Author Topic: Quick-look: GW Instek GPS-3303 Power Supply  (Read 25651 times)

0 Members and 1 Guest are viewing this topic.

Offline bitbangerTopic starter

  • Regular Contributor
  • *
  • Posts: 161
Quick-look: GW Instek GPS-3303 Power Supply
« on: November 01, 2014, 02:47:26 am »
Hi Folks -

I picked this power supply up on eBay for a good price, and just wanted to put some pictures up in case someone else was considering buying one. When I was researching, I couldn't find many first-hand impressions for this model in particular. That said, these along with the MASTEC power supplies have been prolific throughout my career, even school. Some variant of this design always seems to pop up anywhere I've worked. They're cheap, robust, and fall into a good position between programmable and non-current controlled supplies.

This model is a triple output: 5V/3A fixed, and two independent/series/parallel 0-30V/0-3A channels. Full specifications are here:
http://www.gwinstek.com/en/product/productdetail.aspx?pid=38&mid=53&id=131

The layout is pretty standard, but I have to say the displays are great. Very bright, and I appreciate the difference in color between voltage and current.

There's a true hard disconnect switch in front, along with soft control of the supply outputs. Outputs default to OFF when powering the supply on. One really nice feature is when switching between series, parallel, and independent modes, the supply automatically switches the output off before switching. Great to keep you from frying something by accidentally switching modes, or not re-arranging leads before hand.

Also, no need to short the outputs to set your current control: disabling the output shows the current CC setpoint on each display (as seen below):



The fan didn't kick on until I had the supply loaded full out for a few minutes, but it was noisy getting up to speed. I'm not particular about these things, but being a used supply, it may simply need some attention.



The power supply is a nice size (87V for comparison); smaller than it appeared online.
a

Voltage and current readings were spot on throughout both independent ranges, series and parallel. Unfortunately the lack of resolution on the display and lack of granularity on the single-turn control pots makes this supply unsuitable for powering low voltage cores and the like. However for my intended use, it's perfect for supplying regulators, split DC bus rails, etc:



Start-up unloaded and short-circuited was satisfactory. Unloaded:


Shorted:


Nice size multi-tap toroid:


Airflow is directed into the heatsink, around the toroid, and out vents on the two sides, towards the front.


Channel 1 arrangement with rectification, power amps, tap selection relays, etc.


Channel 2, with the addition of a power NPN, presumably for the fixed +5V/3A output. This flying lead arrangement seems a bit lazy, but otherwise the internal construction is on par with the price point this supply hits (a good one, in my opinion).



There are a plethora of adjustments surrounding the front panel PCB, each labeled accordingly. A full adjustment procedure is attached to this post. The supply actually maxes out around 32V/3.3A per channel.







So that's all for now. Hopefully someone will find this of some use, even if just for the pictures. I am still trying to track down a solid schematic and/or service manual, if anyone has any leads!

Best Regards,
James
« Last Edit: November 01, 2014, 04:24:28 am by bitbanger »
 
The following users thanked this post: genghisnico13

Offline Hydrawerk

  • Super Contributor
  • ***
  • Posts: 2598
  • Country: 00
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #1 on: November 01, 2014, 03:01:28 am »
What is the brand of the electrolytic capacitors?
Amazing machines. https://www.youtube.com/user/denha (It is not me...)
 

Offline bitbangerTopic starter

  • Regular Contributor
  • *
  • Posts: 161
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #2 on: November 01, 2014, 03:05:30 am »
Sorry, I meant to mention that above: Jamicon 85C, which have been hit or miss in my (notably limited) experience.
« Last Edit: November 01, 2014, 04:22:57 am by bitbanger »
 

Offline BloodyCactus

  • Frequent Contributor
  • **
  • Posts: 482
  • Country: us
    • Kråketær
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #3 on: November 01, 2014, 01:30:26 pm »
I have the GPS-2303, posted some pics some time back here. I think someone posted a GPS-4303 teardown too.

its been a really great reliable bench psu for me
-- Aussie living in the USA --
 

Offline Hydrawerk

  • Super Contributor
  • ***
  • Posts: 2598
  • Country: 00
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #4 on: November 01, 2014, 04:51:36 pm »
Does this PSU have multi turn potentiometers?
Amazing machines. https://www.youtube.com/user/denha (It is not me...)
 

