Specs for Building a Decade Box

Check the spec of the meters, depending on their output you need to work from there.

The meters specs state - short circuit current is 2mA . For the 5KV IR Meters , does that mean I need 10W resistors

Lets talk about 5kV.

5kV / 2mA = 2.5MOhm

If output short circuit current is 2mA, probably 2.5 MOhm internal resistor is connected between 5kV source and output of Your Meter. This resistor limits output current to 2mA and probably withstands 10W.

So, any resistor connected to output will lower output current below 2mA. Maximum power will be dissipated if You will connect 2.5MOhm resistor to output. Current will be 5kV / (2.5 MOhm +2.5MOhm) = 1mA; 1mA * 2.5MOhm = 2.5kV; 2.5kV x 1mA = 2.5W. This is the maximum power available at the output of Your meter.

In Your series of resistors dissipated power will be lower:

2 Ohms: I = 5kV/(2 Ohms + 2.5MOhm) = ca 2mA ; 2mA x 2 Ohm = 2mV; 2mA * 2mV = 0.004mW. But other Ohmmeters may use 10...200mA, so 1W resistor should be used. But, at 2 Ohms level connector wire can have 0.05 Ohm equals to 2.5% error. Not good enough? Apply Four wire connection. Or thick, short internal wires and connection cables.

200 Ohms: .... 0.4mW ..... as above, better 1W, also due to selfheating error.

... Up to 10kOhm, no problem with Voltage or power, I would use minimm 0.5W

100kOhm: ..I = 1.923mA ; U = 192V; ... P=370mW. Resistor should withstand 190V! See TLC Designer comment.

1Mohm: 5kV / (2.5 MOhm + 1MOhm) = 1.429mA; 1.429mA x 1 Mohm = 1429V; 1429V *1.429mA = 2.041W Here start Voltage issues both with resistors , switches even cable insulation (SAFETY!) - see Electroman & others.

10 MOhm: 5kV/(2.5MOhm+10MOhm) =0.4mA; 0.4mA * 10MOhm = 4kV High voltage issues - cable insulation, safety, High voltage plugs/sockets.

Probably buying old working General Radio Decade Box would be cheaper, than doing Yourself., (to cover 2 Ohms...100kOhms). But for voltages >700V special High Voltage Standard Resistors should be used.

Incidently you didn't answer the question ?Don't Know?

The 2,20,1000 ohm resistor are going to be used to verify the accuracy of Multimeters ( when on the Resistance scale ).

JJJ-

I believe what you are calling an insulation resistance meter is what we call in electrical 'slang' a "megger". Therefore, you are correct in saying you want to measure the resistance of the insulation, not the resistance of the wire conductor inside the insulation. Then we'd be talking about using a simple ohm meter or multimeter - which by the way, you could also test once you built your decade box.

However, not to pour water on your fire, but I think you may find that you are going to spend a LOT of money to get resistors in the wattages and tolerances you need if you really want to test the accuracy of the meter to a reference standard or certification level of quality.

1% resistors in the gigohm range have a pretty wide tolerance, yes? 1 percent of a billion leaves a lot of room for variation. So what this means is you either have to sift through a whole bunch of 1% resistors to find two that match closely enough, or move to 1/2% resistors... and remember, while you are measuring all those resistors to find a matching pair, your meter is probably only good to 1-3% unless you have a dang nice meter. So you will have a really hard time making two of these that are the same, if that is important to you, then I suggest you skip down past all this post to the five suggestions below or check some other posts!

So, maybe it's better to have your meters calibrated by the manufacturer or a lab... or, if that's not possible, if you have two identical meters you can always measure a known conductor and compare readings. If you have two identical meters and one reads 2 ohms and the other reads 10,000 ohms, you either have a dead battery in one meter or a broken meter or both.

Okay - back to building your box -

Yes, 10 watt is the largest wattage you need - you did figure that out correctly, but not all of the resistors need to be 10 watt... at least we would presume that based on the information and the "typical megger" in the marketplace. If you have something odd or custom, well then again, you probably need to skip down to 1-5 below.

You can use Ohm's law to calculate the watts that your 5000V meter will push through the resistor at the maximum current of 2milliamps at the dead short maximum output.

Voltage (E) = Current in Amps (I) * Resistance in Ohms (R)

You will notice right away there's no units for Watts in Dr. Ohm's equation so we have to ask our friend Dr. Joule how to change Current into Watts.

Joule's Law states -

Power (watts) = Current (I) * Voltage (E)

So for your 2 ohm resistor we need to find how many Watts it needs to pass without overheating.

We know your meter claims to push 2 milliamps through a dead short = 0 ohms.

Let's assume for this exercise that it is super-strong and can also push 2 milliamps through 2 ohms. Just in case you don't understand this part - 2 ohms is a HUGE amount more resistance than 0 ohms.

Okay, here we go -

We need to change your meter's 2 milliamps to amps so that's .002 Amps

5000 Volts = .002 Amps * 2 Ohms

And now you say HEY WAIT that does not equate!!!

And you are right! Because your meter can't really do what the label says, and we already agreed that we would assume it could push as much through 2 ohms as a dead short. So, what we learn from this is that at 5000 volts, your meter can't push .002 Amps through a 2 ohm load. BUT, let's pretend it can.

