Author Topic: resistive load bank  (Read 4558 times)

0 Members and 1 Guest are viewing this topic.

kitestrings

  • Hero Member
  • *****
  • Posts: 1376
resistive load bank
« on: April 17, 2014, 08:54:59 PM »
I was looking a bit closer at size configuration and construction of a resistive load bank for our turbine.  I've been thinking the best match would be 3.4 ohms per phase (2 - 6.8 ohm in parallel), 6 resistors total.  They would be controlled ahead of the rectifier; wired in wye.  They're capable of about 800-1,000 watts each, and if my math is right should look something like this:

      (2 in II, 6.8 ohms)         
Vline   Vph   R, ohms   I   watts (tot)watts (per resistor)
80   46.2   3.4   13.58   1,882   314
90   52.0   3.4   15.28   2,382   397
100   57.7   3.4   16.98   2,941   490
110   63.5   3.4   18.68   3,559   593
120   69.3   3.4   20.38   4,235   706
130   75.1   3.4   22.08   4,970   828
140   80.8   3.4   23.77   5,765   961
150   86.6   3.4   25.47   6,617   1,103

In an earlier post http://www.fieldlines.com/index.php/topic,148112.0.html Chris had mentioned using overhead door springs, so I wanted to compare this option, because they are definitely cheaper.  In his case he reported about 3.7 ohms (also wired in wye), with the resistance nearly doubling when they get red hot.

Okay, so a spring for a 120# door is about 4 lbs.  Ours has steel spring wire that measures about .15" in diameter.  With steel/alloy at about .284#/in^3, this translates to a length of about 66'.  And, this is close to what ours works out to - roughly 170 turns of 1.375-1.5" diameter coils.  This was the easy part, estimating the resistance is what I'm having trouble with.

An on-line properties table lists CU at 1.7x10^-7, and steel alloy at 5.94x10^-7.  This would suggest that the steel might be 3.5 times higher in resistance.  If 7 AWG (.1443" dia.) CU is .4982 ohms /1000', does it make sense that my .15 spring steel might be 1.72 ohms/1000'?

If so, a single set of three in wye would yield only .111 ohms /ph, or line resistance of .224 ohms, and even if it doubled under load it would still seem to be way low for our application.  What am I missing, clearly something?

Thanks for any help.

~kitestrings

joestue

  • Hero Member
  • *****
  • Posts: 1759
  • Country: 00
Re: resistive load bank
« Reply #1 on: April 17, 2014, 09:09:49 PM »
you're missing the fun of watching it get red hot while measuring the current.

I heard a story from someone on the dutchforce forum, he used 60 feet of half inch rebar per phase as a test load for a large generator.

i recall reading that steel was 7.5 times the resistance of copper, wikipedia numbers for carbon steel say 8.57, 8.51 fold. not sure which to use myself, i thought it was 7.5.
My wife says I'm not just a different colored rubik's cube, i am a rubik's knot in a cage.

kitestrings

  • Hero Member
  • *****
  • Posts: 1376
Re: resistive load bank
« Reply #2 on: April 17, 2014, 09:23:10 PM »
joestue,

It may be hard simulate a 140-150VDC source into a garage door spring (and not burn the garage down ;)).

That would certainly pull it closer in line.  Here's the source I'd used:

http://hypertextbook.com/facts/2006/UmranUgur.shtml

joestue

  • Hero Member
  • *****
  • Posts: 1759
  • Country: 00
Re: resistive load bank
« Reply #3 on: April 18, 2014, 01:01:43 AM »
the error in your post is the exponent, -7 for copper instead of -8.

anyhow, there are a number of ways to measure milliohms accurately, but you really can't do it with just one meter.

if you want to spend 3$ i can show you how to measure microvolts, and low ohms with one of these digital scales:
http://www.ebay.com/itm/Portable-40kg-10g-Electronic-Hanging-Fishing-Digital-Pocket-Weight-Hook-Scale-TR-/181274427002
i recommend buying two kelvin clips to make the connections easier, but 4 alligator clips are just as well.
My wife says I'm not just a different colored rubik's cube, i am a rubik's knot in a cage.

Flux

  • Super Hero Member Plus
  • *******
  • Posts: 6275
Re: resistive load bank
« Reply #4 on: April 18, 2014, 03:55:09 AM »
Small springs are wound with piano wire, most likely the big ones are a similar alloy.

I found a figure for piano wire of 1.18 x 10^-7.  The figure for copper is 1.68 x 10^-8

Seems to agree with the 7 times copper figure fairly well ( figures assumed at 20deg C).

