HT3525+Lyen124115+30S want more power... safely

24S is way easier to deal with right now. Above 24S and you get higher losses along with reduced power handling. Only run >24S if you REALLY need more speed. Under 100V, you can get a 18 FET from Lyen, and run it at 10kW all day without any real worry. Your HS saturates at something like 7.5kW IIRC anyways, so even a 12 FET 4110 will do the job here. If you want big power, you need a big motor, and an even bigger controller. You can't just toss more power into the motor and hope it puts out more, it basically just gives up right around where you are.

If you want plug and play big power, go 36 FET IRFB4110, 24 FET as a minimum. They're absolutely gigantic controllers, but they should be able to melt any hub motor available.

Honestly, the only real logical step forward from where you are, is 36 FET 4110 / 20AH+ 24S LiPo / 'speed' wind hubzilla (which I think is available). Toss the motor in a 20" wheel with an ML75 tire and call it a day.

zombiess, I was testing with a big damn supply, pulsing it around 50% @ 10kHz or so. Most any FET I ever tossed high current DC into just exploded immediately. It could just be my obsession with really thick wires and tons of capacitance, however. I held it like this for a few minutes before the sink got quite hot and the FET went short. I probably should mount some FET's onto a snazzy liquid cooling setup and see what they can take. I have much snazzier equipment now, so I might be able to make really meaningful conclusions as to the absolute power handling, and thermal limits. I might try this out tomorrow.

HI ZOMGVTEK and zombiess,

Thanks for providing so many details and advice. You have given much more feedback than I expected when posting, and it is a great learning curve for me. Thanks again!

Just let me get some assumptions straight.

About me:
I am a MD with a PhD in neurosurgery. That means that I am perfectly comfortable around highly complicated technologies of which I know nothing about, trusting other people’s skills in making things work that I don’t know anything about, and learning new things everyday. I am very comfortable working manually in a small operation space.

With regards to electronics, I am reasonable comfortable soldering small parts (resistors, caps, fets, even ICs) onto a PCB, but I usually do not understand an electronic circuit even when engineers explain it to me in baby talk. Therefore, generally speaking, I don’t know sh!t about electronics, but I can identify and replace a smoked MOSFET. Furthermore, I have an affinity for flowing things. Flowing body fluids, oil flow, water flow, flowing electrons....etc.

Last but not least, I am reasonably uncomfortable around high voltage setups like my 126V 30S pack. That means that I will protect myself when working on it, will use multiple layers of shrink wrap, and so far managed to get KFF twice. Once from connecting a parallel pack in series (resulting in molten 5mm gold connectors), and once from inverting the two 5S tabs on my 10S packs to the balancing board (resulting in molten JST connectors).

My expectations:
I do not NEED a high power setup (who does?). I “need” an E-bike to do my commute (which will be increasing from 6 to 15km in 2 months due to a move). I “prefer” to tinker with my bike and see how far I can push it, because it is fun. I always try to combine need and fun, so I tried 2WD first. That’s why I have a front 9C 2810 lying around in my garage and a rear HT3525 on my bike. Both were controlled with a Lyen MKII 124115 controller. I did not like it, and continued with the rear motor only.

Now I am trying to push the single motor setup as far as I can. In case I fry the motor, I will continue with another. If I fry the controller, I can continue with the other Lyen MKII. Note that I am not against buying a bigger motor and/or a bigger controller, and I agree that from reliability point of view, that is the best way forward. However, I feel it is kind of a waste (not to mention fun) not to try and tease my current components to their highest performance levels. And that is the reason for this thread. How can I optimize my current setup (rear HT + Lyen MKII 12 Fet + 30S) before having to buy a bigger controller?

Your comments:
“sorry if ZOMGVTEK and I are getting a bit techy and off topic on you, but it's related to your initial question.”
- see above. You both have given much more feedback than I expected when posting, and it is a great learning curve for me. Thanks!

“what is your comfort level with working with electronics? Think you could attempt the same mods I've done (I copied markcycle so it's not my mod idea, I'm just impressed with the results from it)?”
- maybe I could do something similar with my 12 fet controller. I have been reading your threads avidly, and I am very impressed. First thing now is to install a temp probe and record temps in a CA analogger that I just ordered from Grin.

“Why not run an 18 or 24 FET and not have to worry so much?”
- because the “worrying” is my learning curve, and I like it.

“be careful with that pack voltage, treat it with respect.”
- I will. I do.

“if you try running 80A it will probably work for a while, but if it fails don’t be surprised.”
- I won’t. Actually, I am surprised it worked up to now. And I am looking into ways of cooling the controller with oil in order to keep temps low. If there is anything I learned from you guys, it is that I need to keep temps around 25oC to get the most current through my IRFB4115 MOSFETs.

