How to solder the onimous shunt on the KU63

[schwibsi]

Hi guys,

I received my Q100 today and my KU63 controllers (First time ever )

Do the red arrows show, where I should solder?
If so, what should I solder on there?
A little wire?
If so, how thick?
What should I measure to see if I've done it right?

 

[motomech]

YIKES!!, The HAND that ate Endless Sphere. I hope you can size your solder better than your pictures.
Just coat the entire shunt lightly, no globs or build-up.

 

[schwibsi]

Sorry about that.
I resized the pictures.

I'm sorry, if my don't really understand, what I am to do. What exactly is the shunt?
In medicine I know what a shunt in a body is, but in electronics?
Do I put solder on that little circle?

 

[Jeremy Harris]

And be aware that doing this pretty much disables the fault protection circuit in the KU63, so although the controller will now work at a higher battery current limit it will no longer reliably shut down if there is a fault, so may well go up in smoke if you're not careful..............

I don't know how many times we've been through the pros and cons of soldering up shunts on here over the years, quite a few I'd say. Quite apart from the solder having a massively different temperature coefficient from the near-zero coefficient constantan wire the shunt's made from (so introducing a big temperature vs current limit factor) it is also a crude, brute force, way to increase the battery current limit, a way that does the very dumb thing of also shifting the fault protection shut down point.

Although very slightly more complex, adding a resistor to the current sense connection is intrinsically very much better. Not only does it target exactly what you want to do (increase the battery current limit), it also leaves the safety protection circuit working normally and removes the temperature sensitivity that soldering the shunt introduces.

Given that we know pretty much exactly how these controllers work, and that the simple "add a resistor" mod is very well known now, I do wonder why folk are still so enthusiastic about doing caveman mods like adding unpredictable blobs of solder to a shunt.

 

[schwibsi]

So you suggest I solder this part

http://www.conrad.de/ce/de/product/...SHOP_AREA_17442&promotionareaSearchDetail=005

to connect the two spots marked by the red arrows?

Another question.
Am I to solder this
http://www.conrad.de/ce/de/product/...SHOP_AREA_17442&promotionareaSearchDetail=005
where the one in this picture is?

How do I cut the connection? Do I just take an exacto knife and cut a mm into the board?
Sorry for asking questions on the level of a 5 year old, but I've never even seen anybody do lhis.
I feel like a blind man, trying to draw a picture... lol

 

[potatorage]

Yeah basically cut the trace so that you break the connection. It might be easier to cut at the narrowest point instead of the point marked.

 

[Jeremy Harris]

No need to go cutting any traces, just add a resistor across the sense connection, as mentioned above. It's an easy mod, and one that allows you to fit a variable current limit (by using a variable resistor, rather than a fixed one) if you wish (although that needs doing with some caution, as you can set a very high current by accident if you aren't careful).

The details have been documented here (not by me): http://www.avdweb.nl/solar-bike/electronics/ku63-motor-controller.html#h0-1-9-increasing-motor-current and the schematic for the KU63 showing the "tune" resistor (as this chap called it) are here: http://www.avdweb.nl/Article_files/Solarbike/Motor-controller/China-BLDC-motor-controller-36V-250W.pdf

The mod is shown on the schematic in the top right corner. Doing it this way means there's no need to risk cutting traces or doing anything that isn't easily reversible to the controller. It also means that the over current fault protection circuit (the sense pin fed by R78 and C35) remains unaffected, as the shunt value will be unchanged.

 

[schwibsi]

Thanks guys,

Jeremy, yes that is the webpage, I studied before posting here.
I just wanted to make sure, I understood it correctly before breaking something that could have been avoided.
If the 2 resistors will do their job to make the controller put out 20-21A and make the LM78L05 a little healthier.
I wouldn't be increasing the current above that point.

