Tiberius
10 kW
mwkeefer said:
Good work, Mike, well done.
Nick
Making a CC/CV charger
- Thread starter Tiberius
- Start date
Good work, Mike, well done.
Nick
etard
100 kW
Thanks Mike, thats just what I wanted to hear. So it appears the factory Bosch charger takes each battery pack up to 41.5 volts and does so with 3 amps which takes about 40 minutes for a single pack. So I am thinking if I run the 6 in series, and set the CV mode to the same voltage, the power supply would make a good charger that can charge me up in about an hour. The tests done of the these cells indicate a very steady curve with all cells usually identical.
Guidance needed:
So if the power supply puts out 48 volts, and the pot on the supply only changes the HVC for the Cv mode, how do I charge at 42 volts? Will I need as custom shunt to lower the voltage output to 41.5 volts? Or will I need to get resistors to lower the voltage? Or does the adjustment screw do this?
Also, is the voltage steady on these supplies, or do they jump around alot possibly causing harm to my unprotected batteries?
Thanks for the help here, fast charging is one of the biggest problems and you guys are solving it here!! Very exciting!! AND, the charger is portable and cheap.
Guidance needed:
So if the power supply puts out 48 volts, and the pot on the supply only changes the HVC for the Cv mode, how do I charge at 42 volts? Will I need as custom shunt to lower the voltage output to 41.5 volts? Or will I need to get resistors to lower the voltage? Or does the adjustment screw do this?
Also, is the voltage steady on these supplies, or do they jump around alot possibly causing harm to my unprotected batteries?
Thanks for the help here, fast charging is one of the biggest problems and you guys are solving it here!! Very exciting!! AND, the charger is portable and cheap.
Tiberius
10 kW
etard said:
etard,
That's what the CC feature does. A max current is set and if that is reached, it lowers the output voltage.
So to start with the voltage will sit at 41 or 42 V or whatever is needed to put that current into the batteries.
As they charge up, the voltage will rise, until it hits the uper voltage limit; then the PSU changes to CV mode and stops the voltage going any higher. At this stage the current will taper off towards zero.
Nick
heathyoung
100 kW
There are some meanwell power supplies that incorporate a CC mode as part of their protection (rather than hiccup mode).
You still need to do some end-of-charge point termination though.
You still need to do some end-of-charge point termination though.
mwkeefer
1 MW
heathyoung said:
Actually most of the Mean Wells I've looked at have the circuitry to do CC... hiccup mode is just the "hard" setting of the components in the PCB.
I'm still waiting on the manufacturer (not vendor) to send the various configs for end user sale (ie: various voltage ratings and current limiting circuits including 48v adjustable v and c) which we will then be able to scale to whatever voltage we require.
Currently... I'm playing with a set of CellLog 8's from hobbyking, they output the voltage with .000 resolution at a cell level, calculate the delta, etc.. They have a USB based logview interface but the back side of the USB is TTL level serial IO... I am working on capturing that serial IO using atmega via 3 bitbang style serial IO pins (IE: Digital Input Lines) which would allow for up to 24s cell level voltage monitoring.
I realize this isn't "perfect" because it's using the celllog8s for input but... I am also working on the slave module around an atmega48 to do up to 8 (using a single 16ch analog mux per slave) cell monitoring and wanted to get the circuit for controlling the power supply (meanwell) from an MCU before I will be done with the slave units.
-Mike
mwkeefer
1 MW
Hey all,
Just a little update - progress data, whatever:
[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 still working on the 5,12 and 48v supply mods and documentation but have to finish the proper CC circuit to populate the S350 PCB (the circuit is there already,just hard current limited for various models) as we will need to limit current for some voltages to prevent melt down of these supplies.
Hope it helps!
-Mike
PS: This is cross posted in the Review & Mods: Meanwell S-350 thread I started a while back, I will continue to cross post upgrade tech details in both threads (to avoid anyone missing the important stuff).
Just a little update - progress data, whatever:
[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 still working on the 5,12 and 48v supply mods and documentation but have to finish the proper CC circuit to populate the S350 PCB (the circuit is there already,just hard current limited for various models) as we will need to limit current for some voltages to prevent melt down of these supplies.
Hope it helps!
-Mike
PS: This is cross posted in the Review & Mods: Meanwell S-350 thread I started a while back, I will continue to cross post upgrade tech details in both threads (to avoid anyone missing the important stuff).
mwkeefer
1 MW
Hello All,
I have seen Nick's earlier posting about limiting CC mode by removing shunts (perhaps I even suggested it) but since the PCB seems to have positions for this limiting (adjustable) via SVR2 which is missing (along with a transistor, some resistors, etc) I am still trying to determine how to modify this circuit (which I think is feedback loop 2 for the PWM IC (494)) then we should be able to not only configure the brute "Range" of current by switching in various resistor networks but also enable fine adjustment within the current range we have selected - all without removing or bridging shunts (great brute force method which I am about to do to a 48v supply).