So now we need Dr. Joule -

X (resistor watts) = .002 * 5000

and X=10 watts

So for the 2 ohm resistor, you would be right to use a 10 watt resistor.

Now let's see what you need at 100 ohms.

5000 Volts = .002 Amps * 100 Ohms = Wrong again... Right?

To push 5000 Volts through a 100 ohms resistor, your meter would have to generate 50 amps of current. We know that can't happen - its max output is still 2 milliamps.

So we now understand that if the 5000 volts is still our test voltage and 100 ohms is our resistor value - then either the meter is going to put out 50 amps and your decade box and meter and test leads are all going to catch fire, or the meter is going to put out less amperage. And, of course, that's what happens. It puts out less amps as the resistance increases. But by how much???

How do we solve for Watts when we can't determine the Amps unless we can measure the milliamps the meter puts out into the 100 ohm load, and how do we know if we need a 20 watt resistor or a 10, 8, 4, 2, 1, 1/2 or 1/4?

And, we need to know how much current your meter CAN push at the resistance values you want to measure. So, unfortunately, you are now stuck with some choices I can't help you with -

1) Contact the meter manufacturer and get the milliamp ratings at the resistances you want to test, then plug in those amperages at those resistances and calculate your wattage values just like we did in the example.

2) Send meter back for calibration and don't worry about building the box.

3) Test by comparing to another meter and understand that both may be wrong (the man with two watches never knows what time it is).

4) Build your test box using all 10 watt resistors and spend a lot of money for tight tolerance resistors in the higher impedances.

5) Start with a 10 watt 2 ohm, and then a 5 watt 100 ohm, and then a 1 watt 1K and 10K, and then 1/2 watts for the rest... and see if it works or if smoke comes out of something.

Big Giant Disclaimer Inserted Here - don't complain if you get blown up or your meter gets electrocuted!!!

Hope this helps get you started in the right direction.

There are other ways to do what you want to do. I'll wait to see if you are interested in trying something else or putting more effort into one of these before I start making other suggestions. And let's see what the other CR4 gurus come up with. I bet someone has already done this and might give you an example... but I bet you'll have to work for it a little.

I have a couple of points to offer:

1. When checking insulation resistance, you are ordinarily looking for gross resistance changes rather than precise values. Test resistances should be on the high side - say 10's or 100's of megohms or more. Have you ever seen a 10,000 megohm resistor? They are rare and pricey. They are not rated by wattage but rather by voltage withstand capability. A typical 10 gig resistor will be a glass cylinder with an aquadag coating. Standard carbon resistors (note I said "standard") aren't available over 22 megs. This brings us to point number two.

2. A 10 meg resistor will dissipate a 1/2 watt at something over 2200 volts... or will it? Old fashioned carbon composition 0.5 watt resistors were rated, if memory serves, to a maximum voltage of 125 VDC because of their internal construction. A single 1W resistor, used as a high-voltage bleeder, could make a seeing-man blind with the flip of a switch. So we learned to use resistor strings to divide the individual resistor voltages down to manageable levels. Be sure to check on voltage ratings as well as power.

At the low resistance end, do check the specs to see what current your meters will source. A Simpson 260, on the low R scale, would make fuses out of 1N34's all day lomg. As a safety matter, do remember that the IR-meter is generating enough open-circuit voltage to test insulation - and maybe even cause it to fail!

Good luck and be safe.

Quite right mate, the voltage rating of high value resistors is very important.

It it is easily overlooked when calculating high voltage drops, resistor strings is the way to go and carefully testing and correction of such strings, one flash over could lead to a chain reaction.

1. Build a little EXCEL-sheet and evaluate the power of each resistor according to both voltages 500V and 5000V. Then cancel every resistor who`s power is out of imagination because of size and cost.
You will use the low resistors only for resistor meter checks, the mid range resistors with 500V and the high ohm resistors with 5000V in order to keep the power low.

2. Think again about your need of accuracy. Do you really need the information "My isolation resistance is xxx,yy MOhm +- 0.ZZZ%" ??? Normally its enough to know whether the resistance is "more than xxx" to be sure provided that there are no other requirements which force you to be so precise.

3. Getting precise resistor values must not be expensive if you combine, per example, a 1% resistor with the equivalent series resistor which gives you the required value, if necessary.

4. Selecting resistors means to observe the temp-coefficients and the resistor-voltage-dependancy, too.

Just to add to the above comments, remember to check the switch (if you're using one) voltage ratings as well as connecter ratings.

I have built a similar resistance box using good quality GigOhm resistors, if you use a typical rotary switch the contact resistance on the 1 Ohm range is important, for the other end of the range the insulation resistance of the switch is important!!

Most switches can't handle 5kV so try not to use one. Similarly most connectors will not be rated for 5kV, usually for safety reasons. So do check!!

The power rating of the resistors is unimportant, because at 5kV test voltage and a maximum supply of 2mA the maximum power will be dissipated in the 10 MOhm resistor of 2.5 watts.

To get the 5kV voltage rating you will have to buy high (ish) power resistors (larger size).