Review & Mods: MeanWell S-350-24 Supply from SureElectronics

mwkeefer

1 MW
Joined
Jun 14, 2009
Messages
2,263
Location
Malvern, PA USA
Hi Mike,

Jeremy Harris and I have some experience with these PSUs - sold under a different name but I'm sure they are the same. I'm using them to make high current battery chargers.

http://endless-sphere.com/forums/viewtopic.php?f=14&t=6749&p=101832&

I worked out a hack to add a constant current feature. I since discovered that they do have an internal current limit, but its set too high - so high that the PSU blows up if you use it continuously. I did work out a way of adjusting it downwards but I haven't written it up yet.

Just a caution - these things blow up for a hobby; be very careful as you work on them.

Nick
 
Nick,

>
> I worked out a hack to add a constant current feature. I since discovered that they do have an internal current limit, but
> its set too high - so high that the PSU blows up if you use it continuously. I did work out a way of adjusting it downwards
> but I haven't written it up yet.
> Just a caution - these things blow up for a hobby; be very careful as you work on them.
>

I know these have a current limit feature - hiccup mode though I think, but it's okay.... however you telling me that the internal current limit is set too high?

Assuming charging at 20.8 per supply, 350 / 20.8 = 16.83 A @ 100 % efficiency - these are rated at 82% so 13.82 A should be expected at 24v nominal

If this isn't the case, then your right they could work beyond desing...

How many of these have you destroyed?

I would be very intersted to hear the modification or adjustment to limit the current output (so long as it's not through additional heat dissipation) because the last thing I want is to blow these supplies from overload.

Additionally being able to adjust the output current would eliminate the need for me to design a quick and dirty current limiter (this will dissapate heat) in the form of a pair of parallel fets which will switch to vary the available current to the load. This will allow me to dial back the power for 1C charging of 2P packs but also will allow me to vary the curve and taper (since I plan to monitor voltage, current and control the FET gates via an atmel MCU).

Any suggestions, pics or rough info you could post would be appreciated!

Nice thread btw... they do seem similar, have you reviewed the internal PCB pics on my review... are the indeed identical because I received 3 units like the ones with the similar model you have... none of them worked right and were all a bit different internally the designers name is on the PCB I don't remember it but it's there.

-Mike
 
Quick Update: 11/4/2009 12:53 PM

So... we all know by now that I am impatient (even though I like to be methodical to a fault) so I couldn't resist just a test.

I decided to wire up 3 of my 5S 5AH 20/30C Turnigy LiPos in Parallel, Use one of Gary's balance parallel reducers (new name?) and finally connected the temperature sensor and the balance plug to iCharger to monitor cell voltages and pack temps during blind charge.

Power supply A is being used for the test, set to 20.7 HVC and effectively charging the presumed 14.6A rate into a 15AH 3P5S LiPo pack at just under 1C 14.6/3 = 4.8667 A per 5000 mah pack.

Charge process is completed now... SUCCESS !!!

Here are the specifics:

[Pack]
Pack Configuration: 15S1P 5000mah Turnigy Lipo - 20/30C
Pack Max Voltage: 62.25v (4.15v per cell)
Pack Min Voltage: 48v (3.2v per cell)
Pack Discharge Number: 7

[Discharge]
Discharge Method: eBike Use
Discharge Max Current: 35 A
Discharge Amp Hours: 5122 mah
End of Discharge Pack Voltage: 52.524 v (or 17.508v per 5s segment +- .05v)
Average Cell Voltage at Rest: 3.5016v

[Charge]
Pack Configuration: 3P5S 15000 mah Turnigy Lipo 20/30C
Charge Max Voltage: 20.7 (should be 20.8 for 4.15v per cell but figured be safe than sorry)
Charge Current: Unknown - Educated Guess = 16.8269 A @ 20.8V Maximum
Monitor Method: Manual - iCharger1010B+ via LogView

Pack Resting Voltage: 17.508 v (remember converted 15S1P to 3P5S)
Resting Cell Voltage Range: 3.450 v - 3.559 v (already out of balance)

Voltage Maximum: 21 v @ 4.20v per cell x 5S
Voltage Minimum: 15 v @ 3.00v per cell x 5S
SOC 1% Voltage: .06v (because 6v swing from full to empty)

Beginning SOC - 17.508 - 15 = 2.508 / .06 = 41.8 %

SOC @ 15:00 - 19.247v : 19.247 - 15 = 4.247 / .06 = 70.7833 %
SOC @ 30:00 - 19.889v : 19.889 - 15 = 4.889 / .06 = 81.4833 %
SOC @ 45:00 - 20.672v : 20.672 - 15 = 5.672 / .06 = 94.5333 % almost done
SOC @ 46:20 - 20.700v : 20.700 - 15 = 5.7 / .06 = 95.0 % (this is where the supply was set to max out.)
SOC @ 50:10 - 20.751v : 20.751 - 15 = 5.751 / .06 = 95.85 % (our first cell just hit 4.159v)

At the end of charge….

Cell Voltage Range: 4.149v to 4.159v

I ran my test a bit longer to ensure no damage or temperature rises happened… actually I ran it about 30 minutes longer than I needed to… to 1:15 and more. Nothing heated up the cells actually balanced out a bit more in the end.

