Cap banks - the pre-emptive answer

[drwayne]

Sorry for delay, remembered I have a career ... !

fella 1340..
1) friend's fantasm .. which one ?
2) if you have no hole shots planned, the power expectation of circuit is reduced..10S 12S 14S is nice to know, but the total amp draw is the circuit crusher.

I have logged the SF300 to 12S 455amps in peak draw.... =22750W ... your 14000 is quite conservative.
^^ Castle 2028, SF300, 12S1P(12S50/100C5800mAh →29KW avail from packs on peak ), Voodoo 6719/3 !!
Always plan ahead.. expect the unexpected overkill amp draw and if that ever occurs the system will survive to play another day.

5000→6000uf in standard parallel ( as pic below ) is fair for systems not under extreme load.

stdcaps.JPG


If your system is borderline barbeque material, the dual parallel+series array cap banks as detailed originally are more your needs..
With this circuit, the refresh and supply is paramount concern, not the total size of the ripple buffer...
ergo power available on GO, not just ripple reduction..
I aim for 2000→3000uf is this situation ( maths for this are on first page.. or PM for help ! )

8x2capbank.jpg

[Fella1340]

Thanks Doc, I think I can at least make an attempt from here. I think I will end up abusing the boat, increasing kv and motor size so I will go with option two. See if I can't put something together from here. Thanks again,
Jeff

[kevinpratt823]

I have a decent understanding of caps and theory, but bear with me if I might revisit some of your explanations I may not have understood....
In the SprintCat I just built, there is limited real estate that presents a slight challenge. My motor(1527/1Y) is in front, with the ESC(SF240 HV) right next to it, keeping motor leads as short as possible. Long battery wires were kind of unavoidable to the 6s2p/5000 arrangement of (2) 3s in each sponson, and I have little room to install the ETTI 63v/5000uf bank that I had on hand, so I tapped off the ESC leads with 12awg and ran about 14" of wire(the blue/black wires) to get around the motor into the opposite sponson. I could probably make it fit under the ESC shelf, but I was trying to avoid putting any more extra weight on one side along with the servo and ESC. Can you elaborate on the effects of excessive length taps to the bank? Am I causing problems, or just lessening the effectiveness of the caps?
SAM_0394.jpgSAM_0398.jpg

[properchopper]

Seems that single motor cats have this problem of connecting two packs that live in the sponson(s) to a central ESC which can require some long runs [and caps]. I've built a few ; my personal Cheetah has some long wires and an Etti cap board from the CC240 to the motor- so far all is well

DSC05951.JPGDSC05950.JPG

I'm building one now for a customer with a Swordy 220A and came up with a somewhat "shorter-wire" config minus caps.

[Fella1340]

Nice work properchopper. You give me hope I can do it neatly!

[kevinpratt823]

Hello.
The sprintcat cowls have enough 'breathing' space to mount the ESC on top the motor
and it appears you packs are facing towards the transom.. turn them around !

[ATTACH=CONFIG]116307[/ATTACH]

[ATTACH=CONFIG]116308[/ATTACH]

Actually, it's (2) 3s packs end to end. The parallel adapters are a bit long to accomodate the even shorter leads of my other set of TP packs. I really wanted the packs in the sponsons, but with the top deck opening coming in past the tunnel in a sprintcat, the largest packs you can fit in without a shoehorn are 3s/5000. I can fit in 4s packs with some real squeezing, as I want the option of running 8s2p for kicks once I get the boat trimmed up. Next time out, I will try running them standing on end side by side, it may be a better option, and once I figure this stuff out, I will also likely shorten the series connecting leads. I have thought about fabricating something to mount the ESC over the motor, and that will gain me an inch or so, I just want to make sure it doesn't limit my clearance to run the classic hatch forward or backwards. Aside from all that, for my/our own education, would you mind elaborating on the effect of longer tap wires to the ESC?
I am also wondering what you think about the possibility of just running say 2 of these cap banks in series, would that achieve the same results as far as faster charge/recovery rate that you are aiming for with some of your series/parallel setups? Sorry for all the questions, but you are apparently a great resource on this topic.......

