0
\$\begingroup\$

I would like to replace a 500k rheostat with three fixed value resistors. Can I achieve this with the help of 3 sugar cube relays like I've shown in the picture? or is it bad idea?

enter image description here

Those relays are all powered on individually with a separate mcu interface so that I can select which resistor value I want.

\$\endgroup\$
2
\$\begingroup\$

You sure can do it that way. Normally you'd use the powers-of-two for the resistor values. In this case 100-200-400. This will give you 7 equally spaced conductance values. (conductance is 1/ resistance).

Because resistances of 200 are uncommon and 400 are even harder-to-get, especially if you have only three values, you can create such a 1-2-4 sequence from 5 resistors of the same value, for the lowest value, two in parallel, then one single and then two in series.

If you want not the conductance to be linear but the resistance, you would place the resistors with parallel relays all in series. So your 100 Ohm resistor would contribute 100 (relay open) or zero (relay closed) to the total resistance.

So with 100-200-400, you'd be able to make 0 (short all relays), 100, 200, 300, 400, 500, 600, and 700 ohms.

With the given circuit and given resistor values, you get a bit of a weird mixture of values. I've done the math for you and with your circuit you can make: "infinite", 400, 250, 153, 100, 80, 71, and 61 ohms.

Use 200 instead of 250 and you get 400, 200, 133, 100, 80, 67 and 57.

You have the choice: do you want to deal with the edge-case of zero resistance or do you prefer the infinite case.... In both cases you can avoid it with an extra resistor that is not switched.

If you want to aim for 500k max value, theory would dictate that you aim for 143k as the middle resistor value. Depending on your application using 120k as the middle resistor might work too.

\$\endgroup\$
  • \$\begingroup\$ 75/150/300 gives a set of 1:2:4 using preferred (E24) values. \$\endgroup\$ – Brian Drummond Aug 27 '17 at 15:36
  • \$\begingroup\$ Thanks for the reply, I forgot to mention that only one relay will be on at a time so that I just need 3 different resistance value. Also the resistance value mentioned here were just for example purpose, in real project they will be replaced with available values closest to them. \$\endgroup\$ – Kokachi Aug 27 '17 at 15:36
  • \$\begingroup\$ If your 3 different values are a "happy combination" then you might be able to save one relay by using the "two on" combination.... When moving from one value to another one relay will close before the other releases or the other way around. Think about what your circuit will do in that case. If for example you're making a three-possible-voltages DCDC converter, you definitively do not want to have the intermediate "all open" situation. \$\endgroup\$ – user242579 Aug 27 '17 at 16:30
  • \$\begingroup\$ Almost 2 years later, I notice that your diagram has values 100, 250k and 400k.... Probably a typo.... \$\endgroup\$ – rew Mar 31 at 9:46
0
\$\begingroup\$

In complement to user242579's answer:

Since the resistors are very high value, the current will be very low. You must avoid any kind of relays that are designed for power applications, as these have a "minimum contact current" which is required to punch through the oxide layer on the contacts. Without current, these types of contacts will have a random resistance value when closed. This does not matter if you are switching a motor, for example. The tiny spark at contact will clean everything up nicely.

Here you must use "signal" relays like OMRON G6K, which are sealed (thus no contact contamination) and use a contact configuration which is guaranteed to provide a specified contact resistance even without any current.

Also these relays must never be allowed to exceed the datasheet current/voltage spec, and especially not arc over or spark, which would vaporize the silver/gold contact plating...

Also with your high resistors, mind the parasitic capacitance between coils and contacts. This will interfere with your bandwidth, unless it's all DC, in this case no problem. But make sure the coil voltage is clean, because it will couple a little bit into the contacts via capacitance.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.