The Mighty Volt
1 MW
texaspyro said:
Sorry...what I meant to say was 2260mAh, not 2600mAh!!!
LOL, 2600mAh would be nice.
Build your own CD battery tab welder for about $100.00+-
- Thread starter RLT
- Start date
Sorry...what I meant to say was 2260mAh, not 2600mAh!!!
LOL, 2600mAh would be nice.
The same cell cranked out 2440 mAh at 1C rate (6% over spec, don't know how long that will last). They were charged to 3.672V starting at 4 amps (looks like the power supply meter is reading a little bit low).
Looks like you lose around 10-15% of the capacity running them at 4C. The cells came from HobbyKing in A123Racing developer kit packaging.
Looks like you lose around 10-15% of the capacity running them at 4C. The cells came from HobbyKing in A123Racing developer kit packaging.
webfootguy
100 W
In my experience, the actual capacity of M1 cells sensitive to the temperature. They like it hot. If you have just run a discharge/charge cycle at a high enough rate to heat the cell, you will get higher mAh. You can also simulate this by just heating the cell before discharge (lead acid does the same thing and the Electrathon racers know this and heat their batteries in a cooler to 160F before racing. If you want a true test of capacity, charge the cell and let it rest 8 hours and then test it.
The Mighty Volt
1 MW
Just, on a very brief aside, are these cells as good as it get's with LiFePo4? Is this LiFePo4's best answer to Li-Po in terms of power?
The cells that I tested were all at room temp. They were charged outside were it is around 100F. Several of them are running up to 10% over rated capacity. No telling how long until the new wears off. The first cycle or two they pretty much at spec'd capacity.
A123 cells have the highest spec'd LiFePO peak discharge currents out there by a wide margin (way beyond what any ebike would draw). Their internal resistance is good (but not as good as the latest LiPo). Their weightomph ratio is not nearly as good as some of the LiPo cells. Also, they probably won't burn your house down. They tend to be rather tolerant of abuse (way more than what I would tolerate if I were a lithium cell). At 1.3C discharge they have no heating. At 4C the heating is noticeable, pleasantly warm, but not anywhere near objectionable. At 25C, who the heck knows... the heating due to internal resistance is 36+ watts.
A123 cells have the highest spec'd LiFePO peak discharge currents out there by a wide margin (way beyond what any ebike would draw). Their internal resistance is good (but not as good as the latest LiPo). Their weight
omph ratio is not nearly as good as some of the LiPo cells. Also, they probably won't burn your house down. They tend to be rather tolerant of abuse (way more than what I would tolerate if I were a lithium cell). At 1.3C discharge they have no heating. At 4C the heating is noticeable, pleasantly warm, but not anywhere near objectionable. At 25C, who the heck knows... the heating due to internal resistance is 36+ watts.I did some tweaking of the CBA mode on the welder... somewhere 800 more lines of code sneaked in (almost all menu and user interface stuff). I ran into a big potential problem when the code crossed the 128K boundary in the ATMEGA2561 chip. The compiler barfed out bogocode. Dreading the hassel, I installed the latest WINAVR compiler. Absolutely no problems. Hit install and all went well. New compiler code works perfectly. This CAN'T be a good omen. Expect the Sphere to End shortly...
I tried the CBA mode with the welder capacitor bank installed. No problems at all connecting cells. That is a very good thing... it is a bit of a pain to disconnect the caps when you want to test batteries. The one thing that I can't do is test cell internal resistance with the caps installed. But then, I can't do that well with the caps disconnected since I don't have a valid 4-wire connection to the cell terminals. I do have software compensation for the cell connection resistance, but that can change with I^2*R heating of the connections, battery holder, etc. Best to measure internal resistance with the HP3458A...
I came up with some interesting data on how A123 cell capacity changes with charge voltage and discharge current. See http://endless-sphere.com/forums/viewtopic.php?f=14&t=20698
I tried the CBA mode with the welder capacitor bank installed. No problems at all connecting cells. That is a very good thing... it is a bit of a pain to disconnect the caps when you want to test batteries. The one thing that I can't do is test cell internal resistance with the caps installed. But then, I can't do that well with the caps disconnected since I don't have a valid 4-wire connection to the cell terminals. I do have software compensation for the cell connection resistance, but that can change with I^2*R heating of the connections, battery holder, etc. Best to measure internal resistance with the HP3458A...
