Cromotor + MethTek Monster 24FET + 20S lipo build

[hjns]

links to photos restored above

[parabellum]

and in the pictures the copper wire is visible that makes the soldering much much easier

So, you just tie short ends with cooper wire and then sock with solder?

[hjns]

So, you just tie short ends with cooper wire and then sock with solder?

Yes, pretty much.

What I do is the following:

1. Let's take the example of 1x 10AWG + 3x 12AWG. I strip the insulation for about 10 mm from the ends of all the Wires. When soldering both ends, I therefore prefer to have wires of at least 40 mm long, so that in the middle I have at least 20 mm covered with insulation. As the solder tends to flow with the strands to inside the insulation, this gives me some leeway and keep some flexibility.
2. I bind the insulated ends of the 12AWG wires together to prevent any slipping.
3. I then tie the naked ends of the 10AWG and 12 AWG wires together.
4. I put a little bit of solder to a clean tip and start heating the copper wire around the wires.
5. I put the solder (with raisin nucleus) to the copper wire about 1 mm from where my iron touches the copper wires.
6. As soon as my solder starts to flow, I move the solder away from the iron tip, while keeping the solder in contact with the copper wire. The solder will continue to flow, and the heat will continue to spread to the other parts of the copper wire and the encircled large wires, so that in the end the whole connection is so hot that I can see that the 10AWG and 12AWG wires are soaked with solder.
7. Let cool. Put shrinkwrap around. Done

I will post some more pics in this post later today.

[parabellum]

Thanks! I understand the process. It was the missing part in my trials. You are right, it makes it much easier!

[hjns]

So, I screwed up (literally) and attached the freewheel to the Clyte without the proper washer, leading it to bind up. I am in contact with -methods for the appropriate freewheel extractor tool to correct things. He has send one to me free of charge (great service, -methods!!), and it is due to arrive in a couple of days. I will make some pics and post the results. Should not be difficult to solve though.

I now have two harnasses. One for 10S6P which will be used for parallel balanced charging of a maximum of 6 packs of 10S lipos and testing my bike at 37V. Wires are 12AWG. I also finished a harnass for a maximum of 20S3P 13.5Ah pack with 10AWG wires overall. The 72V should give some nice performance to my HT3525. For commuting, 2packs of 10S in series (20S1P 4.5Ah) will provide me with enough power and Ah and charge the lipos in the evening. For long trips during the weekend / holidays, any combination of 4 to 6 packs will provide me with a very nice range, especially in 10S6P 27Ah configuration. Below the harnass for 20S3P, which will be used for commuting in 20S1P config.


The temp sensor has come in, as did another 2 packs of 10S lipos from HK. In a lot of lipo orders from HK during the last 2 years, this is the first time that I received a pack with at least 1 and maybe 2 dead cells. The other pack is fine and well balanced at 3.84 or 3.85V for each cells. However, the pack with the dead cells shows 0.00V for one cell (might be broken balancing wires, because total voltage says 35V) and the other cell shows 1.51V (definitely a loser here). All 8 other cells are nicely balanced at 3.84 or 3.85V. I sent HK a mail and asked for a new Lipo. Let's see how they react.

In the meantime, I have opened controller to take a look. It seems to have 12 mosfets 4110 and 2 caps of 100V. I want to be able to draw minimally 65A and preferably 80-100A. All at 20S lipo (84V max, 72V nom).


I also seem to have found the famous shunt that people seem to fill with solder in order to decrease resistance and allow for higher currents. Can some people confirm this is the shunt?


Now I also noted that people mentioned that the soldered path on the backside of the pcb needs re-inforcement. Can somebody confirm which paths really need another 10AWG wire soldered onto it?


In order to get to 80-100A, should I do something else?

Thanks!

[methods]

I would go to the Technical Reference Area
http://endless-sphere.com/forums/viewforum.php?f=16

It really depends on how hard/how long you are going to run at the elevated currents. Of course you can run 100A for a few seconds as it is - but the longer you run the hotter things will get and the more losses you will have in the traces.

For 60A I would just solder the shunt - remember that you need to have a way to re-calibrate after you do this. A little solder goes a long way too... You dont need to bridge the entire thing.

If you want to go 80A or 100A then I would suggest adding a little more copper to the traces. To determine which need to be beefed up just look at the board. What was already beefed up? But even better... You really should take the time to understand what you are beefing up and why. Once you understand what is going on in there it will make a lot of sense.

