Relay controlled potentiometer

Hi

I am attempting to construct an logarithmic potentiometer using relays and resistors. The goal is to use these circuits to control practically anything with a pot and store the setting for later use. I have successfully constructed an linear one using an article in EDN-magazine as a base. However, a logarithmic one turned out to be more difficult. I used the EDN design as a base for this as well but added some resistors in a T-network (?) fashion. This seems to work but i got some "bumps" when i plotted the graph. Does anyone have a suggestion to what's causing this? I have used Rtot/2^n to calculate the values for the resistors.

EDN article

Schematic of the logarithmic pot:


Graph of the result. y: Voltage x: Step (binary)


Best Regards,
Hampus

Just noticed that the value for R17 was missing an "K" this didn't remove the bumps though.

You might want to look in this extensive thread for some ideas, particularly posts by Mikkel C. Simonsen (mcs):

http://www.groupdiy.com/index.php?topic=655.0

burdij said:

You might want to look in this extensive thread for some ideas, particularly posts by Mikkel C. Simonsen (mcs):

Click to expand...

I have only worked on linear "pots", that can be made log or antilog by selecting the correct steps. But this may not offer steps that are small enough. I have also made attenuators with fixed dB steps, but they are attenuators not pots...

Best regards,

Mikkel C. Simonsen

Hi Mikkel

I sent you an e-mail a couple of hours ago regarding this. So you solved the problem in software then? Can you give me an example of the number of steps you get from, say a six relay configuration? You mention a friend i the other thread, has he made any progress with his pot?

I have now gotten to know the circuit a little better and have fixed most of the bumps in the earlier graph (it is actually a lot smoother now). I also noticed that you can change the curvature of the pot by putting a resistor instead of a link between the middle resistors and one of the terminals on the pot (at the positive side of the battery in the schematic). By decreasing the resistance you get a more linear curve. Now I have fitted an 6.25K resistor here witch smoothen the curve. I will post a graph when I get the time. However, the overall resistance of the pot is now 96k instead of 100k. A 4% error is... okey i guess but i aim for perfection. I think tweaking of this resistor can be fruitful. My biggest problem now is the fact that these graphs take a long time to plot as i have to feed everything from Proteus to Excel manually and do all the stepping by hand. Anyone now a workaround on this?

Analog Devices has this

It would be interesting to know what you intend to do with this pot; someone may come with another idea. The big question is: do you really want a potentiometer or an attenuator?

Dusty Circuit said:

I have now gotten to know the circuit a little better and have fixed most of the bumps in the earlier graph (it is actually a lot smoother now). I also noticed that you can change the curvature of the pot by putting a resistor instead of a link between the middle resistors and one of the terminals on the pot (at the positive side of the battery in the schematic). By decreasing the resistance you get a more linear curve. Now I have fitted an 6.25K resistor here witch smoothen the curve. I will post a graph when I get the time. However, the overall resistance of the pot is now 96k instead of 100k. A 4% error is... okey i guess but i aim for perfection. I think tweaking of this resistor can be fruitful. My biggest problem now is the fact that these graphs take a long time to plot as i have to feed everything from Proteus to Excel manually and do all the stepping by hand. Anyone now a workaround on this?

Click to expand...

It is indeed a known fact that one can "twist" a potentiometer's law by adding a resistor between the wiper and one of the extremities, but it is not seen as a constant resistance anymore.

iomegaman: Thank you for the suggestion but I aim for a purely discrete path.

abbey road d enfer: The applications for this type of pot is quite wide. I want to use it to control EQ:s and basically anything with a pot in it. It is not to be confused with an audio attenuator this i an equivalent of pot (if it works) .

hi dusty
are you proceeding with relay controlled pot?
pcb
parts ecc....
???
cheers
6T9R

Hi!

Progress is very slow right now. I have made some test on board but I run in to problems with varying impedances and all sort of problems. I have however successfully constructed an attenuator with variable curves and no "bumps". This needs to be buffered though since impedance varies with steps. I guess it could be used in a pan-circuit perhaps. It uses two resistors per n-step. I will post some more information about that and a PCB for the linear pot soon.

Best Regards
Hampus