I came up with some interesting data on how A123 cell capacity changes with charge voltage and discharge current. See http://endless-sphere.com/forums/viewtopic.php?f=14&t=20698
I pimped out the battery analyzer mode on the welder some more. What started out as a couple hundred lines of quick code hack for a quick test of a particular battery cell has turned into over 1600 lines of full blown battery analyzer bliss.
rkosiorek
100 kW
texaspyro,
how about adding an EDM mode (Electrical Discharge Machining) to your gadget. something so you could use an electrode to burn out broken taps, bolts and other stuff from blind holes.
or is that too far off base
rick
how about adding an EDM mode (Electrical Discharge Machining) to your gadget. something so you could use an electrode to burn out broken taps, bolts and other stuff from blind holes.
or is that too far off base
rick
rkosiorek said:
Actually that was what I was thinking of before I started playing with it as an anodizer/plater. The main issue there is the need for the CNC stage. Also it would need some short circuit protection (but it would probably just vaporize the EDM wire).
When I added the battery analyzer mode, I wound adding a connection that taps off the capacitor bank so I can now use it a (unswitched, low ripple) power supply. One still has to be VERY careful about short circuiting the leads. The caps can provide 20,000 amps of pulse current. Then the charger will try to charge the short circuit... which will fry the charger fets.... been there, done that. I think I will add an undervoltage lockout in power supply mode. If the cap voltage drops below its nominal regulation threshold, it will trigger a fault and shut down the charger... hopefully before the FETs fry.
Adding current limiting in the charger stage is another possibility, but any current limiting tends to greatly increase the recharge time when you are doing spot welds. When I mounted the unit on a plywood board, going from three feet of 12 gauge wire to 9 inches cut the recharge time in half.
I added some code in the power supply mode to implement a crude current limiter (well more like a crowbar). If the cap voltage falls below the setpoint voltage by more than the regulation threshold (200 mV), it shuts down the charger and kills the power. I had a chance to test it in real life when I accidentally triggered the power supply mode while the 0.2 ohm external battery analyzer mode resistor was connected (50+ amp load) ... seems to work.
There is also a fault shutdown that happens if the caps do not charge or discharge when they should. But that is much slower responding. The first failsafe is only available in power supply mode. The second one is always active. It has to be slow responding since there can be quite a bit of AC ripple on the charger supply when the caps are being charged. Even with a 70,000 microfarad cap on the charger supply, you can see quite a bit of ripple as the 3.7 farad capacitor bank is being charged.
I also noticed that I had another resistor load option in the battery analyzer mode. The capacitor bank has around 800 ohms of capacitor bank bleeder resistance and voltage divider impedance across it. So I added a light load mode. On a single A123 cell that works out to around a 4 milliamp load. I can't switch it on and off, but I can use it as a very low current test load. It came in useful for monitoring the self-discharge of some damaged cells and the voltage recovery after a cell had been drained under load and then had the load switched off.
I also added the ability to plot the voltage/current/power/etc while the discharge is in progress. Before I only allowed access to the plots after the test had completed. I was concerned that the overhead of updating the plots would affect the accuracy of the tests, but it does not seem to be a problem.
There is also a fault shutdown that happens if the caps do not charge or discharge when they should. But that is much slower responding. The first failsafe is only available in power supply mode. The second one is always active. It has to be slow responding since there can be quite a bit of AC ripple on the charger supply when the caps are being charged. Even with a 70,000 microfarad cap on the charger supply, you can see quite a bit of ripple as the 3.7 farad capacitor bank is being charged.
I also noticed that I had another resistor load option in the battery analyzer mode. The capacitor bank has around 800 ohms of capacitor bank bleeder resistance and voltage divider impedance across it. So I added a light load mode. On a single A123 cell that works out to around a 4 milliamp load. I can't switch it on and off, but I can use it as a very low current test load. It came in useful for monitoring the self-discharge of some damaged cells and the voltage recovery after a cell had been drained under load and then had the load switched off.
I also added the ability to plot the voltage/current/power/etc while the discharge is in progress. Before I only allowed access to the plots after the test had completed. I was concerned that the overhead of updating the plots would affect the accuracy of the tests, but it does not seem to be a problem.
Hi,
Im trying to build 1f car cap welder. Tested with 12v PSU , with no good results. So now im trying to raise voltage to 16v with adjustable laptop adapter, which has 15-24v. After i plugged it in, i instantly had a blown adapter fuse, it even didnt charge cap. I wonder if you have any idea how to make adapter compatible with my cap? I'd really appreciate your help.
