Discussion in 'General Electronics Chat' started by sudharshan.panduranga, Nov 22, 2010.

1. sudharshan.panduranga Thread Starter Active Member

Apr 10, 2008
61
1
Hello,
I am planning to do the programmable decade resistance box.
Specification:
0.1Ω to 1MΩ.
In steps of 1%.
Current range 1A.

My doubt is how can implement the variable resistance value in steps of 1% using microcontroller. because 1A i have to achieve.

Regards,
Sudharshana

2. thatoneguy AAC Fanatic!

Feb 19, 2009
6,357
718
There are solutions using JFETs as variable resistors, but the current doesn't reach 1A.

With power being $I^2 \cdot R$, extremely high power resistors will be needed for the 100 -500 ohm range at typical voltages.

3. sudharshan.panduranga Thread Starter Active Member

Apr 10, 2008
61
1
Can you guide me how to use JFET?

4. thatoneguy AAC Fanatic!

Feb 19, 2009
6,357
718
Here is a datasheet

That is essentially the issue, other than current. There are some ICs by MAXIM that give a wide range resistance values based on data input as well.

Apr 10, 2008
61
1

6. tom66 Senior Member

May 9, 2009
2,613
213
Probably best to limit it to watts instead of current, say 0.25W or 0.5W. Then you could probably get away with a 11x5 grid, which could be configured to put resistors in series to get the required value. Each column would have one of each decade - so it would be 10 ohms for one, 100 ohms for the other, and so on. 10 resistors are wired in series letting you select between 1 and 9x multiplication. You then need to wire 5 of these in series again to get 1 Mohms.

It's far easier to use switches.

7. russ_hensel Well-Known Member

Jan 11, 2009
818
47
So 1 amp at 1k is 1kw, get a fan. Really what power are you looking for?

8. tom66 Senior Member

May 9, 2009
2,613
213
Trying to dissipate 1kW in small accurate resistors? I wish the best of luck to the OP.

9. beenthere Retired Moderator

Apr 20, 2004
15,815
282
Gonna be a big box. We had several 1Ω 100 watt resistors. Each was better than a kilo in heft.

10. gootee Senior Member

Apr 24, 2007
447
50
1-Ohm 100 Watt resistors come in MUCH smaller packages than a kilo. The larger values would be the killers, anyway, considering power would be current squared times resistance. His 1Meg resistor, with 1 Amp, would have to dissipate one megawatt! Just so it doesn't get too terribly hot, I'd use a 2-megawatt-rated resistor, there. <grin> Also, the one million volts across the resistor might be somewhat problematical.

11. thatoneguy AAC Fanatic!

Feb 19, 2009
6,357
718
Exactly what somebody needs when doing load testing on a hydroelectric dam.

12. Wendy Moderator

Mar 24, 2008
20,764
2,534
Ok course, instead of a µC, you could use 4 switches per range using basic binary concepts (1, 2, 4, 8) and have 7 ranges.

$10^{-1}, 10^{0}, 10^{1}, 10^{2}, 10^{3}, 10^{4}, 10^{5}$

This will use 28 switches and resistors total, but I think you will have trouble with the really small ranges, since a piece of wire can be 0.1Ω.

Two 5% resistors in parallel can make a 1% resistor using a DVM to measure them.

13. retched AAC Fanatic!

Dec 5, 2009
5,201
312
Trouble with small ranges.

Well, when it comes to test equipment, you have to make a decision on what range will be the most accurate.

If your .1 to 1 ohm range is 5% and the rest of the box is 1%, then that is still quite accurate to real life wiring losses. (ish)

14. sudharshan.panduranga Thread Starter Active Member

Apr 10, 2008
61
1
This is regarding DRB. At 0.1Ω load 1A current should flow. This is maximum current in my project. But my question is, how can increase the resolution steps for resistance value.
In measuring instrument company websites I saw, the resolution is 1mΩ. How is this possible to provide this much of resolution. My intention is to know the design concept behind this........

15. sudharshan.panduranga Thread Starter Active Member

Apr 10, 2008
61
1
I need to flow 1A @ 0.1Ω. But in Measuring instruments company websites, they are giving 1mΩ resolutions. My intention is know the concepts behind to give these much of resolutions.
You can see, www.meatest.cz. In that M602/602A.