So following those traces..... Find the main power in. You have +V and GND. You will find that one of those giant traces goes right to the +V input. The other goes to the GND input. The only other "Heavy" places are where the three phase wires come in. So this tells us that power comes in on V+ and GND, travels over to the row of mosfets, and then out the phase lines. These are the places that need heavy gauge copper. Something equ to 12AWG should suffice.

These come with IRFB4110 fets, 100V caps, 36V to 150V regulator, and built up traces. So besides soldering the shunt and maybe adding some copper to the power traces if you want reliability I would add another 1mF of capacitance to the main rail. You see the biggest capacitors? The ones marked 1,000uF 100V? I would add another one of those (or two). Keep the legs as short as possible. These things are your insurance. You dont need more to make it work but adding more will make it much less likely that you will blow something up.

Make sure any copper you solder to the traces does not contact the housing. Even just adding another 22Gauge wire does a lot - as it gets coated in solder and adds bulk.

-methods

[methods]

Here is an example of what I was doing when I was producing 100A controllers. This is an 18 fet, though you can do it with 12 fets just fine. The reason most 12 fets blow up is a lack of capacitance more than a lack of mosfets.

Capacitors keep the rail steady and filter the huge noise spikes.

Methods 100V 100A Regen Controllers

-methods

[hjns]

Hi methods,

Ah very many thanks for your directions. Yes, I checked the technical area, and found a lot of useful info in the Infineon controller thread. I also remember that I read several threads on CA (re-)calibration (also one of yours), so I will try and dig them up. Several weeks ago, I also read the general introduction into controllers as linked by lfp from the technical area. I also remember your 100V 100A thread, but mainly because of your wedding pics posted there (NICE!). I am fully aware that what I am doing has been done more than 2 years ago by a whole generation of ES modders, and I am thankful for all that experience that I can just retrieve from the ES forum.

Thanks, everybody, for recording that huge collective experience on the forum!!

But -methods, your directions rock, and I beefed up all appropriate traces with twisted copper wire soaked in solder, now that I have at least a limited idea of what I am doing . I do understand that power comes in, needs to go through the fets and then out to the phase lines. I could follow these traces without problem, and indeed the +V and -V were already beefed up with something like a 14AWG copper.

Photo's to follow. I also twisted a copper wire around both shunt wires using micro pliers, and filled the shunt with solder. I will worry about CA calibration later. If anything, the temperature probe on the axis has increased in importance, and I will have to take care that the axis temp does not exceed 100 oF when accelerating or going uphill (or anywhere else for that matter).

I also would like to follow your advice on adding another cap or two. Any suggestions for vendor, because I also remember reading in another thread that the Japanese are the best? BTW, if I blow my controller I assume you can sell me another one, right?

[methods]

Photo's to follow. I also twisted a copper wire around both shunt wires using micro pliers, and filled the shunt with solder. I will worry about CA calibration later.

Bad news friend... like I said it only takes a tiny, tiny amount of solder to greatly increase the current limit. The shunt material is Nichrome - a very high resistance material. Copper is a super, super low resistance material. I am afraid that you will have to remove that copper. Even just a tiny coating of solder will cut the resistance in half.

Also - nichrome is chosen because it does not change resistance over temperature (or at least very little). Copper on the other hand has a huge thermal swing - just FYI.

You will need to calibrate that before riding it or else you will basically be riding with an infinite current limit. That is fine - just make sure you know what you are up for. In reality it will probably be like 150A.

I also would like to follow your advice on adding another cap or two. Any suggestions for vendor, because I also remember reading in another thread that the Japanese are the best? BTW, if I blow my controller I assume you can sell me another one, right?

Anything from Digikey will be fine - at some point I posted the links for fetchter in a thread.

-methods

[hjns]

Bad news friend... like I said it only takes a tiny, tiny amount of solder to greatly increase the current limit. The shunt material is Nichrome - a very high resistance material. Copper is a super, super low resistance material. I am afraid that you will have to remove that copper. Even just a tiny coating of solder will cut the resistance in half.

Also - nichrome is chosen because it does not change resistance over temperature (or at least very little). Copper on the other hand has a huge thermal swing - just FYI.

You will need to calibrate that before riding it or else you will basically be riding with an infinite current limit. That is fine - just make sure you know what you are up for. In reality it will probably be like 150A.

Ouch. Now that is a bit more than what I was aiming for. I guess I will have to take out my solder-sucker and try to remove the copper wire and the solder. The copper wire really was relatively short (6inch) and only there to keep the solder in place. I will try the tiny coating of solder.

Anything from Digikey will be fine - at some point I posted the links for fetchter in a thread.

-methods

Thanks! I will post the results.

[hjns]

Link added in the knowledge base to methods' explanation how to calibrate the CA. I will need it....