Adapter: http://attachments.content4us.com/datasheets/MANUAL_P_SUB_NBT120A_COMP.PDF
Cap: http://www.verkkokauppa.com/popups/prodinfo.php?id=48107
Im trying to build 1f car cap welder. Tested with 12v PSU , with no good results. So now im trying to raise voltage to 16v with adjustable laptop adapter, which has 15-24v. After i plugged it in, i instantly had a blown adapter fuse, it even didnt charge cap. I wonder if you have any idea how to make adapter compatible with my cap? I'd really appreciate your help.
Adapter: http://attachments.content4us.com/datasheets/MANUAL_P_SUB_NBT120A_COMP.PDF
Cap: http://www.verkkokauppa.com/popups/prodinfo.php?id=48107
The problem charging capacitors is that they look pretty much like a dead short when they start charging. A good cap will draw over 10,000 amps if you let it. The only thing that limits the current is the capabilities of the power supply. Your adjustable laptop supply is apparently not protected very well against overloads (it depends upon a fuse instead of electronic overload protection).
Try a 3 ohm resistor in series with your laptop adapter. This will limit the current to around 5 amps. Ideally you want a 100 watt resistor, but the current drops rapidly once the cap starts charging. You can probably get by with a 10-20 watt resistor. The bad thing about using a resistor is it will increase the time to fully charge the cap to around 12-15 seconds (4-5 R*C time constants).
Better yet, build a brute force power supply. My charging supply is shown in http://endless-sphere.com/forums/viewtopic.php?f=2&t=2633&start=585#p280596 It was originally connected to the welder by 3 feet of 10 gauge wire. When I mounted everything on a plywood board, I trimmed the leads to less than a foot. Even that small of a resistance change cut the recharge time in half (well under a second to charge 3.6 farads).
My transformer weighs 8 pounds and is rated at around 400 watts continuous. The diode module is a 60 volt/400 amp dual Schottky diode. The no-load output of the charger is 19 volts. The cap loads down the charger to around 7V for the first 300 milliseconds. The average charge current is around 60 amps. It takes 0.8 seconds to charge 3.65F from 0 to 12 volts.
My welder can run off of 8-30V input. I have run it from a 12V 12 Ah SLA, 4 x A123 cells, a tiny 15V/2 amp wall wart, or my big charger supply. It can regulate the capacitor voltage to any value less than the charger input. That small wall wart takes over 30 seconds to charge the caps.
And see my capacitor rant: http://endless-sphere.com/forums/viewtopic.php?f=2&t=2633&start=525#p267768 The chances of your capacitor being what they say it is are pretty slim...
Try a 3 ohm resistor in series with your laptop adapter. This will limit the current to around 5 amps. Ideally you want a 100 watt resistor, but the current drops rapidly once the cap starts charging. You can probably get by with a 10-20 watt resistor. The bad thing about using a resistor is it will increase the time to fully charge the cap to around 12-15 seconds (4-5 R*C time constants).
Better yet, build a brute force power supply. My charging supply is shown in http://endless-sphere.com/forums/viewtopic.php?f=2&t=2633&start=585#p280596 It was originally connected to the welder by 3 feet of 10 gauge wire. When I mounted everything on a plywood board, I trimmed the leads to less than a foot. Even that small of a resistance change cut the recharge time in half (well under a second to charge 3.6 farads).
My transformer weighs 8 pounds and is rated at around 400 watts continuous. The diode module is a 60 volt/400 amp dual Schottky diode. The no-load output of the charger is 19 volts. The cap loads down the charger to around 7V for the first 300 milliseconds. The average charge current is around 60 amps. It takes 0.8 seconds to charge 3.65F from 0 to 12 volts.
My welder can run off of 8-30V input. I have run it from a 12V 12 Ah SLA, 4 x A123 cells, a tiny 15V/2 amp wall wart, or my big charger supply. It can regulate the capacitor voltage to any value less than the charger input. That small wall wart takes over 30 seconds to charge the caps.
And see my capacitor rant: http://endless-sphere.com/forums/viewtopic.php?f=2&t=2633&start=525#p267768 The chances of your capacitor being what they say it is are pretty slim...
Hi again.
Charging is working well now. I also added one 1f cap in parallel, but welds are still too weak. I'm used 50ria20 SCR, which got fried after 16v. I think 50a was the bottleneck, so i hope you could help me to choose more powerful one. I'm planning to buy it from here:http://oomipood.ee/?t=k_ak&i=FARNELL_GR2332&
Thanks again.
