Novak external Cap Warm !?!

[JV]

The external 680uF electrolytic on the flying wires is getting a bit warm.
Not hot to touch, but noticeably warm.

I noticed because I velcroed it to the top deck of my Yokomo BC Buggy,
and the velcro glue on the capacitor side came unstuck.

Should I stick a few hundred uF of Oscons there instead ?
Maybe on shorter wires?
Or is something wrong with the motor controller/motor ?

[Oscon = ultra low ESR electrolytic capacitors from Sanyo]

I don't think I've ever noticed a capacitor getting warm before.
Or is it getting heated up through the wires? [doesn't seem likely]

My setup is:
Novak 5800, mode 3
Gearing 22/75 [lowish stock gearing] -> 7.5 overall
Battery: untreated GP3300's, via a Tamiya connector
(the Deans connectors are in the post)
Surface is mostly grass.

The Tamiya connector won't be helping the external cap.

I don't recognise the brand of external electrolytic (marked with a capital 'M' in a rounded box), or I'd have looked up the data sheet.

Any thoughts ?

John.

[WHL]

I thought Novak reccomendeds a 9:1 final drive ratio for offroad vehicles?

I believe you might be overgeared.

[webba]

i dont think Novak would have recommended a final drive as this changes by track/motor etc. I also don't think that the gearing would affect the cap. maybe the cap has seen it's last race. I have never noticed my cap getting hot, but never really been checking it.

[WHL]

i dont think Novak would have recommended a final drive as this changes by track/motor etc. I also don't think that the gearing would affect the cap. maybe the cap has seen it's last race. I have never noticed my cap getting hot, but never really been checking it.


They reccomend the following:

Touring cars - 7.3:1

Offroad buggys - 9:1 (This would suggest a 18/19 tooth pinion for the Yokomo)

Offroad Stadium Trucks - 12.3:1

And they also say -"These are gearing recommendations for a fairly large open track. Lower gearing may be needed for: carpet tracks, "tight" tracks, or tracks that have a lot of "short straights" & "hair-pin corners".

The only other compatible motor apart form the SS5800, is the SS4300 which they say is comparable to a stock motor. They also suggest the same gearing for both.

It's all on their website.

[marcoski]

I think the tamyia connector is making that capacitor work really hard due to all the voltdrops it is having to smooth. Deans, or better still corrallies should help to keep it cool. Hard working capacitors do run warm. I am fairly sure the M stands for Matsushita, which is one of panasonics component devisions. They are seen alot in high current switch mode supplies and on the power supply sections of computer mother boards. They are up to the task.

Bear in mind that in unlimited mode, the system will demand over 200A from a standing start! The tamyia, and even a deans (30A continuous) will baulk at carrying that current. The corrally stands the best chance as it will carry 50 amps continuously.

[Remo]

That Banned Word Machine again... ;D

Matsusheeta.

Eat me.

Sorry... :-X

[JV]

Marcus Goulter writes:
> I think the tamyia connector is making that capacitor work really hard due to all the voltdrops it is having to smooth.

> Deans, or better still corrallies should help to keep it cool.

Corralies - the fancy banna-plug things ? I'll stick with polarised connectors. I don't trust myself (or anyone else) to always plug them the right way round !

I've had a bit of a look at industrial connectors (e.g. Rema), but I don't think I've found what I'm looking for. And there's some benefit in staying with the rest of the RC world.

> [...] over 200A from a standing start

... which will heat a 1mm^2 copper conductor at something like 200K/s,
if I've got my arithmetic right.

Is that 200A figure a measurement, or a a calculation ?
Measuring that kind of transient current seems quite tricky.

Maybe deans connectors in parallel ?

John.

[marcoski]

A measurement! Captured on a digital storage 'scope whilst the motor was asked to turn a CS dyno flywheel from a standing start. Accelerating from mid range RPM to full speed asked around 60 to 70 amps.

We use somewhat bigger wire than 1mmsq - most use 12 guage which is around 2.5mmsq and the lengths are quite short (or they should be!) so the volt drops and hence heating are kept within reason, but even the nominal ask of 60A will find any weak links in your wiring and ultimately melt them!

The best solution is to dress the wiring so it is not possible to plug up the wrong way. If this is not possible (preassembled stick packs) then use both male and female corrallies. For instance, use a female on the battery pack + lead and a male on the battery pack - lead. On the speedo you would have a male corrally on the + lead and a female on the - lead. It would be difficult to connect that wrongly, but careful attention to heatshrinking would have to be paid to prevent short circuits.

[MikeS]

With reference to cell connections - I always use Corally tubes on the cells. When I build a pack, I always orientate the label on the cell with the positive connection face down and all the rest face up.
This way, when you pick up a pack you always know that the "odd" cell, which ever way up the pack is, is the positive connection!

Apologies for going slightly off topic!

[JV]

Marcus Goulter writes:
> A measurement!

Yes...

> Captured on a digital storage 'scope

Sure. But scopes measure voltage, not current, and they don't even (usually) have floating inputs.

I guess I'm asking what kind of current probe you were using. I don't think I know of an AC/DC probe with any kind of bandwidth which can handle more than 50A.

The most straightforward answer if we don't have a suitable current probe seems like some king of precision low value resistor - maybe a length of copper wire. But at 200A, the voltage across it would have to be a few mV to stop it heating, which is getting a bit low for many scopes. A differential scope probe might be needed too.