[hjns]

[Drunkskunk]

I think you have seriously overdone it with the shunt. it looks like you have effectivly bypassed them with that much solder and copper.

[hjns]

I think you have seriously overdone it with the shunt. it looks like you have effectivly bypassed them with that much solder and copper.

yeah, looking back I had that idea as well... at that time it looked great and shiny..

as mentioned, I will go in and start sucking away most of the solder. Should not be that hard.

[hjns]

Hi all,

Long time, but a lot of work done, including a lot of soldering connectors, more wires, more connectors, lights, etc. All lipos (4.5kg) are currently in the front handlebar bag, which provides a nice counterweight for the rear hubmotor. I tried to put them in the triangle first, but it won't fit the way I want, and I can not easily extract them for charging. The handlebar bag uses a quickfix system, and it is very easy to take the lipos inside and charge them.

Put Method's controller to work at 84V. Works like a charm. Average 25Wh/km, no pedalling. 1st speed (33%) 23km/h, 2nd speed (66%) 48km/h, 3rd speed (100%) 58.8km/h, awesome acceleration, 90A peaks, 60A sustained. Hot motor, need to watch it. Still need to attach the thermometer that I got from Ebay.

Ordered Lyen's 136V 12fet controller. Put it to work at 84V and 126V. At 126V it works like another charm, but a bit more aggressive, 60A peaks, 45A sustained. 1st speed 30km/h, 2nd speed >55km/h, have not yet collected enough balls to try 3rd speed, but I assume that in any of these days I may ask membership of the 40mph club (65km/h). 120V * 45A = >5kW. Hmm, I do need the thermometer...

Both Methods and Lyen have been a great help and responded to all my questions within extremely short times.

Current setups include either 20S2P 9Ah using 4x 10S 4.5Ah bricks, or 30S1P 4.5Ah with only 3x 10S 4.5Ah bricks. I still charge all of them in parallel (10S 4P) on my 10S charger, and in fast charge mode (7A, not balanced), they are back from 3.70V to 4.17V cell level within 45 minutes.

Now, at these speeds over 40km/h (25mph), I really see the need for a better full suspension frame, better brakes (I use now only one hydraulic front disc brake), better and more torque arms (only use one, and the frame twists when doing regen), and a good way to protect the electronics/motor in the rain, because I commute daily on this bike.

Once everything is complete, I will limit the max speed to 40km/h and try to get a Moped licence in Switzerland. Cops are less likely to stop me when they see a Moped numberplate (obligatory for anything powered on two wheels that goes faster than 25km/h and not faster than 40km/h) on my bike. I am pretty sure my bike would not pass a motorbike test (anything over 40km/h). In town, I now pedal at maximally 30km/h, not to raise suspicion. Feels naughty....

[hjns]

Ok, after reading a lot more, and having experienced >55km/h on my modified cheap Chinese ebike, I found a good Fusion frame with some nice components (Shimano Deore XT hydraulics front and aft, with brake handle shifting). It has a triangle, and 6 bricks of 10S lipo should fit in for either 20S3P or 30S2P configuration.




On my current bike, I carry 4 lipos in a handlebar bag. For commuting this is more than fine. However, handling indeed suffers, more than I was willing to admit...



I measured the triangle of the new frame, and created a paper model of the lipo box that fits 6 bricks together with all wires.




So, I will continue reading and thinking what kind of box is within my capabilities, and will still allow me to easily remove the battery pack from the frame for either bulk charging or balanced charging (once a month). After I have fabricated a good, watertight solution, I will move all components from my cheap Chinese ebike to the Fusion. And yes, the plan is then to limit the CA to 40km/h, and have it approved as MoFa in Switzerland.... We will see.

[hjns]

Some time ago, I blew my methods 72v40A controller. I choose the easy way out and ordered a Lyen 121445 controller that is fit for 150V. I now have one Lyen 121445 controller currently driving my rear HT3525 from -methods at either 20S 84V (in town) or 30S 126V (outside in the hills), and it works like a charm. Another Lyen 121445 controller just came in and will be used to drive a front 9C 2810 from -methods on the same bike. Fun!!

Both -methods and Lyen have been very fast to respond to me with detailed and well-thought-through answers. Very good service, keep up the good works!

Of course, I also blew the Lyen controller. However, with the help from Lyen and some on the forum here, I managed to replace the popped 4115 MOSFETs with 4110 FETs from methods controller and got everything working again. I ordered new 4115FETs and they arrived today. So I can get the Lyen controller back to full 4115 again and do 30S again. I know I should not double post, but I would like to keep a nice log of what I have been doing in this thread, so admins, please forgive me....


