GOPHERT CPS-3205 power supply - really that bad? [Solved, it's not!]

[igendel]

Hi all,

About a year ago I got this cheap Chinese "GOPHERT CPS-3205" power supply for my hobby prototyping.

Recently I had some odd ADC issues, and I finally had the sense to measure actual voltage from this power supply while under load. It turns out that even in "Constant Voltage" state, the output voltage drops significantly with load, while the display still shows me the voltage I defined. 

For example: I set the supply to 5V, 1A, and connect to a 100W 10R resistor.  Current should be 0.5A give or take, right? But the power supply shows 5V, 412mA, and a multimeter across the resistor shows 4.02V.

So first, a sanity check - this is not supposed to happen, right? Constant Voltage mode means I should get an actual constant voltage?

Second, are there reasonably good benchtop power supplies out there for a relatively modest budget (shipping included)? I'll need something that can take 230V/50Hz.

Thanks,

[TheAmmoniacal]

Don't forget the voltage drop across your leads, you can only set the voltage at the output terminals of the power supply - not at the end of your test leads. Remember that your leads are resistors too. If you instead select CC and set it to 0.5 A, and the voltage to the max, you should get 500 mA through the resistor - do you?

[igendel]

Don't forget the voltage drop across your leads, you can only set the voltage at the output terminals of the power supply - not at the end of your test leads. Remember that your leads are resistors too. If you instead select CC and set it to 0.5 A, and the voltage to the max, you should get 500 mA through the resistor - do you?

But the drop across the leads shouldn't be so high, I think.
When I set the power supply to 10V, 500mA and connect the same resistor as before, the power supply indicates 6.16V, 509mA, and the multimeter across the resistor says 4.98V...

[TheAmmoniacal]

Try to measure the resistance of your leads with a DMM, remember to null out anything else.

[igendel]

Try to measure the resistance of your leads with a DMM, remember to null out anything else.

Hmmm, I think I see where this is going 

I disconnected the wires from the power supply; each has a resistance of ~1.3 Ohms according to the multimeter.
If I have a 10 Ohm resistor and I run 0.5A through it, the voltage drop on these wires combined will be 1.3V, not too far from what I got.

So assuming this is indeed the source of the problem, How do I deal with it? Just try to find wires with less resistance?

[Audioguru]

I think the cheap power supply is "really that bad". ebay junk?
There was one from ebay shown on one of these forums recently, 0-30V at 0-3A for only $6.99 and with free shipping. It looked really nice but it had almost no spec's.

[TheAmmoniacal]

I just did the math myself, got to 1.05 ohm each, but 1.3 ohm is fine. And yes, the source of the problem.

I would recommend buying better leads, or even make your own from stripping mains cables. Some of the leads I get cheap from China are in the 1 ohm range, quality ones are in the 10-50 mOhm range. 

I've owned several Gophert power supplies and I quite like them, I currently use the CPS-1620 quite a lot - and I learned quickly how important good leads are at 20 A! The leads I use measure 16 mOhm.

[Audioguru]

For good voltage regulation at high currents you need 4 wires: 2 heavy wires for the current and 2 thin wires to measure the voltage at the load and feed it to the regulator and display circuit.
The poor quality leads were also probably from China. The leads on my Fluke multimeter have a total length of 8.5 feet and a total resistance of 0.3 ohm.

[TheAmmoniacal]

This is why many lab DC power supplies comes with sense terminals, they will ensure that the voltage you set are across the DUT by compensating for the voltage drop across cables. The more expensive lines of Gophert models have it.

[Cervisia]

I disconnected the wires from the power supply; each has a resistance of ~1.3 Ohms according to the multimeter.

What is that, stranded wire with three strands? (Take it apart; some really cheap Chinese vendors actually do that.)

How do I deal with it? Just try to find wires with less resistance?

Yes. Any wire that actually has a specification for its resistance or its cross-sectional area should be fine.

[TheAmmoniacal]

Those leads just use a lot of steel and aluminium, because copper is expensive

[igendel]

I should say that my multimeter is also Chinese, so all the measurements I mentioned are only so accurate 

My power supply doesn't have sense terminals on the outside (yes, it's ebay junk, but more than $7). As a first step I'll try mains wires - at least I have those at hand. If that doesn't work out I'll buy decent cables. Thank you very much for guiding me in the right way!

[TheAmmoniacal]

I am sure you have a computer cable to spare, those IEC cables are everywhere! Should be 16 A rated and copper. Redo the measurement in your first post and see if that helps.

[igendel]

I am sure you have a computer cable to spare, those IEC cables are everywhere! Should be 16 A rated and copper. Redo the measurement in your first post and see if that helps.

Yes, I have all sorts of power cables lying around  What worries me more is fitting them with the alligator clips and banana plugs. I hope that won't add too much resistance!

When I'll find a replacement, I'll open up the old wires to see what's really inside...

[igendel]

Well, you guys are definitely the best!

I made new leads from a power cable. Even with the alligator and banana plug, my multimeter shows 0.0/0.1 ohms for each.
With the same power supply and the same 10R load resistor, I now get 5V and ~500mA on both Constant Voltage and Constant Current modes - and my multimeter agrees!

Also, as a bonus, the leads are now twice as long as the puny originals, that always bugged me too 

BTW The original leads seem to be stranded copper. They're thin, but not ridiculously thin; I'm not sure what their problem is exactly.

[TheAmmoniacal]

Good to know! I would be very surprised if the Gophert power supply was causing issues, I think it's very well made.

[igendel]

Good to know! I would be very surprised if the Gophert power supply was causing issues, I think it's very well made.

They're cheap, so I immediately assumed the fault was in the supply itself 
I changed the title of this thread accordingly.

[mdijkens]

I Use a Gophert CPS-3010 for a couple of years now. Very satisfied.

But because I use it mainly for higher current use, I am searching for a long time now how to modify it to add sense-wires ...
If anyone runs into a schema or has ideas then I am really interested hearing about them...

[ted572]

Hello:
I posted the GOPHERT CPS-3205E Power Supply Calibration Procedure, Specifications, and general comments on easily improving the LED Display Contrast for any of the GOPHERT Power Supplies.
If interested go to -> https://www.eevblog.com/forum/other-blog-specific/new-2016-gopher-cps-3205c-review-teardown/msg1181084/#msg1181084

[Emo]

Hi,

Modifying for sense input configuration is not that difficult. inside the sense lines are mounted right on the terminal. You only need some safety resistors in case the sense line gets disconnected and perhaps a filter cap. Values can be copied from other supplies.

Eric