Regards,
Sudharshana

Dec 5, 2009
5,201
312
17. Wendy Moderator

Mar 24, 2008
20,764
2,534
It sounds good, but the act if connecting wire invalidates a lot of it. I guess banana jacks are out. There are very few connectors that will be under several millohms, not counting wire. Add wire and that goes out the window.

There are applications where 0.1Ω counts, but they are very few and far between, and most of them involve high current.

Of course, if you want to talk measurement, that is a different ballgame. Then you go with a 4 wire setup. The difference between creating a precision 0.1Ω resistance in a box and 10Ω is so great I have to wonder what the OP is thinking. Some of my metrology experience speaking I guess.

Of course, you could make a 4 wire setup in the box for the 0.1Ω and 1Ω ranges. It would be a waste for the higher resistances though. Trying to do such low resistances through two wires is a lot more challenging that I suspect is given credit for.

You would need to focus on the low ends I mentioned, that is where your accuracy will be challenged. The rest is so easy as to be worth putting in a separate box, not to mention extremely simple.

http://www.powerstream.com/Wire_Size.htm

18. sudharshan.panduranga Thread Starter Active Member

Apr 10, 2008
61
1
I did not understand 4wire concept.

Dec 5, 2009
5,201
312
20. Teri Member

Apr 3, 2009
12
0
Hi -
This looks like a fun project but your specifications are not clear.
0.1Ω to 1MΩ. ... ok, but easier if you can reduce the range to .1 to 999999.9.

In steps of 1%. ... unclear -- 1% of .1 ohm = .001 ohm. Do we need .001 ohm steps?
Current range 1A. ... not possible in the real world. 1A through a 1M ohm resistor = 1000000 volts and 1000000 watts of power -- it's just not pratical.
I have some suggestions if you can reduce the specification to practical values:
.1 ohm to 999999.9 ohms
steps .1 ohm.
1 watt (see below if you really do need the 1A specification)

Then I have two suggestions:
The first is to implement the decades as physical decades. That is, each set of resistors can be arranged to provide a one decade range". You will need these 6 sets of 4 resistors, each 1 watt, with each of the six sets connected to a 4-bit port.
.1 1 10 100 1000 100000
.2 2 20 200 2000 200000
.3 3 30 300 3000 300000
.4 4 40 400 4000 400000 ohms
There are other ways to form decades, but it always takes a minimum of 4 resistors per decade.
You will also need 24 relays, and driver and computer interface circuitry. The software will not be complex -- just calculate (or check a look-up table) which relays need to be set at each port to the proper decade to form the desired resistance.

Second option:
The resistors can be arranged as a single binary progression connected to a single 24-bit port, again with the needed relays, and driver and interface circuitry. Software is even simpler -- just set the port to a binary number equal to the desired resistance. The 24 resistors needed to reach 999999.9 ohms are:
.1 .2 .4 .8 1.6 3.2 6.4 12.8 25.6 51.2 102.4 204.8 409.6 819.2 1638.4 3276.8 6553.6 13107.2 26214.4 52428.8 104857.6 209715.2 419430.4 838860.8 ohms.
Actually, the range will will be zero to 1677721.5 ohms in .1 ohm steps.

The hard part is making the resistors because it's unlikely you can buy them.

Precision resistors are hard to find and expensive. I would suggest buying several used decade boxes from ebay just for the resistors. You will need several because much test equipment sold there is from schools where students have overloaded or otherwise abused it.

An afterthought...
I am trying to imagine why you want to do this when a pre-made switched decade box would be so much more simple. So, I suspect it is for some kind of an automatic wheatstone bridge and that is why you specified the "1 amp". If that is the case, then simply determine the wattage requirement of each resistor based on the maximum voltage that will be applied to the bridge. Only the mid-range resistors will need to have larger power ratings.

Another afterthought...
If in fact, you do plan to automatically adjust a precision bridge ciruit, there are other ways to do it. For example you can use a stepper motor to turn a precision potentiometer. The number of pulses to the motor will be directly proportional to the resistance of the potentiometer. Use the bridge null amplifier to drive the stepper motor routine.

A third afterthought...
Again if you want to use relay-driven decades to null a bridge circuit, the relays will be very noisy -- you will need to include a binary search algorithm in the software to minimize the wear and noise on the relays. The binary progression resistor method will converge on the correct value faster than the decade resistor method.

Have fun and good luck with your project,
Teri