Making connection to the spring ends could add a fair bit of uncertainty as there will likely still be a fair bit of heat there and a fair bit of thermal emf wrt copper connectors.

Flux

kitestrings

  • Hero Member
  • *****
  • Posts: 1376
Re: resistive load bank
« Reply #5 on: April 18, 2014, 12:16:13 PM »
Quote
if you want to spend 3$ i can show you how to measure microvolts, and low ohms with one of these digital scales:

I have access to an accurate scale, and I would be interested in this, though I'm still not sure how we account for the effect of heat on the resistance.

I'm starting to think this just may not be the right fit.  Let's assume our spring is 7x higher resistance than CU.  This puts the line resistance at .455 ohms.  At 140V we're looking to apply about a 5-6 kW load (a bit more to be safe).  It looks to me like our initial current would be 180A; in effect applying a load of about 44 kW.  Even if the resistance doubled with heat (or we're off by a factor of 2 in our estimate), we're still looking at 20kW+.  Hard to see that the stator is going to be a lot better off than just continuing to short the windings?

Seems like we're still a long ways from Chris's 3.7 ohms, though I don't know if he meant phase or line resistance.

~ks

kitestrings

  • Hero Member
  • *****
  • Posts: 1376
Re: resistive load bank
« Reply #6 on: April 18, 2014, 12:32:05 PM »
This was the comparative product that I was considering:

http://www.alliedelec.com/images/products/datasheets/bm/OHMITE/70022423.pdf

Flux

  • Super Hero Member Plus
  • *******
  • Posts: 6275
Re: resistive load bank
« Reply #7 on: April 18, 2014, 01:11:20 PM »
Temperature coefficient of resistance for iron is about 5 times that of copper. You mentioned red heat so 600deg C plus so it may be in the right league when hot.

Not sure how inclined it will be to get hot in the short term as the resistance will not rise much until the temperature starts to go up.

Flux

joestue

  • Hero Member
  • *****
  • Posts: 1759
  • Country: 00
Re: resistive load bank
« Reply #8 on: April 18, 2014, 02:25:52 PM »
How to measure low ohms with a digital scale:

1)Disconnect the strain gauge, and acquire two kelvin clips or 4 alligator clips or similar.
2)Connect two of the clips through some test leads to the bridge/sense/difference input, where the green and white strain gauge wires went.
3)Acquire two resistors, for reading 0-1 ohms, you need two 680 ohm resistors. for 0-0.1 ohms, you need two 68 ohm resistors. for calibration purposes you need a few one ohm resistors in series, or a few high value resistors in parallel.

4)solder one of each resistor in series with the terminals where the high and low side of the strain gauge was connected, previously red and black wires most likely.
the other side of each resistor goes to the other two alligator clips.

the way a digital scale is setup, is to read the difference in voltage, generated by the strain gauge's difference in resistance due to strain, but this voltage is compared with the battery voltage, as a ratio measurement.
For almost all strain gauges, full scale on the order of one part in 1250 difference, or about 2.4 millivolts when running from a 3 volt battery.
using just the sense leads, you have a 0- 2.4 millivolt, 3.5 digit volt meter, and you should easily be able to read thermal emfs and whatever.
if you don't apply more than 3 volts to the leads, you won't blow the meter up.

using 68 or 680 ohm resistors instead of a bridge, to push current through the test resistor, generates a voltage, proportional to the battery voltage and test resistance.
the "weight" reading will be proportional to the test resistance.

the digital scale will have a mosfet switch to turn the bridge on and off, so that drops some voltage as well, which is why you may need to fudge the resistors.
also, because the resistors aren't a true current source, the contact resistance of the current sourcing alligator clips does matter, but in practice its negligible.

also, most of those scales null the scale each time it is turned on, so you have to short all 4 clips each time you turn it on.
My wife says I'm not just a different colored rubik's cube, i am a rubik's knot in a cage.

kitestrings

  • Hero Member
  • *****
  • Posts: 1376
Re: resistive load bank
« Reply #9 on: April 21, 2014, 09:43:45 AM »
[quoteHow to measure low ohms with a digital scale:][/quote]

Ah, now I see.  It didn't occur to me that we'd need to alter the 'inards' of the scale.  I wasn't sure where you were headed.  I had borrowed a digital scale from the science dept. at my local high school when we wound the stator coils.  By weighing them I could quickly tell which ones I'd thrown an extra turn into (several as it turned out as I recall).

