How-To: Meanwell 63-84V CV/CC Charger

Kingfish

100 MW
Joined
Feb 3, 2010
Messages
4,064
Location
Redmond, WA-USA, Earth, Sol, Orion–Cygnus Arm, Mil
Goal:
I want an inexpensive charger that is Constant-Current (CC) and Constant-Voltage (CV) for my 15-to-20 cell LiPo configurations, and I want it to charge quickly.

Greetings -

Initially I had thought to append to my earlier topic: Configuring a Meanwell 84V charger, but after my road trip experience my wish is to provide a better solution based upon a more thorough understanding.

ADDENDUM: Related thread - Review: Meanwell HRP-600 Series

Last winter I bought three SP-320-48 genuine Meanwells off of eBay for < $35 each for the specific purpose of quick charging. It turned out that the SP units have a problem with CC or being modified to do so; whatever. The new theory suggests that if you purchase an older S- unit and place it in series you can make your grand arrangement CC. I bought an S-350-27 to complement my array for just that purpose.

With two SP-320-48 units in parallel, followed by one S-350-27 unit in series this will create a 75V setup which can be easily trimmed up to create an 84V charger. The trick was to balance the -48 units to the same voltage first, then trim to the final voltage with the -27 unit.

In the process of trying to leave for my trip the 74V-controller didn’t work out. :cry: My ass was in a bind, and I was just damn lucky I could de-tune the Meanwell setup to support 63V.

Creating the 63 Volt Charger:
Put the S- unit on top (the reason will be apparent very shortly). Tune the voltage down on SP-320-48 units individually first through the SVR POT collocated next to the LED on the rear panel; one unit will be able to go a little lower than the other so you will need to match them to the lowest common setting (mine are set to about 38.5V). Hook up the S-350-27 unit in series and then tune it down as low as it will go which should render the assembly down to about 63V; it is a tiny bit high for 15S LiPo but good enough for use with a BMS.

Shopping thoughts:
Though the -48 units are easy to find, the -27 model took some searching; it wasn’t my first choice but now I wouldn’t trade it away.

Flogging the Thermistor Mod:
There are pages of information given to modifying the Thermistor to modulate the fan on the S- units: Some say remove and short the contacts, other say put a POT parallel. I did the POT thang; it’s hokey but it works. Here’s what you get:
  • Without the mod, when this unit gets hot and turns the fan on to cool down, it will drop the output voltage in half - which IMO it is a terrible design. Adjusting SVR1 to compensate did not change the output; my S-350-27 unit was fixed at 13V.
  • With the mod in place, wait till the unit gets hot and the fan begins to spin. Adjust the 1K POT until the fan stops ~ whalla; end of problem!
  • Or – you can elect to do what I did in the field which is to put the S- unit on the top of the stack and remove the lid when charging which effectively disconnects the fan. Done, no mod required, end of problem.
This now provides you with a CV supply!

Meanwell-6384-Assy0.jpg


Flogging the Current Mod:
This initial mod did not work on the S- unit: Following suggestions I put a 2k resistor in parallel with the R33 to cut down on its’ value to create constant current, however the mod had no effect. In the field my 15S9P battery pack still took 9 to 10 hours to charge and left me quite frustrated: The output of the assembly is 2x320 + 350 watts which should provide close to a kW during charge. However due to the hiccup mode I was lucky to get 500W. A strong angst towards resolving this mystery ensued :evil:

Shortly upon my return home I ripped out the S-530-27 circuit board, opened up Visio, and began to reverse the circuit.

Meanwell-6384-Assy2.jpg

Top-side of S-350 PCB. SVR1 is at the top-left, and SVR2 is at the top-right. The disconnected R33 mod is still visable.
In the center of the PCB you can see the 1k POT mod across the Thermistor.

Meanwell-6384-Assy3.jpg

Back-side of S-350 PCB.

Allow me to state that Visio is a flimsy bit of software for creating schematics; I miss the days of OrCAD, PADS, heck even AutoCAD could do a better job. But it’s the tool that I had on hand. So without further ado, here is the schematic for the S-350 series boards.