“IRFB4110PBF’s are just better for power”
- One day I may replace all my 4115 fets with 4110 fets, decrease voltage to 100V and increase battery current to 120A. But I would like to try cooling my 4115 fets first.

“that is why having a high block time is bad.”
- I understand, and changed block time to 0.1s using the hexedited software.

“18 FET is the minimum for the 4115’s” / “go for the 18 FET controller IRFB4115 controller and set it for 60A battery and probably 140A phase and you'll probably be OK”
- Now I am in doubt with regards to the next controller. However, I would like to postpone that discussion until after I have destroyed my current controllers….

“I'm going to have more high power high voltage controllers than I know what to do with soon LOL, going to have to sell some off on here.”
- one prospective buyer is right here….

“Honestly, the only real logical step forward from where you are, is 36 FET 4110 / 20AH+ 24S LiPo / 'speed' wind hubzilla (which I think is available). Toss the motor in a 20" wheel with an ML75 tire and call it a day.”
- Yeah, you are probably right, and it looks like that in the future I will go that way. However, then I also want a high quality downhill FS frame with a super-duper dual crown 203 fork and the works. So I do not see that coming within the next 6 months….

“I probably should mount some FET's onto a snazzy liquid cooling setup and see what they can take. I have much snazzier equipment now, so I might be able to make really meaningful conclusions as to the absolute power handling, and thermal limits.”

• Now I would like to do something like that but in real life conditions, aka climb a nearby 10-20% hill, record the currents and temps using a CA analogger, and compare the stock controller versus oil cooled controller with a fixed max speed (e.g. 30km/h) on that same hill. That should give me a baseline current/temp curve upon which I can look at the respective changes from baseline from the different mods I plan. The way I see it, my controller needs to be modded first, then the motor. Heat is my biggest enemy, and that will be the main focus of my efforts…
• I was hoping I could copy some ideas from Tomshardware here and from this mineral oil cooled pc site… http://www.pugetsystems.com/submerged.php I should be able to submerge the complete controller into a slightly larger case filled with ATF oil. The shear amount of oil should transfer heat and delay temp increases for quite some time. Otherwise, the addition of a small radiator with a 12V fan would also require a pump to move oil around, but would definitely be more effective in transferring heat. Only one way to find out….

Thanks again for your feedback!

You sound fairly capable of understanding the issues associated with punishing hardware, which is always good. Monitoring the controller temp is less important than the motor temp in your case. Both are near the limits, but the motor will be MUCH more costly to repair/replace, and harder to detect the threshold before failure. At these power levels, a temp probe on the windings of the motor is essentially required.

The rough estimation '100%' value on your controller is 65A battery, 170A phase. You claim to be running at 60A right now as is, which is basically as far as the controller will go. You can try and push it to 65A, but thats going to REALLY punish the controller for minimal gains. If you're determined to cook some FET's, go for it and see what it will take. Just have a spare controller handy. Cooling the controller better will help you run it longer at these limits, but it won't help you push it much beyond 65A. The best way to eek some extra power out of your current setup is to run higher voltage. The FET's should be OK up to about 140 MAYBIE 145V peak. Tag another 4-6S pack on and charge to 4.1V/cell.

In case you're not aware, your motors efficiency decreases exponentially at higher power. It hits a point, where putting another 1kW into it, results to 100W of mechanical output, and 900W of heat. Thats how you melt a motor really fast. The motor can just barely dump a few hundred watts of heat, things can get out of control really fast. You are already at the limits of your motor. Putting more power in will not result in any real noticeable additional output.

Basically, you're past the practical limits already, and riding the absolute limits. About the only way to get much more power involves replacing components.

I should be able to submerge the complete controller into a slightly larger case filled with ATF oil.

Click to expand...

Interesting article on Tomshardware. Can't wait to see the outcome for your controller.
Sidenote: how about lacing up a 16-20" (or equiv metric) wheel?

Cheers,
Kevo

I think im much better killing controllers.
18fet 4110 at 24s 65a main ll unsolder the ca wire from the shunt if i push it in a long hill and the controller is not even harm from the outside.

OK, thanks. So my current priorities are to:

1. take it easy (< 50A) until...
2. add temp probe to motor with life feedback
3. add temp probe to controller with life feedback
4. set some baseline temps at different currents and speeds
5. start improving temp control for both motor and controller (I love the oil bath cooling idea )
6. see how far I can push things without boiling the oil

Maybe somebody will beat me, but otherwise expect to see here the first photo's from a controller in an ATF oil bath in the coming months...

hjns said:

Maybe somebody will beat me, but otherwise expect to see here the first photo's from a controller in an ATF oil bath in the coming months...

Click to expand...

Keep us informed! I am waiting long time for somebody actually try that, it probably will be oil mess in first intent! I oil filled 3 hubs to get 1st right, almost.

