Upload Your Circuit Schematics & Requests Here...

[rf]

LVC board is looking good Gary! If you make extras I'm interested.

Still waiting for parts to try a microcontroller based version.

Spoke with the folks at ebike.ca (the CycleAnalyst makers) earlier this week. Sounds like they might be working on a motor controller of their own. Many things the Crystalyte controller won't do -- like regen braking.

I live up a 2-mile, 800-foot rise hill and my last set of brake pads died after 50 miles. They mentioned that regen does work with a lower voltage battery pack, say 24-volts. Wow! I remember when I was still running 36-volts on my ebike (Crystalyte 408) that I did get some regen and braking. Since I switched to 72-volts I don't get any. Apparently you have to go down hill faster than your top speed. (Top speed of my 408 at 36-volts is about 15mph -- above that the CycleAnalyst shows negative amps and you feel noticeable drag.)

Is there a way to convince the motor controller to do more regen? Or do I need to make my battery pack configuration switchable on the fly? I don't care so much about the actual regen, the braking is what I want. The regen I used to have at 36-volts would save me a lot of brake pad wear.


Richard

 

[The7]

LVC board is looking good Gary! If you make extras I'm interested.

Still waiting for parts to try a microcontroller based version.

Spoke with the folks at ebike.ca (the CycleAnalyst makers) earlier this week. Sounds like they might be working on a motor controller of their own. Many things the Crystalyte controller won't do -- like regen braking.

I live up a 2-mile, 800-foot rise hill and my last set of brake pads died after 50 miles. They mentioned that regen does work with a lower voltage battery pack, say 24-volts. Wow! I remember when I was still running 36-volts on my ebike (Crystalyte 408) that I did get some regen and braking. Since I switched to 72-volts I don't get any. Apparently you have to go down hill faster than your top speed. (Top speed of my 408 at 36-volts is about 15mph -- above that the CycleAnalyst shows negative amps and you feel noticeable drag.)

Is there a way to convince the motor controller to do more regen? Or do I need to make my battery pack configuration switchable on the fly? I don't care so much about the actual regen, the braking is what I want. The regen I used to have at 36-volts would save me a lot of brake pad wear.


Richard

Split the 72V battery into two 36V groups.

Use a DPDT relay (or switch) to switch the groups in series for 72V and in parallel for 36V.

For 72V, you have high power mode.
For 36V, you have low power mode and also regenative braking when over 15 mph.

 

[Jozzer]

Split the 72V battery into two 36V groups.

Use a DPDT relay (or switch) to switch the groups in series for 72V and in parallel for 36V.

For 72V, you have high power mode.
For 36V, you have low power mode and also regenative braking when over 15 mph.

What will actually happen is, if you are travelling over 35ish when you throw the switch, you will feel a reasonable braking force for a few moments that dissipates by the time you have slowed to about 25-28mph. (assuming 26" wheel, and from personal experience).
Switching to 24v with a relay would produce more brake force, and slow you down further, but may produce more reverse current than the batteries (or perhaps controller) can deal with, also perhaps there would be a risk of locking the rear wheel if travelling very fast.
For the amount of time te regen braking is actually working (shaving just a few MPH off your speed before it dries up) you get really minimul power back in the batteries. Hardly worth it I'm afraid, unless you have the longest hill in the world on your way to work :lol:

 

[fechter]

Is there a way to convince the motor controller to do more regen? Or do I need to make my battery pack configuration switchable on the fly? I don't care so much about the actual regen, the braking is what I want. The regen I used to have at 36-volts would save me a lot of brake pad wear.
Richard

Check this out:
http://endless-sphere.com/forums/viewtopic.php?t=583&highlight=regen+crystalyte

I don't think anyone has tried this yet. I have all the parts and will try it at some point. A 3 phase bridge rectifier and a cheap brushed motor controller is most of what is needed.

 

[efreak]

i remember an article in one of Toronto dailies that had a guy using a solar panel on th trailer of his e bike that almost doubled the range of his sla batteries gus the handy man or something like that.


efreak

 

[vanilla ice]

A square foot gets you what, 10w at best?

 

[rf]

Forget the regen. I'm mostly interested in braking -- to extend the life of my brake pads, since I live on a hill and they wear out way too fast. The quick and dirty approach may be good enough.

If I switch out my main pack (72-volts) and switch in a 12 or 24 volt battery that might do the trick. Cost of the switching circuitry will likely be less than one set of brake pads. Question is, how small a battery can I get away with? How do I dissipate the waste regen energy without damaging that battery?


Richard

 

[fechter]

If you can send the regen into all the batteries, then just riding the bike dissipates the energy. Otherwise you have sort of a problem. I suppose you could carry around a discharged battery and pump it up with regen while riding, then discharge it with a light bulb or something at home.