Click on the image to view the full size.

Interestingly one board serves all the models within this wattage range. Please note the board revision as yours may be older or newer. I can’t speak for clone boards; I bought the genuine Meanwell article, genuinely made in China.

:idea: With schematic in hand I now had a map to the device: The first thing I noticed is that there production adjustments to this design; little things give it away such as the ordering of resistor numbers and culling of parts (NS). It’s also possible that there was a cost-reduction performed, though clearly some pieces belong to other models (D14 lends a clue). One item not installed on my unit is the SVR2 and R37 which affects directly how much current the unit produces. Maybe early versions or models didn’t use R33; it’s all speculative. Regardless, after reviewing the schematic and spec sheets for the primary components (U1, U2, RS1, and Q1/2) I set about to put SVR2 back in.

Biked down to Radio Shack and for < $1.75 I picked up a 1k POT. The previous mod for R33 was removed (one or more resistors in parallel). Instinctively I placed a jumper across the empty R37 slot. Then I soldered the 1k POT to the SVR2 location and then made sure the POT was at max ohms which gave a reading across R33 initially at 621 ohms down to about 370. Reassembled the whole unit sans cover, connected the assembly to my commuter battery pack (15S3P drained to about 55.2V), and then plugged it in. Within moments of warming up the Meanwell array was back to pulse-charging: Effectively no change to previous behavior.

Meanwell-6384-Assy1.jpg

SRV2 installed and Jumper placed across R37.

With a thin screwdriver I carefully, very slowly turned SVR2 backwards, reducing the resistance – and almost immediately I began to audibly hear the durations between the pulses grow shorter and shorter, suggesting that the charging was going faster and faster. Using the CA to monitor the present voltage confirmed this point. Cautiously hesitant at first, I proceeded until the unit stopped pulsing altogether. With the CA, I could actually see the battery rising a tenth of a volt at a time over a span of a couple of minutes; it was quite exciting! I triple-checked the present trim-spot and I felt the units for excessive heat (none beyond mildly warm). Satisfied, I made dinner whilst the assembly charged-away.

Within one hour my pack was completely charged. Now I gotta tell you that with my original charger it would have taken 5-6 hours, and with this charging assembly before this moment it would have taken nearly 3 hours. Problem resolved :mrgreen:

FWIW, measuring across R33 with the SVR2 POT in the present setting, the ohms read close to zero.

Musings:
I’d really like to understand why these units were adjusted to pulse-current and not to provide constant current out of the box. Or is it simply the way we employ the device that if goofs with sense circuitry? :roll:

Caveats:
When the charger is going full-blast, the two SP-320-48 fans are loud and going full speed. I can deal with that 8)

Schematic Warrantee:
The diagram provided has not been cross-checked; being human I am certain that there are errors. If someone would be so kind as to check me I would be truly grateful. Also, there is no plan to create a parts list, however if you forward the values for components then I will add them. Lastly, I apologize for the lameness of some of the symbols I used; Visio falls short on electrical shapes and I just did the best of what I could with what I had. All the transistors are presumed to be TO92A and PNP; they need to be checked. Jumpers are not shown on the schematic (there wasn’t a convenient Visio symbol available).

Conclusions:
Finally, I think I achieved what I paid for :wink: :mrgreen: :!:

Worries:
None mate. Between this and dinner, I found myself sated... for the moment :twisted:

Onward, to the next challenge.
Enjoy, KF
 
Kudo - Nice job, especially on the reverse engineering.