In testing I modified an S-350-48:
Stock - Replacement
SVR1 1K - SVR1 10K - 15 Turn
R25 2K - 1.5K
ZD1=62v - 62v + 12 + 12 = 96v maximum (I think this may have been the fatal move, next mod will be 62 + 10 + 10 = 82v cap
So I set the output voltage to 62.5 (full tested range was 14v - 85v, I didn't push it higher and that was no load).
In my previous unmodified tests to 61.8v I noted the current going from 10A to 11 for a brief moment (cells we about 3.7v for 55.5v so I was seeing power levels of 610 watts. That was without modifications, no limiting was kicking in.
The process cotinued until reaching about 4v per cell where the current began to taper down over 20-30minutes (I was charging a 15S1P 5AH pack) - Even once I reached 4v per cell I was seeing 9.5A for 570w.
The next test was post upgrade with the voltage limit set for 62.25 and it popped a fuse!
I'm going to bulk charge @ 10A using my iCharger on a 5S pack to follow the curve - then try to mod these to follow the same curve (not sure how yet but I'll figure it out).
The point of the post:
I see how by feeding back a small current into the feedback loop of the PWM controller you can limit the current... couldn't we make an external version of this adjustable and vary the current delivery of the unit (with maximum potential set to some sane level).
It may be someone already figured out how to adjust the circuit around the non-populate adjustment section or just what value to set which pin of feedback to for various current / wattage limiting?
I have managed to locate the Output Inhibit pin and am working on tying in an exteral remote on/off for this... if I am not mistaken with output inhibit engaged just after power up (without battery connected) it will give the unit a chance to pre-charge the output capacitors which will prevent Arching and Sparks?
Feedback would be greatly appreciated!
Regards,
Mike
I have seen Nick's earlier posting about limiting CC mode by removing shunts (perhaps I even suggested it) but since the PCB seems to have positions for this limiting (adjustable) via SVR2 which is missing (along with a transistor, some resistors, etc) I am still trying to determine how to modify this circuit (which I think is feedback loop 2 for the PWM IC (494)) then we should be able to not only configure the brute "Range" of current by switching in various resistor networks but also enable fine adjustment within the current range we have selected - all without removing or bridging shunts (great brute force method which I am about to do to a 48v supply).
In testing I modified an S-350-48:
Stock - Replacement
SVR1 1K - SVR1 10K - 15 Turn
R25 2K - 1.5K
ZD1=62v - 62v + 12 + 12 = 96v maximum (I think this may have been the fatal move, next mod will be 62 + 10 + 10 = 82v cap
So I set the output voltage to 62.5 (full tested range was 14v - 85v, I didn't push it higher and that was no load).
In my previous unmodified tests to 61.8v I noted the current going from 10A to 11 for a brief moment (cells we about 3.7v for 55.5v so I was seeing power levels of 610 watts. That was without modifications, no limiting was kicking in.
The process cotinued until reaching about 4v per cell where the current began to taper down over 20-30minutes (I was charging a 15S1P 5AH pack) - Even once I reached 4v per cell I was seeing 9.5A for 570w.
The next test was post upgrade with the voltage limit set for 62.25 and it popped a fuse!
I'm going to bulk charge @ 10A using my iCharger on a 5S pack to follow the curve - then try to mod these to follow the same curve (not sure how yet but I'll figure it out).
The point of the post:
I see how by feeding back a small current into the feedback loop of the PWM controller you can limit the current... couldn't we make an external version of this adjustable and vary the current delivery of the unit (with maximum potential set to some sane level).
It may be someone already figured out how to adjust the circuit around the non-populate adjustment section or just what value to set which pin of feedback to for various current / wattage limiting?
I have managed to locate the Output Inhibit pin and am working on tying in an exteral remote on/off for this... if I am not mistaken with output inhibit engaged just after power up (without battery connected) it will give the unit a chance to pre-charge the output capacitors which will prevent Arching and Sparks?
Feedback would be greatly appreciated!
Regards,
Mike
Top work here. I was wondering if there was such a thing as a computer-controlled charger in the EV power range (like the ones the RC people use but at much lower cell counts). Apparenlty not, but using these with a G/F BMS does seem like a top plan in terms of cost-effectiveness.
If I understand the findings so far we can't quite use one of these to charge 72v LiFEPO4 packs (24S, 86.4V voltage limit, because that's just a tony bit higher than they will go?), in which case presumably 2 36V supplies in series is the best plan?
I am currently dithering between buying a 60V or 72V battery pack. It would be nice to buy a charger that could be configured for either case and this looks like a good solution.
I'll buy a couple and join the hack-fest.