 

[d8veh]

If you're going to solder the shunt, don't solder the whole thing. 33% is enough for the Q100. If you're going to do the other mods, make sure that you have a wattmeter to confirm the results, otherwise your motor will smoke before you get a chance to enjoy it. My advice is to solder 1/3 shunt, which is simple, quick and effective. Like this:

 

[Jeremy Harris]

Sorry, d8veh, but soldering the shunt really is a daft thing to do, for the reasons given above. Why some have decided to resurrect this caveman approach, when we've collectively learned the negative effects of doing this, and, what's more, found far better and more reliable ways to mod controllers, is beyond me.

If I can add a resistor costing a few pennies in a few minutes, and get a pretty consistent and fairly accurate new current limit, with no risks of getting things wrong, no temperature effects and no impact on the fault protection circuit, then it makes sense to me to never risk soldering up a shunt like this

 

[d8veh]

I've done it on at least 10 controllers and never had a problem yet. It may be daft, but it's simple and works whatever the theory says. As soon as my next two KU63's arrive in the next few days, I'll be soldering them too. Also, they're not exactly expensive at about £10 each even if something did go wrong.

 

[Kin]

Jeremey, could the variable resistivity of the solder be a pro? If the shunt ever did heat up, the limit would go down .

 

[Jeremy Harris]

The increase in resistance with temperature that the solder introduces does indeed cause the current limit to go down. It also makes the shunt hotter. There's at least one case somewhere on this forum where the solder melted on the shunt in use, and I've had a non-modified shunt burn out on a humble Tongxin controller, so making the thing run hotter isn't wise.

The real problem, and one that won't show unless there is a fault, is that reducing the value of the shunt with solder stops the fault protection circuit (which is completely separate from the current limit circuit) from working properly. This means you run a bigger risk of doing damage to the controller, wiring, or even motor or battery, if you get a fault, as the controller may well not shutdown. It's not just the cost of the controller that matters here, although quite why it's OK to risk even a £10 controller for the sake of a £0.05 resistor is a bit beyond me.

Sure soldering up the shunt increases the battery current, and if done very carefully it sort of works (albeit with some loss of controller functionality and the introduction of a temperature dependent current limit). It's a bit like deciding to increase the performance of your car by drilling big holes in the exhaust, though. That may work, too, but clearly it isn't the best way to go about it.

Given that the resistor needed to do the mod properly is cheap and takes the same (or even maybe less) time than soldering the shunt up, I can't see why people persist in it. It's like going back to the sort of mods we were doing here four or five years ago, before we gained a better understanding on the way these things worked.

 

[Kin]

Well, I was mostly jesting, though I saw it as a reasonable pro- melting the shunt is hardly positive. I suppose I actually hadn't seen the resistor mod before (my recent controllers have been programmable), and originally when I got into all this I think I would have thought the shunt mod is a little more conceptually clear. However, I think I'll look into the resistor info next time I have occasion; thanks.

 

[schwibsi]

One more little question
For how much power are the resistors supposed to be?
I found 100Ohm and 6.8 kOhm resistors with as little as 0.25W and as high as 5W.
I couldn't find any documentation, which ones to use.

 

[Jeremy Harris]

The resistor used to do the current limit increase mod can be as small as you like. An 1/8W or 1/4W would be more than adequate. An 1/8W would be easier to fit, as it has a very small body size.

 

[schwibsi]

Thanks a lot, I suspected, they could be very small, but better safe than sorry.

 

[gsl]

I have a KU63 controller (3 speed switch version) that replaced the defective stoc controller on my eBike. The stock china controller was drawing 15A thus reaching the system's intended wattage and a legal speed of 25 Km/h (27 with newly charged battery but still within tolerance).
The KU63 shuts down at 12.85A thus giving me less acceleration and a top speed of 20 Km/h (with full battery i.e. max wattage)

I want to get 15A out of the KU63 and I implemented this very mod http://www.avdweb.nl/solar-bike/electronics/ku63-motor-controller.html#h0-1-9-increasing-motor-current for increasing motor power, soldering different values of Rtune between the terminal of R45 (2k) farther away from the shunt and negative ground.

For some reason not one of the Rtune values I tried at first (27k, 10k, 6.8k, 2.2k) changed the power drain in the slightest (i.e. no effect). I measured the power intake by a Turingy power meter.
I tried 1.5k then and this increased the current by 0.9A to 13.75A
Next I tried 820 Ohm, but it seems to flatten out. I got the same Apeak 13.75A.