Offline bitbangerTopic starter

  • Regular Contributor
  • *
  • Posts: 161
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #5 on: November 02, 2014, 05:05:14 am »
Unfortunately not, single turn only. This makes it impossible to dial in to the hundredth of a volt. Even with a multi turn though, I suspect stability would be a limiting factor.
 

Offline Hydrawerk

  • Super Contributor
  • ***
  • Posts: 2598
  • Country: 00
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #6 on: November 02, 2014, 06:36:39 pm »
 :( :(
Amazing machines. https://www.youtube.com/user/denha (It is not me...)
 

Offline TheRevva

  • Regular Contributor
  • *
  • Posts: 87
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #7 on: March 19, 2015, 09:49:21 pm »
Hi guys'n'gals,

If ANYONE here has the schematics for this puppy, I'd be most grateful.

I grabbed mine a GPS-3303 months ago and haven't even used it for anything serious.
Yesterday, I powered it up again and...  NOTHING.
A quick check showed that the mains fuse in the IEC socket (3.15A) was blown - rather violently too.
I grabbed a handful of new fuses and it blew the first one immediately proving that there was, indeed a fault.
I disconnected all the connections from the transformer to the power PCB leaving the transformer connected to the front panel 'control' PCB and it blew another fuse.
(I need to buy a few thousand extra fuses... LOL)
My 'gut instinct' at the moment is that I've got a shorted turn (or two) on the mains transformer which will, unfortunately, turn the whole PSU into a paperweight.  I've not yet tried powering it up with the transformer secondaries totally disconnected.  Lack of fuses...

I'll check out the transformer sometime soon and, if my theory is correct, I'll probably build a totally new PSU inside the existing case.

BTW, I assume y'all know that GW-Instek also make a GPD-3303S with a readily available schematic. www.eltest.hu/pdf/GPD3303SM.pdf
The GPD-3303S a totally different beast in all regards.  (Even the power section is different using an IRFP150 as the series pass regulator on each channel)
 

Offline TheRevva

  • Regular Contributor
  • *
  • Posts: 87
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #8 on: March 19, 2015, 11:38:38 pm »
I just pulled the thing apart (mostly out of interest), and took a QUICK look at the front panel...
It's 2 separate PCBs - 1 for 'control' and 1 for 'display'.
What surprised me was the display board which uses FOUR independent ES5107 ADC / LED drivers (1 each for volts and amps * 2 channels = 4)
(The ES5107 is basically similar to the age-old ICL7107)
I was EXPECTING an arrangement with a small MCU having a few ADC inputs and multiplexing the LEDs on some outputs.  Surely it'd be cheaper to produce even though you might need a PAIR to maintain channel isolation?

The control board is designed to also suit the GPS4303 (i.e. 4 channels with voltage controls on Ch3+4)
The opamps used are (wait for it...) through hole UA741. UGH!!!
It looks to be a totally analog design.  Aside from the ADC chips on the display board, the only 'digital' chip I can see is a 4013.
Ch3 (5V 3A fixed) uses another 741 and an LM723

Assuming the transformer IS 'poked' as I suspect, most of this stuff will be relegated to the bin.
Sure, it did work, but it's not what I'd want if I'm starting again from scratch.
I'd use 4 rotary encoders instead of the existing 4 single turn ports.  (It was never 'nice' to set accurate outputs on them, and I just LIKE rotary encoders)
I'd keep similar buttons for the series / parallel switching and output isolation (to minimise butchering the front panel).
The displays would be dumped (and replaced with at LEAST 4 digits LEDs or perhaps even a pair of TFT LCD panels (eg: Nokia phone)

The existing power PCB uses a pair of power transistors per channel (and one for the fixed 5V/3A channel 3 which uses a 2SD1047).  I've not checked to confirm they're also '1047s, but I'd probably replace 'em with FETs anyway.