If I understand the findings so far we can't quite use one of these to charge 72v LiFEPO4 packs (24S, 86.4V voltage limit, because that's just a tony bit higher than they will go?), in which case presumably 2 36V supplies in series is the best plan?
I am currently dithering between buying a 60V or 72V battery pack. It would be nice to buy a charger that could be configured for either case and this looks like a good solution.
I'll buy a couple and join the hack-fest.
mwkeefer
1 MW
wookey,
Computer Controlled in EV power range? RC stuff but lower cell counts? What are you running below 10S (37v nominal lipo) or 12S LiFePo4 (39.6v nominal)?
Personally the iChargers work fairly well for turn key at turn key prices and to be perfectly honest - I wouldn't want a PC monitoring and controlling charge process (but that's what I have now, kinda) because it could crash so as far as I can determine the iCharger and such RC based charge systems are the best blend of offloading the processing time while still giving access to the data back to a computer.
Using the G/F BMS would handle balancing at charge time, high voltage cutout and low voltage cutout (durring operation) and would protect your batteries... still you would need a power supply or two in order to charge this - and likely if LiFePo4 you will need to alter the supply to deliver lower current (fairly simple tweak, replace a resistor)... there are a few other tricks and mods but LiFePo4 (depending on the capacity?) wants what a 1/2C charge rate so a 10AH pack would accept 5Amp charge power... This would require a mod to the S-350-48 unit first to get it to Series voltage (ie: 60/72) then to limit the current to the max charging rate of your lifepo4 pack.
Lets look at the worst case scenario here - S-350-48 meanwell power supply has stock range: 39.1v to 58.1v so this must be adjusted upwards (for either 60 or 72 and I've never tried a single supply to 72v - not sure it will work) 72v x 5A = 360 watts - this is good, it means your within reasonable operating range for a single unit current output (I"ve run mine to 480 watts for testing but they will live longer properly configured) so a single 48v supply could be modified for your charger. I have and do daily 15S2P 10AH charges using my 48v modified Meanwells (63v max, 62.25 configured for 4.15v per cell and current set at 5.5A standard for 1/2C + charge rate) for weeks now to charge LiPo packs and have only managed through ignorance and experiments to burn one unit up (didn't catch fire, just pops fuses).
So the final response - yes you should be able to use bulk style charger with LiFePo4 and the G/F BMS to manage the system and handle the delicate bits.
Honestly you can just set the proper voltage and current for your pack, be conservative with the voltage... 95% charge max (LiPo is 4.2v I go to 4.15)... and just run it till the voltage of the battery goes to 62.25 and the current drops to nothing - even without the BMS it will work (unless you have a damaged or lazy cell).
Hope this helps - post back more info about your pack (capacity, rated C, charging rate max?) and I or someone can give you a better idea of the mods needed.
-Mike
Computer Controlled in EV power range? RC stuff but lower cell counts? What are you running below 10S (37v nominal lipo) or 12S LiFePo4 (39.6v nominal)?
Personally the iChargers work fairly well for turn key at turn key prices and to be perfectly honest - I wouldn't want a PC monitoring and controlling charge process (but that's what I have now, kinda) because it could crash so as far as I can determine the iCharger and such RC based charge systems are the best blend of offloading the processing time while still giving access to the data back to a computer.
Using the G/F BMS would handle balancing at charge time, high voltage cutout and low voltage cutout (durring operation) and would protect your batteries... still you would need a power supply or two in order to charge this - and likely if LiFePo4 you will need to alter the supply to deliver lower current (fairly simple tweak, replace a resistor)... there are a few other tricks and mods but LiFePo4 (depending on the capacity?) wants what a 1/2C charge rate so a 10AH pack would accept 5Amp charge power... This would require a mod to the S-350-48 unit first to get it to Series voltage (ie: 60/72) then to limit the current to the max charging rate of your lifepo4 pack.
Lets look at the worst case scenario here - S-350-48 meanwell power supply has stock range: 39.1v to 58.1v so this must be adjusted upwards (for either 60 or 72 and I've never tried a single supply to 72v - not sure it will work) 72v x 5A = 360 watts - this is good, it means your within reasonable operating range for a single unit current output (I"ve run mine to 480 watts for testing but they will live longer properly configured) so a single 48v supply could be modified for your charger. I have and do daily 15S2P 10AH charges using my 48v modified Meanwells (63v max, 62.25 configured for 4.15v per cell and current set at 5.5A standard for 1/2C + charge rate) for weeks now to charge LiPo packs and have only managed through ignorance and experiments to burn one unit up (didn't catch fire, just pops fuses).
So the final response - yes you should be able to use bulk style charger with LiFePo4 and the G/F BMS to manage the system and handle the delicate bits.