Nekkid Lyen 124115 controller: non-modified traces on the backside. From here I checked all MOSFETs. The black lines indicate shorted MOSFETs.


Nekkid insides. Lots of wires, not a lot of room, and definitely no room for a small soldering iron......hmm


Closeup of the popped MOSFETs. I can not see any difference with the other MOSFETs....


Screws removed of all Phase C MOSFETs.


I put a tiewrap through every hole on top of the MOSFETs. Then I turned the controller and put it on its side. Then, while gently pulling on the tiewrap of one MOSFET, I heated the three pins on the backside of the board until all solder was flowing, allowing me to pull out the MOSFET. Cleaning up the holes to allow the new FETs in was a pain in the neck.


4 new candidates at lower voltage (4110 is good for 100V max)


I took the opportunity to change my frame. This is the conventional version. It has a cheap Fox front fork, which is actually not bad at all.


And this is after modification to electric. I tried putting the lipos in the triangle, but it completely destroys the stealth look. With the batteries in front of the steering bar, handling is not that bad on this bike. Much better than on my cheap chinese ebike. So there they are again. I also replaced the front fork with my Fox TALAS FIT RLC 140mm, and the rear brake disc has been enlarged to 203mm. Thinking of doing 203 in front as well, but that has time. Also need mud protectors.

The ride is very very nice. Having front and aft suspension and a Dutch saddle makes for a very smooth ride, even at 55 km/hr. I am one happy camper at the moment....

... cant wait for a 2WD at 126V, though...

A MW clone 36V 9A PSU arrived. I will be using it with 2x 48V 7A PSUs to bulk charge 30S at 7A. I will need to pack the PSUs in a box in order to prevent my kids from being electrocuted.... cute..

My Schwalbe Ice Spiker Pro 26x2.10 just arrived. Impressive number of spikes for an impressive price (EUR 80 per tire).


Photo of what I actually received will follow tonight.

I will only ride these when it has actually snowed on the road. As Hugues mentioned that he lost 30% of the spikes, I will be driving very carefully with them, and hopefully they will last a couple of winters.

When it is not snowing and no ice, I just use these specific Ebike tires: Schwalbe Energizer 47-559 26x1.75.

[hjns]

OK, I have been keeping track of my Wh expenses for the last three days. Here are the results:

Day 1
Total distance: 11.36 km
CA volt at end of day: 77.4 V
CA ah used: 3.147 Ah
Calculated Wh (End of day voltage * ah): 243 Wh
ah put back in by the charger at 10S: 7.141 Ah
Calculated Wh as returned by the charger at 37V nominal: 264 Wh

Day 2
Total distance: 12.04 km
CA volt at end of day: 77.4 V
CA ah used: 3.234 Ah
Calculated CA Wh: 250Wh
ah put back in by the charger at 10S: 7.155 Ah
Calculated Wh as returned by the charger at 37V nominal: 264 Wh

On Day 3 (today), I remembered that the CA keeps track of the actual Wh used at the varying voltages during discharge, and that by pushing the button, I can actually see it. This is of course much more accurate than calculating Wh with the end of day voltage. Now I still have to find a way to see on the iCharger how much Wh has been returned into my 4x 10S1P lipo bricks, so that I do not have to calculate that as well. At this moment I calculate with the nominal voltage (37V), so I probably underestimate the Wh that the iCharger put back into the lipos. Anybody has a good idea how to solve this?

Day 3
Total distance: 14.18 km
CA volt at end of day: 77.0 V
CA ah used: 4.078 Ah
Wh as monitored by the CA: 318 Wh
ah put back in by the charger at 10S: 8.291 Ah
Calculated Wh as returned by the charger at 37V nominal: 307 Wh

edit: charger results added. Looks like the CA sees much more accurate Wh than when I calculate it. Obviously, the charger puts more into the lipos than calculated at 37V. Just assuming a linear drop from 42V to 37V will give me 327 Wh (using 39.5V as basis for calculation), but we all know that Voltage tends to drop fast from 4.2 to 3.9 V at cell level, and then slowly goes down until 3.6 or so. Well, whatever.
At least, I have a good idea of how much I can bike. With this config (20S2P 9Ah) I should be able to take a route of at least 24km and still have 20% left in the lipos.

[hjns]

Today I opened 2 throttles, removed the Hall sensor from one, and transplanted it into the other throttle, as originally posted by methods some years ago. It was frighteningly easy. Results are great. I connected the hall to a 5V source. Turning the throttle shows a voltage between 0.85 and 4.08v for the first hall, and between 0.81 and 3.87v for the second hall. So, not too bad.

