending on temperature and noise. Unit should be powered at least 30 minutes. All measurements should be done in one session under idally the same conditions.

- More important, if you do a series of measurements in the same range one after the other like the 8 CV-outputs, errors will be static (=the same).

 

So i adjusted the trimmers slightly below their target value, set calibration output to 4v and trimmed just untill the reading jumps from 3.999 to 4.00.

This way there might still be a little error, but it should be as low as it gets with a multimeter like this, AND, most important, all 8 outputs will be very much on the same level.

 

I would say relative accuracy is most important as the VCOs that are being controlled have trimpots to set their scaling and will get along with a 1,005V/Oct or 0,991V/Oct CV just fine.

I used my ears and a tuner-plugin from logic and was able to get 6 octaves of almost perfect tracking on my 3 music from outer space VCOs.