Honestly you can just set the proper voltage and current for your pack, be conservative with the voltage... 95% charge max (LiPo is 4.2v I go to 4.15)... and just run it till the voltage of the battery goes to 62.25 and the current drops to nothing - even without the BMS it will work (unless you have a damaged or lazy cell).
Hope this helps - post back more info about your pack (capacity, rated C, charging rate max?) and I or someone can give you a better idea of the mods needed.
-Mike
317537
10 kW
- Joined
- Oct 19, 2008
- Messages
- 939
http://www.reuk.co.uk/LM317-Current-Calculator.htm
http://www.reuk.co.uk/OtherImages/lm317t-current-regulation-circuit.gif
A simple lm317t for 1.5 amps or LM338 for 5 amp for constant current regulation. This will only shunt but can use voltage drops to regulate current potentials.
When your battery is in full float. (EG) The batteries voltage no longer rises or falls on the charger, take a current reading from the battery chargers input to the battery and you will see some form of current being supplied to keep the battery in float mode. Take the voltage when disconnected and multiply how much current when connected by the original voltage to gain how much watts it takes to get your pack into float mode.
Though constatnt current can be cool. You only need a volt or two to charge a battery pack. By SLA standards two hundred miliovervolts at 3 to 5 amp per cell supply would charge a battery.
I did one charger based on a LM338 and higher than float voltage spec of the battery and adjusted voltage drop on the adj until this voltage stayed steady when charged. To my surprise it drew 2.3 amps initial and took only a few hours until the amps being drawn was limited to the voltage drop of the battery, to tests on a smart charger the batteries rocketed up to full charge in seconds.. You see the battery will never stay at float voltages with no power and this area can be exploited as charging mode for good solid charging.
Maybe use a transistor resistor combination instead of the resistor in LM the current shunt circuit. Could put some control in there.
http://www.jaycar.com.au/images_uploaded/LM138.PDF with a
http://www.reuk.co.uk/OtherImages/lm317t-current-regulation-circuit.gif
A simple lm317t for 1.5 amps or LM338 for 5 amp for constant current regulation. This will only shunt but can use voltage drops to regulate current potentials.
When your battery is in full float. (EG) The batteries voltage no longer rises or falls on the charger, take a current reading from the battery chargers input to the battery and you will see some form of current being supplied to keep the battery in float mode. Take the voltage when disconnected and multiply how much current when connected by the original voltage to gain how much watts it takes to get your pack into float mode.
Though constatnt current can be cool. You only need a volt or two to charge a battery pack. By SLA standards two hundred miliovervolts at 3 to 5 amp per cell supply would charge a battery.
I did one charger based on a LM338 and higher than float voltage spec of the battery and adjusted voltage drop on the adj until this voltage stayed steady when charged. To my surprise it drew 2.3 amps initial and took only a few hours until the amps being drawn was limited to the voltage drop of the battery, to tests on a smart charger the batteries rocketed up to full charge in seconds.. You see the battery will never stay at float voltages with no power and this area can be exploited as charging mode for good solid charging.
Maybe use a transistor resistor combination instead of the resistor in LM the current shunt circuit. Could put some control in there.
http://www.jaycar.com.au/images_uploaded/LM138.PDF with a
mwkeefer. Sorry I wasn't very clear in my question. I have previously used things like the peak prodigy charger:
http://www.eamodels.co.uk/o-ip2801-pro-peak-prodigy-ii-charger-1-14ncd-1-5li-1349-p.asp
to charge small packs like 4.8V, NiMH and 14.4V Lithium Ion. They are nice because it's computer controlled - you select a chemisty and a charge rate (and maybe number of cells) and it just gets on with the appropriate CC, CV, cutoff voltages, temp compensation or whatever your chemistry prefers. It would be really nice to have this tech for our higher-power chargers. Because ultimately all chargers actually do the same things - control current, control voltage, check pack voltage, compensate for pack/ambient temps. So if you have a controllable circuit one charger can do a very wide range of things.
Given that such a thing does not appear to exist currently, your meanwell-based charging look good.
I currently have 48V 30Ah SLA and a 300W (5A@58V) charger. I plan to replace the dying batts with LiFEPO4 DIY pack around 20Ah capacity. It could be 48, 60 or 72V. Ideally it would be switchable between 48V 30Ah and 72V 20Ah (both 48 10Ah cells) so I could choose between range or speed. I'm trying to work out which meanwells to buy to charge that. Normal config will probably be 72V 20Ah.
http://www.eamodels.co.uk/o-ip2801-pro-peak-prodigy-ii-charger-1-14ncd-1-5li-1349-p.asp
to charge small packs like 4.8V, NiMH and 14.4V Lithium Ion. They are nice because it's computer controlled - you select a chemisty and a charge rate (and maybe number of cells) and it just gets on with the appropriate CC, CV, cutoff voltages, temp compensation or whatever your chemistry prefers. It would be really nice to have this tech for our higher-power chargers. Because ultimately all chargers actually do the same things - control current, control voltage, check pack voltage, compensate for pack/ambient temps. So if you have a controllable circuit one charger can do a very wide range of things.
