Trouble diagnosing controller problem.

[jamiejackherer]

So about a year ago I bought a 48v 1000w ebike kit (the controller is 28a~35a) made a 48v battery from 20ah eig Li-Polymer cells (I chose these as they deliver 100a continuous and 200a for 10 second bursts) and put it all together.

I've had no problems for the past year and have been thinking about upgrading the bike to a 72v battery/controller. But I thought I would try to mod my current setup for higher amps, so I performed a shunt mod to my controller.

This worked for the initial test and really improved the torque of the motor by a dramatic amount! So I decided to make it permanent and a more proper upgrade by adding a wire to each of the phase connections so I had 6 (2 x 3) phase wires coming from the controller to allow more current to pass through, I also used 100a (4 awg) welding wire for the main, long phase wires. I also replaced the phase wires into the motor with the thickest wires that could fit into the motor hub axle.

I set off for my second test and the bike rode for around 200 meters and the bike stopped giving power then the voltage dropped out. I assumed the MOSFETs on the controller would have blown but after testing they are all ok. I ordered a new 48v 50a 1500w controller (next day delivery) anyway and hooked it up. But the voltage just drops out when I connect the controller.

If I use a multimeter to test my battery (via the BMS) it reads 51.4v (full charge minus the initial testing of the shunt mod). But when I connect the controller, the battery reads 39.5v (the lowest voltage for my battery).

I have checked inside the motor and everything is fine there.

I am now stuck as to what is happening.

Any suggestions?

Sent from my SM-G970F using Tapatalk

 

[amberwolf]

This worked for the initial test and really improved the torque of the motor by a dramatic amount!

That's because the controller can no longer know how much currnet is actually flowing, so it has no way to protect itself or the motor against overloads, etc. So it basically gives as much current as the parts are physically capable of, since the shunt is now just a wire and doesn't have enough voltage drop across it at the currents used to tell the controller to stop doing that.

Gives more performance, but removes the protection against overcurrent/overload. Sometimes it works great...sometimes it results in blown FETs or gates (or both), or overheated motor / damaged windings.

I set off for my second test and the bike rode for around 200 meters and the bike stopped giving power then the voltage dropped out. I assumed the MOSFETs on the controller would have blown but after testing they are all ok. I ordered a new 48v 50a 1500w controller (next day delivery) anyway and hooked it up.

Does the motor turn normally by hand, both when off the controller and when connected?

Does the problem occur without the motor connected (just the controller ot battery)?

Does the battery have *any* electrical connection to teh bike frame?

Does the controller have *any* electrical connection to the bike frame?

Does the problem happen if the controller and battery are removed from the bike and connected together with nothing else involved?

But the voltage just drops out when I connect the controller.

If I use a multimeter to test my battery (via the BMS) it reads 51.4v (full charge minus the initial testing of the shunt mod). But when I connect the controller, the battery reads 39.5v (the lowest voltage for my battery).

This indicates that the controller/motor system is drawing so much current that the battery's BMS is shutting down to protect itself and the cells. It might be because a cell is dropping below LVC, or because teh BMS overcurrent is tripping.

If you monitor each cell *at the BMS wire inputs* (one at a time if you have to) (not at the cells!) with a voltmeter while connecting the controller, youc an verify if any cells are dropping in voltage during this event.

If any are, then check that same cell's voltage *at the cell* under the same conditions. If ti does not drop there, then you havea connection problem to the BMS at it's sense wire.

 

[jamiejackherer]

Thwnks man. The first part of my post is just to round up the story and give some insight I to my project, so it less important, but that is when the problem started and the problem is the same after a new controller. I did suspect the BMS. I will check the cells and BMS this morning as you described. In my experience, if the motor and controller try to pull too much current the BMS will protect the system and shut off or possibly blow out. The BMS is delivering the correct voltage to the voltmeter when disconnected but drops out when connected to the controller only.

There are no wires touching the frame at all. The problems occur even when the controller, battery and motor are out of the bike. The problem of the voltage drop happens when only the controller and battery are connected by Pos+ and Neg- only.

I know this voltage drop behaviour means that voltage has to be going somewhere to drop out.

Sent from my SM-G970F using Tapatalk

 

[jamiejackherer]

Ok so I managed to get the motor running!!

After managing to whittle it all down to the BMS, I simply bypassed the BMS discharge to be direct from battery to controller. The BMS will still balance charge but will allow full discharge (yes, yes I know!!) So I have to be a bit careful, although the controller should also cut off at 39v anyway.

