pfm2 with yellow OLED : high frequency solution

[Xavier]

Situation : Yellow OLED display have a big inductor that sends a 13.5Khz sin wave in the preenfm2 audio path.

Yellow OLED solution

I spent some time trying to find a solution to this problem, and i have somthing really efficient.
Look at the After and Before screenshot.

If you have a PCB R5d (or +) : solder a 22ohm at the "oled" position and go to step 4 directly... DON't do step 1,2,3.

If you have a PCB R5c (or -) : step 1,2 and 3 is for you.

The idea is to put a resistor 22 (ohm) between the LCD and the preenfm2 5V.
As there's no direct possibility, we'll use the Led+ pins (pin 15 of the LCD) that is not used with OLED display and where we can insert a resistor.

Step 1 : put a 22 ohm in the 68 ohm slot... this resistor is between the 5v and the pin 15 of the display.
Step 2 : cut the pin 2 of the OLED so that it's disconnected from the preenfm2 5V.
Step 3 : Link the pin 2 and 15 on the Display. Now the Oled is powered throught the 22 ohm resistor and get around 4v.
Step 1 to 3 makes the parasit frequency much more sensitive to shield. I think it removes almost totally the electronic noise but the magnetic noise is still there.

Step 4 : build a shield with many alumnium sheets that you use for food. you may have that in your kitchen.
The more you put the more efficient it is.
Take a 22cm x 22cm, fold 1 time then 2 times horizontaly, same thing vertically. You get 16 layers.
Do that with 4 sheets to get at least 64 layers.
Put that in a plastic bag (antistatic may not hurt) and add some tape. Be sure no aluminium will touch the contact.

Step 5 : put your shield under the screen.

Step 6 : Let me know

Xavier

[Xavier]

Details screenshots....

[clément]

Oh nice Xavier! Gotta try it out soon thanks for taking your time looking into this

[Đorđe Golubović]

Done and done, 64 layers of alu foil, wrapped in antistatic bag twice and then sticky tape (not sure if antistatic bag can damage something, just to be sure).

High pitch noise is drastically reduced just by adding this shield. (sadly, I didn't try to make a shield as proposed by clément, I suppose I was too lazy to cut and solder perfboards )

I didn't hear any difference with or without 22 ohm resistor, but I was only using my ears for testing, as I don't have oscilloscope... After soldering, I used soundcard oscilloscope (didn't come to my mind earlier), and peak is much lower with shield.

I am thinking about making (some day) shield out of thin steel sheet, like on RF tuners, this could maybe lower magnetic noise...

Thanks very much,
Djordje

[Xavier]

The 22 ohms resistor alone does not change anything.

You have to cut the pin 2 of the OLED header then link pins 2 and 15... See the red wire on the picture above and the missing pin between the OLED and the preenfm PCB.
I remove the whole pin but simply cutting it on the OLED side will work.

Then yes it's really worth it...

[Đorđe Golubović]

Yes, I did cut the pin from oled. I forgot to say most important thing, that I did all modifications

I was comparing (by naked ear) first only with shield, and after that shield+resistor+modified oled pcb. It was silly to compare it, because it is impossible for me to determine volume of sound of such high frequency by ear...
What I wanted to say is that I was pleased with results even when using only shield. (but did all of the modifications nonetheless)

Anyway, after playing with it for a few hours, I can confirm that I can no longer hear high pitched sine through my speakers, but I can hear it from preenfm2 box (I have moved box 3 meters away from speakers to be sure).
This is what clement mentioned in other thread, probably coil oscillating.

As high-pitched noise is practically eliminated in signal, I am no longer motivated to try a steel shield.

Thanks,
Djordje

[Xavier]

From my experience, with the shield only I could decrease the parasite by 10dB at best.
With the full modification i achieved a 35dB improvement.

You can check easily check with voxengo SPAN.
You should place your shield with SPAN active because a 3mm position move can easily lead to a difference of 6dB or more.

Xavier

[clément]

I gave it a shot and the results seem to be great indeed, thanks Xavier!

[Đorđe Golubović]

Tested in rehearsal studio situation (finally), works perfectly, no high-pitched noise here. I've turned the mixer volume as high as it goes, to the point where I hear (relatively) loud white noise, and there is no ~15kHz peak.

I haven't yet come to measuring it with real tools, but I am perfectly satisfied with this, as it can be used in live situation, as far as me and my people are concerned.

Thanks again,
Djordje

[solipsvs]

could it be that i dont have the same issue for some reason?  i have the yellow OLED too

[Xavier]

could it be that i dont have the same issue for some reason?  i have the yellow OLED too

It's more likely that you don't hear it.
Most people don't hear a low level 15Khz sin (I don't hear it at all).

[lokki]

hi xavier,

i put a 68ohm instead of 22ohm in there, is it also ok? everything seems to work, so i thought i ask why exactly you choose the 22ohm value? (brightness maybe?)

also note that this "mod" cures my problem with power feeding from the macbook air (if you remember). so it seems the display was the current hungry thing that made the macbook air struggle.

cheers

[Xavier]

Hi,

With a 68 ohm resistor, the OLED is darker, and it is powered by around 2.5V although the OLED requires between 3.3 and 5.0V.
If it works for you and if you like dark OLED, that's fine 

With this resistor, it's possible that the overall current requires by your preenfm2 is smaller. So it's now small enough to be powered by your macbook air.

Xavier

[lokki]

ok, thanks for the clarification. strange, it does not look darker at all! surely perfectly useable for my purposes.

the aluminium shield / resistor mod combo works wonders, thanks for this simple fix! i put 128 layers of aluminium between the display and the pcb and i cannot see (or hear) any spike in voxengo span. this is with the internal input (gain maxed out) on a macbook pro. with 64 layers i could still see a small spike.

[tubeohm]

Details screenshots....

 Hi Xavier - i noticed this peak on - 70 dB . This you normally can't hear . Have you set the Slope on the FFT to zero ( linear )
Specially this FFT had a slope function that enhances   the high frequencies  up to 6 dB  and its is on - on default .