GGoodrum
1 MW
fechter said:
Okay, now you have my attention...
How would this work?
Upload Your Circuit Schematics & Requests Here...
- Thread starter xyster
- Start date
Okay, now you have my attention...
How would this work?
OneEye
100 W
Cool, thanks for running a test. 120mA is in the ballpark of a desirable cutoff, right? What would you want to use as the CC/CV end-of-charge termination current? Lets say for a 10Ah-20Ah pack of LiFePO4 or LiMn(Co/Poly/whatever).
I browsed a few SCR spec sheets of 3A - 12A SCR's while trying to figure this stuff out. I'm going from memory, but I think Ih max was in the 5mA to 50mA range for them. I don't have anything to go on, but those holding currents seemed too low to use for a reliable shutoff.
Some heavy duty SCR's apparently already have a gate to cathode resistor built in, I wonder if that is the case for yours. Supposedly if you can detect continuity in both directions between Gate and Cathode it has an internal resistor.
I'm a bit suprised a direct short of gate to cathode didn't affect the shutoff current. Is it possible there is a built in resistor between the gate on the chip and the gate's lead to help prevent unwanted latching?
I see I've piqued Gary's interest. That's encouraging--maybe I'm not a complete nut for hoping this could be of use. 8)
I browsed a few SCR spec sheets of 3A - 12A SCR's while trying to figure this stuff out. I'm going from memory, but I think Ih max was in the 5mA to 50mA range for them. I don't have anything to go on, but those holding currents seemed too low to use for a reliable shutoff.
Some heavy duty SCR's apparently already have a gate to cathode resistor built in, I wonder if that is the case for yours. Supposedly if you can detect continuity in both directions between Gate and Cathode it has an internal resistor.
I'm a bit suprised a direct short of gate to cathode didn't affect the shutoff current. Is it possible there is a built in resistor between the gate on the chip and the gate's lead to help prevent unwanted latching?
I see I've piqued Gary's interest. That's encouraging--maybe I'm not a complete nut for hoping this could be of use. 8)
the optos you mention have a fairly low CTR (current transfer ratio). the ones i specified i chose specifically because they have a 500% CTR so that a milliamp into the led can cause the output transistor to sink the necessary 5 ma.rf said:
if you could find a multi-channel opto with a 500% CTR it would probably be fine.
GGoodrum said:fechter said:
Okay, now you have my attention...
How would this work?
Just put the SCR in series with the bulk charger and trigger it. As soon as the current gets down to 120ma, it will shut off and stay off. The downside is the SCR has a voltage drop, so the charger would need to be set about a volt higher to compensate.
I wonder if you could use one of those solid state AC relays to get the same effect?
actually i was wrong about using the quad optocouplers richard suggested. there is one version, the -C version, that does have a minimum CTR of 200%, which would be ok. the -a version i looked at first has a minimum of 50% so it would not be suitable.
in the interest of minimizing the current drain when the low voltage detector turns on the opto we chose a resistor that will provide about a milliamp to the led. On the crystalyte controllers i have seen the ebrake is pulled up to 15v by 10k, so it takes 1.5 ma to pull it down. this means an opto with a minimum CTR of over 150% would be fine.
the latest data sheet richard posted shows the ACPL-827C as having 200-400% and the latest one i found for the CNY17-4, the original opto i specified shows 160-320%, so either one would work.
i was looking at an old data sheet when i said the CNY17-4 had 500% typical current gain. I'm not sure if they changed the part or if this was an error. I calculated the values based on the minimum gain of 160%.
thanks richard for pointing out the error. it might make the layout simpler to use the quad optos. my original intention was to mount the tc54 and opto directly on each cell and just tie the outputs together.
in the interest of minimizing the current drain when the low voltage detector turns on the opto we chose a resistor that will provide about a milliamp to the led. On the crystalyte controllers i have seen the ebrake is pulled up to 15v by 10k, so it takes 1.5 ma to pull it down. this means an opto with a minimum CTR of over 150% would be fine.