For schematics - Express SCH isn't too shabby. And its free :)
 
Kudo,

Thanks for the schematic work... odd that your mods didn't work like the rest of us, perhaps you did the current limiting before the fan mod... the fan does draw current screwing with the compensation amps (op amp inputs)

I'll double check your schematic against my S-350-48 to confirm they are close or identical - document the variances and post and delta to what you have already :)

-Mike
 
I love my s-350-27's ! I was able to cut one of the three shunts, this dropped current down to 12.5A , roughly what I was looking for!!! Another meanwell data point is my new S-350 48v, which I have hooked in series for an 84v 20s charger, are providing 345w each out of the box, with no mods !!! Well I did the fan mod , replacing the thermister with a 1000ohm resistor. Did a few cycles with the stack and works perfectly!!!!! the ebay item is here http://cgi.ebay.ca/48V-DC-7-3A-350W-Switching-Power-Supply-New-S-350-48-/180406617796?cmd=ViewItem&pt=LH_DefaultDomain_0&hash=item2a011286c4 the seller is rcnewworld. very nice charger so far!!!

I charge alot so I need a fast charger!! wow KF your running 5p thats a big pack !!! oh I see its NINE P!!!! :mrgreen: Im jelous!!
 
Nice job, thanks for taking the time to make this diagram. I've spotted one snag though, LF1 isn't a transformer, as drawn, but a line filter. If you turn it 90 degrees in the drawing, so that the line and neutral terminals connect to the switch part then that would fix it.

Jeremy
 
Nicely done. :)

The Meanwell's do make a fantastic charger when setup correctly. A hell of a lot better design for charging Lithium than 99% of the POS chargers that come with ebike batteries, and sometimes cheaper as well. :)
 
heathyoung, thanks for the suggestion; I will certainly check that out :)

Jeremy Harris said:
Nice job, thanks for taking the time to make this diagram. I've spotted one snag though, LF1 isn't a transformer, as drawn, but a line filter. If you turn it 90 degrees in the drawing, so that the line and neutral terminals connect to the switch part then that would fix it.

Jeremy

Many thanks Jeremy! I know it's still kinda goofy to look at, but how's this? :)


Click on the image to view the full size.

And thank you to all for your support, KF
 
Great effort! And you can now charge fast enough to use that connection you found behind the soda machine on your trip!
 
A quick question. I am currently running 2x meanwells, with thermister replaced with a 1000ohm resistor. This keeps fan always on. I did notice the fan speed goes to a slower speed when the voltage is almost equal (tapering off the charge at the end). Would I be corrent in surmising that since i changed the thermister to a resistor, I may lose the top end thermal shutdown? say for example my fan fails, or gets a chip stuck in in , will the unit respond as it should by limiting current in an over over temperature condition? Or is there some other devices in there that will detect this?

HOW DANGEROUS IS THIS MOD, will it start a fire in the box if the unit say is placed ontop of the fan hole?
 
Hydro,

The answer to your question is: I pulled a “Kirk” and changed the game towards my advantage. :lol: Allow me to direct you to the image below:

Meanwell-6384-Assy4.jpg

Left: Ready to operate. Right: Collasped for travel.

Firstly, I don’t care about the thermistor now because the lid is always off ~ as in forever, except when traveling (and even then it is not screwed down least it bar me from siphoning from a Pepsi machine…but I digress).

Secondly, I do not allow the vents to be covered: To ensure this, my assembly travels intact and ready to operate for any clandestine operation – I affixed four flat aluminum bars, two to a side as indicated in the image on the left, and allowing for a generous 1/2 inch air gap in-between. When traveling, to reduce space the whole assembly collapses nicely as indicated in the image on the right. Placing a small sheet of closed-cell foam between the units when collapsed reduces vibration damage.

And there you have it: Simply, effective, avoiding any possibility of fire, and yet still entertaining the hazard of electrocution which leaves some modicum of thrill when manipulating the device whilst it is energized. Hmm, yes, exciting is it not? :twisted:

The aluminum bars were made from one single piece of flat stock purchased at Ace Hardware, AND they have a good selection of metric fasteners. Best to take one of your units in for a proper fit (which is what I did). The screws are thumb-tight. Winged- or knobbed-screws would be a better choice than my dinky machine screws; meh ~ it’s on my Christmas list.

