New GW Instek GDS-2000E oscilloscope released

[wraper]

Mine also has the whining noise; the problem is likely poor regulation / feedback stabilisation in the PSU. Maybe I'll hack that some day when the warranty has expired. The beta version in my scope has some fixes for signal display problems. I'll test what you describe with my scope in a few days.

I think it's not psu. Considering what causes the noise to change, I suspect ceramic capacitor(s) singing in the DC/DC converter.

[ali80]

Mine also has the whining noise; the problem is likely poor regulation / feedback stabilisation in the PSU. Maybe I'll hack that some day when the warranty has expired. The beta version in my scope has some fixes for signal display problems. I'll test what you describe with my scope in a few days.

Yeah it is likely in the PSU since when the backlight of the device turns off after some idle time, the noise increases a lot,
I'm guessing it is fixable even by simply adding a power resistor on the supply rail.
Also can you provide me with the beta version of the firmware, or would the final version release anytime soon?

[ali80]

Mine also has the whining noise; the problem is likely poor regulation / feedback stabilisation in the PSU. Maybe I'll hack that some day when the warranty has expired. The beta version in my scope has some fixes for signal display problems. I'll test what you describe with my scope in a few days.

I think it's not psu. Considering what causes the noise to change, I suspect ceramic capacitor(s) singing in the DC/DC converter.

Different sampling rates and memory depth could change the power consumption of the main processor.

[billfernandez]

They really do need a good industrial designer to improve the looks of their scopes. They are just so, bland 

I'm an industrial designer (studying other stuff now, be yes, I am one, did mostly furniture for shops and the shops themselves for over 10 years):
And I think they have a design partner, but the priorities are wrong.
Let me explain:

I'm a user interface architect and I second your comments 

[nctnico]

2. The scope doesn't show the waveform right! how to reproduce it: same settings as above just change time base to 500ns and record length to 100k, and you see the 1khz calibration signal jumping around although the trigger settings are right.

I have not been able to reproduce this but the firmware I have a newer version which also fixes some signal display issues. I'm wary of handing out this new firmware as it is a beta version.

I also investigated the PSU noise a bit further. The noise is coming from the transformer in the power supply. I measured the output on one of the rectifier diodes on the secondary (safe) side and there is a peak around 5kHz. When looking in the time domain I could clearly see the switching frequency jumping all over the place. The power supply is a standard off-the-shelve unit made by Lytec (a Taiwanese company) : http://www.lytec.com.tw/?avada_portfolio=lp6304

[tautech]

I also investigated the PSU noise a bit further. The noise is coming from the transformer in the power supply. I measured the output on one of the rectifier diodes on the secondary (safe) side and there is a peak around 5kHz. When looking in the time domain I could clearly see the switching frequency jumping all over the place. The power supply is a standard off-the-shelve unit made by Lytec (a Taiwanese company) :

Is the PSU shielding insufficient?

[nctnico]

In this case it is noise you can hear with your ears 

[ali80]

2. The scope doesn't show the waveform right! how to reproduce it: same settings as above just change time base to 500ns and record length to 100k, and you see the 1khz calibration signal jumping around although the trigger settings are right.

I have not been able to reproduce this but the firmware I have a newer version which also fixes some signal display issues. I'm wary of handing out this new firmware as it is a beta version.

I also investigated the PSU noise a bit further. The noise is coming from the transformer in the power supply. I measured the output on one of the rectifier diodes on the secondary (safe) side and there is a peak around 5kHz. When looking in the time domain I could clearly see the switching frequency jumping all over the place. The power supply is a standard off-the-shelve unit made by Lytec (a Taiwanese company) : http://www.lytec.com.tw/?avada_portfolio=lp6304

I think in the case of light loads the power supply may be having some kind of pulse skipping feature, anyway having a switching frequency in audible range is anoying.

[wraper]

If it is quasi-resonant topology, there will be no constant frequency without any skipping. Also skipping is used only for really low output power, usually (almost) zero.

[nctnico]

From what I have measured I see the switching frequency varying a lot. It is definitely not skipping pulses so I guess it is a quasy-resonant PSU. Perhaps adding a capacitor on one or more of the outputs helps to even out the load peaks.

[ali80]

From what I have measured I see the switching frequency varying a lot. It is definitely not skipping pulses so I guess it is a quasy-resonant PSU. Perhaps adding a capacitor on one or more of the outputs helps to even out the load peaks.

Can you give us the controller IC's part number or a high resolution photo of the PSU board?

[markone]

From what I have measured I see the switching frequency varying a lot. It is definitely not skipping pulses so I guess it is a quasy-resonant PSU. Perhaps adding a capacitor on one or more of the outputs helps to even out the load peaks.

Hi nctnico, have you notified the problem to GW Instek and seller ?

If you start to play around the PSU you will lose the warranty leaving soldering signs, plus you will risk to frie everything at the first mistake.

I'm not saying that you have not the capabilities to facing this problem, but for the price that you paid you deserve a decent equipment, if the whining noise is at a boring level i think you have the right to ask for a free fix.

[nctnico]

I agree but since this PSU is clearly an off-the-shelve item GwInstek buys from another company it is very unlikely the issue will be fixable by them without designing a new PSU from scratch. I already have complained about the whining noise. It could also be a combination of the scope's hardware and the PSU since it seems GwInstek uses this PSU in all their scopes.

[markone]

I agree but since this PSU is clearly an off-the-shelve item GwInstek buys from another company it is very unlikely the issue will be fixable by them without designing a new PSU from scratch. I already have complained about the whining noise. It could also be a combination of the scope's hardware and the PSU since it seems GwInstek uses this PSU in all their scopes.

Have you figured out how many voltage lines does it provide ?

[nctnico]

It has 6 outputs according to this PDF:
http://www.lytec.com.tw/wp-content/uploads/2014/10/LP6304.pdf

[markone]

30W total power, most spent on 5V lines, all the rest is quite low current, not a nightmare to rebuild from the scratch at worst, maybe with better ripple specs.

Have you already measured the input power consumption ?

The typical cause of whining noise is underloaded PSU output, but 30W is a quite small amount and i would guess that this device pulls at least 20W from wall socket.

If this part was available directly from east at cheap, buy one and make some "modding" could be a way to proceed, keeping the original in case of repair under warranty.

[markone]

If the whining noise is from inductors and not from transformer, you can try to put some glue on them to damp mechanical vibrations, many DSO's PSU have it from factory, dunno if your is without like i saw in Dave's other GW-Instek videos.

[nctnico]

I fixed the whining noise problem! I have put two ordinary 4700uf / 10V capacitors across the - what I think are- 5V and 3.3V outputs and the noise is gone. I also checked the PSU switching frequency using the FFT function and the harmonics around 5kHz are gone. The fix is easy to implement as well (see the pictures):

[markone]

I fixed the whining noise problem! I have put two ordinary 4700uf / 10V capacitors across the - what I think are- 5V and 3.3V outputs and the noise is gone. I also checked the PSU switching frequency using the FFT function and the harmonics around 5kHz are gone. The fix is easy to implement as well (see the pictures):

Hi, so you dared to surgical operate on the PSU, i guess the acustic noise level was quite boring.

The instrument was already out of warranty ?

I can see that you placed them in place of free PCB position for C31 and CN5, is the jumper wire underside the PCB put by yourself ?

Anyway,  why you say "what I think are- 5V and 3.3V outputs" ... i would assume that you measured that lines  before to insert those capacitore polarized caps.

4700uF is a quite invasive value, they will pull high inrush current during the PSU startup, especially if are low ESR type, take care about that.

[wblock]

Panasonic FR series are very good, but very low ESR.  These seem a bit large for the purpose, both physically and capacitance value.  But then, it's hard to put a scope on a power supply when the power supply is out of the scope...

[nctnico]

I fixed the whining noise problem! I have put two ordinary 4700uf / 10V capacitors across the - what I think are- 5V and 3.3V outputs and the noise is gone. I also checked the PSU switching frequency using the FFT function and the harmonics around 5kHz are gone. The fix is easy to implement as well (see the pictures):

Hi, so you dared to surgical operate on the PSU, i guess the acustic noise level was quite boring.

The instrument was already out of warranty ?

My ears are already producing a high pitched sound by themselves; I really can do without the extra. I hope I don't need warranty ever but I'll cross that bridge when I get there. Sometimes I just can't leave things as they are and I fix them. Oh and I did remove the flux before putting it back together!

I can see that you placed them in place of free PCB position for C31 and CN5, is the jumper wire underside the PCB put by yourself ?

Anyway,  why you say "what I think are- 5V and 3.3V outputs" ... i would assume that you measured that lines  before to insert those capacitore polarized caps.

4700uF is a quite invasive value, they will pull high inrush current during the PSU startup, especially if are low ESR type, take care about that.

I put the jumper wire there myself so I could place capacitor into existing pads. For determining which rail is which I looked at the outputs providing the highest current (most pins on the connector) and decided to put the capacitors on those. Maybe smaller values might do as well but since my guess is that the load just varies too much for the PSU to regulate properly I opted for bigger=better. The output capacitors are already in the order of 330uf so a significant amount of extra capacitance is needed to make a real difference. The large capacitors smooth the current and therefore (should) improve regulation. The rush-in current should be handled properly since most PSUs have soft start and current limiting so it should be fine. After all this is a general purpose PSU which should be able to deal with various kinds of loads.

[ali80]

I fixed the whining noise problem! I have put two ordinary 4700uf / 10V capacitors across the - what I think are- 5V and 3.3V outputs and the noise is gone. I also checked the PSU switching frequency using the FFT function and the harmonics around 5kHz are gone. The fix is easy to implement as well (see the pictures):

Nice Fix, Happy to see it and know that an easy workaround is available, forward it to GWINSTEK
I dont know if I should void my warranty, lets see how much more I can keep up with that noise.
I'm still waiting for that firmware update, since the beta is out, hope that it would be sometime soon.

[markone]

I fixed the whining noise problem! I have put two ordinary 4700uf / 10V capacitors across the - what I think are- 5V and 3.3V outputs and the noise is gone. I also checked the PSU switching frequency using the FFT function and the harmonics around 5kHz are gone. The fix is easy to implement as well (see the pictures):

Nice Fix, Happy to see it and know that an easy workaround is available, forward it to GWINSTEK
I dont know if I should void my warranty, lets see how much more I can keep up with that noise.
I'm still waiting for that firmware update, since the beta is out, hope that it would be sometime soon.

There is a warranty sticker on the scope's cover ?

If not, you can build a caps filter wired board with flat cable connectore and put it in series between the PSU and the scope mainboard, so you can reverse the mod in case of warranty repair.
A bit of work, but would worth if the noise is actually boring for you (usually they are a lot...).

You can consider to use a PI C-L-C filters to lower a lot the cap values and reduce space and PSU startup stress.

[nctnico]

I doubt filters will work because the problem is obviously loop stability in the PSU. The in-between-cable could be an option but you'll have to find a good mounting point for it. In case of issues it isn't hard to put the PSU back into it's original state but I doubt the service techs will even notice the extra capacitors. It would be interesting if someone experimented with the values and see if only the 5V needs extra capacitance or only (what I think is) the 3.3V line.

[wraper]

IMO this could be solved without any big caps. Feedback loop could be modified a bit, for example, by adding some small capacitor somewhere before the optocoupler.