The pitfalls of using cheap chinese power supplies

[fasteddy]

I tried using cat6 cable, no difference, I tried disconnecting the ground on the data cables, no difference, i tried making the ground common between PSUs no difference, I tried turning everything off in the house and only leaving one section on and one PSU connected and no difference, there are countless things that were tried but with no success. Its definetly a noise issue and its all pointing to the power supplies, but with all noise issues finding the source can be a very time consuming process, expecially if you dont have any diagnostic tools on hand

 

[TimW]

Eddy - I was also having problems earlier with flickering on fades of 3ch dmx strips supplied using ray 12v supplies.

What fixed it for me was shorter 12v supply runs. My theory was that it wasn't EMI.. but voltage conditioning/voltage drop that was causing the problem between the supply and the module. I was not using low enough gauge wire... (?)

How long are your supply runs from psu to strip?

Currently not having the problem (but aren't fully deployed yet... will get back to you)

 

[fasteddy]

Eddy - I was also having problems earlier with flickering on fades of 3ch dmx strips supplied using ray 12v supplies.

What fixed it for me was shorter 12v supply runs. My theory was that it wasn't EMI.. but voltage conditioning/voltage drop that was causing the problem between the supply and the module. I was not using low enough gauge wire... (?)

How long are your supply runs from psu to strip?

Currently not having the problem (but aren't fully deployed yet... will get back to you)

I have a mix of long runs and short runs and some where the controller and the power supply are in the same box. If you have seen this issue and you havent loaded up the power supply sufficiently yet then i would recommend you do this as in all my testing in the garage i did not see this issue, it was only when it was up and running that it becanme a major issue. I found that having the DR4 located right next to the controller help dramatically minimise this issue as it appears the cat5 and even the cat6 cables i tested with seem to be acting as a big antenna.

I maybe having a combination of issues here, but trying the new power supplies should at least be a good test and is what im pointing to. It will be nice to confirm the power supplies once and for all as i still do have some flickering but its not the major issue it was before.

an example was i have a matrix which is now running from a E680 controller, i have issues with the third strip and it was only when i either held the cable in my hands that the interferance reduced or when i re routed the cable by adding an extension it was reduced . But there is still some interferance but not to the level i had which was making the show unusable and an eye sore.

 

[ryanschristmaslights]

As David mentioned, I had major problems with just one "knock off" power supply, which I bought through another display around the corner from me in the one order. The other display uses several of the knock off's doesn't appear to be having issues.

The level of EMI that was being emitted last year was heavy enough to cause permanent damage to my modem, where it now loses sync when I'm blinking AC channels. In 2008 and 2009 I used the same AC controllers and same modem without problem. This is enough proof to me that there's been permanent damage here with the modem.

 

[TimW]

Ok.... some more investigation from my side...

• I have a bethlehem star, approx 10m of standard rgb strip (30 l/ m)
• It is controlled by 4 x 3ch dmx controllers, mounted on the star.
• Fed by 12v dc from the ground. In the garage, lets say its 5m of lead in wire.
• Ray wu psu 12V 120w (smaller than eddy's?). I plan to use multiple smaller supplies to reduce voltage drop

I found that on whites there is some "flicker" on the dimming. The flicker can occur on loads <1A. So the problem I'm seeing does not qualify as 'loading' the supply yet! (not to say that the underlying cause isn't the same though)
The flicker does not happen on single RGB channels... but there's less variation on load for single channel pwm.
If i disconnect dmx the controllers run standalone and flicker still occurs (so I don't suspect DMX interference)
If I replace the power supply with my bench supply the problem remains
If I shorten the power wire the problem goes away
If I parallel power cables the problem goes away
If I put a 2200uf electro cap just before the controllers (on the star) the problem goes away
Looking at the power feed with a cro shows the voltage at the psu is constant (with a bit of ripple). Towards the controller it goes crazy when the flicker starts. I think its a load problem at the controller. To generalise... the pwm from the controllers creates a variable load.
What I don't know is what would happen if I loaded the supply further (say 5A or more). I suspect the problem gets worse and the cross impacts to other feeds gets worse... to the final result that Eddy sees....

What I'm thinking though is that it may be premature to blame the supply. It may be the load we need to work on.... (eg put capacitors close to the load.)

 

[David_AVD]

I understood that Eddy tried many things and what helped was reducing the length of the DMX cable. As far as I know all other parameters stayed the same.

 

[Superman]

I have one window with a 27ch and 12v ray power supply that is suffering the same problem. I will be changing the PSU with a spare one and having a look at it tonight as well

 

[random]

Ok.... some more investigation from my side...
If i disconnect dmx the controllers run standalone and flicker still occurs (so I don't suspect DMX interference)
If I replace the power supply with my bench supply the problem remains
If I shorten the power wire the problem goes away
If I parallel power cables the problem goes away
If I put a 2200uf electro cap just before the controllers (on the star) the problem goes away
Looking at the power feed with a cro shows the voltage at the psu is constant (with a bit of ripple). Towards the controller it goes crazy when the flicker starts. I think its a load problem at the controller. To generalise... the pwm from the controllers creates a variable load.
What I don't know is what would happen if I loaded the supply further (say 5A or more). I suspect the problem gets worse and the cross impacts to other feeds gets worse... to the final result that Eddy sees....

What I'm thinking though is that it may be premature to blame the supply. It may be the load we need to work on.... (eg put capacitors close to the load.)

