TheBackShed.com

Forum Index : Electronics : MPPT question for Poida

Page 2 of 2

Author

Message

Murphy's friend

Guru

Joined: 04/10/2019
Location: Australia
Posts: 671

Posted: 04:32am 08 Dec 2021

Copy link to clipboard

Print this post

poida said

I think the huge EMI radiated by the choke and power board couple well
with the metal case, causing all sorts of unplanned voltages on anything metal
inside.
We have a Faraday Shield with the radiation on the inside!

Yes, I thought that might be the case with my unit. Fortunately just a 2mm gap between the LCD face and the cover cutout lets some of the EMI escape

, the rest of it no doubt finds a way out at the fan and top cooling holes.

Its a 40 degree day here in Perth and the MPPT unit has no trouble with that. We'll see how it goes through the rest of the Summer.

You done a great job designing that Poida, it feels so much better to have something one built himself working for off grid power

poida

Guru

Joined: 02/02/2017
Location: Australia
Posts: 1432

Posted: 05:14am 08 Dec 2021

Copy link to clipboard

Print this post

Many here helped design it. And critique my ideas.
But it is so very repairable and that is what I wanted from the exercise.

Recently I saw here a post showing a disassembled Victron 150V 45 Amp mppt.
To open the case means you rip the chokes out of the PCB due to them being glued
in place. And now the PCB is damaged beyond repair.
This is not good enough. And unnecessary too.
This particular model sells for about $850 AU

I have blown a mppt board about 4 times in testing this and that.
(the pulse by pulse over current control, not implemented yet, was a difficult birth)
It's so easy to debug and repair.
Easy to replace the two FETs, easy access to signal paths, cheap to make,
performs well enough to take on equivalent mppt controllers costing
near $1,000 on DC-DC efficiency.

wronger than a phone book full of wrong phone numbers

wiseguy

Guru

Joined: 21/06/2018
Location: Australia
Posts: 1206

Posted: 01:17pm 08 Dec 2021

Copy link to clipboard

Print this post

I believe the noise injection into the LCD is what is referred to as differential noise. There is noise existing between the front panel and the LCD metal surround - this noise (no doubt emanating from the MPPT switching noise) and can get coupled (injected) into the data/control lines of the LCD and upset its internal workings.

The cure I have used in the past is to turn the noise into common mode noise, This involves coupling the LCD ground trace to the LCD metalwork if the metalwork is "floating" and not already connected to ground on the LCD PCB. If the metalwork was already connected to the ground of the LCD PCB the next step is to then couple the ground of the LCD PCB to the front panel, usually a 10n ceramic multilayer capacitor is sufficient.

The noise that was being picked up by the control lines on the LCD, is now the same noise that is present on the ground of the PCB so their is no differential effect and normal operation usually resumes.

The noise is often present on all the power and data wires connecting to the LCD and originates from the noisy MPPT, the fast edges often get into everything, putting a few turns of the cable to the LCD through a ferrite core can attenuate the noise difference between the LCD and front panel, that is why it may help enough to effect a cure. For more stubborn cases the above approach usually sorts it out.

For Murphs "cure" the increased distance between the case and the LCD is less capacitive coupling of the noise between the LCD and front panel the EMI didn't really escape it is just having less detrimental effect.

If at first you dont succeed, I suggest you avoid sky diving....
Cheers Mike

poida

Guru

Joined: 02/02/2017
Location: Australia
Posts: 1432

Posted: 10:33pm 08 Dec 2021

Copy link to clipboard

Print this post

WG:
I think I will try the 10uF cap and see if it works.
It would be really neat if it does.

I don't have the knowledge to tackle problems like this properly
so it's great to have your insight here.

wronger than a phone book full of wrong phone numbers

Murphy's friend

Guru

Joined: 04/10/2019
Location: Australia
Posts: 671

Posted: 05:21am 09 Dec 2021

Copy link to clipboard

Print this post

Thank you, wise guy, for your explanation. This is something a non engineer like myself could never have thought of

.
All the metalwork in my unit (base plate, heat sink, choke frame, cover) are connected to a ground stake but not to battery ground.

I will try the 10nf capacitor cure if I have further screen blanking problems.

I did try "putting a few turns" of the cable from the nano to the LCD through a tubular ferrite core but that screwed up the display (some text unintelligible). What I do use is very short (~50mm) wires between them and the SDA & SCL wires have two small ferrite beads over them. The nano brain board and the display are piggy backed, making a very short connection possible.

Page 2 of 2

Print this page

To reply to this topic, you need to log in.