GW Instek GPD-3303S - slow voltage drop

[hellzakk]

Hi all, when i have inverted the polarity connections of a 12V fan ( who burned with much smoke ) in a circuit the psu have started to disconnect and reconnect continously the output (i think this is because of the short protection) and after about 2/3 seconds i have powered off.
After that i have disconnected the circuit and powered on the psu, without the output active i can correctly select the voltage, but when i set the output on the voltage raise or drop (and change over the time) of some mV. (for example, if i set the CH1 voltage to 16V when i set the ouput on the reading is 15.997V and drop of 1mV after some time).

I tried to recalibrate the unit but the service manual i've found online dont show the correct procedure, so... anyone have an hint on how to resolve this problem?

[JFJ]

... if i set the CH1 voltage to 16V when i set the ouput on the reading is 15.997V ...

That 3mV error appears to be well within the meter specification:

... and drop of 1mV after some time...

Most power supplies require a warm-up period before their meter readings stabilize.

[hellzakk]

Hi JFJ thank you for the answer, yes i know it's within the specs but before this... "accident" ... the reading with the output active are the same as with the output off.
I've checked the output of the psu with my keithley 2000 (after a warm-up time of 1 hour) and the reading on the psu display are pretty the same (difference in the order of 1mV)

 

[Kleinstein]

A different fan could cause higher temperature inside and this more warm up effect.

Another point is a possible damage to the internal supposedly 12 V supply used for the fan. This might also be used for the regulator. I have not checked the diagrams though.

[hellzakk]

Hi Kleinstein, the blowed fan was inside a circuit powered by the psu

[hellzakk]

update, this is the circuit part the psu was conneted to (Scullcom Electronic DC Load http://www.scullcom.uk/design-build-an-electronic-dc-load-part-1/)

DC input was connected to the psu ch1 with 16v 1a output, arduino is used to control the start and stop of the fan (temperature threshold)
After the LM7812CT the +12V is used as input in a LM7805CT for the +5V line

This is the psu power section


And this is (a parto of) the psu control section


Link to the complete service manual (schematics at pages 77-89 )http://www.eltest.hu/pdf/GPD3303SM.pdf

I cant understad why after i switch the output on the voltage readings constantly decrease, it seems that the psu cant keep it at a costant value even without any load... this behavior was not present before the accident and i cant figure out what cause this

[coromonadalix]

If you have 1mv drop at the output  it is very good,  dont forget you have an mcu with DAC's and tons of circuit to do the regulation ...


When the fan busted it may have created some problems on the +12v line, i would have simply changed U1 (lm7812) the D2 diode (1n4007) and the TIP122 transistor.



I honestly don't know why did you think of connecting an DC load on this circuit   Simply checking the parts specs could have told you some things.

Ua7812 lm7812 are normally up to 1.5 amp regulators, 1n4007 are 1 amp diodes, tip122  an 4 amp transistor.

Normally you should not derive power of this circuit other than supply the fan and the power supply circuit, it was calculated that way.  If you add stuff you may add problems too.

[hellzakk]

Hi coromonadalix, the dc load is the circuit i have powered with the PSU, the fan is used for cooling the MOSFET. I got no problem with it.

[Ingo]

Hi,

today I got a brandnew GPD-3303S. First time I swiched it on the fuse at home (not that inside the power supply) failed. I repeated to switch on with the reinstalled (reswitched) fuse the device worked. Now I swithc on and off and the fuse is ok.

Then I found the following effect: I set a voltage at CH1 for exampple 32.000V and the press "OUTPUT", the voltage dropped to be 31.982V. If I set it to 10V it drops to 9.984V after pressing "OUTPUT". The same test at CH2 leads to a drop of 1 ... 3 mV.

The second effect I have is that if I set the maximal current at both channels to 1 mA and the voltage at both channels to 32.000V and no load at both outputs then Uch1=9.7...9.9V (floating) Uch2=12V ... 12.3V (floating) - both LEDs red for current reg. If I increase Imax to 2 mA then after some seconds (to load the electrolyt capacitor at output I think) I get Uch1=31.983V and Uch2=28.8V ... 29.2V (floating) CH1-LED after 4 minutes green, CH2-LED remains red.
Increasing Imax to 3 mA at both channels leads to both LEDs to light green and Uch1=31.982V, Uch2=31.998V

Until now all tests were performed without any load at the output jacks.

Can anyone confirm similar behaviour? Can I change that behaviour by calibration (as found in service manual)?

Regards

Ingo.