The digital stuff would probably be 'loosely' based on the design by Gerry Sweeney but (as mentioned) will use FETs on the outputs.  (I never agreed when he switched from FET to PNP in the power stages.  If HP / Agilent / Keysight [pick a name at random] like them, then so do I!)

It might sound weird, but I'm actually starting to HOPE that the mains toroid is truly dead!
 

Offline TheRevva

  • Regular Contributor
  • *
  • Posts: 87
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #9 on: March 22, 2015, 07:53:14 am »
A quick 'follow up'...
It seems I was premature in the fate of my GPS-3303...  It's alive!!!
The fuses that I'd bought to confirm it was dead were all of the FAST BLOW type.  I now realise that they are supposed to be SLOW BLOW (T3.15A).
The inrush current to the mains torroid on switch on was enough to vaporise them instantly.
However, the slow-blow fuse seems to last JUST long enough to overcome the large inrush current.
So why did it blow in the first place I hear you ask???
My _theory_ is that the fuse is actually under-rated for the initial power on surge and it's heating up quite a bit on every power up.
Whether this is 'boiling off' metal from the fuse wire or just invoking metal fatigue isn't known.
However, it reached the point where it was incapable of 'surviving' that one time and, in the process, blew metal all over the glass.
The same value quick-blow replacements that I tried simply couldn't handle the surge current and immediately 'let out the magic smoke'
Maybe it could do with some form of inrush current limiting?
Needless to say, I've thrown a half dozen new 'spare' fuses inside the case.

The weird part is that I am kinda annoyed that it's fixed.  I suddenly lost some of my motivation to build a replacement PSU.

Footnote:
The beast is also a little more 'forgiving' than I would have expected.
Some dumb idiot (yep, that's me), accidentally connected the channel 3 (Fixed 5V / 3A) output directly to the tranformer winding it's supposed to use as INPUT.
I thought it seemed funny that the fan turned on so early and then noticed a brightly glowing power resistor along with that 'Oooops' smell.
I shut it down and connected it back up the RIGHT way expecting that I had killed Channel 3... Nope, it's still PERFECT. It's still a rock-solid 5.00 volts from no load to the 3A max.
(The Ch3 overload indicator comes on at 2.98A - I'm not going to adjust it!)

Next time I take it apart, I'll be replacing the 4 * single turn pots with ten turn pots.

I guess this means I better give GW a decent 'thumbs up'
 

Offline LaurenceW

  • Frequent Contributor
  • **
  • Posts: 258
  • Country: gb
    • It's Time, Jim, but not as we know it
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #10 on: March 22, 2015, 07:38:17 pm »
One suggestion, Revva. Unless you KNOW you have a particular need for a precision adjustable current source (in additition to the more normal need of a precision adjustable voltage source), then I wouldn't bother with ten-turn pots for the current - it may make the unit HARDER to use.

The logic is this. You probably need to be able to set the power supply accurately to, say, 1.8V, or 2.5V or 3.3V or 5V, etc., but then you'll just "rough set" the current limit to "a bit more than I expect" and switch on. In other words, you'll tend to use the current limiting feature just to LIMIT the output to some safe level, and not to some PRECISE level. If you'd previously had your ten-turn current pot at max and now need to throttle down to 50mA, that's a lot of nobb tiddling to do...

That said... My single-turn current limits on my triple output TTI monster analogue supply are at least LOG in value, so setting lower currents precisely is easier than setting higher currents with equal precision. But in practice, the latter is generally unimportant (to me), and the supply retains maximum usability.

Current displayed to three decimal places, and voltage to two, certainly makes for a nice supply to user - you're not diving for an additional meter every so often, to cross check the power supply's settings!

A 723 linear regulator, eh?  Ah, yes, I remember sitting on my grand-dad's knees when I were a lad, and him telling me all about those...
If you don't measure, you don't get.
 