the latest data sheet richard posted shows the ACPL-827C as having 200-400% and the latest one i found for the CNY17-4, the original opto i specified shows 160-320%, so either one would work.
i was looking at an old data sheet when i said the CNY17-4 had 500% typical current gain. I'm not sure if they changed the part or if this was an error. I calculated the values based on the minimum gain of 160%.
thanks richard for pointing out the error. it might make the layout simpler to use the quad optos. my original intention was to mount the tc54 and opto directly on each cell and just tie the outputs together.
shinyballs
1 kW
both the shunt and the meter are marked 50 mv 100A so that means they are matched so that 100A through the shunt will cause a voltage drop of 50 mv and the meter will hit full scale at 50 mv.
connect either wire from the current source to one of the heavy terminals on the shunt then the other side of the shunt to the load. the small terminals are for 2 wires to the meter. the end of the shunt closest to the + terminal will go to the + of the meter.
the voltmeter does not go inline. both + and - go to it
the voltmeter is not marked so it may just be a voltmeter. try connecting a low voltage to it and see. some meters may have a voltage scale but respond to low current so a resistor is needed to calibrate them, so if you put a low voltage on the meter and it pegs that will be the case.
connect either wire from the current source to one of the heavy terminals on the shunt then the other side of the shunt to the load. the small terminals are for 2 wires to the meter. the end of the shunt closest to the + terminal will go to the + of the meter.
the voltmeter does not go inline. both + and - go to it
the voltmeter is not marked so it may just be a voltmeter. try connecting a low voltage to it and see. some meters may have a voltage scale but respond to low current so a resistor is needed to calibrate them, so if you put a low voltage on the meter and it pegs that will be the case.
the hold current of an scr is specified as Ih and is one of the standard parameters given for a device. the STNx10 family of 10A scrs have a max hold current of 30 ma., the STN2540 is 150 ma., the STYN640 is 75 ma. so you can pick the hold current you want and this looks like an excellent way to terminate fast charge on LiFePo4 with a constant voltage source, and when the pack hits the terminal voltage the charge will continue until the current drops below the hold current. Ih is the max current, so it might be less.
a simple control system like gary's board that cuts off the high rate when the first cell hits 3.7v could be used with the scr then supplying the rest of the charge at the desired lower rate.
good idea, oneeye! attached is an app note on scr hold current from ST
good simple tutorial on scr's at
http://www.americanmicrosemi.com/tutorials/scr.htm
a simple control system like gary's board that cuts off the high rate when the first cell hits 3.7v could be used with the scr then supplying the rest of the charge at the desired lower rate.
good idea, oneeye! attached is an app note on scr hold current from ST
good simple tutorial on scr's at
http://www.americanmicrosemi.com/tutorials/scr.htm
shinyballs said:
I have the same ammeter on my scooter. The shunt goes inline, and the ammeter goes across the shunt. The wires from the shunt to the meter can be very small. The voltmeter goes across the battery (not inline).
I used a piece of plastic tubing to cover the shunt so it doesn't short out on anything. That shunt is really huge.
RLT
10 kW
fechter said:
Is the lead length an important factor? Plenty of room in the stock box If I can have the caps on leads about 2 inches long, maybe on a 'daughter board or something; That is the cheapest & easiest route, although if the leads need to be short I can put the controller in a taller box or something.
Thanks!
The resistance of the leads is important, so you'd want to extend the leads with some heavy wire. I don't think the length would matter much otherwise.
shinyballs
1 kW
thank you bob, fecht! yea shunt is really huge. is it ok if I use 22g wire for both meters? power source I'm using is 72v 14A
shinyballs said:
No problem. I think that's what I'm using. Just use something with a rugged jacket so it doesn't get damaged easily.
RLT
10 kW
Ok, I need help again. (Fechter !?!?!?! )
In my 48V brushed controller, the voltage across the resistor to the 7815 regulator is 22.8V (+~ 1.0V depending on what I do with the throttle) with 52.4V (no load) coming from the batteries.