Here are some pics how I did it. Thanks again to methods, who showed me the way with his post from 2009: viewtopic.php?f=2&t=12356&p=220262&hilit=dual+throttle+methods#p220262
edit: after the fact, I found some more pics from Graykard 2010 who did the same mod: viewtopic.php?p=226193#p226193


As mentioned, I removed the hall sensor + wire from another throttle. Then I drilled a hole next to existing hole and pulled the hall sensor through.



After estimating the amount of space available for the 2 sensors, I decided to glue them together. The end of a plastic zip tie is CA-ed to the back of both sensors to keep them in place.


Here it is obvious that the sensors+wires need to be fixated perpendicular to the moving magnet and at an angle to the entrance.


I removed part of the sensor-holder to make space for the twinsensor and first kept the twinsensor in place with some tape. I then tried if the throttle functioned electrically as it should before CA-ing in place, as described above. Not visible is the single zip tie around both wires to protect the connections/sensors from pulling forces.


Completed throttle with twinsensors and both wires "disguised" in some shrinkwrap.

Checked electrically by:

• Connect a small power source of 5.00volt to red wire, with black to ground.
• Connect voltmeter to white wire, with black to ground.
• With throttle in base position, hall 1 gives 0.85v and hall 2 gives 0.81v
• With throttle in full position, hall 1 gives 4.08v and hall 2 gives 3.87v

[hjns]

I added a child seat fixated to the frame, to bring my 4y/o to school everyday. Initially, I would carry her in the trailer behind my bike, but this is much more convenient. Of course, it completely throws off the balance of my bike, but with my daughter in the back there is no technical riding anyway.

As I had my 20S3P pack on the back rack, I had to move that in order to make place for the child seat. The lipo pack is now temporarily hanging in a pannier bag just under the child seat, on the left side of the rear wheel. I had to reduce the pack to 20S2P, because otherwise it would not have fit the pannier bag, which is ok because I only use it for commuting my daily 11km. It also throws the balance of the bike way to the left. I am in the process of modifying the Falcon triangle bag into a saddle-pack bag that I can fill with lipos and throw over the triangle and fixate with the stock bands. The whole thing needs to be strengthened with alu plating to provide a good fixation of the lipos to the frame and to protect the lipos. The Falcon bag is covering all that, so that one does not look directly upon the alu plates.

Furthermore, I replaced my hydraulic Shimano Deore XT with avid Juicy Seven calipers and handles. I like the way one can adjust the Juicy Seven's sensitivity, and I can come down from 50km/h to full stop with only 1 finger left and right. Awesome.

Lastly, I replaced the curved steering bar with a straight one, added bar ends in the middle like Zombiess has shown somewhere else here on the ES forum, and added a larger mirror. As my weight is now much more to the front, no more wheelies, and I feel much more in control especially at high speeds. I took the opportunity to clean up the wires, and I am continuing to prepare for the final touch: 2WD.

[hjns]

This morning I opened up the Falcon triangle bag, and put it on the table butterfly-like. I inserted a metal frame to hold 3 packs of 10S1P within each wing of the butterfly. So I now have a "saddle-bag" that I can throw over the triangle, with 3packs on either side connected in series, giving me a total of 20S3P 74V 13.5Ah to play with. It is still in design fase, but I can clearly see where this is going. The weight is much much better and the bike feels nicely balanced. Next is to create a better Lipo holder with a similar design that is also watertight.

[hjns]

Without Falcon triangle bag


With Falcon triangle bag


Some more pics

[parabellum]

What rear rack are you using? How many kg or lbs does it support?

[hjns]

What rear rack are you using? How many kg or lbs does it support?

Hi parabellum,

I bought exactly this one: http://www.bikester.ch/fahrradteile/gep ... 11524.html

It will carry 10 Kg.

The rack is attached to the seatpost tube with a 4-screw clamp. Very good. My experience is that with a lipo pack of 9 Kg, there is still absolutely no vertical movement in the rack. There is some sway horizontally, but that has more to do with the full suspension frame than with the actual rack.

Originally, I had bought this rear rack bag with it:
http://www.bikester.ch/fahrradzubehoer/ ... 16098.html


The clamp on the bottom clicks very nicely and very sturdy to the rack. And removing is instantaneous. But then I can not add a child seat on top. Oh, and in case you were wondering, the child seat is not supported by the rear rack, but by the stock clamp that attaches to the frame itself. Like below:

[parabellum]

Oh, and in case you were wondering, the child seat is not supported by the rear rack, but by the stock clamp that attaches to the frame itself.

Yes, I was wondering. Thanks.