So to completely answer your questions:
I never place the cover on the unit when it is charging so it never gets hot enough to attract a moth. OK – I lied: I did find a dead moth snuggled up tight next to the heat sink when I went to extract the PCB for the shots above, but it was his fault for falling asleep and not mine. :roll:

If you block the hole whilst it is charging and you walk away… then someone should take your toys away. I mean it. :p

CYA: I never leave the room when the unit is charging even when the lid is off. Luckily I haven’t had any issues with batteries or electronics going south; when the unit is charging I am as attentive as a mother hen watching over her chicks, and this includes frequently checking the batt-pack voltage. When it is done, I unplug it. End of story. This is the way I use the device, and it is pretty fool-proof. :wink:

Since the mods, I have rapid-charged about every day; it is wicked-fast! I don’t let it approach the top voltage limit though; I don’t want to leave the unit running to find out if it will stop on its’ own. A smart BMS will help here. :idea:

Safe charging, KF
 
Schematic update:
L1 is linked to the center-tap of T1 (per suggestion by commanda)

Meanwell_S-350-27-v1.2.png


Enjoy, KF
 
When the fans run inside the Meanwells...which way to they work..pull air in through the fan or draw air out of the unit?

I am mounting my two Meanwells inside a case...to which I want to mount another fan..to keep inside of the outer case cool, so want to make sure the aditional fan I am adding is working in the same direction as the Meanwell fans
 
The fan, mounted in the cover, pulls air out.

Interestingly, only the middle unit is turning (top of the pair of -48s). I have a spare -48 so perhaps a swap is in order to determine if there is a fluke in the assembly.

The fan(s) do turn faster when under load.
~KF
 
Great, so I'll set my 'outer case' fans to blow air in to the housing, and the Meanwell fans will blow out

Thanks
 
I'll repeat here what I posted in another thread on this topic.

The best way to have the fan run constantly is simply to pull Q5 & place a link Collector-Emitter.
This way the thermistor over-temp protection remains intact.

Amanda
 
Hi Amamda
yes, I had see that and it was going to be one of the mods, I have read all that thread you refer to (Switch Mode PSU Mod thread). Will look at the board and the schematic and see which is collector and emmiter when i get the boards out.
Rather than removing Q5 and linking the collector - emmiter holes on the board, I did wonder about attaching flying leads to the transistor legs and running them to a simple switch, so the permanent fan on can be enabled/disabled, just put a switch across C-E rather than a hard link.

Other mods were going to be :
SVR1 switchable between two values to give 38 volt (running iCharger) and 41.5 volt for direct charging
R37 SVR2 resistor chain also switchable between two values to give two current limit values, relating to the two different voltages.

Seem to remember my present R33 is 330 Ohm, so the value of R37 SVR2 for one current limit was around 2k..(.combined 282 Ohm) and 7k to give 315 Ohm. giving current limits of 9.4 amp at 36 volt and 8.4 amp at 41.5 volt. If I use a 4PDT switch for this, I can switch bot PSU at once, giving instant switching between voltage and current limit.
May add a CC/CV board in future from ttpacks when they become available ..again switchable in/out to use when I have the two PSU's in series ( 2 x 41.5) to act as direct 20s pack charger
 
Update

For the past 11 months I have been using the Meanwell assembly as previously described with satisfaction for the configuration of which it was intended: Reasonable recharge times for Commuter and Cross-Country configurations.

However this year I am planning on going farther and faster for Cross-Country, and this may entail twice the number of batteries over last year. The spec prior was 15S12P = 63V @ 60Ah; expect this year to be close to 63V @ 120Ah. Well, simple math says this will take twice as long to charge. I decided to see if the charger assembly could be improved. :idea:

To lend a bit of perspective, when I first acquired the assembly and applied initial mods, the rate of charge from a nearly depleted battery was as:

  • Starting: 52.5V @ 7:45PM (0:00)
  • Ending: 63V @ 5:15AM (9:30)
We knew something was profoundly wrong with this setup, and after the road trip corrections were made (as noted at the start of this thread) to essentially cut that time in half. :)