So, the resistor is dropping right around 30V; right where you predicted. I know enough about Ohm's Law to calculate the value and the wattage of the resistor to modify it to control 120-132V..... except for one thing:
What do I use for the current value? The miliamps that the hall effect throttle requires? The one or so Amp the motor draws at no load? The 20 or so Amps that I plan to draw at maximum? The maximum of the MOSFETS ? Something in between?
Reverse calculating the 'factory' voltage drop of around 30V and the 1200 ohm resistor that is in there I come up with .025A Could that be right? If so, I should need 4400 ohms at 3 watts. ... Or am I totally wrong???
The new mosfets, schottkys and caps are on the way from digikey; I should already have whatever value of resistor necessary in one of the parts boxes around here, so hopefully by early next week, I'll have a working 130V controller if someone will help me figure out the resistor.
Thanks!
)In my 48V brushed controller, the voltage across the resistor to the 7815 regulator is 22.8V (+~ 1.0V depending on what I do with the throttle) with 52.4V (no load) coming from the batteries.
So, the resistor is dropping right around 30V; right where you predicted. I know enough about Ohm's Law to calculate the value and the wattage of the resistor to modify it to control 120-132V..... except for one thing:
What do I use for the current value? The miliamps that the hall effect throttle requires? The one or so Amp the motor draws at no load? The 20 or so Amps that I plan to draw at maximum? The maximum of the MOSFETS ? Something in between?
Reverse calculating the 'factory' voltage drop of around 30V and the 1200 ohm resistor that is in there I come up with .025A Could that be right? If so, I should need 4400 ohms at 3 watts. ... Or am I totally wrong???
The new mosfets, schottkys and caps are on the way from digikey; I should already have whatever value of resistor necessary in one of the parts boxes around here, so hopefully by early next week, I'll have a working 130V controller if someone will help me figure out the resistor.
Thanks!
RLT said:
That sounds about right.
The .025A is what the hall sensors and control chips take. This load is not too dependent on motor load. The idea is to keep the input to the 7815 at about the same voltage.
RLT
10 kW
Thanks Fechter;
I hope everyone here appreciates your regular (and prompt) advice and assistance as much as I do.
I hope everyone here appreciates your regular (and prompt) advice and assistance as much as I do.
RLT
10 kW
YAY! It works!
The 130V modification of the 48V Yi-Yun (TNC Scooters) brushed motor controller works, first try!
(I thought I fried it when hooking everything up, but all I destroyed is a three dollar HF multimeter.)
Boy, eleven 12V SLAs in series makes much bigger sparks than just four or five of them.
I'll post photos of the modification if anyone wants.
All that was necessary in doing this modification was:
Replace the four MOSFETS with Fairchild 2532s,
Replace the three freewheeling Schottky diodes with 20200s
Replace the four 470mfd 100V caps with seven 330mfd 200V ones*,
Replace the 1200ohm power resistor with ~4500 ohms.
All the necessary parts from Digi-key or Mouser would run about 25 or 30 dollars; I spent more because I ordered several extra of each component, expecting to destroy several before I got it working... But I lucked out and didn't mess any of them up.
Since the 36V and the 48V (and maybe even the 24V ????) Yi-Yun brushed controllers are essentially the same, as far as a base for modification; yet the 48V one costs about 50% more. If you are doing the modification, you might as well buy the lower voltage controller (The 36 V one costs $28 vs. $45 for the 48V one at TNC).
So, this project should only cost about $60 + plus about 2-5 hours work... Not bad for a 120-130 V 60+A brushed motor controller, IMHO.
There wasn't quite enough room inside the box to put all of the caps, so I drilled some holes and have three of the caps on the outside. The smart thing to do would be put the whole thing in a slightly bigger project box, but for now, I'm just going to keep it kind of crude, at least until the rest of the project proves itself to be workable.