I checked the Throttle voltage: supply voltage for the hall sensor is 5.3V. No-throttle voltage output is 1.2V and full throttle voltage is 4.9V . Seem all right
I bridged the shunt with a heavy gage copper wire. I got an current intake of 30.53A at full throttle and full load, so there is no external limiting factor.

Am I understanding something wrong, or what could be the problem?
Any thoughts would be greatly appreciated.

 

[d8veh]

As I've said before, and jeremy will hate me for it: Just add some solder to the shunt like the photo in my previous post. You'll be done in about 1 minute, and you'll have what you want. Maybe some other controllers might get some theoretical problems, but i personnally guarantee that you won't get a problem on a KU63. If you do, I'll send you a new one.

Looking at your photo, it looks like the bare leg of the resistor is touching the shunt. If you look at the schematic, that end of the resistor should go to the other side of the 2K resistor and not directly to the shunt.

 

[Jeremy Harris]

As I've said before, and jeremy will hate me for it: Just add some solder to the shunt like the photo in my previous post. You'll be done in about 1 minute, and you'll have what you want. Maybe some other controllers might get some theoretical problems, but i personnally guarantee that you won't get a problem on a KU63. If you do, I'll send you a new one.

That's way OTT, d8veh! I really object to the word "hate" being used here, when all I've done is (correctly) point out that soldering the shunt disables, or partially disables, the controller over-current fault protection circuit. If people are happy to accept that they are defeating part of the controller safety functionality by adding blobs of solder to the shunt then that's fine. The important thing here is that people need to be aware of the impact of their actions, so they can decide whether or not they are comfortable with the extra risk.

 

[gsl]

Dear Dave, dear Jeremy!

Thank you for your comments and sorry I'm replying so late. Somehow I didn't get an email notification and thought noone was replying.

Dave:
1) I agree with Jeremy in princople but in the meantime I used a more conservative way to change the shunts resistance. I coiled a copper wire around the half the length of the shunt. This way I can reverse the mod by removing the copper "snake" around the shunt wire and second, I can change the resistance and thus the power by pulling the coil apart, or squeezing it tighter without soldering and filing and also the temperature coefficient is not so express.

2) The reason I soldered Rtune to the negative base of the shunt was because Albert van Daal, who published the mod in the first place, thought it might make a difference (althought I don't see how because both are obviously (according to the schematic) negative ground.

3) An additional reason I would want the mod to work with Rtune is that by using a 3way switch I can have 3 power ranges on the controller. No Rtune (i.e. 13A) for power conserving operation, Rtune1 (15A) for normal operation and Rtune2 (17-18A) for emergency "throttle through the gate" operation. For my needs and taste that would be ideal in any controller.

The question remains, why can't I get Rtune to work as intended? But if you could help, you would have done so...
I have read here that this topic was discussed on many occasions but I don't seen to be able to find those discussions in the search. Could you maybe point me to relevant threads so I can better understand the mod or find alternatives?

Thanks

 

[gsl]

Looking at your photo, it looks like the bare leg of the resistor is touching the shunt. If you look at the schematic, that end of the resistor should go to the other side of the 2K resistor and not directly to the shunt.

On second reading, d8veh, I find I do not quite understand your comment. Which schematic are you refering to?
I am trying to implement the mod according to this schematic http://www.avdweb.nl/Article_files/...ller/China-BLDC-motor-controller-36V-250W.pdf

Here Rtune goes from the junction of R45, R44, C44 and CPU9 to negative ground, not parallel to the 2k R45 (that is how I understand "hat end of the resistor should go to the other side of the 2K resistor").

Are we on the same page here, or am I not "getting" it?

regards

 

[gsl]

Have I fallen from grace?

I am sorry if I did or said something wrong!

 

[d8veh]

I just looked at it again, and I think I got it the wrong way round. The leg of the resistor is touching the shunt but I forgot that it's grounded, so it looks like you're correct.

 

[gsl]

Thanks D8veh!