So its now clear what happened; I modded the first controller to deliver more amps than the BMS could handle. I think the BMS is rated at 35amps so could probably handle maybe 50a~80a; the first controller has/had 80a MOSFETs in so could pull up to 80a via the 35a BMS. Clearly the BMS could not handle that current and although it still allowed voltage to pass through it seems any kind of current caused it to shut off.

Looks like I'm BMS shopping for a high amp BMS, I'll need at least a 100a BMS as my new controller is capable of 110a!! Or should I include a BMS bypass switch to the system? That way I can achieve full amperes from the battery when I want it and the rest of the time everything is BMS protected.

Sent from my SM-G970F using Tapatalk

 

[jamiejackherer]

After blowing a BMS by performing a shunt mod to my old 35A controller I managed to get the bike running again by bypassing the BMS discharge (see pic). I bought a new 50A controller (thinking it was the 35A controller that blew), which is cutting out under any decent load, the battery is capable of delivering 5C (100A) continuously and 10C (200A) for 10 second bursts. The BMS was previously preventing the battery from delivering much more than 35A. And so now the battery will hand out 100A+ all the time but the controller is now preventing that power getting to the motor.

Basically I am trying to find out how much power I get out of this system, which is as follows;

Battery - 48v 13s 20Ah EIG cells with a ~35A (discharge) BMS.
Motor - 48v 1500w rated voilamart rear hub motor.
Controller - 48v 1500w 50A (which makes no sense as 48v x 50A is 2400w lol).

I am ordering a 120A BMS asap, but I still want to improve what I have and learn as much as possible along the way.

So I know my battery will give 100A but the controller will only handle 50A as it is. Clearly this isn't mixing well.

So what are my options for stopping the bike cutting out under load?

1. New 100A BMS & 100A controller (best option, most expensive)
2. Just a new 50A BMS (cheap & easy)
3. And this is the fun option!!! I'm thinking I should shunt mod the 50A controller to be able to handle the 100A from the battery. If I "trick" the controller shunt to thinking it's a 100A not 50A it won't cut out? You might think that the MOSFETs will blow... But it's ok because they are max. rated for 110A, so in theory, so long as I don't mod the shunt to deliver over 100A then everything will be good??

Does anyone have any experience/suggestions/thoughts on this?

This will be an incredibly powerful machine if I am able to get 100A constant current to the motor! And even more so if I can get the full 200A for the first 10 seconds of setting off!!! I will upgrade the battery to a 72v battery in the future and will 100% need a better motor, something in the 15Kw range.

Sent from my SM-G970F using Tapatalk

 

[amberwolf]

First, your new thread's title "Controller cuts out under high load?" indicates that you are still having problems directly related to the stuff in your previous thread, and that you probably haven't done the battery tests I suggested in your first thread, so your battery is still dropping in voltage enough to trigger the controller's LVC when you put a high enough load on the battery to cause enough voltage sag. Since the problem you are having now is directly related to the other, I merged the threads to keep teh whole problem together.


The most likely problem with the BMS that you say is "blown", is that the battery cells are sagging under the new heavier load after you basically wired around the shunt in the controller, so that the controller can't limit current anymore, and the batteries are now being hammered much harder than previously. So the battery cells sag in voltage below the BMS's cell-level LVC. But now you've wired around the BMS's protections too, so nothing is now protecting your battery from you.

If you don't want to test the battery, and instead just hammer it even harder, that's ok...but it will probably be bad for your battery.

You may not even have an actual cell problem--it could just be a connection problem. But it's much more likely that you are now pushing your cells too hard so they can't do what you want (even though they could do what was asked of them previously at the lower current demand).

Next, some replies to the post text itself below:

Battery - 48v 13s 20Ah EIG cells with a ~35A (discharge) BMS.

Does this mean you've replaced the blown BMS with a new one that can output only 35A max? If so, you'll have to replace it with a different one that can supply the current you desire. If not, you'll have to state exactly what you do mean. If you have wired around the BMS, then you dont' have a BMS at all, as far as your new desire is concerned, as it is not protecting against any overcurrent and isnt' a bottleneck for current.

Motor - 48v 1500w rated voilamart rear hub motor.
Controller - 48v 1500w 50A (which makes no sense as 48v x 50A is 2400w lol).