Hi all,

I need some help here.

The idea is that I install temp probes in my reserve equipment first (testlab / sandground). Aka first in my front 9C and extra Lyen controller on a second bike, to make sure that when I break something I can still commute with my main equipment. After making sure the temp probes work, I will create a baseline for both, running up a big 10-20% hill with (1) a fixed speed (WOT with speed fixed in CA), measuring max amps and max temp, and (2) with fixed amps (WOT with fixed current in the CA), measuring avg/max speed and max temp. Doing this at multiple speeds (e.g., 20 km/h and 20 mph) and multiple amps (e.g., 30A and 50A), and repeat every individual variation so that I have minimally 2 readings for each parameter at each variation, should give me a nice standardized but real life environment for comparison. Then add the oil and do it again. Then do some more tweaking if possible, unless ZOM is right and there is nothing to tweak anymore...

I don't intend to try WOT without CA limitation in fear of burning something until I am reasonably sure the motor/controller can take it. Only then will I transfer the methods to my main commuter.

However, my current issue is a bit stupid...

I wanted to open up my 9C front motor to try and install a temp probe. Unfortunately, I seem unable to remove the cover from my 9C. After removing the screws I have put a chisel into the groove and hammered around. The cover can be lifted 1.5-2 cm, and is then held by something around the axle. I don't know what.

Can someone provide some guidance how to remove the cover?

Thanks!!


Various states of undress:


Using the chisel


Only able to open for 2 cm....



Peeking inside through the hole.



Around the axle, what did I miss?

(edit: put images inline)

The bearing is a rather snug fit, and often takes some 'persuasion' to get it off.

I just whack the opposing side axle on the ground while pulling the cover up. Might not be the recommended method.

ZOMGVTEK said:

The bearing is a rather snug fit, and often takes some 'persuasion' to get it off.

I just whack the opposing side axle on the ground while pulling the cover up. Might not be the recommended method.

Click to expand...

Make sure when whacking that opposing axle on the ground you have a nut on it that's flush with the top to distribute the force more evenly or possibly suffer bent threads on the axle!

ZOMGVTEK said:

The bearing is a rather snug fit, and often takes some 'persuasion' to get it off.

I just whack the opposing side axle on the ground while pulling the cover up. Might not be the recommended method.

Click to expand...

oh.. thanks ZOMGVTEK and zombies.

I found a nice thread by Auraslip here on opening the hubmotor. I still feel very uncomfortable putting my weight on the motor and/or axle like suggested. The thread is great, though. Thanks Auraslip!

I read the threads, watched the vids, and I believe I just need to buy a 3 jaw pulley.

OK. I conquered my fear of damaging the motor, stepped with both feet on the wheel, and pulled with all my might on the cover. It came off !! Thanks again for the encouragement!
I may still order the 3-jaw puller...

Next steps:

1. Add temp probe
2. Have a good look at the phase wires
3. Increase heat tolerance for the hall wires (to be replaced? Or just add heat shrink?)
4. Wait for the high temp silicon to come in
5. Drill oil filling hole
6. Drill vent hole
7. Seal everything with silicon
8. Close and start taking baseline measurements

View of the hall wires:


Nicely sealed wire entry into the axle:

Any suggestions as to which combination of hall wires + phase wires + temp probe wires will fit through that tiny tiny axle hole?

I am looking at heatshrinking individual hall wires. That would leave room for thicker phase wires, also to be heatshrinked individually. 12G?

Also, the hall wire pcb is seated on a small piece of insulation paper. What will happen if that drains with oil? Should I replace it with something?

Good thread here.

Always useful to see real world experience at melting stuff. Reminds me of Luke's experiments. We all learn from them.

Im grtting pretty inspired readingthis thread. I wanna open my motor now! However, what do you do with the bolts that are too tight to come off? Or what would you do with bolts that are too tight to come off? I heard the bolts are made of some soft metal that is easily stripped....

All the bolts on my 9C were tight, but not too tight to come off. Just make sure you have the correct hex/allen key that fits the bolt perfectly. I have both metric and SAE (american) sizes.

However, I find that the bolts are not an issue. The soft motor cover threads are an issue, especially for the bolts that fixate the disc rotor. I have destroyed the threads of at least one hole when trying to tighten it up to specs.

Sorry, no updates today. I broke two more spokes yesterday, and needed the whole evening to replace the spokes and true the wheel again.

As a separate topic, after some great advice from MadRhino (thanks!!) I ordered a new rim and spokes. I describe the details in my build thread not to clutter this thread here.

I bought some extra heatshrink and some 400oC isolation tape. I am still in doubt how much wire can pass through the axle. I will post pics later. Still waiting for the high temp silicon gasket.