When I originally bought the units I had planned on using this in the 84V configuration, and well now it escapes me why I bought Qty-3 of the SP-320, however I ended up with one spare. In another thread I did a quick study on the efficiency of the assembly and noted it was still not putting out the intended wattage. Today I decided to apply a bit more diligence, and to see what – if any – effect could be had by adding the 3rd SP-320 unit in parallel. Herein are the results:

MeanwellChargerStudy0.png


Comments:
  • As the Battery pack (15S3P) reaches near the terminal voltage I noted that at 62.5V the charger assembly was pulling 4.5A (about 540 Watts). I am not sure how this affects the overall function but it is a curious result.
  • As the voltage approached 63V the current rapidly began to drop below 4A, and because of this trend, reaching the terminal voltage takes quite a bit of time relative to the rest of the charging cycle. Example: It took 15 minutes to drop 0.1V to reach 63.1V, the current is down to 2.0A, although the charger assembly is noticeably quieter.
  • When they reached 1.7A there was a profound audible change as each unit independently tuned their fan rotations out of sync. With the current now down below 1.2A it is taking forever (relatively speaking) to reach 63.2V, though upon arrival the maelstrom of fans is eerily quiet.
  • It took 1.5 hours to go from 58.3V to 63.0, and yet it takes another 45 minutes to squeeze the last 0.2V! The quick conclusion is that we need another governing device: Set the assembly to a higher voltage and let her rip, though use another device to manage the cutoff/conditioning period. Regardless, I am dubious that adding the 3rd SP-320 enhanced the charging performance beyond contributing to the din of operation.
  • With that, I shall do another test in the reverse wherein I shall use only one SP-320 in series with the S-350 to determine if the “paralleling SP-320s is additive to the overall current output” theory is a bust.
Other notes: I was surprised to discover all the three A/C legs pulled current. Perhaps my low-tech methods are flawed, although I recorded the following:
GND = 0.04A; could be field leakage from the adjacent leg, and N = 0.1A less than L of which I did not expect.

Curiously concerned, KF
 
Addendum

Upon reflection, I had difficulty understanding why the current measurements came out the way they did, so I drew up a little schematic to explain to myself how and why. There’s a slight flaw in my spreadsheet, however the bottom-line remains unchanged.

MeanwellChargerStudy0-schematic.png


  • Point-1: Total A/C current was measured.
  • Point-2: S-350 current was measured. This is incorrect: The measurement here is the entire SP-320 subassembly current.
  • Point-3: SP-320 (2nd in line, designated as “B”)
  • Point-4: SP-320 (3rd in line, designated as “C”)
Redoing the math:
If we presume that the 3.01A erroneously assigned to the S-350 unit was instead the total SP subassembly then it is clear that the S-350 unit consistently pulls about 1.5A (180W). Still well below the rated current. :wink:

In the earlier measurement at Point-3 where the value was 0.15A, I presume that the “B” unit never engaged, leaving the “A” unit to pull the load. When we added the “C” unit and placed a load onto it, the load could be split between B & C, without A carrying any. Regardless, none of these units are operating at rated currents.

:?: Why? :?:

Second Test

I went digging around and found my other battery pack, the one I strap on if I know I’m going 100 miles etc. This pack is 15S4P, and it still had a good charge on it since the New Year, though there was enough off the top for a quick test of the theory.

MeanwellChargerStudy1.png


Just imagine the schematic above with the SP-320 units B & C disconnected. Made the inital measurements without a load and all was fine. As soon as I connected the battery pack the SP-320 (Unit-A) fan was off on a tear! The S-350 unit doesn’t have a fan; the cover is removed permanently. I had enough time to make the loaded measurements – but just barely, beginning with the total A/C current of the assembly. The battery was at 62.3V. By the time I got back to checking the total A/C current it had already dropped to 3.3A – indicating to me that the pack had already reached 63V and was now in the final conditioning phase. 8)

The behavior from this point forward was not at all like the previous test; the assembly continued to push hard to charge the pack and aggressively dropped the current. I noted at 0.54A that the fan had finally slowed down – maybe 15-20 minutes since charging began, and five minutes later I shut the whole process down as the fan turned to a whisper with the current drawing little more than 0.36A.