If the controller is actually marked as 1500w, it almost certainly means that regardless of what combination of voltage and current you use with it, it's maximum continuous capability is only 1500w, so it doens't overheat or blow up, even though it can handle peaks of the higher current. That's common.

I still want to improve what I have and learn as much as possible along the way.

Then the first thing I highly recommend is that you *test* the battery cells, so you don't damage them in the process.

So I know my battery will give 100A

Are the cells brand new? (not just new to you, but brand new from the factory). If not, they may no longer be capable of that. Even if they were unused, just sitting around they will age and decrease in capacity and capability. When used, each charge/discharge cycle takes away a little capacity and a little capability. The harder they are used, the faster this happens.

Even the voltage they are charged to makes a difference. For instance, a 13s pack of EIG cells doesn't charge up as high as a regular 13s pack, as their max charge voltage is only 4.15v (pack voltage of 53.95v). So if your charger (which you don't list) is charging them to 4.2v per cell (what a typical "Li Ion" BMS balances to) for a pack voltage of 54.6v, then they're repeatedly being overcharged, which wears them out faster than if they weren't.

but the controller will only handle 50A as it is. Clearly this isn't mixing well.

The controller doesn't "handle" 50A. It *draws* only that much from the battery, and then only when the physical load on the motor is enough, and the throttle input is asking for it, to cause it to happen.

3. And this is the fun option!!! I'm thinking I should shunt mod the 50A controller to be able to handle the 100A from the battery.

Unfortunatley this is probably a great way to make smoke out of the controller. The reason the shunt is there is to allow the controller to protect itself against currents so high that they will blow parts up. They will usually handle some amount higher than the overall rating, but that is intended to be a safety margin.

When you modify the shunt you remove the ability of the ocntroller to even have any idea how much current is flowing, so then it cannot protect anything, and then its' up to you to stop when you smell smoke (it's usually too late for the controller at that point). :/

It is very common for people to modify controller shunts to get more power...but it is also very common to turn them into metal bricks.

There are controlled ways to get more power, by adding one more shunt wire, either taken out of a similar controller, or bought new and installed, etc.

Just adding solder to the shunt basically turns the shunt into a wire, so the controller can't read a useful voltage drop across it, and it then cant' tell the current that's flowing....

If I "trick" the controller shunt to thinking it's a 100A not 50A it won't cut out?

It will cut out worse after you do this, because the battery voltage will sag even more than it did before.

You might think that the MOSFETs will blow... But it's ok because they are max. rated for 110A, so in theory, so long as I don't mod the shunt to deliver over 100A then everything will be good??

No.

That rating is per FET, for the *phase* wires, which do not see the same currents the battery does. The battery side has much lower currents than the phase currents, so the FETs will see much higher than 100A when you do this.

This will be an incredibly powerful machine if I am able to get 100A constant current to the motor!

You won't get any constant current to the motor. It will depend on the load on the motor whether that much is even drawn, and the current will vary depending on the actual load, vs the throttle setting, vs the battery voltage sag,

Also, as noted above, the current seen at the battery is much less than that at the motor in the phases. This is just how these things work, by converting the voltage and current from DC to pulsed AC you can't direclty compare or link what is on the battery side to the motor side. They work out as equivalent *power*, but the voltage and current aren't measured the same way anymore.


Regarding modifying controllers and how this stuff works, there are dozens of threads about shunt mods, etc., and many will have good info in them, if you take the time to look them up and read them.

Regarding voltage sag and battery problems...same thing.

Regarding the general way a system interacts, so that you can understand what is going on in your system, and what you might need to do to upgrade to get the results you are after (which we don't know, only you do), I recommend going to http://ebikes.ca/tools/simulator.html , reading the entire page so you know what everythign is and how it works, then setup different systems and different conditions and experiment to see what is like what you had, then what you have now, and then what you want to end up with.

 

[amberwolf]

BTW, I would guess that the "white residue" / corrosion you see is actually where you touched the white thermal / heatsink paste that would be on the metal bar the FETs are bolted to, and then touched all the areas you see the residue.

If not you, then whoever built or last worked on the controller.

 

[jamiejackherer]

I'm about to go to work but just wanted to reply to your first paragraph.

Firstly, I did test the voltages in each cell, and each cell was testing at 3.76v.

Also the voltage is not sagging much at all, in fact with the BMS it would sag from 54.1v (full charge) to 49.9v, and at 47v it would sag to 44v. After bypassing the BMS it sagged from 51v to 50v under high load!