Charging is working well now. I also added one 1f cap in parallel, but welds are still too weak. I'm used 50ria20 SCR, which got fried after 16v. I think 50a was the bottleneck, so i hope you could help me to choose more powerful one. I'm planning to buy it from here:http://oomipood.ee/?t=k_ak&i=FARNELL_GR2332&
Thanks again.
The Mighty Volt
1 MW
Hi Bull3T, please read back through my posts for information on all the capacitor and power-supply issues.
You will need 2.0 to 3.0F of genuine capacitance, as supplied by the recommended Scosche, Rockford or Monster capacitors.
Anything else, other than those 3 brands, or the genuine computer grade electrolytic capacitors, and you are wasting your time.
It is advisable to get a power supply with a monitor, so you can perform test welds and keep performance data notes and it also allows you to visually note the capacitors recharging.
Joining the capacitors, in parallel, negative to negative, positive to positive, with copper bus bars is highly recommended. I noticed a massive difference between 8ga 200Amp rated copper jump leads and quarter-inch thick, inch wide bus bars. The bus bars are king. Basically, with audio caps, even the very good ones, you need every ounce of capacitance you can get your mitts on, no point in spending money on good caps and then seeing all your precious amps get frittered away on poor connections and poor connectivity.
You will need 2.0 to 3.0F of genuine capacitance, as supplied by the recommended Scosche, Rockford or Monster capacitors.
Anything else, other than those 3 brands, or the genuine computer grade electrolytic capacitors, and you are wasting your time.
It is advisable to get a power supply with a monitor, so you can perform test welds and keep performance data notes and it also allows you to visually note the capacitors recharging.
Joining the capacitors, in parallel, negative to negative, positive to positive, with copper bus bars is highly recommended. I noticed a massive difference between 8ga 200Amp rated copper jump leads and quarter-inch thick, inch wide bus bars. The bus bars are king. Basically, with audio caps, even the very good ones, you need every ounce of capacitance you can get your mitts on, no point in spending money on good caps and then seeing all your precious amps get frittered away on poor connections and poor connectivity.
See my "loss budget" post for my FET based welder at http://endless-sphere.com/forums/viewtopic.php?f=2&t=2633&start=555#p273672
I've improved my welder some since then (better hand selected caps, bigger wires), but the gist is tiny, tiny, tiny resistances can sap huge amounts of energy from your welds. You need to pay attention to the details, not cut any corners, And if you are not using GOOD caps that meet their published specs, you might as well be welding with watch batteries.
My capacitor bank now has around 3.7 true farads and 200 micro-ohms ESR... hand selected Monster Cable caps.
The Rockford RFC1 caps have the most consistent measurements... VERY little spread between units. If you are hand selecting from a larger group of caps, Monster Cable caps tend to have more caps that exceed their specs. Also the older Rockford Punch 1.0 caps tend to have units with better than normal specs. I have not tested enough Scosche caps to determine their spread.
And for what it's worth, the Rockford caps are heavier than the Monster caps... Both appear to be made by the same company.
I have had more experience running them at over 16 volts. I have had no failures. The caps have a working voltage rating of 16V and a "surge" rating of 20V, but the caps are rated to work at 90 degrees C. Since they are being used at room temp, they seem to not be at all stressed at 18-20V. That said, I have not had to go over 16V to weld anything, I have left the caps charged at 19V for weeks on end.
I've improved my welder some since then (better hand selected caps, bigger wires), but the gist is tiny, tiny, tiny resistances can sap huge amounts of energy from your welds. You need to pay attention to the details, not cut any corners, And if you are not using GOOD caps that meet their published specs, you might as well be welding with watch batteries.
My capacitor bank now has around 3.7 true farads and 200 micro-ohms ESR... hand selected Monster Cable caps.
The Rockford RFC1 caps have the most consistent measurements... VERY little spread between units. If you are hand selecting from a larger group of caps, Monster Cable caps tend to have more caps that exceed their specs. Also the older Rockford Punch 1.0 caps tend to have units with better than normal specs. I have not tested enough Scosche caps to determine their spread.
And for what it's worth, the Rockford caps are heavier than the Monster caps... Both appear to be made by the same company.
I have had more experience running them at over 16 volts. I have had no failures. The caps have a working voltage rating of 16V and a "surge" rating of 20V, but the caps are rated to work at 90 degrees C. Since they are being used at room temp, they seem to not be at all stressed at 18-20V. That said, I have not had to go over 16V to weld anything, I have left the caps charged at 19V for weeks on end.
The Mighty Volt
1 MW
bull3t said:
You would need 3 of them.