Then, I found four 680mF 200V caps of reasonable size in my 'recycled electronic components' box AFTER I finished soldering the others in place; I think I could have gotten all four of them inside the box, and kept everything neater, and still functional, if I had found them sooner.
Had to jury rig the regulator power resistor, because I didn't have the right value laying around, so it is kind of frankenstein-ish right now. What I think I'll do is either rig up sort of an external socket mount so I can just swap resistors if I decide to run at, say, 48 or 60V, or maybe install both resistor values and a switch to make the change. (With the ~4500 ohm resistor, it seems to work OK up to 144V and down to 72V)
I haven't tested it under real life circumstances though. I don't have any rolling platforms ready yet, and not set up to put much of a load on the system, 'on the bench' (got it up to about 3A @ 132V, is all I could do by hand) . But 'on the bench' everything works much better than I expected, and I don't really expect any significant problems when I transfer it to a real world application.
Thanks Fechter and everyone else who provided advice, directly or indirectly.
* the 36 V Yi-Yun controller uses seven 330mF 63 V caps compared to the four 470mF 100V caps that the 48V controller uses, and when I was ordering parts, the 330mF 200v caps seemed like the best way to go by guesstimation; I'd probably do it differently if I do it again..... But it DOES seem to work just fine.
The 130V modification of the 48V Yi-Yun (TNC Scooters) brushed motor controller works, first try!
(I thought I fried it when hooking everything up, but all I destroyed is a three dollar HF multimeter.)
Boy, eleven 12V SLAs in series makes much bigger sparks than just four or five of them.
I'll post photos of the modification if anyone wants.
All that was necessary in doing this modification was:
Replace the four MOSFETS with Fairchild 2532s,
Replace the three freewheeling Schottky diodes with 20200s
Replace the four 470mfd 100V caps with seven 330mfd 200V ones*,
Replace the 1200ohm power resistor with ~4500 ohms.
All the necessary parts from Digi-key or Mouser would run about 25 or 30 dollars; I spent more because I ordered several extra of each component, expecting to destroy several before I got it working... But I lucked out and didn't mess any of them up.
Since the 36V and the 48V (and maybe even the 24V ????) Yi-Yun brushed controllers are essentially the same, as far as a base for modification; yet the 48V one costs about 50% more. If you are doing the modification, you might as well buy the lower voltage controller (The 36 V one costs $28 vs. $45 for the 48V one at TNC).
So, this project should only cost about $60 + plus about 2-5 hours work... Not bad for a 120-130 V 60+A brushed motor controller, IMHO.
There wasn't quite enough room inside the box to put all of the caps, so I drilled some holes and have three of the caps on the outside. The smart thing to do would be put the whole thing in a slightly bigger project box, but for now, I'm just going to keep it kind of crude, at least until the rest of the project proves itself to be workable.
Then, I found four 680mF 200V caps of reasonable size in my 'recycled electronic components' box AFTER I finished soldering the others in place; I think I could have gotten all four of them inside the box, and kept everything neater, and still functional, if I had found them sooner.
Had to jury rig the regulator power resistor, because I didn't have the right value laying around, so it is kind of frankenstein-ish right now. What I think I'll do is either rig up sort of an external socket mount so I can just swap resistors if I decide to run at, say, 48 or 60V, or maybe install both resistor values and a switch to make the change. (With the ~4500 ohm resistor, it seems to work OK up to 144V and down to 72V)
I haven't tested it under real life circumstances though. I don't have any rolling platforms ready yet, and not set up to put much of a load on the system, 'on the bench' (got it up to about 3A @ 132V, is all I could do by hand) . But 'on the bench' everything works much better than I expected, and I don't really expect any significant problems when I transfer it to a real world application.
Thanks Fechter and everyone else who provided advice, directly or indirectly.