Given that such a thing does not appear to exist currently, your meanwell-based charging look good.
I currently have 48V 30Ah SLA and a 300W (5A@58V) charger. I plan to replace the dying batts with LiFEPO4 DIY pack around 20Ah capacity. It could be 48, 60 or 72V. Ideally it would be switchable between 48V 30Ah and 72V 20Ah (both 48 10Ah cells) so I could choose between range or speed. I'm trying to work out which meanwells to buy to charge that. Normal config will probably be 72V 20Ah.
mwkeefer
1 MW
wookey,
thanks but I can't take credit for the meanwells, the credit for these goes to Jermey Harris, Fechter, Methods and so on!
What your looking for doesn't exist to my knowledge (you can link 2 hyperion automatic chargers / balancers but only to 12S I think) - you could use a pair of iCharger 1010B+ or iCharger 208B+ depending on how many cells in the series you will have. This would require the packs be split down in half their serial length but it's how I managed to change for several months (actually on a single charger with all packs in parallel).
So yes and no on the computer controlled charger - that said... I am working on adding an MCU to the meanwells for controlling current and voltage. This will include cell count detection, internal and external temp monitors and automatic adjustment to a specified wattage of output. The units will also have 0 output until a load is connected, making them more charger like.
Although this covers alot of what you were meaning by "computer controlled" (Actually micro-processor controlled) but doesn't address cell level balance issues - for now I leave that to Gary / Richard and any other ideas (you can use hobby king balancers / dischargers - they take forever but they work)... also doesn't address cell level HVC / LVC - again I leave this to Gary and RIchard (or you could do it the hard way and use a few cell log 8s configured for pack monitoring and HVC/LVC).
Hope this helps!
-Mike
thanks but I can't take credit for the meanwells, the credit for these goes to Jermey Harris, Fechter, Methods and so on!
What your looking for doesn't exist to my knowledge (you can link 2 hyperion automatic chargers / balancers but only to 12S I think) - you could use a pair of iCharger 1010B+ or iCharger 208B+ depending on how many cells in the series you will have. This would require the packs be split down in half their serial length but it's how I managed to change for several months (actually on a single charger with all packs in parallel).
So yes and no on the computer controlled charger - that said... I am working on adding an MCU to the meanwells for controlling current and voltage. This will include cell count detection, internal and external temp monitors and automatic adjustment to a specified wattage of output. The units will also have 0 output until a load is connected, making them more charger like.
Although this covers alot of what you were meaning by "computer controlled" (Actually micro-processor controlled) but doesn't address cell level balance issues - for now I leave that to Gary / Richard and any other ideas (you can use hobby king balancers / dischargers - they take forever but they work)... also doesn't address cell level HVC / LVC - again I leave this to Gary and RIchard (or you could do it the hard way and use a few cell log 8s configured for pack monitoring and HVC/LVC).
Hope this helps!
-Mike
methods
1 GW
I have an old Sorenson linear power supply, which is adjustable in voltage from 0 to 60V (65-ish, actually), and has a current limiter that can be set for anything from 0 to 1.5A (will go to 1.8A but I wouldn't do it for long, even with hefty forced cooling).
If I need them, it has terminals on the back for "programming" various functions via external controls, including the voltage and current limits.
Since it has this current limit built in, would I need to use the circuit created in this thread to program it as it charges either a LiPo pack or a NiMH pack?
Or could i simply set it to the max charge voltage I want the pack to reach, and the max current that I want it to have, and just charge the pack (via a diode, as usual)?
I can also use a thermal-sensing circuit to shut off current to the pack should it go beyond safe limits as read by a thermistor or other sensor inside the pack.
I *think* I can do it the above way, without any extra stuff, but I prefer a sanity check just in case.
If I need them, it has terminals on the back for "programming" various functions via external controls, including the voltage and current limits.
Since it has this current limit built in, would I need to use the circuit created in this thread to program it as it charges either a LiPo pack or a NiMH pack?
Or could i simply set it to the max charge voltage I want the pack to reach, and the max current that I want it to have, and just charge the pack (via a diode, as usual)?
I can also use a thermal-sensing circuit to shut off current to the pack should it go beyond safe limits as read by a thermistor or other sensor inside the pack.
I *think* I can do it the above way, without any extra stuff, but I prefer a sanity check just in case.
mwkeefer
1 MW
amberwolf said:
Amberwolf - I think you should check to see if there is some form of limiting diode for voltage if it caps at 65... still if you wanted to charge well... nearly anyhting up to 15s this should work (in theory).... I sure would have tried it!
But remember I was the moron who used to put normal CC/CV SLA chargers onto 3S packs and just disconnected them as they reached charged voltage levels (20A against 8 3S packs in parallel) but it worked !