Therefore, I must conclude that paralleling Meanwell PSUs to create additive current does not pan out, and in fact may work against the individual sense circuitry to bring about a quick and resolute end to the charging cycle. :(

This brings me to my next hypothesis:
I no longer care about the thrift of my charger when the utility of the concept is an impediment to the goal and ultimate personal economy, the cost of which is measured as “my Time”. This leaves me free to entertain replacement in whatever form can be had.

I am still keen on Meanwell; they have a great product line and to date have served me well enough considering the pseudoscience of their employment in the present configuration (paralleled; that I can assure will change). However I am affirmed that placing these devices in series does work, though maybe not to the fullest extent by reasons yet to be discovered. Nonetheless, I am going to bite on this lead and configure anew with just two units, money no-object. It’s now a question about weight, size, speed, quality, balance, and cutoff – with alarm optional. That’s what I want, and I’m going to get it before I leave on my summer trek. :)

Stay tuned. KF
 
dbaker said:
What are your summer trek plans?
Hi db :)

I haven’t started the thread yet, but it will be similar to the Going to California 2010 thread – except this time I plan to do the entire journey by ebike.

In brief, I start in Redmond – Washington, head south along the backbone of the Cascades (Hwy 97), pick up Hwy 89 near Shasta, follow that to Tahoe, pick up Hwy 88 and head west over the Sierras to Jackson, where again I head south along the old Forty-Niner Hwy down to Fresno. For you see it was there that in my youth I leapt outta the ground like a durn weed, pulled up my gypsy roots and started wandering this old Earth.

Once I am good and watered-up with my ol’ pals and kinfolk I’ll head due west and pick up Hwy 1 in Watsonville, then head all the way north along the Coast Hwy, through San Francisco, Mendocino, The Redwoods, Eureka, Coos Bay, and on up till I can hang a reggie over to Portland. From there it’s the STP in reverse back to Redmond. In all, about 2000 miles.

That to me would be a satisfying holiday. I don’t think there’s room for the guitar though. Harmonica maybe. Jew’s Harp for sure. “The Plan” is mainly to just be “on the road”. I expect to camp overnight at RV parks possibly up to a handful of times, though motels figure big when they can be had; I like a soft cozy bed as well as anyone else.

~KF 8)
 
Back to the Meanwells

Seems to me there are basic flaws in how you have them setup, and how you are measuring

First off, you have nearly 1000w capacity with the 3 in parallel. You then go and stick a 350 w unit in series with the parallel output, limiting the output of the 3 in parallel to that of the single 350.
The idea of putting them in parallel is to give greater current, then you have stuck it in series with the 350, so the max current through the parallel string is never going to be more than what is going through the 350. The NEW one in parallel will achieve nothing

Secondly, you are measuring input current. Measure the output current instead. I have two Meanwell S -350-48 in parallel, and the only way to get them to load share nicely is by setting them up monitoring output current of the two simultaneously while under load and tweak the output voltage of one unit till load between them is balanced.
You may find if you have not modified the voltage adjustment pot on the front for a multiturn unit , then it is almost impossible to balance the current. This is because the adjustment is just too course and inprecise with the single turn trimmer pot. I have 16 turn cermet pots on mine and even using them, a quarter turn or less either way, can be enough to push the voltage of one unit up enough to make it take all or almost all the load.

