GW Instek GPP-4323

[Martin72]

Hi,

GPP-4323

Four outputs, fairly low ripple, attractive exterior/clean display...
Are there any reasons not to buy this power supply?
I haven't found a showstopper yet and would like to replace it with four separate power supplies for space reasons.

Martin

[KungFuJosh]

That model is well regarded. Mike likes it, and it's discussed here too: https://www.eevblog.com/forum/testgear/is-there-a-better-power-supply-option-for-$700/msg5257575/#msg5257575

[dexar]

I have had mine for about four years and I am pleased with it.
Quiet and really small but still easy to operate. The interface is well laid out.

A nice touch is the optional load on one of the ports even though I have never used it.

As far as I know there are only one negative thing with it and that is the numbering of the outputs.
Look at the image of it and you will understand what I mean. Why the did that is a mystery for me.

It also has recessed binding posts. This may be a minus or a plus for some users.

[Antonio90]

I've been thinking about buying it for a few years, but I already have another GW Instek PSU more than good enough with a few QOL  problems. I might get to sell a few things and buy that one sometime.
Also, you might already know, but LAN is optional and installed at the factory, if RS232/USB is not enough for your use case.
This thread might be of interest: https://www.eevblog.com/forum/projects/firmware-mod-for-gwinstek-gpp-series-lab-bench-power-supply

There is another thread complaining about the current readback being slow and only showing the instantaneous ADC value, and not the average between samples. (Not useful due to rather slow speed).

[Antonio90]

I have had mine for about four years and I am pleased with it.
Quiet and really small but still easy to operate. The interface is well laid out.

A nice touch is the optional load on one of the ports even though I have never used it.

As far as I know there are only one negative thing with it and that is the numbering of the outputs.
Look at the image of it and you will understand what I mean. Why the did that is a mystery for me.

It also has recessed binding posts. This may be a minus or a plus for some users.

It seems they put the two full-spec channels around the mid-point, as seen on the two-channel version.
Maybe the additional channels are an afterthought in design, or there are mechanical/thermal reasons for the placement. The GPD-4303S has the same layout.

[nctnico]

It also has recessed binding posts. This may be a minus or a plus for some users.

AFAIK there are two version available. One with recessed banana sockets and one with regular binding posts which also accept a wire.

[Martin72]

Or you can use these for wires:
Wago Cage clamp

They are simply good, up to 2.5mm diameter.

[Lexy]

I have this model for ~8 months and is a good supply. The reason for me to buy this is it's low ripple, 4 channels and small formfactor.

There are however some things i don't like. The UI is a little to small for my old eyes. The most irritating thing are the (lack of) binding posts. Because they are recessed, the annotations and red/black colors for plus/minus are not clearly visible and i find my self trying to find an angle to see them.

But for the price it's good.

PS.: The PC Software is a piece of crap..

[dexar]

It seems they put the two full-spec channels around the mid-point, as seen on the two-channel version.
Maybe the additional channels are an afterthought in design, or there are mechanical/thermal reasons for the placement. The GPD-4303S has the same layout.

Even if channels 3 and 4 is an afterthought when it comes to the electronics inside
is it such a big job to just rename them on the GUI and on the labels?

[dexar]

It also has recessed binding posts. This may be a minus or a plus for some users.

AFAIK there are two version available. One with recessed banana sockets and one with regular binding posts which also accept a wire.

I think the model for the US market is with regular posts and the recessed for the European market but I may be wrong.

[Domitronic]

We have 3 of them at work. 4323L and 3323L variants. I have used both models and like them. But i have to say that i did not use them for anything special. Just the regular 24V, 5V and so on which pretty much any power supply could do. My colleague used the load function as well for debugging of a small buck converter. He liked the device as well since he could use one channel as power supply and the other channel as load.

But i would recommend to get the variant with LAN interface. Just in case you want to remote control it or do some automation in the future.

https://eleshop.de/gw-instek-gpp-4323-programmable-power-supply-lan.html

[Caliaxy]

I have been using one for almost four years now. A few minor questionable interface choices (to my taste…) but other than that really nice.

I like the feature that allows you connect the two main power channels in series or parallel internally, to extend the voltage or the current range. When you do so, Ch. 1 becomes the master - both channels are controlled by what you enter in Ch. 1 (Ch. 2 is grayed out). If you need 42V, for instance, you have to set the output voltage to 21V, which is displayed on each channel. If you need to increase the output to 43.2V you have to set the output voltage to 21.6V. Not sure if other manufacturers do it the same way, but this doesn’t feel quite natural to me. Basic math though, you get used to…

It perfectly meets its specs, so you don’t expect a great current read-back accuracy at low currents (mA to tens of mA). You might get a read-back current of a few mA even if the actual current is 0. On my copy, one of the main channels reads 3.3mA, the other 2.2mA even if nothing is connected to the output. Again, perfectly within specs (you have to pay much more for better low-current read-back accuracy…), but I have cheaper power supplies that show 0mA when nothing is connected to the output.