I'd be afraid to switch a 12v battery straight to a fast moving hub motor (or any motor). Something could break, blow up, or get really hot. You need some kind of current limiting.

If you just want electric braking only with no regen, then you can use a large resistor switched across the motor. You could have multiple taps on the resistor to change the braking force.

I made a resistor by wrapping a bunch of wire around a metal tube. You could possibly use the bike frame. The wire gets really hot, but having the wind blowing on it keeps it from melting. The more wire you use, the greater area is available to dissipate, so you can keep the temperature reasonable.

 

[rf]

I have a $4 microcontroller I'm hoping to use as the core of my BMS. It's got 20+ lines of digital I/O and six 8-bit A/D converters, and some other stuff. (Got the MCU in kit form with a tiny board and 'glue', for $11.) Works fine on my proto board and accepts programs from my PC no problem. Very easy to use.

I'm hoping to use the onboard A/Ds to monitor individual battery cells. I've located a 4-channel Differential Analog Multiplexer chip (26-cents from Mouser) that I think will work to interface the A/Ds to the battery cells:

http://www.fairchildsemi.com/ds/MM/MM74HC4051.pdf

Looking to use two or three of the battery pack's cells as power for everything as you folks are doing. The MCU board has a small 5-volt regulator and will accept a beefier one if necessary. The multiplexer's input range is defined by the power supply voltage. Five volts range seems plenty for the A123 cells.

This should be sufficient to create a programmable, adjustable, cut-off signal upon low voltage on any cell. With additional parts it should be able to balance the pack too.

Does this make sense, or am I missing something?

Thanks!

Richard

 

[Mark_A_W]

If you can send the regen into all the batteries, then just riding the bike dissipates the energy. Otherwise you have sort of a problem. I suppose you could carry around a discharged battery and pump it up with regen while riding, then discharge it with a light bulb or something at home.

I'd be afraid to switch a 12v battery straight to a fast moving hub motor (or any motor). Something could break, blow up, or get really hot. You need some kind of current limiting.

If you just want electric braking only with no regen, then you can use a large resistor switched across the motor. You could have multiple taps on the resistor to change the braking force.

I made a resistor by wrapping a bunch of wire around a metal tube. You could possibly use the bike frame. The wire gets really hot, but having the wind blowing on it keeps it from melting. The more wire you use, the greater area is available to dissipate, so you can keep the temperature reasonable.

I've heard of regen current being dumped into really, really bright braking lights - like a couple of 150w spotlights (pointing down I guess...or you'll dazzle the car behind....thump.."What was that?")

 

[fechter]

This should be sufficient to create a programmable, adjustable, cut-off signal upon low voltage on any cell. With additional parts it should be able to balance the pack too.

Does this make sense, or am I missing something?

Sounds like it should work in theory. The input voltage range of the mux is limited to around 5v, so you will need to divide the voltages from the batteries to keep the inputs happy. You'll need precision resistors for the dividers to keep it accurate.

How about some kind of display option? If you have a weak cell, it would be good if it told you which one it was.

OT, yes 150w brake lights would slow you down. I was thinking more like 30-40 feet of 18ga wire wrapped around the downtube. You could have both a resistor and a really bright brake light in parallel.

 

[rf]

How about some kind of display option? If you have a weak cell, it would be good if it told you which one it was.

Good idea. All kinds of things become possible with the MCU. Displays from an LED or two, to multi-line scrolling displays. Cheapest is to connect the MCU's serial programming port to your PC or palmtop and see what it has to say. An interface to accept a thumb/flash drive for logging might be nice too -- haven't figured the cost on that yet.

OT, yes 150w brake lights would slow you down. I was thinking more like 30-40 feet of 18ga wire wrapped around the downtube. You could have both a resistor and a really bright brake light in parallel.

Very cool idea. Would be very impressive to see regen energy doing tangible, useful work.

Thanks

Richard

 

[Beagle123]

My First Attempt at a Circuit

Update: This circuit has been replaced by the one on my next post. Don't bother with this one.

Recently I purchaced this power supply to charge my batteries:

<a href="http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&rd=1&item=290174751148&ssPageName=STRK:MEWA:IT&ih=019">0-10v, 50A power supply</a>

However, I can't just plug it in and forget about it because it won't shut off when the batteries are fully charged. I'd like to have the convenience of having it shut-off by itself.

So I designed this circuit. I want to use relays to disconnect the AC current to the power supply when the current to the batteries gets close to zero. The way I want it to work is to use a 50 amp current sensor to read the current going to the batteries. This current sensor will output a voltage that represents how much current is flowing. The voltage is sent to a transistor that holds the relays closed. When the batteries are charging there will be plenty of current and therefore voltage to the transistor and therefore current to the relays, so the power supply stays plugged in. But as the current to the batteries drops, the power to the transistor and relays will be starved of voltage/current until the relays shut-off. And when the relays shut off, the power shuts off to the whole circuit because the circuit will get its power from the 4.2v current to the batteries.