Because I am not using 18650's and I am using lithium polymer cells my low voltages are a little different. The cells have a recommended low voltage of 3.0v per cell so that's 39v for my 13s battery. But, the cells are capable of discharging to 2.5v which is 32.5 for a 13s. The BMS would LV cut off at 41.3. I have a voltmeter throttle so can keep an eye on the battery voltage.

I also know the technical data for my cells very well.. the high voltage for my eig cells is 4.2 not 4.15 (see spec sheet in pic) you are both assuming a lot and wrong a lot.

The cells were new when I bought them, I have ran around 250-300 cycles and the cells are rated as 80% efficiecy after 1000 cycles, so I am around ¼-⅓ of the way into its decent life. The cells are all still balanced perfectly and are very high quality cells. They are designed specifically for EV applications. Unlike most 18650's which are very rarely designed for high discharge. I really dont know why you are dead set on degrading everything I have.

The batteries can give out 5C no problem, and 3C is recommended! They can deliver 10C for 10 seconds.. I cannot see them being faulty tbh pal. They have been better than any other battery that me or anyone I know has used. And now I'm unleashing there true power I am finding the battery is obvs giving out high amps because the controller keeps cutting out.

Bare in mind I don't really care if everything breaks, if my battery fails etc because I am already planning on building a new 72v battery with new cells. I just want to learn a bit more than I already know.

Also you said about the controller being rated at 50A, and that it has a safety margin, so you're saying that a 50A controller has a safety margin of 60A? The fets are rated at 110A so the controller should handle 110A, and even still I am perfectly capable of testing and replacing, even upgrading, blown MOSFETs in a controller.

Shunt modding does not remove the ability for the controller to read the current, it measures the voltage drop across the shunt and so determines the max current based on that voltage drop. By increasing the voltage drop the board is *tricked* into believing the current is lower than 50A and so will allow more current to pass through. You say it like performing the shunt mod removes all that voltage measurement etc when this is not the case.

Sent from my SM-G970F using Tapatalk

 

[jamiejackherer]

Also the white residue is not thermal paste pal. I'm sorry but you are coming across very patronising mate. It's clear you know a good amount of information but you really taking as if I have no clue at all. I appreciate your input, and you have actually helped me with the BMS crisis, but I don't appreciate being spoken to like I don't know wtf I'm on about at all.. I mean I built the whole thing from scratch, I soldered the components onto the BMS when I built my battery etc. There are just a few gaps in my knowledge that's all.

Sent from my SM-G970F using Tapatalk

 

[amberwolf]

Firstly, I did test the voltages in each cell, and each cell was testing at 3.76v.

Under load, or just sitting there not connected to the controller? If under load, what is the exact current reading (amps)?

(the complete conditions of a test need to be stated along with the results, or the results aren't useful to help you without making more assumptions that have a high probability of being wrong.)

Also the voltage is not sagging much at all, in fact with the BMS it would sag from 54.1v (full charge) to 49.9v, and at 47v it would sag to 44v. After bypassing the BMS it sagged from 51v to 50v under high load!

What exactly is "high load"? Is it the same exact load that the other voltage measurements were done under? (it's not possible to compare the one to the other directly if the load is not the same. If it's different, then one can only make assumptions about the differences in behavior, unless the entire range of behaviors at varying loads is characterized and graphed.)

Because I am not using 18650's and I am using lithium polymer cells my low voltages are a little different. The cells have a recommended low voltage of 3.0v per cell so that's 39v for my 13s battery. But, the cells are capable of discharging to 2.5v which is 32.5 for a 13s. The BMS would LV cut off at 41.3. I have a voltmeter throttle so can keep an eye on the battery voltage.

You can discharge any cell to a lot lower voltage under carefully controlled lab conditions (which is what spec sheets are written from; many spec sheets also give those conditions but some don't) than you can in actual usage situations. Also, those cells tested to those conditions aren't typically expected to be used for the same number of cycles that you would expect to get out of a pack in actual usage.

However, the main problem with using an LVC below the "recommended" voltage is that it is a lot harder on the cell to do that. Typically, the higher they are charged to, and the lower they are discharged to, and the more often these are done, the faster any particular cell's capacity and capabilities degrade.

I also know the technical data for my cells very well.. the high voltage for my eig cells is 4.2 not 4.15 (see spec sheet in pic) .