So, monitor output current, drop the new supply, it is not helping. In fact having even two in parallel with the one 350 is a bit of a waste. Try two series strings in parallel
 
Alright Neil, let’s discuss:
We need to make some basic clarifications here…

  • Neither the Meanwell S-350 or the SP-320 are rated or certified for parallel operation. For some reason this became a large red flag when I went digging for alternatives.
  • The S-350 isn’t being driven to full potential. We can see that with the input current in comparison with the spec. Given the amount of time it takes to charge at voltage, we can also ascertain that the unit is underperforming, significantly – possible greater than 50%. Imagine it’s pulling 1.5A @ 120V; how does 180W of VAC magically translate to 350W at 63V? What have I missed?
  • Adding the 3rd SP unit confirms that I did not have bad SP units; that the performance was exactly the same. In addition, I was able to glean in detail the performance during the final 0.2V of the charging cycle. The final test with just one SP unit confirms there is contention between the paralleled units in the sense circuitry during the termination phase. Do you agree?
  • With regard to the SP-320 & Imbalance, I can see and accept your point easily; that seems quite reasonable, and would explain the artifacts of the observation. But then consider that if the stock POT is so delicate and difficult to bring into balance – then the entire assembly comes into question and is hardly road-worthy.
In fact Neil, you have stated clearly a trait that should have been mentioned back when people first began tinkering with these components: It takes a skilled technician with professional equipment to properly dial these devices into balance. I could not ask for any better validation. You Sir deserve a gold star! (sincerely) In other words, the modest tools at my disposal, with the resolution and measurement that I have, it is still not enough to bring these simple low-cost devices into balance; that we shouldn’t be paralleling these devices anyway - unless someone can come up with a “unified” sense circuit like in the HRPG, PSP, RCP, RSP models, and so forth. :)

Two series string in parallel:
That is a good idea. It is ultimately the direction I decided to go, except that instead of buying another S-350, thus having four units to lug around, I went the other direction because space is not my friend when travelling cross-country, which profoundly is the reason... the impetus for the investigation: To seek the reduction of weight & volume, yet enhance performance.

I already know that my battery array is large enough to handle a 2kWh fire hose; that’s like 1C to my Commuter pack. The challenge is at the other end of the telescope: Maximizing the Source. Therefore I began looking at putting together two matched units in series that could operate within the 15-20A household breaker thresholds. Interestingly enough – there is a performance/consumption leap between the 320/350W range and 600W, with choices in-between being less advantageous.

These are the other units I reviewed:
  • HRP-600: Weight = 1.5kg; the best of the lot. Considering I am replacing three units weighing 1.1kg each (or potentially adding a 4th) I am actually coming out weighing less by grams and with less complexity. There is no support for current-sharing, though that is not my application; however it does CC like the S-350. Meets and exceeds the performance specifications of the newer SP series, and with better efficiency. Price begins around $168/unit; not the cheapest alternative, though everything else is positive.
  • HRPG-600: 1.58kg. Current-sharing up to 2400W, CC, good specs on par with the HRP. Price begins around $185.
  • PSP-600: 1.9kg. Current-sharing up to 2400W, CC, specs less whelming than the HRP, although the price is about $145.
  • SE-600: Didn’t seem much different than the SP series. Weight = 2.1kg. No CC. Price begins around $110.
  • SPV-600: Weight = 1.9kg. Not yet available.
Others I looked at were farther out of spec and more costly. There was a 500W unit but it was a power-sucking mad-heifer. A 450W unit was in the same bloat. Clones abound, but so does my fear of them. Lastly, the competition couldn’t compare on price.

Can you guess which way I went? :)
~KF
 
Kingfish said:
  • Neither the Meanwell S-350 or the SP-320 are rated or certified for parallel operation.

  • Agreed...The bigger ones are..at least some of them...the 1.5 or 3 kw versions certainly do have a parallel facility

    Kingfish said:
    [*]The S-350 isn’t being driven to full potential. We can see that with the input current in comparison with the spec. Given the amount of time it takes to charge at voltage, we can also ascertain that the unit is underperforming, significantly – possible greater than 50%. Imagine it’s pulling 1.5A @ 120V; how does 180W of VAC magically translate to 350W at 63V? What have I missed?

    I am thinking maybe be what you have missed, and I also missed, ( and this is possibly borne out by what you say later), is that maybe the current limit is one of the 320 units that you have in parallel.