I hope to find a current limiting component that I can use between the 4.2v power leads and the circuit to limit the current to the amount of current the realys need to stay closed. I don't know how you guys figure this stuff out, but I'm just going to test the relays when I get them.

I ordered these:

http://www.allelectronics.com/cgi-bin/item/RLY-642/500/6VDC_DPDT_RELAY,_PC._MT._.html

I just noticed that they have a 50 ohm coil, so using ohms law it should be in the neighborhood of 80mA?

Do you guys think its ok to just pull-the-plug on the power supply? Would this be a bad idea?

One potential problem with this circuit is that when the current is flowing freely to the battery, the relay may be getting too much current. Hopefully, a current regulator will solve that problem.

Alos a problem stems from the fact that the transistor isn't an on/off switch--it will give a variable amount of current. I want the relays to turn off when the sensor is near zero volts, but I'm sure the relays will turn-off well before then. So I need somthing that will maintain the open connection until the voltage gets near zero. Are there transistors that work like an on/off switch?


Am I missing anything? suggestions?

 

[Beagle123]

Fixed the Design

I figured out how to make the circuit correctly. I made a voltage divider ussing a 100K resistor and a 1K resistor making a contstant voltage of 0.5v. Then I used an op amp to compare that voltage to the output of the current sensor. When the current gets really low, the voltage will drop below 0.05v which will activate the transistor to turn off the relay.

And the best part is that it works! I set it up on my breadboard. The only problem I was having was that the transistor was only providing 20mA of current. I'm guessing that another transistor will be able to output more amps.

Should I use a FET instead of a transistor?

Have a look. What do you think?

 

[fechter]

I'd recommend a NPN transistor to reduce the base current.

The diode needs to go backward across the relay coil to conduct the voltage spike when the current drops.

You might have trouble finding a relay that runs on 3v. They make solid state relays that would work. Those could run straight off the op-amp output so you can skip the transistor and the diode. You can get a solid state relays pretty cheap surplus (highly recommended).

You need a switch to get the thing started.

If the battery is completely dead or disconnected, it won't start (good).

You need to check the specs on the current sensor. It needs to be powered. It might not work correctly on an unregulated supply. Most of them take at least 5v to be in spec. You might need a small separate power supply to run the current sensor and op-amp (recommended). It might be OK the way you have it.

The output of the current sensor might be different than 0v at zero current. Use a 10 turn trimmer pot instead of R1 and R2 to make it adjustable.

 

[Beagle123]

Thanks for the good advise fetcher.

I'd recommend a NPN transistor to reduce the base current.

I tried that first, but the PNP works much better because the op amp didn't have enough current to power the relay on. It put out decent voltage (4v), but when it was connected to the transistor, it sagged to 3v which caused the transistor to not be powerful enough to hold the relay open. However, the op amp does have enough power to shut off the PNP transistor. Its just easier to thurn the power off than hold it on.

The diode needs to go backward across the relay coil to conduct the voltage spike when the current drops.

Excellent! I was wondering about that.

You might have trouble finding a relay that runs on 3v.

I bought these:

http://www.allelectronics.com/cgi-bin/item/RLY-642/500/6VDC_DPDT_RELAY,_PC._MT._.html

They're 6v, I'll bet they'll activate on 4v. The solid state ones don't tend to handle as much current.

You need a switch to get the thing started.

I'm amazed that you noticed that

You need to check the specs on the current sensor. It needs to be powered. It might not work correctly on an unregulated supply.

I'm totally unclear on powering components. It seems strange to hook a tiny little component into a line thats outputting 50 amps. But it seems like it would work fine in some cases. The relay for example has a 50 ohm coil, so it you connected it to the 4v only 0.08 amps should flow, so I think that will work. But what about connecting my little op amp chip or the transistor? (I guess the transistor would be ok because it only uses the same 0.08 amps as the relay). I'm guessing the voltage divider would be ok because it has so much resistance that barely any current would flow. The chip and the current sensor are my only questions.

I looked up the specs on the current sensor, and you were right on all counts. 0.6v = no current, it can only handle 10mA on the Vcc, and its minimum current Icc is 4.5v. However, I bet it will run on less. I don't need it to be an accurate measuring device, I just need it to signal when its low current.

So, I tested a 470 ohm resistor across a 4v battery, and there is no voltage drop. This would limit the current to about 8 mA as the current sensor needs. I guess that's how you do it?