I've been using the same cells (EIG C020 NMC 20Ah, right?) for about a decade on three different ebikes (one medium-cargo bike, one light-cargo trike and one heavy-cargo trike), so I have a tiny bit of experience with them. It's possible it is not as much as you have, but I will make the assumption that it is significantly more.

Given the image you posted, I did a bit of poking around.

I've attached the PDF spec sheet from EIG for these cells, where it specifies the maximum charge voltage as 4.15v. So maybe EIG has two different specifications for the same cells, implying either they revised the spec in light of new data that neither of us has (I don't know which spec is newer) or they have two (or more) versions of these cells and one of these applies to one version, and one to another...but there's no data on them to say which would be which.

EIG doesnt' appear to be still around to ask
https://www.komachine.com/vi/companies/eig/
says "this manufacturer closed down"
with the WM's last archive from about 2013 indicating it's been quite some time, but I found a version of their pages on the wayback machine:
https://web.archive.org/web/20130407092202/http://www.eigbattery.com/
Since it uses flash, I can't get a direct link out of that, but if you click the Cell picture in teh scrolling display, it gets you to the page where I long ago got the same spec sheet that I attached (though the one taht's attached came from the Ebaracus site since that was easier than locating the actual file I have on my system). On that new page clicking the C020 download column link gets you a jpg version of it
https://web.archive.org/web/20130408041849fw_/http://www.eigbattery.com/eng/product/3.jpg
The other site
https://www.komachine.com/vi/companies/eig/products/56934-C020/
doesn't have the complete specs, but those it does have match the archived EIG sheet.

I didn't search for very long but I couldlnt' find a version of the spec sheet that looks like the pic you posted, to see if either had publication date or other info on it that would help figure out what the right specs for these cells are.

I do note there are some significant discrepancies between terminology and numbers on some parts of the one from the other, such as the second Energy Density listed on your image, which seems to correspond to Power Density on the spec sheet from the EIG site archive. I'm not sure if it's just a typo on one sheet or the other, but I suspect it's correct on the EIG archived version simply because the units aren't what is used for Energy Density, but are for Power Density.

My personal suspicion, based on experience with released spec sheets vs preliminary spec sheets, is that the pic you have is from the latter, and the archived (and attached) version is the former, so that the cells actually out there are supposed to be used with the version as attached. Given that it is more conservative, it's safer to go with that, so that's what I'll continue to do.

Why, exactly, they give either number, I don't know, as I am not a battery chemist. But this is what they spec'd their cells to operate at in the spec sheet I've always run across (as attached), so that's what I use them at, because other people have run other cells meant to be used in a lower voltage range at higher ranges, and gotten less useful life out of them than they should have based on spec sheet expectations...and probably would have gotten the full life if they hadn't "overcharged" them.

There's not enough energy in the portion up there between 4.1 and 4.whatever to make enough of a difference in capacity between 4.15v and 4.2v, so if given the choice between a teensy bit more capacity and a potentially significantly longer lifespan, I'll pick the latter.

So, you can use your cells at any voltage range you prefer, as they *are* your cells. And I will continue to use mine at less (4.12 is what mine charge to with my present 14s2p pack on the SB Cruiser trike).

you are both assuming a lot and wrong a lot

I have to assume a lot because most people that post here don't post complete info, and won't answer questions completely (if at all). After years of trying to help people here with incomplete info, I've given up on expecting it, and learned to just accept that some people wont' like my style of help, but to go ahead and help them anyway by making the necessary assumptions to post conclusions that might set them on the path to fixing their problems. It's ok with me if you don't like it, either--I'll still help as much as I can, unless you prefer me to stay out of your threads...but all help (from anyone) will still be full of assumptions until you provide all of the necessary data to not have to do so.

And anything based on assumptions is bound to be wrong a lot.... But it is what I can do with what's provided so far.


One example of this that happens frequently with all sorts of problems people post here: if the results of tests that are suggested aren't provided, then anyone reading has to assume either the tests weren't done, or that if they were done they wont' have any idea what they were. Thus they have to either drag that info out of the person with the problem, or just make assumptions and suggest further tests or potential problems that are based on the info that *has* been provided.



Since everything you've provided so far points to a problem with voltage sag, as the symptoms described for both BMS and controller are all exactly that kind of result and cause, then without those voltage results (some of which you've now provided but only a little with context of some test conditions, but not all test conditions and not for all results) the only thing I can do is assume you have a battery problem, or that you have a connection problem (and the voltage is dropped across that).