[coromonadalix]

Maybe there is a minimum current to have to make the psu switch between constant voltage (cv)  and constant current (cc), i would say 1ma  is very low output,  and without a load the supply may hesitate ?

Some fixed 3 pins regulators ask a minimal current consumption to operate efficiently ?

For the small drop in voltage vs the read values, i would say you may do a recalibration,  but i think youll have to settle with this.  I do think theses psu's with many other models of the same league uses 8 bits mcu's or dac's, you may hae some loss in the designs of the circuits. You can have some discrepancies with the voltmeters or ammeters in some models, even an perfect calibration will have some tolerance to deal with.

Any GPD-3303S owners there to confirm ?

[Ingo]

Thanks for your Answer.

In between I ordered an inrush current limiter, because sometimes when I switch on the PSU the fuse for the mains socket switches off and I sit in the dark. I think the peak current in the moment of "switch on" must to be high - surprisingly for a 250 W PSU.

Ingo.

[coromonadalix]

For an exaample of load and line regulation  my Kikusui Pad 70-5L   an old linear model is 1mv,  yours is supposed to be 3mv,  as i said  added digital circuits / converters  add some loss.   

Just saw you have 16 bits resolutions DAC's, and a Pic 18f85j10 and 16 bits ADC,   pretty good stuff in there, i'm impressed 


In smps when the fuse generally blows up, many transistors in the switching section has gone bad, its a last protection measure.  Still havent found my HY3010E-2 troubles, the main smps section has severly busted ?  lots of precision resistors and mosfets later, still a paper weight.

But yours is based on a toroid transformer a good design choice,  love to see x-former secondary relays switching, to minimize the heat dissipation.

I think you have the service manual, just foun out this ?  : http://www.eltest.hu/pdf/GPD3303SM.pdf

On page 89  we see the input circuit,  i dont see any  EMI/RFI filter ?!    Not sure you could cram one in there easily ?

Would be nice to measure the inrush current when you start your supply,  you could change the fuse for an slowblow model ?

[coromonadalix]

Oh  just saw this here : https://www.eevblog.com/forum/testgear/thoughts-on-gw-instek-3303s-supply/

Kleinstein did have a go  (reply #26 - on the last post) to limit some inrush current,  send him an email to check if it was worth it ??

[Ingo]

Thanks again.
Yes I found the service manual and had a look at the schematic. The relays that switch the transformer to reduce power dissipoation works well but I found that the hysteresis between the switching point wile increasing and drcreasing the output voltage is only 0.6V. This could lead to fast switching realays if constant current mode is set and the load resistance varies. therefore it might be worth to increase that hysteresis to about 2V or to implement a delay that delays the switching of the relays while the output voltage decreases.

I think I will add a relay in series to the mains switch. The contact of the relay that is in series with the mains line should have a parallel resistor of abou 22 Ohm. The relais gets its energy from any of the DC circuits at the secondary side of the transformer. The relais contact shorts the 22 Ohm resistor. This way the initial current is limited.

Ingo.

[coromonadalix]

Or an timed relay / resistor on the main ac, done with an 555 ??

[JFJ]

I think I will add a relay in series to the mains switch. The contact of the relay that is in series with the mains line should have a parallel resistor of abou 22 Ohm. The relais gets its energy from any of the DC circuits at the secondary side of the transformer. The relais contact shorts the 22 Ohm resistor. This way the initial current is limited.

Alternatively, you could just fit a 20 Ohm Inrush Current Limiting (ICL) NTC thermistor in series with the line supply, e.g. something like this:

https://uk.farnell.com/epcos/b57236s0200m000/thermistor-ntc/dp/9751890

[Ingo]

Hi JFJ,

yes this would also solve the problem but the NTC would remain hot while using the power supply and its temperature and resistance would depend on the load at the PSU output.

The PSU is in standby mode every time you switch it on - no output is activated. So it does not need much current at primary input of the transformer in that case. A 22 Ohm resistor would limit the input peak current to lower than 10 A in every case and otherwise allow that all voltages needed to start the PSU will have the right values.
If I then power a relay from a raw 24V line of one of the 15V auxiliary voltages then this relay shorts the resistor. Then not extra heat is produced. I can buy for example a 24V relay that can switch 20A with a coil resistance of 1440 Ohm resulting in only 17mA current (0.4W power). In series to the relay coil I can place a resistor with a big parallel electrolyt capacitor (for start up) to reduce the DC current through the relay to a value that keeps it on. This way the 24V line is not loaded too much and the relay keeps cool.
The relay switches on if the secondary voltage is high enough - so the input peak current of the transformer is gone - there is no additional delay necessary.

Ingo.