So, I'll pick up one of the cheaper scales that you suggested and see if I can run through this.  As I said, at a glance it looks like the resistance may be too low, and I'm not sure I'll gain much (considering the added labor) if I have to start putting pairs in series or some such.  I am going to check some different spring s sets though.  I looked at the normal (9' x 7' door) stock springs, perhaps a longer one with more turns would be a better match.

Quote
Making connection to the spring ends could add a fair bit of uncertainty as there will likely still be a fair bit of heat there and a fair bit of thermal emf wrt copper connectors.

I haven't wired many toasters or Amish, 'energy-saving' space heaters lately, and I agree I think it will be the anomaly that we reach, and sustain high temps.  But, I assume with something like this one might go to higher temp wire/insulation (450 degC?), and steel connectors?  Are there any other precautions that make sense, and what is the preferred terminal connection method - compression, brazed? - I assume you don't want soldered connections, or maybe just a solder with a higher melting point and/or map gas?

Thanks for the feedback.

~kitestrings

Flux

  • Super Hero Member Plus
  • *******
  • Posts: 6275
Re: resistive load bank
« Reply #10 on: April 21, 2014, 12:47:54 PM »
I would be tempted to weld or braze substantial flat steel strips on the ends sufficient to get the temperature down to the point where you can bolt copper crimps on. Copper oxidises very badly if you braze it directly to heating elements at that sort of temperature.

As a compromise you could use shorter steel strips and braze the copper to them and use glass sleeving or something. I would keep the copper below 200C if possible for a brazed joint and even lower for a bolted crimp.

Small commercial heaters are often crimped on to stainless wire but I regard that as a cheap production method and wouldn't try it on this scale of wire.

You are out of the range of soft solder whatever grade you look at.

Flux

Mary B

  • Administrator
  • SuperHero Member
  • *****
  • Posts: 3169
Re: resistive load bank
« Reply #11 on: April 21, 2014, 04:23:10 PM »
Between 80-150 volts I would be very tempted to use an electric baseboard heating element...

kitestrings

  • Hero Member
  • *****
  • Posts: 1376
Re: resistive load bank
« Reply #12 on: April 22, 2014, 08:29:10 AM »
Quote
Between 80-150 volts I would be very tempted to use an electric baseboard heating element...

I've considered some additional "opportunity load" using baseboard or a space heater through my inverter (winter); perhaps ice making summer.

The problem I see with baseboard is the stock convection heater sections run 250w per ft. (at 240V).  For a 4', 1 kW heater that puts the resistance at 58 ohms.  At 150V then, it would pull only 2.6A, or 390w.  We're looking to potentially dump 5-6 kW, but likely at a voltage of only 150-160.

The of course make 120V heaters in all sorts of configuration, but here I think the voltage would be too high on the high-limit(s) of things.  Maybe I've overlooked another option?

We're diverting to a pre-heat water tank now.  this works great, but it is not enough load and without alteration is not 100% dependable.  I have also thought about using a 3-phase electric water heater, and setting it up to intentionally discharge unused hot water at a safe margin below the limits of the tank.  I seem to keep returning to wanting to keep it as simple as possible, and effectively shutting down the turbine when power is not needed.

~ks

Mary B

  • Administrator
  • SuperHero Member
  • *****
  • Posts: 3169
Re: resistive load bank
« Reply #13 on: April 22, 2014, 07:03:45 PM »
they do make 120v baseboard

kitestrings

  • Hero Member
  • *****
  • Posts: 1376
Re: resistive load bank
« Reply #14 on: April 23, 2014, 01:18:21 PM »
MaryA,

Well, I looked this (and tried to say it above, typo not withstanding).  They make 120V units up to 1,500w.  At that size, I make the rated current at 12.5A, and resistance at 9.6 ohms.  So, in a wye configuration we'd need to parallel pairs of them, and even with six units we're still at only a bit over 4.5 kW at 150V.  In delta it's be a better match if the upper limit were 120V, but at 150V we'd be at 125% of rated current.  Other lower wattage combinations exist, but I don't see improvement.  Please let me know if you are thinking of a combination that I haven't considered.

Either way baseboard is 250w/LF; 6' for a 1,500 strip.  There are fan/coil units that are more compact, but if we have to parallel units the costs rises quickly.

The best match seems to be low resistance, high power (& ideally low price).  The wire-wound resistors, or the garage springs, would allow you to move along the element to select intermediate resistance value.

~kitestrings

Mary B

  • Administrator
  • SuperHero Member
  • *****
  • Posts: 3169
Re: resistive load bank
« Reply #15 on: April 23, 2014, 04:04:02 PM »
I was thinking along what you came up with, parallel strings of heaters. They run cooler than a spring might so less of a fire hazard was my thoughts.