Since a connection problem dropping that many volts would be really hot, and possibly smoking or at least melting or deforming insulation or plastic around it, it is a fair assumption that it's not a single-point connection problem. But I still have to make assumptions...because the data provided is not complete.

 

[HrKlev]

Where are you measuring voltage sag? You might have a lot of sag in the wiring between battery and controller. Those wires in the pictures does look small for a high amp application.

Btw, if you are using the generig chinese BMS, go bigger than what you need. Better safe than sorry, and thise BMS's are a little overrated. You can connect the biggest BMS you can find to a small controller without any compability issues, the controller only draws as much current it needs.

 

[Blacklite]

By increasing the voltage drop the board is *tricked* into believing the current is lower than 50A and so will allow more current to pass through. You say it like performing the shunt mod removes all that voltage measurement etc when this is not the case.

Don’t you mean decreasing the voltage drop?

 

[amberwolf]

Also the white residue is not thermal paste pal. I'm sorry but you are coming across very patronising mate. It's clear you know a good amount of information but you really taking as if I have no clue at all.

As noted before, I'm basing my assumptions on the little data provided, along with many years of experience both with what other people have acutally seen and done, and what I myself have personally seen and done.

Anything I say that's wrong is usually because there's not enough information provided by those asking the questions.

For example, I can't know that it is or isn't thermal paste, corrosion, silicon sealant, powdered sugar, or anything else, because all I have is photos of things that appear to show a white substance that seems to have impressions of smeared fingerprints. There are a lot of things that could be, but since the areas with it on there are all items that normally have thermal paste applied to them in the assembly process, and it is normal for that to also get on people's fingers (or gloves over the fingers"

If you know for certain that it isn't thermal paste, then you know much more information than you have provided to us, and should have provided that information to begin with--at the least, you should have said that you know for certain it is not thermal paste.


But I still bet that if you did a spectroscopic analysis of a scraping of the stuff on taht FET, for instance, that you would find a high concentration of titanium dioxide.

If you really want to know what it is, then posting a couple pics of it and then complaining that someone that makes a very educated guess from experience is definitely wrong isn't going to get you an answer. You'll have to do some chemical testing or analysis of it to find out for certain. Or spectroscopy. Assuming there is enough of the substance to be able to test it that way (I have no idea).

I appreciate your input, and you have actually helped me with the BMS crisis

Which I was able to do quickly (rather than over days of Q&A) by making a whole bunch of assumptions.

, but I don't appreciate being spoken to like I don't know wtf I'm on about at all.. I mean I built the whole thing from scratch, I soldered the components onto the BMS when I built my battery etc.

There's more information you didn't provide. Almost everyone that comes here with a problem did not build any part of any of their stuff themselves, and knows absolutely nothing about any of it. So it is a fair assumption that anyone else coming here with a problem also is similar, unless they say otherwise.

It's up to the people providing hte information to provide it completely, if they dont' want people helping solve a problem with that information to *have* to make assumptions, or to first ask a hundred pages of time-wasting questions before they can begin to give *any* answers.

Since almost every time the answers I give based on assumptions made from experience are good enough to get the people with a problem to the point where they can solve that problem, and only rarely are they completley off-base, then there's no point in me wasting both my time (whcih is volunteered here because I like to help people fix things and solve problems) and the people with a problem's time by asking every possible question about every possible condition, and waiting until they have answered in complete and perfect detail (which never happens) before I then give any advice at all.

But we can do that if you want--I expect it would be even more annoying to you than my previous assumptions, for you to have to answer all the questions and perform all the tests that would be required.

 

[amberwolf]

By increasing the voltage drop

Decreasing, if you mean the voltage created across the shunt by any particular current flowing thru it, vs what it was before the modification. (lower resistance x same current = lower voltage)

the board is *tricked* into believing the current is lower than 50A and so will allow more current to pass through. You say it like performing the shunt mod removes all that voltage measurement etc when this is not the case.

I've already seen the results of this many times. But if you want to find out for sure how your specific system responds to the specific mod that you performed (each mod is different because it is not a controlled manufacturing process), then you can see for yourself whether it is so greatly reduced that it is practically removed with a relatively simple test, if you have a voltmeter that can measure accurately in the millivolt range, or you can build a small op-amp circuit to magnify the tiny voltage across the shunts to a range a common multimeter can read. YOu mgiht be able to read it better at the output of the opamp circuit between the shunt and the MCU, if you trace out that circuitry to find it.