It doesn’t quite replace four separate power supplies, as Ch. 3 and Ch. 4 are limited to 5V and 15V, respectively (both max 1A) - but fully controllable.

I used the load capability much more frequently than I expected. It is fully featured (CV, CC, CR), limited to max 3A. Very nice.

I got mine with the LAN option (paid extra $25 at the time), which appeared as an extra-item on the order (as if it was a separate part). However, I believe it must be installed in the factory, you cannot buy the part alone and install it yourself at a later time.

Mine came with binding posts (like all the other power supplies bought in US).

All in all, I like it.

[Martin72]

Hi,

It doesn’t quite replace four separate power supplies, as Ch. 3 and Ch. 4 are limited to 5V and 15V, respectively (both max 1A) - but fully controllable.

This.
Most power supply units have these outputs, so this is a nice bonus.
I also have a few other power supplies up to 0-60V.
My decision has been made, the GWInstek will be purchased soon.

[Fab6657]

Hello,

Firmware 1.23 is available since few days. I was looking about changelog description but was not able to find any information.
Anyone know something about the changes in this new release for the GPP-4323?

Thank a lot

[Hexley]

Firmware 1.23 is available since few days.

Found a bug in FW_GPP-X323_V1_23 firmware for the GPP-4323 which I downloaded and installed today.

Sequence to reproduce:
SYSTEM
VERSION, then RETURN
VERSION, then RETURN (repeat for a second time)
RETURN

The system now becomes totally unresponsive. No buttons work. The beeper emits a continuous sound. The GPP-4323 must be powered off and back on to recover.

This has been reported to Instek.

[Eric-H]

There is another bug in firmware version 1.23 on my GPP-4323. Channels 3 and 4 show no voltage and current setpoints when selected. When you turn these channels on, they do output the last voltage.

I reverted to version 1.21

[Fab6657]

Hey Community,

As far I bought this power supply 2 weeks ago, I am struggling with the function powerup voltage and current settings.
My understanding is that the powerup parameter with last settings recall the defined voltage and current per channel.

But on my equipment, voltage and current always set at 0V and 0A at power up.
Is it my mistake to misunderstand this option, is it a bug (in 1.22 and 1.23 same), or it is something else?

Thank a lot for help.

[Fab6657]

Another question related to firmware, GW Instek provides only latest firmware actually (V1.23)
Does a place exists where I can download older version like V1.21 (which seems have less bugs than latest one)?

Thank

[Eric-H]

I saw that firmware version 1.23 is no longer available on the gwinstek site. They went back to version 1.22. However, this version also has a serious flaw.

When you use the display mode that only shows the settings of one channel (mode 4) the voltage and current setpoint of channel 3 and 4 are shown incorrectly. When you first select channel 1 or 2 and then switch to channel 3 or 4, the setpoints of the previous channel are displayed! When you than activate channel 3 of 4, the output voltage does not match the displayed setpoint. E.g. the actual value shows 5V while the setpoint shows 3V (if that was the setpoint of the previous channel).

In the display mode that shows all channels, the values are shown correctly for all channels. But I prefer using the mode that shows one channel because of the larger font.

So I will continue using firmware version 1.21, that works correctly.

[Fab6657]

The downgrade is recent, it is from today (I am checking mostly all days )

Right now I have not find version 1.21

Could someone share ?

Thank a lot,

[Eric-H]

Here is firmware version 1.21

[Fab6657]

Thank a lot Eric

[jwise]

One thing I really wish the supply's firmware would let you do is put the CH1/CH2 relay to parallel while one is configured as a load and the other is configured as a source.  This is particularly useful for using it as a 2Q power supply; you can do it with external jumper wires, but it would be more convenient to do it without.

I spent a long while reverse engineering mine, and I have a pretty good understanding of the internal architecture of it, and a decent understanding of how some of the GPIOs are wired up on the application core (a lpc177x / lpc178x).  If anyone wanted to pick that up, I dumped a whole bunch of state into a github repo here and then kind of lost interest: https://github.com/jwise/gpp-hax

The highlight so far is that I got it to do the two standard stupid tricks for a device with a microcontroller, a display, a USB port, and a piezo:





I guess they weren't kidding when they advertised it as a 'programmable DC power supply'.