Offline SeanB

  • Super Contributor
  • ***
  • Posts: 16265
  • Country: za
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #11 on: March 22, 2015, 07:59:00 pm »
Instead of a ten turn pot for current get a 3 turn one. Similar in ability to set precisely but only 3 turns end to end, so faster. They are quite nice if you do not really need 1% resolution, and are happy with 3% instead.
 

Offline Lightages

  • Supporter
  • ****
  • Posts: 4309
  • Country: ca
  • Canadian po
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #12 on: March 23, 2015, 02:04:21 am »
There is probably another benefit to a good 2 turn or 10 turn pot. I modified one channel of my PS with ten turn pots and the settings are much more stable over time and temperature.
 

Offline TheRevva

  • Regular Contributor
  • *
  • Posts: 87
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #13 on: April 28, 2015, 09:20:56 pm »
All of the above are quite valid suggestions (so thanx for them).
I did a little 'scanning' on Digikey et al and noticed that the multi-turn pots aren't exactly cheap and, perhaps most importantly, they simply wouldn't FIT because they're physically somewhat larger.
The existing 2 voltage control pots (1 per channel) measure out at just under 6k (which is a 'weird' value, but could possibly be affected by the in-circuit measurement) and the 2 current control pots measure just under 1k.
They also look to be linear pots (although I've not actually confirmed that).  Swapping them all for corresponding log pots would definitely help setting low values more accurately.
Other options would be to investigate dual-gang, concentric-shaft pots (thereby allowing 'coarse/fine' adjustment).  (There's simply not enough front panel area to mount an additional 4 independent pots and the front panel PCB would be in the way anyway so they'd have to be concentric-shaft units)

On a side-note, there's a new 'fault' that I've just started to notice in the unit.  The PSU has 4 * 3-digit LED displays - Volts + Amps for the two primary channels.  Each of these uses its own LED driver (which look to be ICL7106/7 clones), but the Ch2 voltage display is somewhat erratic on initial power-up and also appears more prone to 'drift' over time.
The actual output voltage of the channel seems to be fairly solid, so I'm guessing that the issue is totally with the display board.  My guess being that I have a 'drifting' reference voltage on that display.  (Since the two channels are 'symmetrical', I should be able to simply swap the V/I display pairs to the opposing channel as an initial confirmation of my theory).

If I had to guess, I'd say that the original design (many many years ago) employed analog meters and the 7-segment displays were simply 'tacked on' as an afterthought.

One of these days, I'm still hoping to complete designing / building my own lab PSU.
I've viewed dozens of different schematics (some commercial, some DIY) and I'm wanting to merge all the 'features' I like from each into a single, cohesive design.
For example, I like the digital control section of the GerrySweeney PSUs and the electronic transformer tap switching inherent in many of the HP / Agilent / Keysight E36xxA PSUs.
At present, I'm leaning towards quite a 'modular' approach by starting with a single channel (with an available totally isolated serial digital interface).
A simple 16x2 line LCD for the display and a couple of rotary encoders for manual V/I setting.
With two (or more) such units in a single chassis, I'd then eliminate the independent LCD / encoders of each channel and control them all from an 'upper level' CPU via the isolated digital interfaces of each channel.
(Don't laugh, but I'm seriously considering an Arduino Due with a 3.5" touchscreen LCD as the upper level controller.  It just works out to be amazingly cost effective and should have more than enough 'grunt'!)
If nothing else, it's been a great 'learning exercise' thus far.
 

Offline Zbig

  • Frequent Contributor
  • **
  • Posts: 927
  • Country: pl
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #14 on: April 28, 2015, 09:51:38 pm »
Wow, that looks surprisingly similar to the Siglent SPD3303 series: https://www.eevblog.com/forum/testgear/siglent-spd3303d-review/msg288251/#msg288251 inside. It seems Siglent took the analog (power) section, even the case, with few modifications and went with totally different front panel and control PCB, added USB, etc.
 