[kevinpratt823]

As far as something like a SF300 with 2 sets of batt leads that have to be joined, I am curious what you think of the way I did it. I ran a #12 from one esc batt lead, straight through the bank and back to the other esc batt lead. The connection between the 2 esc leads serves mostly just to equalize the voltage I believe, and I now basically have a set of paralell #12 to feed the cap bank. It's been working great for me, but by all means I welcome your input......SAM_0315.jpgSAM_0312.jpgSAM_0313.jpg

[kevinpratt823]

I will rework the layout in the Sprincat. I am well schooled in electrical theory(above average electrician), and I understand capacitors and inductance, what they do, how they add up in series/parallel, etc... just not to the level of detail of some of the effects/side effects of different arrangements and applications like PWM as you obviously do. For my own knowledge can you tell me the problem of having the tap wires too long to the caps....... Obviously there is a little more resistance, just wondering if there is anything else, and how big a deal it is.

I think you understood the SF300 wiring, but here is a diagram. Yes, the wire that passes through the cap bank unbroken is also the conjoiner, and FWIW the wires are equal length in reality, and the I am feeding each set of independent Fets with it's own 10s1p supply, with the capacitor lead also serving to join the two and equalize the voltage. It runs form one +, across the cap bank(unbroken), back to the other +.

That is a nice piece of work you've got there with the dual output shafts!
I still have some to learn about monitoring and correcting ripple current, as well as what is too high, and what is good or acceptable, but I do love data logging! I have always been intrigued with electrical theory and understand it well, but a lot of this branches into electronics as opposed to my work application, so I haven't delved into it as deeply in the past. Thank you for your time and attention.

[kevinpratt823]

I also am a bit of a Geek when it comes to this stuff, so I will continue to entertain these discussions with the hopes of elevating my knowledge, until you feel that I'm taking away from the thread, so feel free to stop me.........
How could one of the sets of fets possibly see a dropped voltage from one set of packs when the feeding wires are essentially common via the loop that just happens to also pass through the cap bank? The 2 sets of packs, as well as the esc wires, are essentially parallel, they just happen to have a few inches of wire doing the job, instead of bringing the 2 directly together. Even if one side drops, the total voltage the esc sees on each set of fets should be the same aside from a minute voltage drop across said wire, even as one pack "charges" the other pack to recovery, no? In your picture of the sf300, you have the pairs of #8 coming together before the cap bank, and what appears to be single #8 running through from there. My parallel #12's are not that far off from a single #8 in terms of resistance/ampacity I don't think, assuming the ripple and discharges are relatively even between the 2 sets of fets. Mine just happens to have a little more resistance because of the few extra inches of conductor. In the case of the SF300, I really don't see a benefit of bringing the 2 esc wires down to one when they are going to split back to 2 to feed a 10s2p setup anyway.

I thought the firing of the fets was based on the signal from RX, I didn't know the voltage would play a part anyway other than affecting the output, but then again, my understanding of PWM is mostly based on inquiring to a startup rep for large Variable Frequency Drives. I'd love to see you start a thread explaining the ins and outs of fets and the sequences of PWM as it applies to our uses.........
I certainly will be considering all of this stuff, but in the end I am not at the level where this stuff becomes as critical to my sport applications, I have barely broken 70mph as of now, but I still want to learn. I am also not at a point where I want to completely fabricate cap banks like these impressive systems you are displaying, for now I'm just looking for better ways to install what I can readily get to work with. Thank you for giving me food for thought. In a nutshell I will do my best to lower the resistance to my cap banks(and everything else of course) and allow for the highest ampacity as well now that I am more aware of the importance.

[kevinpratt823]

Thanks Doc. I had considered those potential failures when I did this, but I think when I took into consideration that I wouldn't be drawing much over 200 amps and I just wanted the boat in the water I went for it. I will reconsider and put it on my list of weekend morning hobby projects. The smelly, smokey learning curve is still just beginning to level off for me.

[ray schrauwen]

Every time I start reading in this thread I get overwhelmed and almost throw up. I'd love to build one of those cap banks but, it's not going to fit under the cowl of my Whiplash 40.
So Dr. Wayne is it true you deigned or had some part in designing the SF esc's??

I almost cooked one at the nats, tough esc. I added more cooling and I have one of those smaller Etti Cap Banks for LV or at least up to 35Vdc. I only run 6S2P in my Whip.

I'm hoping the cap bank I bought will help some??

[drwayne]

I do think it's time this line of postings was audited for content.
Lots of not really relevant stuff now floods this area.

Ray
240HV and 120LV had fingers like mine through them in earlier days, yes.
Still beta dabbling with new, and twisting through code for performance/longevity updates..