I'll look up the specs on the op amp, and see if it needs current limited too.

current sensor specs:
http://www.allegromicro.com/en/Products/Part_Numbers/0755/0755-050.pdf

 

[Beagle123]

I just noticed a problem: When the power supply is cut-off by the transistor, the voltage on the gate will drop, and the PNP transistor will open the relays using the batteries' power. Its essentially a zombie circuit that you can't kill. I guess I'll have to use a NPN transistor after all.

 

[Beagle123]

I figured-out that I had the NPN transistor connected backwards. When I switched it, it was able to hold the relay open. However, I'm getting unexpected behavior from the op amp. I connect the power pin of the op amp to 4.65v, but when it sends output to the transistor its only 3.8v. Since the transistor is variable, it doesn't send out the full voltage--it outputs about 3.6v. Since the coil of the relay uses current, it drops the voltage across the coil to 2.85v.

This could mostly be solved if the op amp didn't have so much resistance. Its droppong the voltage 0.8v. Is this normal? Perhaps my op amp is wrong?

 

[fechter]

The output of the op amp will be a bit lower than the supply unless you have a "rail to rail" output type. You could use a different op amp.

You could possibly find a high current op amp that could drive the relay directly without a transistor. This would further reduce the voltage drop.

Otherwise, you could put the NPN transistor on the low side of the relay coil and drive it as a saturated switch, so there would be less voltage drop.

 

[Beagle123]

I found another solution. I'm using a comparator instead. I was missing a resistor when I tried it the first time. It needs a "bring-up" resistor (or something like that). Now it has 4.1v to the relay, and 3.2 when under load (relay on).

New circuit:

 

[fechter]

You're getting there.

You might consider putting a diode between the battery and the rest of the circuit so when the charger goes off, the battery can't discharge back into the circuit. If you do this, the charger output voltage will need to be slighty higher to compensate for the voltage drop of the diode.

Hmm.... seems to me if you used a 5v supply and a plain old diode, you'd wind up right about at the desired voltage. This might make trimming the output voltage of the supply easier.

You could also put some kind of meter on the output of the current sensor to monitor charging current. A reasonably sensitive analog meter with a series resistor could be calibrated about anywhere. A silicon diode in series with the analog meter will take out the 0.6v zero offset.

If you used a solid state relay, you could drive it direcly from an op amp and skip the transistor. $3.50 surplus:
http://www.allelectronics.com/cgi-bin/item/SRLY-22/500500/SOLID_STATE_RELAY,_AC_CONTROL_VOLTAGE_(USED)_.html
some others here:
http://www.mpja.com/prodinfo.asp?number=17154+RL

 

[Rob Mcgregor]

Hi Folks,

I'm new to the board and I've been fascinated with your bikes and knowledge. I want to thank you for all your sharing and community.

I was thinking about starting a new thread but this thread has "schematic requests" in the title so I thought I might start here. I hope I'm not changing the subject too much.

I was interested in modifying two DC controllers to over 48V. I was noting that DC Crystalyte motors and controllers could be very efficient, weigh less, less freewheel resistance, and most importantly- cost less. I was noticing if I could run it at 72V, two dc motors in 26" wheels and two controllers would run under $500 and get me to theoretically 47 mph at 75% efficiency, 29 lbs total, and 100 N-m torque. If I did the same w/ 20" wheels, I should theoretically be able to go 43 mph at 83% efficiency with 393 Newtons force forward. Did I get these numbers right? I also heard the Killacycle, 0-60 in .97 seconds and 1/4 mile in 8 sec 158 mph, uses a dc motor.

So can the 48V 35A Crystalyte DC controller be modified for 72V? The controller presently tops out at 48V. Could I drop in 4110s or are they not compatible w/ the DC controllers? Where can I find the schematics of these controllers?

Also, I imagine 72V will cause the brushes to go bad sooner. Are the replacement brushes easy to find? Are they hard to change? I don't mind changing brushes if they are easy to find, not too expensive, and not too hard to change.

You guys are awesome,
Rob McGregor

 

[fechter]

I assume you're talking about brushed motor controllers.

They can be modified for increased voltage, but their layout is different than the brushless controllers.

I've never seen the insides of the ones you are talking about, so I'm not sure what would be involved.

Worst case, you'd need to replace the FETs, the freewheel diodes, the main capacitors, and deal with the voltage regulator.

You might find cheap 72v controllers at TNC scooters.

 

[Rob Mcgregor]

Sorry, yes I mean brushed controller and motor. Thanks for the TNC idea. However, they don't carry 72V controllers. I especially like their lockable bike luggage trunks for $15. There must be some brushed controller schematics online somewhere.

Thanks so much,
Rob McGregor

 

[vanilla ice]

I wonder how high you could go with a 48v'er from TNC? Maybe they will do 60 or 72v right out of the box??