* the 36 V Yi-Yun controller uses seven 330mF 63 V caps compared to the four 470mF 100V caps that the 48V controller uses, and when I was ordering parts, the 330mF 200v caps seemed like the best way to go by guesstimation; I'd probably do it differently if I do it again..... But it DOES seem to work just fine.
TylerDurden
100 GW
RLT said:
Ummmmmm...... YES! (ahem) Yes, please.
OK, now you can directly drive one of those surplus treadmill motors.
The real test will be when you draw enough current to hit the limit.
The real test will be when you draw enough current to hit the limit.
RLT
10 kW
fechter said:
Yes; That's what I intended to do. I bought a couple of the '3/4 HP' ones two years ago to replace the motors on my mini-lathe and on another tool, , but never got around to adapting them; That's what I tested it on. (that, and a 480 watt rated @ 36V Amtek motor at 72V) I have one of the 2-1/2 HP" treadmill motors on the way, should be here in a couple of days.
Now all I gotta do is build something to use it on.... and make it all work. Probably not a bicycle. 10-12 SLAs are a little bulky for that. I'm thinking something with three wheels; half way between a scooter and a narrow go cart, with wheels big enough to do light off-pavement.
------------------------------
Tyler; I'll try to get the photos up tonight.
WARNING: It isn't pretty with 3 caps and Frankenresistor dangling from the bottom.
RLT
10 kW
Ok, here are the images for the conversion of the Yi-Yun BRUSHED motor controller (from TNC scooters) to 130VDC. See 3 posts above for some of the earlier info on the modifications, and page 13 of this thread for photos of the stock controller boards.
My reason for doing this conversion is to be able to use the cheap surplus treadmill motors for e-bike, e-scooter, e-go cart propulsion.
http://www.surpluscenter.com/item.asp?UID=2007112602564110&item=10-1783&catname=electric
http://www.surpluscenter.com/item.asp?UID=2007112602564110&item=10-2326&catname=electric
The parts you need to do it are:
Replace the four MOSFETS with Fairchild 2532s,
Replace the three freewheeling Schottky diodes with 20200s
Replace the four 470mfd 100V caps with seven 330mfd 200V ones*,
Replace the 1200ohm power resistor with ~4500 ohms (at least 3 watt).
All the parts above should be available from mouser.com or digikey.com for US$25-$35
Maybe a larger project box if you want to keep things neater and safer.
(*You might also be able to substitute something like four 470mF - 680mF 200V capacitors for the seven 330mF ones, and fit everything inside the original box.)
Larger view here: http://www.az123.com/130v_cont/step1.jpg
------------------------------------------------------------------------------------
Larger view here: http://www.az123.com/130v_cont/Step2.jpg
--------------------------------------------------------------------------------------
Larger view here: http://www.az123.com/130v_cont/Step2B.jpg
-------------------------------------------------------------------------------------
-------------------------------------------------------------------------------------
Larger view here: http://www.az123.com/130v_cont/Step3b.jpg
-------------------------------------------------------------------------------------
Larger view here: http://www.az123.com/130v_cont/Step3c1.jpg
-------------------------------------------------------------------------------------
Larger view here: http://www.az123.com/130v_cont/Step3c2.jpg
-------------------------------------------------------------------------------------
Larger view here: http://www.az123.com/130v_cont/Step4.jpg
--------------------------------------------------------------------------------------
Larger image here:http://www.az123.com/130v_cont/Step5.jpg
--------------------------------------------------------------------------------------
Pretty simple. If I can do it, anyone can.
Just remember that this modification hasn't actually been tested 'on the road' yet. It may not work in a practical application. ( I have no reason to believe that it won't work well on the road, but until it does, don't expect too much.) And that 120 - 140VDC from SLA batteries make some big, hot sparks. that will arc for inches if you make a mistake.