Offline TheRevva

  • Regular Contributor
  • *
  • Posts: 87
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #15 on: April 30, 2015, 12:04:53 am »
Wow, that looks surprisingly similar to the Siglent SPD3303 series: https://www.eevblog.com/forum/testgear/siglent-spd3303d-review/msg288251/#msg288251 inside. It seems Siglent took the analog (power) section, even the case, with few modifications and went with totally different front panel and control PCB, added USB, etc.
It's definitely quite a similar design... But even the power PCB layout on my GW-GPS-3303 differs a bit from the pix Frank shows.
Furthermore, my little unit uses a pair of 2SD1047 NPN transistors for the Ch1 / Ch2 series pass transistor and a single 2SD1047 for Ch3 (fixed 5V / 3A).
(Frank mentioned that the Siglent uses IRFP150N MOSFET devices)
Another thing is that my GPS3303 only uses Jamicon 4700uF/63V caps as the primary Ch1/Ch2 bulk caps versus the Rubycon 6800uF/63V shown in the Siglent.
I've oft heard it mentioned that we need somewhere close to 2000uF per output Amp as a good 'rule of thumb'.
(I guess I should throw a scope on those caps sometime when I am sucking out 3A into a dummy load?)

Still, the devices are so strikingly similar that you have to wonder at the level of incest that was involved.
You'd have to start thinking that there was only ONE way to layout the internals of a PSU?
I am willing to bet that the same 0-6-18-30 windings exist on the xformer for Ch1/Ch2?
 

Offline Zbig

  • Frequent Contributor
  • **
  • Posts: 927
  • Country: pl
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #16 on: April 30, 2015, 08:40:29 pm »
It's definitely quite a similar design... But even the power PCB layout on my GW-GPS-3303 differs a bit from the pix Frank shows.
Furthermore, my little unit uses a pair of 2SD1047 NPN transistors for the Ch1 / Ch2 series pass transistor and a single 2SD1047 for Ch3 (fixed 5V / 3A).
(Frank mentioned that the Siglent uses IRFP150N MOSFET devices)
Another thing is that my GPS3303 only uses Jamicon 4700uF/63V caps as the primary Ch1/Ch2 bulk caps versus the Rubycon 6800uF/63V shown in the Siglent.
I've oft heard it mentioned that we need somewhere close to 2000uF per output Amp as a good 'rule of thumb'.
(I guess I should throw a scope on those caps sometime when I am sucking out 3A into a dummy load?)

Still, the devices are so strikingly similar that you have to wonder at the level of incest that was involved.
You'd have to start thinking that there was only ONE way to layout the internals of a PSU?
I am willing to bet that the same 0-6-18-30 windings exist on the xformer for Ch1/Ch2?

I think that's perfectly reasonable decision to reuse the proven design and not to reinvent the wheel just for the sake of it. Provided none of them just ripped off the other's design, that is (and there's no reason to suspect they have). It just felt a bit weird for me to find this as I spent some time tinkering with my unit (see Frank's thread). Kind of like getting in someone's car and realizing it's basically the same design as yours, only with the little different body and shape of switches on the dashboard, I guess. I believe mine (Siglent SPD3303S) has 4 gears... err, I mean transformer taps ;)
 

Offline jjustin

  • Newbie
  • Posts: 1
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #17 on: May 02, 2015, 11:34:33 pm »
I picked up this same model from ebay but which also has problems.   CH1 basically works fine but CH2 seems to have a voltage regulation problem and the max voltage adjustment does nothing to the output (which ranges from 0-42V max depending on voltage knob position).  Were you ever able to obtain a schematic from GW Instek?  I emailed them and am waiting to hear back. 
 

