MCP42100

[metal]

Hi,

I am trying to find how much current can pass through the wiper of this potentiometer? I can't find the clue from the datasheet, unless there is sth I did not uderstand :S

[UncleBog]

The output voltage at the wiper is set by pull up and pull down resistors in the IC, if you load the output you are putting additional resistances in parallel with the voltage setting resistors and will therefore change the output voltage. You must connect the wiper to a very high impedance input such as an Op-Amp so that any parasitic resistance is negligible compared to the output resistance of the pot'

You know the IC internal resistor values that you are setting so you can calculate how much the output voltage will change (and how much current will flow) when connecting a particular external resistance.

[pickit2]

Data sheet page 14, 4-2 says not to exceed 1mA

[UncleBog]

The +/- 1 mA wiper current limit is documented in the DC characteristics but this refers to rheostat mode. As metal specified potentiometer I think that he's interested in potentiometer mode and it's likely that he'll have a significant output voltage error long before he sinks or sources 1 mA.

[flyback]

In my opinion, it should be in rheostat mode for Metal's application. In potentiometer mode, even with the lowest R device (10k), the current flowing in the ePot will never exceed Vdd_max/R_min = 7V/8k = 0.875mA.
br

[metal]

Actually, I am trying to run NSL-32 LDRs. I tried proteus simulation and thought that things will be just fine for the attached schematic.. When I assembled the circuit, I discovered that MCP42100 potentiometer wiper lowest output is not 0v at digital value 0, which caused a lot of troubles; the current passing through the LED of the NSL-32 was 100uA at digital value 0 which resulted in LDR resistance of 2K, this is not acceptable at all.

Two days ago, I tried connecting the NSL-32 directly to MCP42100 in potentiometer mode through 180R resistor, it works OK, but the max voltage at the wiper is 3.5V  at digital value 255 and the max current passing through the NSL-32 LED is 9.6mA. The resistance of the LDR is approx. 65R, but I am unable to find out whether I am killing the MCP42100 or not, it has been working for tow days already, and I can increase/decrease the volume without any issues. What frightens me the most is that the max wiper output is 3.5V instead of 5V. MCP supply is 5V, of course, along with the upper potentiometers pins as well.

Any clues on how I can solve the problem, what other solutions I can use, all but PWM, because audio freaks will not accept 62.5KHz PWM injected into the LDR from the LED!

[crunx]

The root cause of you problem in other words is that there is a small offset voltage at output of the digital potentiometer?

That might be easiest compensated by applying a small positive bias to the op-amps negative input through a large resistor and a trimmer potentiometer. You should a fairly stable voltage to the "upper" end of that trimpot, and the other end is  grounded.

As your negative feedback resistor seems to be 10k, for example a resistor of 1M and voltage of 12 V over a trimmer potentiometer (of say around 100k - value is not ultra critical) would give compensation of offset from zero  up to 120 mV

Of course, if the digipot's offset is fairly constant, you might just use a fixed bias instead of the trimmer pot.

[metal]

Thanks for the help. It turned out that PBx is to be connected to GND, while PAx has to be connected to Vcc!
I was connecting them the other way around. While I was reading about shutdown pin, datasheet says that PAx is disconnected and PBx is shorted, it came to my mind that PAx is connected in a different way other than PBx, I then looked deeper, and found figure 2-25 on page 11 which supported my thoughts while reading, I reversed the connections, the lowest output is now 10 times smaller, 3.8mV!

[motox]

 “Resistor terminals A, B and W have no restrictions on polarity with respect to each other.” But in your case you have to connect B to ground because:
-   In shutdown mode this insures zero current in the current source.
-   “Code 00h effectively connects the wiper to the B terminal” this insures zero current in the current source.

Code FFh does not connect W directly to A! There is one resistive element between.
There are 256 resistive elements. This results in 257 wiper connection possibilities. Because the digital codes are limited to 256, the top connection (i.e. W to A) is not physically implemented. That’s why you couldn’t get 0V output when A was at ground potential.

You say that the max wiper output is 3.5V when the A pin is at 5V with code FFh? This should not happen. The data sheet says that the maximum voltage at A, B and W pins is Vdd, which is 5.5V. Maybe you are overloading the wiper pin?...

I see a problem in your schematic. I know is only a simulation, but I notice that you are powering the MCP6021 with 12V. The maximum is 5.5V. Or you use a 12V opamp or you change the voltage to 5V. But the last option brings another problem. The maximum voltage in the sensor resistor (270 Ohm) will never be 5V! Also R4 is decreasing even more the maximum current of your current source. If you remove R4 (it's not doing anything useful there) the maximum voltage at R1 should be around: 5-0.7=4.3V.

Maybe when you say 3.5V maximum at wiper voltage is actually the maximum voltage at R1?...

[metal]

motox, read my last two posts again please :--/ when I said 3.5V, I wrote that I was connecting the LEDs of NSL-32 directly to MCP42100 through 180R limiting resistors and the current was 9.6mA. Then I returned to the original schematic and noticed a problem caused by me not paying enough attention to the datasheets: PAx was connected to GND, I got 23.8mV, then I looked again at the datasheets and corrected that as I stated in my last post. Actually things were not really solved at this point.. I kept getting strange behavior, read below, this is an email I sent to my colleague yesterday, this was two hours later from my last post here:

"LDR behavior is correct now. There were two problems that I found and solved in my prototype:
1. PBx should be connected to GND, while PAx pins to +5V, my prototype had these connection reversed.. resulting in 23.3~23.8mV when the POT is at value = 0.
2. I noticed that LM358 is not working as expected by the voltage measurements across the 270R resistors that I did this evening. So, I replaced it with an OPA2134 as this is what I had available for the moment. I don’t understand LM358’s opamps behavior; while both opamps inputs having the same exact potential level, they suffer 27mV difference at their outputs.. this was causing the whole project to fail, indeed.. This problem was a hidden one, and appeared after I fixed point 1.
3. I removed 1K5 resistors  between transistors' bases and GND, this did not make any differences, but I can’t see why they are soldered in the first place... May be we decided on sth, but I am unable to remember what it was.."
 

My friend replied that he doesn't think that the OPAMP is the cause of this problem, anyway, it is working for me now. What I understand not at the moment, why did LM358 behave this way..!

I still have another problem

My colleague ran a test with a normal potentiometer, and the lowest value he gets across both LDRs is 7K, but when using MCP42100 and OPAMP circuitry, he gets 300R, he says that this will ruin the signal source... Now, I have to solve this by code, I am wondering how to do it :--/ Very simple project I thought, it turned out to be a real PITA for me, being away for a very long time at work.

[motox]

Don’t connect the LEDs directly to the MCP42100, you could destroy them. Also the excessive heat will create nonlinearities to the pot. response.

First let me see if I understand correctly.
1:
Your system has two channels and you measure a difference of 27mV in opamps outputs between channels with LM358? This offset disappears with OPA2134?

Remember that OPA2134 is not suitable for this application because the common-mode input voltage is out of range. The minimum DC voltage that you can put in their inputs is (V-)+2.5 V.

2:
You want to limit the minimum LDR resistance to 7k. So you have two options to limit the maximum current in the LEDs: increase R1 or limit the maximum code of the digipot. The last solution decreases the system resolution.