Next test will be logged using my new atmega based BMS PC Interface (thank god for 16x analog duplex multiplexers - monitor any cell, up to 16 with just 2 A2D pins and 3 digital IO pins (to control which cell you are sourcing analog input from) -

Some things I should mention… I hope to soon determine if Nicks power supply is a knock off or off label mean well…

Knock Off = Lower Quality Imitation, ie Fake.
Off Label=Same Quality, Components and QC but branded to a different manufacturer for resale.

I am hoping that they are Off Label because I would love to determine how to control the output current of the supply. The thought of using FETs to switch at 63v and to dissipate 10A (the supply will be more than that) 630w of heat? NO WAY!!! With some luck his mod will be adaptable to MCU control.

Some good news… there are 3 current sensing shunt resistors in the Meanwell (see the PCB pics with review) which means I can connect an appropriate voltage divider and then input the pre/post shunt voltage readings into the atmega I plan to use for control / management of this thing in the end… now I can calculate the current, since I know how to limit voltage by replacing the pot with a kind of digital pot controlled via atmega … it would be perfect to be able to manage the current produced by the power supply also… Then I could adapt to charge various chemistries and capacities (various AH) while managing the curve and use a FET just to disconnect the supply from the battery for an instant so I can sample the battery standing voltage before engaging the FET again to allow power to pass… this would likely be checked each 3-10 seconds (configurable) to ensure battery within range and determine where we are in the charging cycle.

Well… the supply does what it's supposed to do and seemingly well. The outside got a bit warm but I expected that. Fan kicked on and ran for a while, until I opened the window (it's really cold here today).

All in all and assuming these units last for a considerable length of time (ie: the life of a well maintained pack, 2-3 years) and the type of abuse and use they will receive from me… I am very satisfied with the product and plan to purchase MANY MORE!

I am at this point pleased with sure-electronics also, and plan to place orders for several additional units of various size and capacity in the next days. I will be doing reviews on each of them also.

[Todo]
Complete full 15S2P charge test with manual monitoring of cells and temp, post results with video and pictures
Post schematic to provide analog 0-5v control of output voltage (interface to PC or MicroController)
Mod units to allow MCU programmable control of output voltage and monitoring of voltage and current
Get or find mod to control output current and post it with review
Adapt current control mod for MCU or PC control.
 
Hey all,

Many thanks to Hyena for inspiring the theoretical (yes, still needs to be tested - tommorow) method of variable current limiting... can't believe I didn't think of this sooner, it was right there in front of my face the whole time and I just couldn't see it = )_+

Has many other potential applications... I will post more later / tommorow.

Regards,
-Mike
 
Nick,

Thanks for that!
I checked your pics, they are essentially the same (would love to compare macro of trace and component side to trace differences. My unit is 24v factory 350w (400 really)... I was going to mod the pot anyway to do away with the piece of crud it comes with.
Yours has a single shunt, mine has 3 in parallel (I think, need to spot them out becuase they may be for different purposes...
Also the input caps on yours are much larger than mine, what are they rated for.

Thanks to you and to J Harris too, I wouldn't have a Wye/delta hub or tower pro without his discoveries - nor would I have functional hall sensors in my tower pro which works nice with infineon contorllers = )_

-Mike
 
Okay....

I'm not done my testing but have some early configuration info which might be helpful:

[S-350-24]
Rated: 24v, 14.6 A @ 350w
Normal Voltage Range: 18.4v - 27.4v
Normal Components (R25; SVR1; ZD1; Output Caps): R25=1.8k (measured 1.63k); SVR1=1k - 1 ohm; ZD1=27v Zener; CAPS(35,36,38)=35v caps 2200uf
Upgrade Components (R25; SVR1): R25=1.2k; SVR1=10k - 1k
Upgrade Voltage Range: 6.9v - 29v

[S-350-36] - *thanks to Nick for this! - (I don't have a 36v model but I have all the rest)
Rated: 36v, 9.7A @ 350w
Normal Voltage Range: Unknown
Normal Components: R25=3.9k (measured unknown); SVR1=1K; ZD1=39v, CAPS=50v
Upgrade Components: R25=2.2k; SVR1=5K Pot; ZD1=Additional 9v in Series for 48v; CAPS=NoChange
Upgrade Voltage Range: 22.7v - 49.4v

I am currently sorting out the current regulation circuit (see what happens when I try to be funny) for these meanwell and clones before I continue with voltage modifications on the remaining models (5v, 12v, 48v) I have here to experiment with. Since this is a weekend without my little ones, I expect to finish the r&d and testing over the next day and post accordingly.

The end game is as it aways was, to allow for very wide range of voltage adjustment and sufficient current limit adjustment to allow safe use of expanded voltage range (keeping it below 350/400w for these models).

Hope this helps!

-Mike

PS: Thanks to Gary for pushing my ars to get moving on this stuff... I have too many projects, but I like that way, when one becomes too frustrating I just flip to another for a while to cool off... Anyway thanks again for the fire bud.
 
Back
Top