Thinking about it. It is really the amount of copper that can pass through the axle hole that matters most. Once in the motor there is enough space again for thick insulation and loads of heatshrink. Anybody tried passing bundles of enameled motor wires through the axle only? Would that save space as compared to nekkid wire covered with heatshrink?

hjns said:

Anybody tried passing bundles of enameled motor wires through the axle only?

Click to expand...

Not only but with comparably thin shrink tube wrap, you do not want shorts there, right?

zombiess said:

One of the first things I do is turn block time down to 0.1 Sec (need to hex edit the software to do this) to help keep it alive. my own controller for high voltage setups to make sure everything is exactly as I want, but it can be tedious with the mods I do.

I'd suggest dropping your block time down to 1.0S, don't hammer it from a dead stop and drop your battery current down to 45A and phase amps to 125A if you'd like to see it live a while. You might get lucky and be OK, but your certainly pushing the controller really hard right now.

Click to expand...

i just found this very interesting thread. I hope you won't kill me if i ask what block time is?

I have just had similar drama trying to pick 4110 or 4115 for my back up controller. At 100v Cell_man explained me that 100 v is actually between those two. Complete max for 4110.. and "not enough" for 4115.

So in terms of safety, i would either go down below 100v and run 4110 or keep 100v with 4115 and reduce A to 70...

My logic was to use 26'' wheel with 5405 [to increase speed (bigger wheel) and keep high torque [5 windings] - but i can read often here that 20inch is easier on the setup, stronger [stiffer] and on the frame?

My other setup will be 5403 with 16'' moto rim [20'' bicycle] and 16'' front wheels in future.

block time setting is the time before the current limitation of the controller sets in. E.g., if you have programmed your infineon based controller to a maximum of 50A battery current and a block time of 1.0 second, then in the very first second the controller will not limit the current. The current limitation will only start after 1.0 second.

This setting seems to be introduced for lower power setups, to allow temporary higher currents to the motor to make starting from complete standstill a bit easier, and still remain within legal requirements.

Read more here.
To get an idea how to play with it, see here.

I always run my block time at 0.1S on all my controllers. I hex edited the software to allow this + mod several other settings. Old school hex editing skills come in handy

zombiess said:

I always run my block time at 0.1S on all my controllers. I hex edited the software to allow this + mod several other settings. Old school hex editing skills come in handy

Click to expand...

Hi zombiess,

Yes, you did describe that several times. However, do you think this is also necessary for the more mainstream "intermediate power" setups, like between 1.5kW and 5.0kW? Seems to me that 0.1s block time may be overkill, and >2.0s may be too long. However, this is only my gut feeling talking. I don't have the popped MOSFETs to prove it... yet.
Also, what else did you modifiy? Regen? Cruise control? Must be nice to cruise at 10kW....

I just ordered a bunch of high temp wires, including 0.25mm2 (23AWG) for hall/temp wires (180oC CHF 0.70/m), and 6.0mm2 (10AWG) for phase (180oC CHF 3.55/m), and just for the looks I also ordered dual 6.0mm2 for battery (85oC CHF 10.95/m). For the EU colleagues who are curious where I sourced it: I went to Conrad.ch.

Even better block time thread here.

hjns said:

zombiess said:

I always run my block time at 0.1S on all my controllers. I hex edited the software to allow this + mod several other settings. Old school hex editing skills come in handy

Click to expand...

Hi zombiess,

Yes, you did describe that several times. However, do you think this is also necessary for the more mainstream "intermediate power" setups, like between 1.5kW and 5.0kW? Seems to me that 0.1s block time may be overkill, and >2.0s may be too long. However, this is only my gut feeling talking. I don't have the popped MOSFETs to prove it... yet.
Also, what else did you modifiy? Regen? Cruise control? Must be nice to cruise at 10kW....

Click to expand...

I can modify any of the settings I want and even figured out how to expand some of the tables to add more choices. Works great so far. I choose 0.1S block time to make sure I don't over shoot the FET control algorithm for too long so I don't have to worry so much about an over current condition which could hurt the FETs. At my power level block time just isn't needed. I've tired programming it for 0.0S but it doesn't seem to make any difference vs setting it for 0.1S. I've also programmed my lower powered bike to 0.1S block time and it too works fine, I figure why risk the FETs more than I have to. I barely notice any difference in the settings unless I really lower the power down. Ebikes + nerd skills are such a great match. This is by far my favorite hobby and soon to be business. I'm in this for the long haul and trying to learn everything I can.

BTW, 10-12KW is just absolutely nutty fun to ride around on when you learn to control the power, just make sure you carry a large battery pack and a good controller (and spare fuses). I'm hoping I can get my throttle interface refined some more tomorrow so more people can enjoy this kind of fun with cheap Xie Chang controllers, they work great for most hub motors but their problem has been crappy throttle control which I'm working hard to fix.