Or a transformer ? Can't do DC though. Could be OK for transients.

Regards,

John.

[marcoski]

The scope has floating inputs, though how that is relevant I fail to see!

The current shunt, which was in series with the positive battery wire, is rated at 250 amps and gives 1mV per amp. So we are looking at 200mV initially. The scope will go down to 5mV per division so plenty of sensitivity there. The shunt is just a slab of copper and is non inductive so there is loads of bandwidth, certainly enough to capture the event in question.

An alternative probe would be to use a hall probe rather like a DC clamp meter. I don't have one of those.

[bob burr]

Can i ask what any off this is to do with Brushless ? ???

[JV]

Bob Burr writes:
> Can i ask what any off this is to do with Brushless

Well, if you want to know what your brushless motor is really doing,
you'll have to measure it, ideally putting the thing on a dyno, and instrumenting the
motor + controller as Marcus describes.

I guess there isn't anything very specific to brushless in the instrumentation,
but there doesn't seem to be an 'instrumentation' forum...

I think I'll be done with this thread soon anyway, if it bothers people.

Regards,

John.

[JV]

Marcus Goulter writes:
> The scope has floating inputs, though how that is relevant I fail to see!

Well, if you want to make measurements in more than one branch of the circuit,
and the scope has all the input grounds wired together, there's a measurement
problem. And maybe a repair bill...

In that respect a Scopemeter seems more handy than a TDS3054.

> The current shunt, which was in series with the positive battery wire,
> is rated at 250 amps and gives 1mV per amp.

> So we are looking at 200mV initially.

Righty ho. I'm more a digital person, and when I've poked into power stuff (CPU power and the like) in the recent past I've had nice wideband hall-effect current probes to play with. So I'm not so familiar with the kind of resistive shunt you seem to be using. I think I get the idea after a bit of googling, though.

> and is non inductive so there is loads of bandwidth

Right.


Can you suggest a couple of manufacturers to look at ?

Regards,

John.

[marcoski]

That would be tricky - mine was not off the shelf! It has been calibrated at my local calibration house.

I suppose your first port of call would be someone who makes shunts for DC ammeters, like Hobart. However, a hall probe is a MUCH better way of doing it.

Re Bob, at the moment it is difficult to get a quantitative analysis of the brushless motor. The CS dyno is not man enough to cope with the start up current so alternative means of evaluation need to be found. I have tried many times to convince the CS to take sensible readings of the BL system, so far to no avail. Even with 4 packs of Dooms finest in paralell the CS still insists the cell voltage is too low. This is because there are so many losses in the current measuring circuit and cables. I am sure that we all would like to see a dyno print of a BL system on something like the CS so we can compare it to the motor technology we are used to.

[Dave G]

i believe its very close to the 200watt mark marcus.

dave

[bob burr]

Blimey Marcus are you trying to damage my brain ? im a old fasioned motor mechanic ??? wheres my big hammer and chisel that will fix it. ;D ;D

[JV]

Chisel ?? A screwdriver and a hammer work for me.

Anyway...

The supplied capacitor got hotter still, even after lowering the gearing to 8.5 or so.

So I replaced it with a pair of 18 milli-Ohm 16V electrolytics in parallel. [see below]

They get just noticeably warm. They certainly don't melt velcro glue. Not after some 10's of hours running.

I think the tamiya thing was a red herring. While the relatively high contact resistance of a Tamiya connector will make life a bit harder for the capacitor, the resistance of the Tamiya connector is a small fraction of the internal resistance of the battery.

I suspect that gearing, the nature of the surface & the track, and the state of the batteries have more to do with it, as some folks suggested.

Depending on the kind of surface, and the state of the surface, and how twisty the track is, it seems to me that like overall ratios from 7 - 11 might make sense. Then gearing is not something that I think I'm good at making sense of.


Capacitor stuff (Or what I made of it, anyway):

At the switching frequency of the ESC, the capacitance of the external capacitor is a lot less important than the ESR ("Effective Series Resistance"), which is roughly equivalent to the internal resistance of a battery. That's because at the switching frequency (~30kHz), the ESR dominates the impedance of the capacitor to the extent that the impedance of the capacitor is largely resistive. Capacitance matters more at the lower switching frequencies of brushed motors.

So for brushless the ammount of capacitance doesn't really matter. If the ESR is low enough, there will be enough capacitance.

I never figured out exactly what the supplied capacitor was. From the size of it, I'd expect the ESR to be in the 20-50 milliOhm range.
But I could be wrong.

If the ESR is a lot bigger than the battery internal resistance (at the switching frequency), the capacitor isn't going to do anything useful at all. Diminishing returns apply as the ESR gets much lower than the internal resistance, as the charging of the capacitor during the off-time of the speed controller becomes limited by the battery's internal resistance.

So assuming the battery has an internal resistance in the 12-20milliOhm ballpark, It seemed a capacitor, or set of capacitors in parallel, with a combined ESR in the 5-10milliOhm range would be good.

The lowest ESR capacitors I could easily get hold of at the time were Nichicon HD series UHD1C222M, a 16V 2200uF capacitor, 12.5x25mm, with a maximum ESR of 18milliOhm. (Available from Farnell).

Two in parrallel seemed enough. Three or more seemed a bit too bulky.

Much better capacitors are available, like Nichicon HN series, where the same size of capacitor has an ESR below 8 milliOhms.
--
"That statement is no longer operative"