What's the voltage drop across the shunt before the mod, for say, 10A? (assuming you measured it before modifying it)

What's the voltage drop after, for the same current?

If you didn't measure beforehand, then since there appears to still be one unmodified shunt, you can measure just that one (by first disconnecting the other two), then do the test with just the unmodified shunt. (removing the mod from the other shunts doesn't restore them to original condition, so doing that will not create a valid test).

Then you can disconect that shunt and reconnect the modified pair, and do the test.

Compare the results to see the scale of the difference, which should be very large.

The typical resistance of solder is very low. At a guesstimate, there is at least 1mm width of solder over about 1cm between the shunts. Assuming around .017241microohm per cubic meter from a randomly selected internet calculator
https://owenduffy.net/calc/WireDcResistance.htm
(so I don't have to get Kirin off my legs and go dig out a book)
that comes out to around 0.00022ohms, or 0.22milliohms. The original total shunt resistance of a controller like that is probably just under 1.5milliohms, based on several randomly selected threads here on the forum that provide that info for similar-capability controllers. Since the resistances are all paralleled in this case, then that gives a new resistance of less than 0.2milliohms. If there is more solder width (greater cross section area) than 1mm, the resistance is even less. Either way, it means that there is approximately just over 1/10th the original resistance, so the new current limit is ten times what it was designed to be. That's effectively no current limiting at all, as far as the design of the controller goes, and what it was meant to deal with.

That all makes a bunch of assumptions due to lack of exact information, and I'm terrible at math (I can hardly ever get 2+2=5 right ), but you can run all the numbers yourself with the more complete data you may have available to you with the controller right there.

If it is useful, there are much more controlled ways to do a shunt mod. If you trace out the amplifier circuit between the shunt and the MCU chip, you'll probably find a voltage divider at the input to the opamp. If you put a potentometer in place of one of the resistors (or use one *as* the voltage divider) of the right resistance to give you a range that includes the original resistance up or down to what's needed to change the current to the magnitude you would like, then you can finely adjust this until it gives just the right amount for you, or the controller blows up first...but at least you can easily change this at any time, as precisely as you like. If you prefer, you can use the potentiometer across the opamp in place of it's gain resistor, and change the gain of the whole circuit instead of the input voltage--this may work better for some circuits, but I don't know how it will affect the MCU input's behavior. Depends on the rest of the circuit's design.

 

[jamiejackherer]

....................................

 

[Vbruun]

Changing thickness Will change KV. Do you want that?

Else, stick with the same turn count and thickness

 

[jamiejackherer]

....................................

 

[Thecoco974]

The wire thickness does not mean anything, current carrying ability is all about total cross-sectional area and total lengh of wire. Also, more thinner wire are going to be easyer to wind. You can estimate the lengh you need by measuring one of the phase you removed, factoring the increase of turn count and taking 50cm extra for margin of error.

Envoyé de mon Redmi Note 9S en utilisant Tapatalk

 

[jamiejackherer]

....................................

 

[Thecoco974]

The number of parallel strand time the sectional area of one strand gives the phase sectional copper area S. You can then get the phase resistance with the resistivity of copper Rho and the phase length L with this formula :
R=rho *(L/S)
With a lower the resistance you can push more AMP without overheating (there is a max value where the iron is satured ). There is no difference between two wind with the same section one with 1mm and the other with 0.5 and more parallele wire. It's mostly a matter of ease of winding.

Envoyé de mon Redmi Note 9S en utilisant Tapatalk

 

[stepus]

Its better to use more paralel wires, its easy to hand winding, better use space in slots, and higher efficiency because skin effect at switching freq.

 

[jamiejackherer]

So I am about to rewind a Voilamart 1500w motor I burnt out by putting 5600w through it (the motor was pulling 105 amperes @ 54 volts with a shunt mod on a 48v 30a cheap controller).

The motor was handling 4500w (85a @ 54 volts) easily with no damage, but that extra kW was too much.

Anyway to the point [emoji1787]:

The original motor configuration was as such;

51 slot stator
0.5mm wire
11 strands
5 turns

As seen in the pic of the motor being unwound, the windings are done in 5 turns around the stator in a config of 3-3-3-4-4. (3 + 3 + 3 + 4 + 4) x 3 = 51.

I'm curious why the 3-3-3-4-4 configuration?