Hope it helps!
-Mike
PS: I don't ever use diodes, but that's just me!
Since they're linear supplies, the transformer in there (several pounds and about 3 cubic inches) is the main limiting factor on the voltage and current. There is a significantly smaller one in the lower-current 40V 0.75A versions, and a similar sized one in the 40V 2A version.
I haven't checked what the raw voltage out of the transformer/rectifier is, but I'd bet it's no more than 70VDC.
These things are fairly old; at best from the early 80s, and more likely mid-70s.
As for experimentation--I cannot afford to replace anything I blow up in the way of batteries. I can probably fix the power supplies themselves, but the batteries are a one-shot thing, unless people start just donating them to me
(as does sometimes happen with SLAs locally, but most often those are overly-well-used).
Keep in mind that as of yet I have no BMS, no shunts, and no cell monitors. Before I do the charging in series, I will end up at least making some sort of HVC for each cell, so that I can have the charger shut off once the first cell reaches that level, and beep at me until I turn the beeper off.
I haven't checked what the raw voltage out of the transformer/rectifier is, but I'd bet it's no more than 70VDC.
These things are fairly old; at best from the early 80s, and more likely mid-70s.
As for experimentation--I cannot afford to replace anything I blow up in the way of batteries. I can probably fix the power supplies themselves, but the batteries are a one-shot thing, unless people start just donating them to me
(as does sometimes happen with SLAs locally, but most often those are overly-well-used). Keep in mind that as of yet I have no BMS, no shunts, and no cell monitors. Before I do the charging in series, I will end up at least making some sort of HVC for each cell, so that I can have the charger shut off once the first cell reaches that level, and beep at me until I turn the beeper off.
methods
1 GW
amberwolf said:
Sorenson supplies have a true CC limit so if you set the voltage of the supply to the target HVC voltage and hook up your pack it will simply ride the CC until it reaches your target voltage then level out to 0A. So yes - you can just use them and I dont think you will have any problems. I have a few Sorenson supplies at work (40V 15A 600W) and I like them better than most of the expensive crap. Really nice fast recovery on mine.
-methods
Thanks for that answer! I was fairly sure it should work that way, but I know that theory and actual practice dont' necessarily agree, depending on the design of the electronics (and sometimes the phase of the moon ).
I like these Sorenson's mostly because I already have them, salvaged from stuff a friend was going to have to toss out. I have them on a table trying to get back to reverse-engineering them and figuring out which parts are bad in each one, compared to the others, so I can get them all working again.
The 60V one smoked a resistor and transistor just recently but had worked fine up to then. The big 40V turns on but won't go up all the way in voltage or current--just stops trying past about 30V and a couple hundred mA. The two smaller ones have disintegrating power cords that I have to replace before I can safely test them fully, but they do at least turn on and go up to 43V or thereabouts.
If I stick a diode in there, I can also parallel them for higher current charging, if I need to. That way if the two smaller ones do actually work once cords are replaced, I can use them to charge at up to 1.5A (750mA each).
).I like these Sorenson's mostly because I already have them, salvaged from stuff a friend was going to have to toss out. I have them on a table trying to get back to reverse-engineering them and figuring out which parts are bad in each one, compared to the others, so I can get them all working again.
The 60V one smoked a resistor and transistor just recently but had worked fine up to then. The big 40V turns on but won't go up all the way in voltage or current--just stops trying past about 30V and a couple hundred mA. The two smaller ones have disintegrating power cords that I have to replace before I can safely test them fully, but they do at least turn on and go up to 43V or thereabouts.
If I stick a diode in there, I can also parallel them for higher current charging, if I need to. That way if the two smaller ones do actually work once cords are replaced, I can use them to charge at up to 1.5A (750mA each).
mwkeefer
1 MW
AmberWolf,
On the one's which go to about 30v and a few hundred ma, check the protection/limiting zener diode... I've had a similar problem with the meanwell power supplies and after a week of trouble shooting, I decided to pull a hail mary and just replace the ZD1... problem solved.
May not be the same for you, but it's worth testing with a DVM to ensure proper bias still exists.
-Mike
On the one's which go to about 30v and a few hundred ma, check the protection/limiting zener diode... I've had a similar problem with the meanwell power supplies and after a week of trouble shooting, I decided to pull a hail mary and just replace the ZD1... problem solved.
May not be the same for you, but it's worth testing with a DVM to ensure proper bias still exists.
-Mike
That's an idea, except I'm not sure they even use a zener anywhere in them. I dont' recall seeing one.
mwkeefer
1 MW
My bad brother, I misread - thought you were talking soniel style charger... dyslexia and speed reading at 2AM just don't add up.
I'll see if I can find a schematic for that model, I may (emphasis on the may) be able to figure out what's bad...
