0
\$\begingroup\$

I have one of these power supplies:

enter image description here

I also have an MCR-4V data logger:

enter image description here

If I connect wires directly from the power supply to the data logger (red to red, black to black), then the voltage output on the power supply closely matches the measured voltage, e.g. a 10.00V shows consumption of 0.000A (i.e. < 0.0005A, very low resistance), and measured voltage is about 9.95V.

Now I also have a homemade hot plate. Hooking this up directly to the power supply, at a constant 10.00V it draws about 2.000A (hence 20W total).

However if I now connect my measurement wires to measure the voltage between the black and red terminals, the device no longer reads 9.95V but rather 9.70V.

This is well outside the stated error margins of the measurement device (it claims ~0.3%, or 30 mV).

I'm thinking the most likely culprit is that at a higher load, the power supply's display is lying. Is that indeed the likeliest possibility? What other sources of error can I look for? I would tend to trust the measuring device as it's much more expensive and also appears to be high quality. But what I trust least of all is my circuit building skills!

Update

  1. If I connect the datalogger directly at the power supply terminals, I get the 9.96V reading.

  2. If I connect the datalogger to the end of the alligator clips coming out of the power supply leads, I get the same ~9.96V reading:

enter image description here

To me this indicates the extremely low resistance of the circuit up to this point.

  1. Now if I connect my hot plate wires later on the alligator clip, keeping all else the same, the reading on the data logger now drops to 9.688V:

enter image description here

  1. I get this same ~9.67V reading if I try the following connector options too:

a. Join the data logger and hot plate wire, clip in the middle:

enter image description here

b. Clip the alligator clip to the end of the hot plate wire, and wrap the data logger wire a bit before:

enter image description here

What puzzles me is why the sole act of hooking up the hot plate wires is changing the voltage here. The resistance of the measuring circuit is very small. As I understand it, this forms a parallel circuit -- one going to the data logger, and another going to the hot plate -- where each branch should have the same voltage as the whole circuit, namely 9.96V as measured at the power supply leads.

\$\endgroup\$
2
  • 3
    \$\begingroup\$ You need to show your precise set-up including wire lengths, wire types and exactly where you connected the logger. \$\endgroup\$
    – Andy aka
    Commented Mar 27 at 15:20
  • \$\begingroup\$ @Andyaka: I've updated the question \$\endgroup\$
    – DaleCloudman
    Commented Mar 27 at 23:22

2 Answers 2

Reset to default
1
\$\begingroup\$

If you measure at the power supply terminals, you will read the same voltage on both instruments. If you measure at the hot plate, you will measure lower than what the power supply displays. This is because voltage is dropped across the resistance of the conductors connecting the power supply to the hot plate due to the 2A load.

While measuring the voltage at the hot plate, try sweeping the voltage down toward zero and you will see the displays will tend to "converge", meaning, the error is proportional to the load current.

High-end power supplies (especially high-current versions) mitigate this problem with what is known as Kelvin connections in which high-impedance "sense" leads monitor the voltage at the load instead of at its own local output terminals. This puts the error, due to the lead resistance, into the control loop and the power supply regulates accordantly by compensating for the lead loss.

\$\endgroup\$
6
  • \$\begingroup\$ I believe I was measuring at the terminals but getting a different result. I'll double check! \$\endgroup\$
    – DaleCloudman
    Commented Mar 27 at 16:31
  • \$\begingroup\$ @Claudiu The theory is still the same - there are leads inside the power supply that connect the user-accessible terminals to the actual power conversion stage of the power supply. The power supply display is sensing the voltage on the PCB instead of at the actual terminals. So there will still be a little bit of error that's not accounted for. It's done this way because it's easy to design and manufacture. And that residual error is usually negligible and not important for a benchtop supply - it's not intended to be a precision voltage source. It's more of a "close enough" PS. \$\endgroup\$
    – MOSFET
    Commented Mar 27 at 16:38
  • \$\begingroup\$ Exactly as you described - 9.96V at the terminal, 9.6V at the hot plate, and the error is proportional to the load! Amazing. If I connect just the ends of the alligator clips from the power supply to themselves, I get 0.43 volts at the max 5.2 ampere it supplies, which means 0.083 ohms resistance. At full load, (0.083 ohm * 3.145 amperes) gives my missing 0.26 volts!! Thank you! \$\endgroup\$
    – DaleCloudman
    Commented Mar 27 at 23:35
  • 1
    \$\begingroup\$ @Claudiu Glad to help. A few points: You are very fortunate because the numbers don't usually work out that perfectly with these power supplies. There usually another error on the display called "offset error". Your not seeing this because the technician at the factory did a damn good job trimming that out. Almost all of my supplies OCV never match precisely. Another thing, the readings can drift when the power supply heats up (especially with "analog" supplies with linear pots on the front. Quite normal. Always measure with an accurate DMM at the load if the volage/current is critical. \$\endgroup\$
    – MOSFET
    Commented Mar 28 at 3:45
  • 1
    \$\begingroup\$ @Claudiu Wait until you have your experiment perfectly set up, you leave it alone for a few hours, come back to the lab, and you notice the voltage has changed. No one else was in the lab. It will be then, when discuss long-term stability and drift... \$\endgroup\$
    – MOSFET
    Commented Mar 28 at 3:48
2
\$\begingroup\$

I'm guessing your data logger is connected at the load. If that's the case, you are probably seeing the resistive drop of the leads from the power supply to the load based on the current you're drawing.

Some power supplies allow you to mitigate that by running additional sense lines from the load back to the supply. (I don't see any connections for sense leads on your supply, unless they're on the back.) This allows the supply to accurately measure the load voltage (since the sense lines conduct very little current) and compensate its regulation to adjust the voltage at the load rather than at the supply.

Without such a feedback mechanism, the only place the supply has to measure its output is at the supply terminals.

\$\endgroup\$
2
  • \$\begingroup\$ Forgive me if this is an inept question, but if that were the case (ie no feedback mechanism), does that mean the power supply would be thinking it's pushing 10.00V, but really it's only pushing 9.70V, and the measuring device is getting the correct actual voltage? And does it mean the power supply's listed current may also be wrong? \$\endgroup\$
    – DaleCloudman
    Commented Mar 27 at 16:33
  • 1
    \$\begingroup\$ It means that the supply is pushing 10V into its load, which includes the connecting wires. 0.3V of that is dissipated by the current running through the resistance of the wires (and connectors), and that leaves only 9.7V for the hotplate. \$\endgroup\$
    – Cristobol Polychronopolis
    Commented Mar 27 at 16:35

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