I would agree with this, PWM causes a horrible load condition and any distance of power cabling is resistive so the PSU has no hope of maintaining regulation at the load in this situation. A chunky (preferably low ESR, but probably any would do) capacitor at the load will help keep the supply voltage stable. You just need to be careful about how much capacitance you're adding to the system, as it will cause quite an inrush when you first turn the system on. If you have a CRO then you could try different cap sizes and see which one gives you adequate performance (voltage ripple is low enough not to upset your lights) without going overboard (too much inrush for the PSU).
I have a hard time believing that DMX won't run over long distances, RS485 is specified as 1MBps @ 1200m using 120 ohm cable (CAT5 is 100 ohm, close enough). Eddy's problem may be that his data cabling "antenna" is picking up interference from the power cable which is caused by the PWM activity. All those sharp edges will be creating a lot of EMI/RF harmonics.


There are a few things to try:
1) Added capacitance at the load, to smooth current draw and voltage at the load.
2) Use twisted cabling for power (this will reduce the amount of EMI radiated)
3) Add terminating resistors to each end of the data cable (100 ohms or 120 ohms). Occasionally this makes things worse, suck it and see. In theory it should make it better as an unterminated data line reflects signals that reach end of line. Terminating resistors go from DMX+ to DMX-, and you only put one at each end.
3) Use ferrites on power and data cabling. Even better if you can put a couple of turns of wire through, ferrites attenuate RF/EMI proportional to the square of turns through the middle. If you go overboard on the data cabling it may soften the edges of your data signal too much, you would need a CRO to see this but experimentation would give you an idea of what's possible.

 

[Mike]

I was under the impression that most if not all switching psu's have or can cause some form of EMI. Regardless of make.
I know some will be better than others and I am sure the meanwell ones are superior to the ones Ray sells.

Will it justify the cost of the upgrade or would a simple EMI filter work better and cost less.
Time will tell I guess

 

[TimW]

I would agree with this, PWM causes a horrible load condition and any distance of power cabling is resistive so the PSU has no hope of maintaining regulation at the load in this situation.

And often the power for the logic and PWM control is also copping this lack of regulation at the load.... because its some distance from the psu in configurations like the one for my star.... and it uses the same voltage source.

Anyway.. some good options to try now (in addition to swapping out the supply)

 

[fasteddy]

The controllers that seem most effected are the ones where the controller and the power supply are installed in the same steel B&R enclosure so the distnace between PSU and controller is about 1/2 metre, I now have a DR4 in those 2 panels which did reduce the amount of flickering considerably. It is strange that the only fix i found was to reduce the DMX cable but the dmx cable acting as an atenna is the only reasoning i could come up with and this was even when i tried a CAT6 cable

I tried a 20 metre cable (was the initial install cable), 10 metre, 5 metre and a 2 metre cable the shorter the length from DR4 to controller the less flickering. i still see minimal flickering but nothing like before


Im now searching for some 1000mf capacitors to put on the output

Another thing to note was when i changed the way i had things setout one of the only controllers which is a TP3244 which was on its own universe didnt show any signs of flickering, since i re-routed this data cable (same data cable just a different place it is fed from) it now exhibits some flickering, so its the same power supply, all that changed was the way the cable was run and its not run anywhere near any mains power except for where it enters the panel. but that was no different to the original run

 

[Superman]

Probably unrelated. but, I have fixed mine.Mine seemed to be a rj45 connection going into a CCR that was plugged in line before the Flickering DMX 27ch controller. I changed my PSU and it was no different, My problem solved itself when a disconnected the CCR and after a few more plug In's it seemed to be OK again. That's as technical as I can be with it.

Good Luck Eddy.,

 

[random]

I was under the impression that most if not all switching psu's have or can cause some form of EMI. Regardless of make.
I know some will be better than others and I am sure the meanwell ones are superior to the ones Ray sells.

Will it justify the cost of the upgrade or would a simple EMI filter work better and cost less.
Time will tell I guess

All switching PSUs will generate EMI internally, by the nature of turning an output on and off many thousands of times per second. The difference is in how the manufacturer mitigates it both in circuit design and in the device enclosure (eg metal frame) in order to pass mandatory legal requirements for this type of emission. For Australia this is C-Tick, Europe is CE and US is FCC. All of these have hard limits on the amount of electrical radiation a device can emit.


In many cases, cheap hardware has no hope of compliance which is where the myth of CE actually standing for "China Export" started. You can be sure that any company knocking off a Mean Well supply and making a similar logo doesn't give a damn about copyright and doesn't give a damn about EMI either.


It's simply a case of "you get what you pay for". A well designed supply (Mean Well) will be well behaved. A cheap supply might work OK if you have a simple resistive load but when you ask a bit more of it (PWM load) then it gets ugly.


Eddy if you have problems where the controller and PSU are in the same enclosure then there is probably radiated emission from the PSU itself. You could test this by increasing the distance from the controller to the PSU, and see if it helps. Pull the controller outside the steel enclosure temporarily, shut the enclosure with the PSU inside.

 

[TimW]

Im now searching for some 1000mf capacitors to put on the output

Might be overkill.... ? :-\ (1000uf I would understand)
I'm going with the radiated emi from proximity to the supply theory . I tried the AM radio trick. Its screaming for 30cm from mine! Holy Farady cage batman!

 

[fasteddy]

yes correct i did mean uf :-[ unfortunatly i didnt find any