What esc had you almost toasted but then resurrected ?

[kevinpratt823]

Kevin .. Ive been ruminating over this for 10 days. ... lol
What is ampacity.. do you mean amperage ?

Ampacity is the amount of current something is rated to withstand.
"The ampacity of a #12 thhn copper conductor is 25 amps" (but it can only be fused at 20 per electrical code)
Think "amperage capacity"
Sent from my SCH-I535 using Tapatalk 2

[ray schrauwen]

I almost toasted a new SF200 8S esc in my Whip 40 Sport Hydro on 6S2P. I have yet to load test it again but, it's fine on the bench.

It has gotten up to 175F at least on one run, the caps appear to be fine physically so hopefully it still works when I run it again soon. Added an extra cooling plate and the Etti LV cap bank.

I do think it's time this line of postings was audited for content.
Lots of not really relevant stuff now floods this area.

Ray
240HV and 120LV had fingers like mine through them in earlier days, yes.
Still beta dabbling with new, and twisting through code for performance/longevity updates..

What esc had you almost toasted but then resurrected ?

[kevinpratt823]

Wow. :shock:
proof positive I dont know everything !
hahaha

Shhh ... If you spread that admission around I will never hear the end of it

I now understand 'ampacity' in domestic ac circuits.. mainline 110AC for eg.
However an ac cable suitable to power my 2500W toaster oven on 110AC ( 22.7A) with length of say 6' from the power socket.. would over heat and melt on 25VDC at same 22.7Amp .( just 567W )...
Automotive headlight cables are minimum 5mm (1/5 of 1") to meet approval in Australia .. else they melt.. for a maximum 150 W globe ...

I see from here that the listed 6AWG(AC) of 4.11mm diam is much smaller than the AWG6 of 6.61mm diam used in our DC circuits.
LOL now that's confusing !!!

http://www.keywolf.com/American_Wire_Gauge.php

Ive a little book that determines wattage capacity for DC cables based on wire diameter, number of strands, quality of copper, quality of insulation and voltage through the wire.
it's 11pm here Wed night .. I will search Thursday daytime for online the same details.
regs
W

As far as I know, Doc, AWG refers only to the total circular mil area of the conductor and is the same for all conductors, AC or DC. The diameter in the chart could be for a single solid conductor or one of fewer strands than the one you referenced. A multi stranded conductor like we use in boating will have a larger physical diameter because of the air gaps between strands, but still has the same cross section of copper. Even with AC wiring, some wire strippers will recommend for example to use the 12 AWG hole to strip 14 AWG stranded. Also, the ampacity of a conductor is primarily limited by heat, so while the #10 with common THHN(90 deg C) insulation that we use in buildings that has an ampacity of 35 amps(but only to be fused at 30), can be bought with much higher temp insulation which gives it a rating of 70 amps. Not sure how that applies to DC, but most of the wire we are using in boats is rated for very high temps, I'll have to look in my code book and see what a 200 deg insulated #10 is good for....
Acceptable voltage drop in my world is typically figured as a percentage of the voltage, so yes, you have far less room for loss with lower voltages, as the resistance is still the same per foot regaurdless of the voltage applied.

For the sake of discussion, in the case of your toaster, the heating element is really just a pure resistor. The 2500 watts is what it is rated to consume at 110v, making it essentially a 4.85 ohm resistor, pulling 22.7 amps at 110v. If you only apply 25v, it's only going to draw 5.15 amps, and you aren't making much toast at 128 watts, never mind melting the wires.


So, in your hypothetical situation, I think if the wire only carried a fixed 22.7 amps, it likely wouldn't heat up any more at 25v DC than at 110v AC, could you explain to me why you are saying otherwise? Amps are amps, and unless the current increases over 22.7 to compensate for voltage drop while trying to still supply a fixed wattage load, I don't see why it would heat up any more. I am vaguely familiar with the "I squared R losses". iI's been a long time since I've been in a theory class and it's not something that regularly applies to my work, but voltage is not part of that equation. If current is still 22.7, the I squared R loss is the same regaurdless of voltage.

In my work, most of our wire insulation is rated for 600v, the ampacity ratings do not change weather it's 24v or 480v, and I didn't think it changed for AC or DC either, but I don't deal with much big DC. Looks like I'll be flipping though the code book tomorrow. I am not familiar with ratings of "wattage capacity for DC cables", I am interested to understand more....
BTW, that sounds like a serious toaster!