Again my appologies,
Mike
I'll see if I can find a schematic for that model, I may (emphasis on the may) be able to figure out what's bad...
Again my appologies,
Mike
No apologies necessary--these things are quite old, and should be easy to troubleshoot if I can find enough time when i'm not worn out or sick to do it (along with a zillion other things).
The friend that gave them to me has the schematics and manuals for them, too, but until we get far enough in rescuing/moving all his stuff from a storage place that's going away to the house he's staying in, I won't have access to them.
The friend that gave them to me has the schematics and manuals for them, too, but until we get far enough in rescuing/moving all his stuff from a storage place that's going away to the house he's staying in, I won't have access to them.
mwkeefer
1 MW
Hey all,
Just a quick update on the S-320/SP-320/S-350 model issues and differences...
I obtained this information via reverse engineering and also from (of all places) MeanWell Intl.
Mind you, I am not... nor have I ever claimed to be an EE... I have literally forgotten (I was a tech class ham at 12) over the past 24 years more than most "average" normal HS graduates know about electronics - still I am no engineer and I invite the experts with more knowledge about these topics, no I beg of them - where (if) I am wrong, correct me.... where I am not wrong, but unclear - please elaborate for not just the community sake but also for my own.
Thanks in advance for your cooperation!
Discovery #1.)
There are various versions of the MeanWell power supplies floating around out there - beyond just the S or SP or 320/350 (these are for now the ones were most interested in) but then we have knock offs and now I come to find out - the SP-320's I have (which are ROHS I think, but for sure AU/EU/CN certified and perhaps UL (can't confirm yet) - well they are designed back in 2005 and are known as PFC + PWM style SMPS supplies.
I have also received confirmation that the NEWER (2009+) SP-320 do have proper current limiting – or at least the 48v I received images of from a User had 3 shunts in place and was dated 4/2009 – obviously a newer revision (I think he ordereded direct from Jameco).
Now here’s a funny bit more – the manuals for the S-350 and S-320 models boast over current protection at 130% while the SP-320 units boast over current protection at 135% - 5% higher but then the protection which kicks in is hiccup mode.
Discovery #2.)
There seems to be no easy way to offer variable current levels… changing resistors introduces some form of buzzing and crackling into the system. Additionally, in the earlier SP-320 models (now seemingly flooding the market since there is now a demand for them and people had old stock + new models are coming the later model SP-320 with the shunts).
Finally the shunt mods – sure they work but to get variable shunt action you need to somehow DPDT remove shunts from the circuit to raise the current or add shunts to increase current (relays or HD dpdt switches).
Realization #1.)
We are wasting our time… the majority of required mods are simple, the voltage modifications work (within reason) and alone they introduce no buzzing sounds from any of the models (within so far as I have safely adjusted a 48v 350 to 83v max CC/CV cutoff.
In order to facilitate all those with the hiccupping mode Meanwells (SP older series) and any future models they may come up with…. The idea dawned on me… A BMS current reducer is basically a PWM using paralleled FETS (think IRFB4110 – 100v/100a, and yes in a heat sink) – the same basic thing as a BLDC or Brushed Motor Controller.
I began searching out current limiting circuit designs, attempting to find one which only required a sense resistor and would allow current to be set via a 0-5v input line (currently controlled with variable LM317 but intended for MCU control)… I found a few and realized quickly that substituting their gate drivers and misc components (AC rectifier and transformer) that I could turn up a 48v meanwell to about .7v over what I needed and then simply run the output through the limiter with a sense resistor…. When the value gets too high, engage PWM to limit current.
The beauty here is I have a rough circuit for a 1-20A limiter controlled by 0-5v input… This opens all new possibilities and makes even hiccup mode supplies useable for CC/CV supplies.
If a SP-320-24 can handle 312w and we reconfigure it to 31.25, run it in series with another S-320-24 then set the limiter to (312 / 31.25 = 9.9 A) then it will prevent the supplies from shutting down or pulsing into hiccup mode. It also negates the need for a current resistor or shunt mod on the S-350 series and will likely eliminate the buzzing as we will control the load.
I have no idea what the requirements will be for heat dissipation but heck, maybe an old controller body and sink will suffice, wired to a PCB modified for just this.
[Another Idea – OT]
I’ve been thinking about a different balancer type – yes still using a micro-processor but in a more active role:
1.) MCU would monitor cells and even during discharge if a cell becomes too low, the system would open a circuit path to the highest cell in the pack… voltage would need to be stepped up (or simply use 2 adjacent cell) which would be used to charge a capacitor. Circuit would then disconnect from high cell and then reconnect to lower cell and dump the available current in (CC/CV style). This would (in theory) allow higher cells to transfer their power not just during charge but at discharge which would I think, cause more efficient and even balance at LVC and thus a healthier battery pack.