Is it simply to distribute the windings across the stator into the 3 phases? If so then why not use a stator with 72 slots and do a 4-4-4-4-4 configuration? Or 45 slots and 3-3-3-3-3?

______________________________

In order to be able to put the 100 amperes through the motor I've decided to go with a different set up when rewinding... I've bought some high quality enamelled copper wire that is 0.75mm and plan to do the following;

0.75mm wire
11 strands
6.5 or more turns (as many will fit in the slot)

Im aiming for high torque and less RPM that can handle 100+ amperes or around the 5.5kW mark.

I'm thinking the increase in wire thickness (34.44% more copper in each slot) will be able to handle the amps, the more turns will give the higher torque and lower RPM?

I am also not going to be putting any plastic or insulating paper in the motor as it can never withstand the high temperatures in the motor. That cheap insulating paper the Chinese motor manufacturers use is sooo terrible!! Twice I've had it melt onto motor windings and causes the enamel to burn off.. only where the insulating paper is though (see pic)! So I am using Ambersil electronic varnish to coat the stator and Kapton tape to cover any sharp edges.

Any input would be appreciated.

Sent from my SM-G970F using Tapatalk

 

[jamiejackherer]

So I burnt yet another motor out today [emoji91][emoji91]. It's a Voilamart 48v 1500w 31a with internal controller. Luckily I only paid £50 for it.

I amped up with shunt mod (not the easiest to do a built in controller covered in gel) but was still very slow. It was very quiet, as in silent! But as soon as I took it to my testing hill (40 metres with inclines of 30% and 45%) it burnt out by the time I was at the top (see pic).

I'm wondering what people think about whether to rewind the motor or try a bit Ambersil varnish on the damaged area?

I cannot get anymore of the melted plastic off the windings and think that this will just get hot and do the same thing again but worse.. on the one hand, it's easy to varnish but will most likely fail deeming it pointless. On the other hand, it's not a quick and easy job to rewind the motor.

Update: the thread used to secure the windings has also started to melt.. a quick search revealed that proper stator lacing string is made of polyester which has a melting point of 260°C!! So my motor got to about that temperature.. wasn't expecting that sort of heat tbh! I think the mix of motor heat + controller heat in the hub chamber is not a good idea at all. Also after more reading I realised that the proper insulating paper is called Chinese Nomex paper (I suggest what's in these motors is not real Nomex paper though) and has a max temp of 220°C.. so the motor must have gotten between 220°C and 260°C.

Sent from my SM-G970F using Tapatalk

 

[e-beach]

Why rewind a motor not suited for your hills? Sounds to me like you should just get a motor, controller and battery suited for your hills. It most certainly will cost you more then 50 pounds, but at least it will last.

Oh, and BTW, controller shunt mods will increase torque, but not top end speed. (RPM's)

 

[jamiejackherer]

Why rewind a motor not suited for your hills? Sounds to me like you should just get a motor, controller and battery suited for your hills. It most certainly will cost you more then 50 pounds, but at least it will last.

Oh, and BTW, controller shunt mods will increase torque, but not top end speed. (RPM's)

I am going for torque over RPM mate.

I don't have a spare £500 lying around for a 3-5kW (or more) motor fella! I wish I did and I've been begging my friend to sell me his old 3kW motor but he won't [emoji1787][emoji1787]. I also could rewind the motor in one day.. but a £500 motor would take weeks to arrive from China.

I am just trying to learn as I go along and I'm the sort of person that always gets inside stuff to see how it works and how I might improve it (the weeks following Christmas as a kid must have been horrendous for my parents when they realised I tore everything apart to see how it worked).

The new copper costs around £30 and the 2 broken motors I now have spare will only sit there gathering dust. The way I see it is; why buy a new motor when I can build my own for 10% of the cost.

I want to discover how much power I can put through a 1500w motor. I now know the max wattage for the standard cheap 1500w China hub motors is about 4500w! Triple the rated value! But 5500w is too much, so if they can handle 4500w out of the factory, then what if I modify the motor for my specific usage (high torque, low top speed).

I want a bike that is good off road, climbing hills and pulls away really quickly, but I'll be happy to go 20-25mph at the top end, basically swap out 5-10mph for extra torque.

On top of that, if I can build a motor that can handle 5-7kW I will be able to make a lot of money where I live as ebikes have become very popular and the local lads are always wanting upgrades and mods etc.

Plus 1500w motors are very cheap, small and lightweight Vs their 3000w cousins.

Sent from my SM-G970F using Tapatalk