I have 4 half-farad Caps and one 1F cap, Rockford Fosgate "PUnCH" giving me 3F nominal capacitance. Works fine.
steveo
100 kW
Hey Everyone,
I've been reading up on this tread and i've been seeing alot of good pics on a123 battery tab welding...
I have an extremly big project that needs tab welding to make it happen and i'm turning to those here that have experimented & can recommend a good setup to tab weld my ev car pack
That pack consits of appox 1408 a123 cells in total!...
i have an scr at home already ... i think its like 250 amps ..
i still need alot of parts...
PARTS TO BUY LIST
-molybdenum rod .25" x 12" ( no more melting electrods (i think thats what there called!)
if anyone could please list out there setup/links for parts used... and there sucess . with A123 Cells. I would like to follow in someones footsteps if they have had good results.
Please refer here for photos on the batteries i have & am looking to construct into ev car pack..
http://endless-sphere.com/forums/viewtopic.php?f=34&t=21472&start=0
thank you
-steveo
I've been reading up on this tread and i've been seeing alot of good pics on a123 battery tab welding...
I have an extremly big project that needs tab welding to make it happen and i'm turning to those here that have experimented & can recommend a good setup to tab weld my ev car pack
That pack consits of appox 1408 a123 cells in total!...
i have an scr at home already ... i think its like 250 amps ..
i still need alot of parts...
PARTS TO BUY LIST
-molybdenum rod .25" x 12" ( no more melting electrods (i think thats what there called!)
if anyone could please list out there setup/links for parts used... and there sucess . with A123 Cells. I would like to follow in someones footsteps if they have had good results.
Please refer here for photos on the batteries i have & am looking to construct into ev car pack..
http://endless-sphere.com/forums/viewtopic.php?f=34&t=21472&start=0
thank you
-steveo
The Mighty Volt
1 MW
SteveO- the Molybdenum isn't necessary, as such. It's also very expensive. I have belted out thousands of welds with the basic copper tips and they work fine. However, if $ is not an issue, and you are going all out, then Molybdenum is as good as it gets
liveforphysics
100 TW
steveo said:
steveo
100 kW
Hey guys,
I'll probably do copper electrodes...
what do you recommend for thickness & length of the electrodes?
Also ... for the scr ... when you say 2,500amps are you refering to peak?.... i think mines 235-250amps ... but i'm guessing peak is 2000-2500amps...?
i just checked an scr that was 110amps ... and it peaks at about 1,400... ...
Any recommends for capasitors?... how many farads?; voltage to use for A123?
-steveo
I'll probably do copper electrodes...
what do you recommend for thickness & length of the electrodes?
Also ... for the scr ... when you say 2,500amps are you refering to peak?.... i think mines 235-250amps ... but i'm guessing peak is 2000-2500amps...?
i just checked an scr that was 110amps ... and it peaks at about 1,400... ...
Any recommends for capasitors?... how many farads?; voltage to use for A123?
-steveo
Attachments
liveforphysics
100 TW
Yeah, only the peak rating matters for a CD welder.
The energy storage requirement is quite low, so the amount of capacitance doesn't matter much. The rate of discharge is the key. I think 0.25F would be plenty if the ESR were low enough. For example, 50F of ultra caps would be useless for a CD welder, yet a bank of 20mF caps could be hundreds of times less capacitance, yet make a kick-ass welder.
The energy storage requirement is quite low, so the amount of capacitance doesn't matter much. The rate of discharge is the key. I think 0.25F would be plenty if the ESR were low enough. For example, 50F of ultra caps would be useless for a CD welder, yet a bank of 20mF caps could be hundreds of times less capacitance, yet make a kick-ass welder.
liveforphysics said:
Beg to differ... For welding thick nickel tabs or copper, you need quite a bit of energy. 600 watt-seconds or so can be useful. To get there with car audio caps that works out to around 3 farads at 20V. My welder now has around 3.75F with an ESR of less than 250 micro ohms (via three hand-selected Monster Cable capacitors). For lighter duty welding, 1F can be plenty. Just make damn sure your caps are actually what you paid for. Almost all are utter, fraudulent crapassators (with 1/10 the rated capacitance and 5 times the ESR)
The trick in a good welder design is to pay VERY close attention to the details of getting that energy out of the caps and into the weld. You will be lucky to get 20% out the probes...
And I do have a CD welder with 58F/30V of ultracaps... works well for what it needs to do... long, pulses of around 1000-2500 amps. The Monster cap welder can dump its 750 Joules in a very few milliseconds... 20,000 amps, 0.4 megawatts...