Just my .02c
-Mike
Just a quick update on the S-320/SP-320/S-350 model issues and differences...
I obtained this information via reverse engineering and also from (of all places) MeanWell Intl.
Mind you, I am not... nor have I ever claimed to be an EE... I have literally forgotten (I was a tech class ham at 12) over the past 24 years more than most "average" normal HS graduates know about electronics - still I am no engineer and I invite the experts with more knowledge about these topics, no I beg of them - where (if) I am wrong, correct me.... where I am not wrong, but unclear - please elaborate for not just the community sake but also for my own.
Thanks in advance for your cooperation!
Discovery #1.)
There are various versions of the MeanWell power supplies floating around out there - beyond just the S or SP or 320/350 (these are for now the ones were most interested in) but then we have knock offs and now I come to find out - the SP-320's I have (which are ROHS I think, but for sure AU/EU/CN certified and perhaps UL (can't confirm yet) - well they are designed back in 2005 and are known as PFC + PWM style SMPS supplies.
I have also received confirmation that the NEWER (2009+) SP-320 do have proper current limiting – or at least the 48v I received images of from a User had 3 shunts in place and was dated 4/2009 – obviously a newer revision (I think he ordereded direct from Jameco).
Now here’s a funny bit more – the manuals for the S-350 and S-320 models boast over current protection at 130% while the SP-320 units boast over current protection at 135% - 5% higher but then the protection which kicks in is hiccup mode.
Discovery #2.)
There seems to be no easy way to offer variable current levels… changing resistors introduces some form of buzzing and crackling into the system. Additionally, in the earlier SP-320 models (now seemingly flooding the market since there is now a demand for them and people had old stock + new models are coming the later model SP-320 with the shunts).
Finally the shunt mods – sure they work but to get variable shunt action you need to somehow DPDT remove shunts from the circuit to raise the current or add shunts to increase current (relays or HD dpdt switches).
Realization #1.)
We are wasting our time… the majority of required mods are simple, the voltage modifications work (within reason) and alone they introduce no buzzing sounds from any of the models (within so far as I have safely adjusted a 48v 350 to 83v max CC/CV cutoff.
In order to facilitate all those with the hiccupping mode Meanwells (SP older series) and any future models they may come up with…. The idea dawned on me… A BMS current reducer is basically a PWM using paralleled FETS (think IRFB4110 – 100v/100a, and yes in a heat sink) – the same basic thing as a BLDC or Brushed Motor Controller.
I began searching out current limiting circuit designs, attempting to find one which only required a sense resistor and would allow current to be set via a 0-5v input line (currently controlled with variable LM317 but intended for MCU control)… I found a few and realized quickly that substituting their gate drivers and misc components (AC rectifier and transformer) that I could turn up a 48v meanwell to about .7v over what I needed and then simply run the output through the limiter with a sense resistor…. When the value gets too high, engage PWM to limit current.
The beauty here is I have a rough circuit for a 1-20A limiter controlled by 0-5v input… This opens all new possibilities and makes even hiccup mode supplies useable for CC/CV supplies.
If a SP-320-24 can handle 312w and we reconfigure it to 31.25, run it in series with another S-320-24 then set the limiter to (312 / 31.25 = 9.9 A) then it will prevent the supplies from shutting down or pulsing into hiccup mode. It also negates the need for a current resistor or shunt mod on the S-350 series and will likely eliminate the buzzing as we will control the load.
I have no idea what the requirements will be for heat dissipation but heck, maybe an old controller body and sink will suffice, wired to a PCB modified for just this.
[Another Idea – OT]
I’ve been thinking about a different balancer type – yes still using a micro-processor but in a more active role:
1.) MCU would monitor cells and even during discharge if a cell becomes too low, the system would open a circuit path to the highest cell in the pack… voltage would need to be stepped up (or simply use 2 adjacent cell) which would be used to charge a capacitor. Circuit would then disconnect from high cell and then reconnect to lower cell and dump the available current in (CC/CV style). This would (in theory) allow higher cells to transfer their power not just during charge but at discharge which would I think, cause more efficient and even balance at LVC and thus a healthier battery pack.
Just my .02c
-Mike
Hey Fechter ! am Trying to make your plan work on dell PSU's and is showing promise, after stabimg a bread board up with random stuff around . Know doult it will work. I need tips on components and how to set gain etc.fechter said:
Funny, bike fully charged easy to run. Charging a pain.
I did make a bit more sophisticated limiter circuit designed for Meanwell supplies, but it would work for most supplies if you can find the right spot to make the voltage sensing connection. http://endless-sphere.com/forums/viewtopic.php?f=31&t=21768
The simple transistor version should be good enough for battery charging, again if you can find the right spot to tie in.
The simple transistor version should be good enough for battery charging, again if you can find the right spot to tie in.
Similar threads