    You next say

    Kingfish said:
    [*]Adding the 3rd SP unit confirms that I did not have bad SP units; that the performance was exactly the same.

    If adding the third unit made no difference..then maybe you only have one of the 320's taking al the load and that is the one limiting the current


    Kingfish said:
    In addition, I was able to glean in detail the performance during the final 0.2V of the charging cycle. The final test with just one SP unit confirms there is contention between the paralleled units in the sense circuitry during the termination phase. Do you agree?


    Yes, I have noticed that with two S-350's in parallel, they only stay in balance and equally sharing the load at a particular voltage...above and below this they do drift out...but not by enough to be too concerned



    Kingfish said:
    [*]With regard to the SP-320 & Imbalance, I can see and accept your point easily; that seems quite reasonable, and would explain the artifacts of the observation. But then consider that if the stock POT is so delicate and difficult to bring into balance – then the entire assembly comes into question and is hardly road-worthy.


Ah...I was not implying that the stock pot is delicate, it is just that the adjustment is very coarse because it is less than one turn...maybe 270 to 300 degrees so it is therefore difficult to adjust accurately, so setting an accurate voltage and matching it to another can be tricky. Consider then a 16 turn pot, ...same value 2k or 5k whatever..but adjustable over 16 turns...far finer control is possible.

I have two meanwells...each with 4 pots all 16 turn cermets. I replaced both the stock VR1 and also added SVR2 the current limit pot. The reason I have 4...they are switchable, so I can have two preset voltage limits, and two preset current limits. My set up consists of two S-350-48's 8 cermet 16 turn trimmers, and two 4 pole change over switches, so i can switch between parallel and series, and also switch between two voltage/current settings.... 41.5 volts and 8.5 amps and 36 volt and 9.7 amps. When I have it switched to 41 volts, I also switch them in series for bulk charging direct to the pack. Or in parallel a the lower voltage to run the charger for a balance charge

Kingfish said:
In fact Neil, you have stated clearly a trait that should have been mentioned back when people first began tinkering with these components: It takes a skilled technician with professional equipment to properly dial these devices into balance.

have to disagree there. Correct that they are not designed for parallel operation, but if I can get them running balanced in parallel, anyone can. With the 16 turn pots and two cheap 10amp DVM's I can easily get mine in balance. I bought 4 of the el cheapo DVm's from e-bay for $0.99 . They are only 10 amp, but for my 350's 10 amp is fine. All I did first off was ran all 4 of them in series, with my pack and the Meanwells, and chose two that gave he same current reading. The put them in sereis with each parallel PSU output, load up the pair of them with a battery, and then adjust the voltage of one or the other to equalise the current between the two


Kingfish said:
I could not ask for any better validation. You Sir deserve a gold star! (sincerely)

Very kind, but hardly deserved.


Kingfish said:
In other words, the modest tools at my disposal, with the resolution and measurement that I have, it is still not enough to bring these simple low-cost devices into balance;

get a few of those cheap DVM's, they are good enough for this job..


Kingfish said:
unless someone can come up with a “unified” sense circuit like in the HRPG, PSP, RCP, RSP models, and so forth. :)

Now, that is the real answer.


Much earlier in this thread, you mentioned using the thermistor and the pot mod for the fan control. Far better is bridging the collector-emmiter of the Q5 transistor. I put a switch across mine, so I have always or full auto.


Well which ever way you decide to go with those other supplies, let us know.

It is a shame you are limited over there to what is it? 110 or 120volt AC in the household supply

Regards

Neil
 
How about these:

http://cgi.ebay.com/20-x-Cherokee-International1200W-PSU-CAR1248FPB-1A-/320694703436?pt=LH_DefaultDomain_0&hash=item4aaae4d54c
 
dbaker said:
How about these:

http://cgi.ebay.com/20-x-Cherokee-International1200W-PSU-CAR1248FPB-1A-/320694703436?pt=LH_DefaultDomain_0&hash=item4aaae4d54c

Could you be a bit more definitive?

~ KF :?
 
Back
Top