My reason for doing this conversion is to be able to use the cheap surplus treadmill motors for e-bike, e-scooter, e-go cart propulsion.
http://www.surpluscenter.com/item.asp?UID=2007112602564110&item=10-1783&catname=electric
http://www.surpluscenter.com/item.asp?UID=2007112602564110&item=10-2326&catname=electric
The parts you need to do it are:
Replace the four MOSFETS with Fairchild 2532s,
Replace the three freewheeling Schottky diodes with 20200s
Replace the four 470mfd 100V caps with seven 330mfd 200V ones*,
Replace the 1200ohm power resistor with ~4500 ohms (at least 3 watt).
All the parts above should be available from mouser.com or digikey.com for US$25-$35
Maybe a larger project box if you want to keep things neater and safer.
(*You might also be able to substitute something like four 470mF - 680mF 200V capacitors for the seven 330mF ones, and fit everything inside the original box.)
Larger view here: http://www.az123.com/130v_cont/step1.jpg
------------------------------------------------------------------------------------
Larger view here: http://www.az123.com/130v_cont/Step2.jpg
--------------------------------------------------------------------------------------
Larger view here: http://www.az123.com/130v_cont/Step2B.jpg
-------------------------------------------------------------------------------------
-------------------------------------------------------------------------------------
Larger view here: http://www.az123.com/130v_cont/Step3b.jpg
-------------------------------------------------------------------------------------
Larger view here: http://www.az123.com/130v_cont/Step3c1.jpg
-------------------------------------------------------------------------------------
Larger view here: http://www.az123.com/130v_cont/Step3c2.jpg
-------------------------------------------------------------------------------------
Larger view here: http://www.az123.com/130v_cont/Step4.jpg
--------------------------------------------------------------------------------------
Larger image here:http://www.az123.com/130v_cont/Step5.jpg
--------------------------------------------------------------------------------------
Pretty simple. If I can do it, anyone can.
Just remember that this modification hasn't actually been tested 'on the road' yet. It may not work in a practical application. ( I have no reason to believe that it won't work well on the road, but until it does, don't expect too much.) And that 120 - 140VDC from SLA batteries make some big, hot sparks. that will arc for inches if you make a mistake.
The four caps inside the box should be enough (although more is better).
You might be able to skip the ones coming through the bottom.
If you look around, you might find a 5w resistor that has an aluminum body. These kind can be bolted to the heatsink so you could mount it inside the box too.
You can get the FETs with the original style heatsink tabs if you specify the right case style. Should be a TO-220AB. The ones I had came with tabs. Clamps are fine too.
If you keep the outside caps, be sure to glue them to the case with silicone or hot melt glue to prevent the leads from bending and breaking off.
Now for some 120v batteries! Danger! Three DeWalt packs in series?
You might be able to skip the ones coming through the bottom.
If you look around, you might find a 5w resistor that has an aluminum body. These kind can be bolted to the heatsink so you could mount it inside the box too.
You can get the FETs with the original style heatsink tabs if you specify the right case style. Should be a TO-220AB. The ones I had came with tabs. Clamps are fine too.
If you keep the outside caps, be sure to glue them to the case with silicone or hot melt glue to prevent the leads from bending and breaking off.
Now for some 120v batteries! Danger! Three DeWalt packs in series?
RLT
10 kW
fechter said:
Yes; I already have some on order. Couldn't find any the right value, so I'm putting two 2.2K ohm in series.
fechter said:
Before I 'hit the road' I'm either going to graft on another box to cover the external parts, or get a bigger box, OR replace the seven 330mF caps with the four 680mF ones I found in my parts box... I think I can squeeze all four of them inside.
fechter said:
I'm going with SLAs for now; Since this is more of a 'proof of concept' thing at this point, no point in spending the extra money on exotics.
Besides, I already have the SLAs.
I suppose I could use the 32650 "D cell" 5Ah LiIons that I got for my 'stealth e-bike' project... But that project is going to (hopefully) be my primary EV, so I don't want to waste them on this one.
The 32650s are planned to be inside the (larger diameter than normal) frame tubes; three parallel strings to give me 15 Ah / 44V of 'invisible' power.