Offline TheRevva

  • Regular Contributor
  • *
  • Posts: 87
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #18 on: May 04, 2015, 02:46:29 am »
I picked up this same model from ebay but which also has problems.   CH1 basically works fine but CH2 seems to have a voltage regulation problem and the max voltage adjustment does nothing to the output (which ranges from 0-42V max depending on voltage knob position).  Were you ever able to obtain a schematic from GW Instek?  I emailed them and am waiting to hear back.
Sorry, but I don't have a schematic of the GPS-3303.  (The GPD-3303S schematic is available on the net from various sources, but it's quite a different design being a digital programmable PSU).
I'm assuming you're talking about the 'calibration' process described in the 'verification / adjustment manual'? (Specifically, VR301 which is the bottom populated trimmer location on the left side of the front panel of a GPS-3303).
I'm assuming you've verified that the problem is on the front panel PCB (by swapping the relevant Ch1/Ch2 connections on the power PCB).

From what I recall, there's no 'exotic' components on the front panel PCB.  (The closest to 'exotic' are the 4 * ICL710x clones used to drive the LED displays, but they're on a standalone PCB!).  The opamps on mine are all 741 (UGH!)  Analysing your fault should be VERY straightforward since you have a 'clone' right next door to compare readings against.  (The supposedly 'working' channel).
If you think it'll help, I'm willing to replicate any readings on mine as another check

My 'gut instinct' would be to follow Dave's 'first rule of faultfinding'.  Check that the internal supplies are good.  (There looks to be 2 windings from the toroid to drive the front panel.  These go to the plug on the top/center of the front panel PCB).  If the 'control' supply isn't present, then the control system will TRY to derive some control from the POWER circuitry.  Obviously, this is unable to supply the needed voltage ABOVE the + output terminal so the opamps will be saturated in a big hurry (and thus entirely non-linear).  You MIGHT even be able to notice a different brightness between the Ch1 & Ch2 LED displays? In fact, the displays of the 'faulty' channel might almost turn off altogether when that channel is set for MINIMUM voltage output on the front panel (or at least become VERY erratic).
Next up, I'd confirm the Ch2 output with a 'scope to make sure it's not oscillating like mad.  (If so, check the solder joints on the output cap next to the Ch2-minus banana jack).  If you don't have a 'scope, just whack a decent (1000uF/50V or more) capacitor across the Ch2 output terminals.  If that manages to bring VR301 back to life, then it was almost certainly oscillating!
Given that the V-Max trimmer is doing nothing and is closely coupled with the I-Max trimmer, I'd seriously avoid drawing more than a couple hundred miiliamps for now...  It's entirely plausible that it's not setting any I-Max either so you can risk frying the series pass transistors with any decent loads.  (The 42V max output you're seeing is the limit imposed by the rectified transformer output 12+12+6 = 30VAC -> about 42VDC)
 

Offline CCB

  • Supporter
  • ****
  • Posts: 17
  • Country: nz
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #19 on: October 26, 2016, 07:49:25 am »
I've got a couple of these power supplies and they are great. Here is the main board view as it may help someone else. Definitely no room for multi-turn pots..
 
The following users thanked this post: Samba

Offline electricar

  • Regular Contributor
  • *
  • Posts: 89
  • Country: ch
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #20 on: November 29, 2016, 07:20:23 am »
Hey folks,

I got the GPS-4303 and can hear something like a silent transformer humming as soon as I power the unit up (with and without load). I looked at another GPS-4303 at work and can hear the same thing...
Can someone confirm this issue?
 

Offline calin

  • Regular Contributor
  • *
  • Posts: 238
  • Country: us
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #21 on: February 20, 2017, 06:43:36 am »
Sorry to revive this old thread. Does any of you managed to get their hands on the schematic for 3303 or 4303? I have a slightly sick GPS-4303 and I did send a mail to Instek but they answered me with a half arse incomplete schematic for CH2 and a typical "recommend you go to a technician"  :)
 

Offline Bear55

  • Newbie
  • Posts: 2
  • Country: ua
Re: Quick-look: GW Instek GPS-3303 Power Supply
« Reply #22 on: June 05, 2019, 09:49:26 am »

And I have PSU GW 3303C without relay Rl501,502,503,504. What are they doing? I want CH 3 (5 volts) to be adjustable, like the GW 4303C, but I can’t find how it is done with the GW 4303. Maybe someone has this part of the circuit. Please, help.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf