mwkeefer
1 MW
Hello all,
I finally got to crack open the 48v unit and mod it tonight for bulk charging of my 15s packs... I've limited the cutout to 61.8v which is only 4.12v per cell just because I don't have the right Zener Diode on hand.
There is no more time for working on documentation tonight (though I do have pics) so I figured I'd just dump my entire "notes" - even though I may refer to users and people, these are just rough draft of a review on this product and seperate mod thread for all meanwells which I am in the process of writing (about 50% complete). I have only 1 supply available to me to review still the 145w 24v model which I'm looking at as a slow 15s charger (really slow) for me and some others.
Below is a literal, copy and paste from OneNote where I keep all my engineering and R&D notes - they aren't pretty and probably really "train of thought" but there have been so many questions about how to reverse engineer these lately to get a desired output voltage and this "engineering" (I use the term loosely) notes may help people to do just that... in a sane and well thought out mannor (well.... not my notes literally but I do cover all the caveots - I think).
I will update the following when I craft the final review / how-to:
Hope this helps!
-Mike
PS: If this should be somewhere else besides Reviews and Testing (Like rantings of a lunatic) please let me know... I'll gladly remove and repost in a more appropriate category.
----
S-350-48 from SureElectronics.net
Friday, November 13, 2009
6:18 PM
Initial impressions:
Quality piece - it worked better than expected!
Out of box voltage: 47.8v
Out of box range: 38.8 - 58v (almost enough - 4 more volts is all I need for 15s - 62.25v)
Disassembly Notes:
1.) One screw seemed to have threads which were off a bit… I had to use a special screw driver on this one and be very, very careful. Drilling it out would have been the solution if the head had gone bad (or notching with a dremel and using a flat head).
[Voltage Limiter Circuit]
R25: 2k, 5%, 1/8th watt - (red-black-red-gold)
R61: 10k, 5%, 1/4 watt - (brown-black-orange-gold) *Note the 1/4w - first meanwell with this large R61
C35,36,38: 100v330uf - (highest voltage of any meanwell I've seen yet and should handle to 85v fine.)
ZD1: 1N4759 - 62v, 4mA Zener Diode - this will work for 15s without mod @ 4.15v per cell = 62.25v
Shunts: 3x ?? - ** These are the same as in the 24v model but they have been "smashed" with a hammer to raise their effective resistance by narrowing the pathway - interestingly enough, this could be a calibration method used by the manufacturer but my guess is it's the builders attempt at being as frugal as possible… nice part is now we know it's cool to do this = )_
SVR1 - 1k - .6 ohm or 600 milliohms
Replaced with precision SVR1 - 10 turns - 1K, top voltage=58.1v (so range is same but precision - 10 turn is far better… I wonder what going to 5k will result in for range?)
If I test for resistance on the between leg 3 of the pot and the far side of R25 I measure 2.xK with the pot reading milliohms (no resistance) this is the point I get the highest voltage… everything else is lower. This means to increase the top or maximum voltage the supply can achieve I must lower the value of R25 some… how much is the question?
Since I'm flying blind here… lets turn the pot all the way up (down) and test the resistance again: 2.82k max resistance @ 39.1v
Now assuming a linear scale… for a voltage output swing of: 58.1 - 39.1 = 19 volts we have a maximum swing of 2.82 - 1.92 = 0.9 k or 900 milliohms call it 1K for argument sake.
19 / .900 = 21.1111 milli ohms per volt so…
63 - 58.1 = 4.9 * 21.111 = 103.4439 more milli ohms reduction to R25 (Correct me on this if I am wrong or doing the math wrong)
So we are trying to find a resistor of 1.5k actual value (1.6 measured - 100k), time again to turn to my trusty bin of scrap resistors:
Looking for: 1.5k(brown, green, red), 1.2, 1.8k resistors to find the best actual value match (closest to 1.6k)
Changing to a 5K pot (10 turn) will really fix the bottom end range… I will do this once I test the 1.6k upgrade for 63v (roughly)… In my 24v supply tests this change gave me a 6.9v low end (that's range).
Ive changed my mind… for the first test phase run of this "charger" in brick mode… I will set the cutout to 4.1v per cell just to keep a margin of safety coming from the charger so my actual output will be: 61.5v which will still require the mod and even falls within the rated spec of the diode (62v) which I am counting on some tolerance with so I can get to 62.25 without the limiting effect kicking in. There is 0 danger of popping these in no-load testing, only when they are asked to limit and there is current in place will they pop.
Out of circuit the resistor R25 tested to 1.98k - well within tolerance… adjusting the assumption above:
Found a 1.5K which tests out of circuit to: 1.46k, close enough for my needs… just need to set the pot to highest resistance position before power up.
2.35k - 1.41k is my new resistance range… in theory, this will give me far past 62v output voltage range and will limit the low end voltage (I've only shifted the scale by changing just 1 resistor, changing the SVR1 for a 5 or even 10K will extend rather than move the range of voltage).
I preset the variable network to 1.95 K - just a little more resistance than our previous lowest… that will put the voltage on power up around 57-58v and I can adjust up from there gradually while monitoring other components… time for power up:
Starting Voltage (with 1.95k of impedance in the voltage regulator network): 56.5v (@ 1.95k calibrated)
So I nudge it up a bit… 58.1 now passing the delivered range…
59.1v - grinding my teeth (you would be too if you realized im running this PCB on styrofoam board - elmers style).
60v - I could stop here and consider it a successful first test but… nope, lets go.
61.5 the max I need for now… keep going 15 * 4.25 = 63.75
62.2 - 62.3 reached which is 4.15v per cell and meets my actual maximum needs… however, lets push to the max.
Hmmm… interesting 62.5 seems to be the limit - probably the Zener diode protection kicking in so lets dial back to 61.6v cutout (4.1 without balancing seems safe - I have active hvc for this method at cell level.).
Now… I begin to think about modes / gears like the chinese charger manufacturers speak of… just switching between resistors would do this - so why won't they offer an adjustable pot in addition to the gears - sales volume, plain and simple.
So… once the power was removed and the LED died out (caps drained, always wait for this before touching) I measured the impedance of the voltage regulator (I am assuming voltage of 61.55v as it would teeter between 61.5 and 61.6 on my crappy volt meter) to be:
** Another trick I've learned… disconnect and reconnect power without changing pot once you have desired voltage showing on volt meter… make sure it powers up to the same voltage - if not… adjust the circuit and try again.
61.55v (4.1033v per cell) : 1.81K Ohms
61.8v (4.12v per cell): 1.81K (my meter isn't getting any better at measurements than this.).
62.2v (4.1467v per cell): 1.82K Ohms
62.5v (4.1667v per cell): 1.79K Ohms
63v (4.2v per cell): not possible - diode override I believe… there is still adjustment range in the circuit.
So I need to add some diode, not much 5v in series or upgrade to a single 1N4760 (68v, 4ma zener) which would allow full adjustment to 63 and then some (could be dangerous). I think instead I will combine 2 in series for a perfect 64v maximum cutout (true 15s overvoltage protection)… I will need to test the cutoff curve of these to be sure… for now I will live with 62v 4.1333 maximum charge per cell. This should be safe since the supply is not really operating far beyond it's values… I will test a quick "burst" in a moment with the eagle tree and a pack which I have run down to 3.3v per cell (it's out of balance too) but it's a single 5A pack. In theory the current limiter designed in this unit feeds back into the TL494CN and should automatically adjust based on the new voltage limiter impedance to it's factory preset which in the 48v supply is: 48v @ 7.3A or 48 * 7.3 = 350.4 watts so we can expect (hope) to see the limiter keeping the supply current at start of charge 15 * 3.3 = 49.5 : 350.4 / 49.5 = 7.0788 A and when we reach HVC the supply itself should only provide 350.4 / 62 = 5.6516 A (still 1C hence the happyness with 4.13 cutout until I log this properly and see the batteries reduce their consumption as they reach HVC) - otherwise I believe 4.13 is still below the 90% mark and I know these will fast charge until 90% (as seen in the past on other experiments) capacity or so… that should make this safe(ish).
I set the actual supply output voltage to 61.8v to keep it under the Zener Diode rating since I don't feel like replacing it yet (but I should since I have the PCB out of the case and have to reseat the heat sinks in either condition). That keeps the evenly (hopefully) distributed cell voltage at cutout to be 61.8 / 15 = 4.12 v per cell (very safe) and I have cell log 8 connected with an active HVC interface on each of the 3 5S packs in series which I will be test charging.
*** Be sure to power off the unit and repower to recheck the output voltage… the first time I did this on 61.8 it came up at 61.6v - I turned it up, unplugged… waited for v to hit 0 on the DVM and plugged in again to retest… tada 61.8v = )_
So… time to button this supply back up (oh yea… have to measure impedance for reference:
1.81K Ohms = 61.8v calibrated output.
One thing I should note… this is HIGH VOLTAGE technically and not only voids the warranty but could cause all sorts of issues including YOUR DEATH if your not careful or don't know what your doing. To be perfectly honest… I will not run my own supplies with this mod alone… since it's over the 60v mark I have a fear of arching and shorts… so I will line the inner case (below the PCB) with non-conductive material and take other required high voltage precautions. I will post the final spec which should be "safe" but I still don't recommend anyone else follow this thread - it is for edumactational purposes only. If you don't know what your doing with power / ac, then go to an electricican or TV repair shop with a power supply and print out of this thread… he should be qualified and competent to make the modifications needed - can't say he/she will be willing!
Next installment, making sure the current and voltage remain within spec - WITH LOGGING and pretty GRAPHS.
I finally got to crack open the 48v unit and mod it tonight for bulk charging of my 15s packs... I've limited the cutout to 61.8v which is only 4.12v per cell just because I don't have the right Zener Diode on hand.
There is no more time for working on documentation tonight (though I do have pics) so I figured I'd just dump my entire "notes" - even though I may refer to users and people, these are just rough draft of a review on this product and seperate mod thread for all meanwells which I am in the process of writing (about 50% complete). I have only 1 supply available to me to review still the 145w 24v model which I'm looking at as a slow 15s charger (really slow) for me and some others.
Below is a literal, copy and paste from OneNote where I keep all my engineering and R&D notes - they aren't pretty and probably really "train of thought" but there have been so many questions about how to reverse engineer these lately to get a desired output voltage and this "engineering" (I use the term loosely) notes may help people to do just that... in a sane and well thought out mannor (well.... not my notes literally but I do cover all the caveots - I think).
I will update the following when I craft the final review / how-to:
Hope this helps!
-Mike
PS: If this should be somewhere else besides Reviews and Testing (Like rantings of a lunatic) please let me know... I'll gladly remove and repost in a more appropriate category.
----
S-350-48 from SureElectronics.net
Friday, November 13, 2009
6:18 PM
Initial impressions:
Quality piece - it worked better than expected!
Out of box voltage: 47.8v
Out of box range: 38.8 - 58v (almost enough - 4 more volts is all I need for 15s - 62.25v)
Disassembly Notes:
1.) One screw seemed to have threads which were off a bit… I had to use a special screw driver on this one and be very, very careful. Drilling it out would have been the solution if the head had gone bad (or notching with a dremel and using a flat head).
[Voltage Limiter Circuit]
R25: 2k, 5%, 1/8th watt - (red-black-red-gold)
R61: 10k, 5%, 1/4 watt - (brown-black-orange-gold) *Note the 1/4w - first meanwell with this large R61
C35,36,38: 100v330uf - (highest voltage of any meanwell I've seen yet and should handle to 85v fine.)
ZD1: 1N4759 - 62v, 4mA Zener Diode - this will work for 15s without mod @ 4.15v per cell = 62.25v
Shunts: 3x ?? - ** These are the same as in the 24v model but they have been "smashed" with a hammer to raise their effective resistance by narrowing the pathway - interestingly enough, this could be a calibration method used by the manufacturer but my guess is it's the builders attempt at being as frugal as possible… nice part is now we know it's cool to do this = )_
SVR1 - 1k - .6 ohm or 600 milliohms
Replaced with precision SVR1 - 10 turns - 1K, top voltage=58.1v (so range is same but precision - 10 turn is far better… I wonder what going to 5k will result in for range?)
If I test for resistance on the between leg 3 of the pot and the far side of R25 I measure 2.xK with the pot reading milliohms (no resistance) this is the point I get the highest voltage… everything else is lower. This means to increase the top or maximum voltage the supply can achieve I must lower the value of R25 some… how much is the question?
Since I'm flying blind here… lets turn the pot all the way up (down) and test the resistance again: 2.82k max resistance @ 39.1v
Now assuming a linear scale… for a voltage output swing of: 58.1 - 39.1 = 19 volts we have a maximum swing of 2.82 - 1.92 = 0.9 k or 900 milliohms call it 1K for argument sake.
19 / .900 = 21.1111 milli ohms per volt so…
63 - 58.1 = 4.9 * 21.111 = 103.4439 more milli ohms reduction to R25 (Correct me on this if I am wrong or doing the math wrong)
So we are trying to find a resistor of 1.5k actual value (1.6 measured - 100k), time again to turn to my trusty bin of scrap resistors:
Looking for: 1.5k(brown, green, red), 1.2, 1.8k resistors to find the best actual value match (closest to 1.6k)
Changing to a 5K pot (10 turn) will really fix the bottom end range… I will do this once I test the 1.6k upgrade for 63v (roughly)… In my 24v supply tests this change gave me a 6.9v low end (that's range).
Ive changed my mind… for the first test phase run of this "charger" in brick mode… I will set the cutout to 4.1v per cell just to keep a margin of safety coming from the charger so my actual output will be: 61.5v which will still require the mod and even falls within the rated spec of the diode (62v) which I am counting on some tolerance with so I can get to 62.25 without the limiting effect kicking in. There is 0 danger of popping these in no-load testing, only when they are asked to limit and there is current in place will they pop.
Out of circuit the resistor R25 tested to 1.98k - well within tolerance… adjusting the assumption above:
Found a 1.5K which tests out of circuit to: 1.46k, close enough for my needs… just need to set the pot to highest resistance position before power up.
2.35k - 1.41k is my new resistance range… in theory, this will give me far past 62v output voltage range and will limit the low end voltage (I've only shifted the scale by changing just 1 resistor, changing the SVR1 for a 5 or even 10K will extend rather than move the range of voltage).
I preset the variable network to 1.95 K - just a little more resistance than our previous lowest… that will put the voltage on power up around 57-58v and I can adjust up from there gradually while monitoring other components… time for power up:
Starting Voltage (with 1.95k of impedance in the voltage regulator network): 56.5v (@ 1.95k calibrated)
So I nudge it up a bit… 58.1 now passing the delivered range…
59.1v - grinding my teeth (you would be too if you realized im running this PCB on styrofoam board - elmers style).
60v - I could stop here and consider it a successful first test but… nope, lets go.
61.5 the max I need for now… keep going 15 * 4.25 = 63.75
62.2 - 62.3 reached which is 4.15v per cell and meets my actual maximum needs… however, lets push to the max.
Hmmm… interesting 62.5 seems to be the limit - probably the Zener diode protection kicking in so lets dial back to 61.6v cutout (4.1 without balancing seems safe - I have active hvc for this method at cell level.).
Now… I begin to think about modes / gears like the chinese charger manufacturers speak of… just switching between resistors would do this - so why won't they offer an adjustable pot in addition to the gears - sales volume, plain and simple.
So… once the power was removed and the LED died out (caps drained, always wait for this before touching) I measured the impedance of the voltage regulator (I am assuming voltage of 61.55v as it would teeter between 61.5 and 61.6 on my crappy volt meter) to be:
** Another trick I've learned… disconnect and reconnect power without changing pot once you have desired voltage showing on volt meter… make sure it powers up to the same voltage - if not… adjust the circuit and try again.
61.55v (4.1033v per cell) : 1.81K Ohms
61.8v (4.12v per cell): 1.81K (my meter isn't getting any better at measurements than this.).
62.2v (4.1467v per cell): 1.82K Ohms
62.5v (4.1667v per cell): 1.79K Ohms
63v (4.2v per cell): not possible - diode override I believe… there is still adjustment range in the circuit.
So I need to add some diode, not much 5v in series or upgrade to a single 1N4760 (68v, 4ma zener) which would allow full adjustment to 63 and then some (could be dangerous). I think instead I will combine 2 in series for a perfect 64v maximum cutout (true 15s overvoltage protection)… I will need to test the cutoff curve of these to be sure… for now I will live with 62v 4.1333 maximum charge per cell. This should be safe since the supply is not really operating far beyond it's values… I will test a quick "burst" in a moment with the eagle tree and a pack which I have run down to 3.3v per cell (it's out of balance too) but it's a single 5A pack. In theory the current limiter designed in this unit feeds back into the TL494CN and should automatically adjust based on the new voltage limiter impedance to it's factory preset which in the 48v supply is: 48v @ 7.3A or 48 * 7.3 = 350.4 watts so we can expect (hope) to see the limiter keeping the supply current at start of charge 15 * 3.3 = 49.5 : 350.4 / 49.5 = 7.0788 A and when we reach HVC the supply itself should only provide 350.4 / 62 = 5.6516 A (still 1C hence the happyness with 4.13 cutout until I log this properly and see the batteries reduce their consumption as they reach HVC) - otherwise I believe 4.13 is still below the 90% mark and I know these will fast charge until 90% (as seen in the past on other experiments) capacity or so… that should make this safe(ish).
I set the actual supply output voltage to 61.8v to keep it under the Zener Diode rating since I don't feel like replacing it yet (but I should since I have the PCB out of the case and have to reseat the heat sinks in either condition). That keeps the evenly (hopefully) distributed cell voltage at cutout to be 61.8 / 15 = 4.12 v per cell (very safe) and I have cell log 8 connected with an active HVC interface on each of the 3 5S packs in series which I will be test charging.
*** Be sure to power off the unit and repower to recheck the output voltage… the first time I did this on 61.8 it came up at 61.6v - I turned it up, unplugged… waited for v to hit 0 on the DVM and plugged in again to retest… tada 61.8v = )_
So… time to button this supply back up (oh yea… have to measure impedance for reference:
1.81K Ohms = 61.8v calibrated output.
One thing I should note… this is HIGH VOLTAGE technically and not only voids the warranty but could cause all sorts of issues including YOUR DEATH if your not careful or don't know what your doing. To be perfectly honest… I will not run my own supplies with this mod alone… since it's over the 60v mark I have a fear of arching and shorts… so I will line the inner case (below the PCB) with non-conductive material and take other required high voltage precautions. I will post the final spec which should be "safe" but I still don't recommend anyone else follow this thread - it is for edumactational purposes only. If you don't know what your doing with power / ac, then go to an electricican or TV repair shop with a power supply and print out of this thread… he should be qualified and competent to make the modifications needed - can't say he/she will be willing!
Next installment, making sure the current and voltage remain within spec - WITH LOGGING and pretty GRAPHS.