EEVBlog #497 - Siglent SDG5000 Function Generator Teardown

[SeanB]

I can see the reason for this chassis being like it is. The die used to punch it out of the sheet makes both the inner case and the central power supply divider out of the same sheet, complete with cutouts and the broaches for the cable ways. Then a second machine finishes the broaches to make non sharp edges so as not to cut cable insulation, and does the bends to make the case assembly. A third die then finishes the bends, so that the case is basically finished. Sharp edges are a given with new dies, so somebody went over them with a file to dull them, though this would have been better with a different method after welding. Then they were welded most likely with a MIG welder, and then a spot of protective paint was applied over the welds. Better would have been to have welded, then tumbled to smooth them then wash, acid etch and electrogalvanise them in a bath to provide a complete cover that would not rust easily. Then the stand off posts were placed and riveted in in a press.

Good way to do a case though, as it uses the outer thicker case to provide rigidity, all the internal frame has to do is to keep the boards in line and transfer the forces to the outer shell.

[Rasz]

Better would have been to have welded, then tumbled to smooth them then wash, acid etch and electrogalvanise them in a bath to provide a complete cover that would not rust easily. Then the stand off posts were placed and riveted in in a press.

but that would cost them an equivalent of a big mac, cant do that! Just like washing soldered PCBs before shipping costs something like 0.2 cents so they stopped doing it and almost all Chinese products look inside like someone fished them out of a piss pond.

[GoosHänchen]

Differential DAC output converted to single-ended by the Intersil EL5166 then entering the relay selected reconstruction filter (Elliptical for sin output and Bessel for everything else I would guess). 

Looks like they left room on the PCB for increasing the order of the elliptical filter (currently populated is 7th order) in case it was needed or maybe they use a 9th order for the high frequency models.

Interesting design choice as opposed to filtering deferentially. 
Definitely costs less and doesn't seem to hurt their distortion specs.

[ByteMe]

Just to sum up the whole rust point. They obviously used a painted sheet of metal and did punching and folding directly onto that. If you ever want to play with the big boys, EVER, then just don't do any surface coating before all the mechanical forming was done. I don't care if the case is stainless steel, aluminium, magnesium, galvanized or even spray-painted. It just doesn't make any sense to form cases from surface treated metal. And Siglent is just embarrassing themselves by not reacting on this issue! Microchip has shown how to win the EEVBLOG's flying pig award, Siglent better put in some effort too!

[Rufus]

If you ever want to play with the big boys, EVER, then just don't do any surface coating before all the mechanical forming was done.

From what I have seen and read it is common to punch and form steel sheet with no further processing provided the sheet has some form of coating offering cathodic protection (galvanising or zinc loaded paint).

I have had a cheap GW Instek generator for several years which is of similar construction. The sheet is a bit thicker with screwed not welded fastenings. I looked at it carefully and the punched edges are bare steel. In one place there was a slight hint of discolouration which might have been rust otherwise it was fine.

Plasma cutting (if it was plasma cut) is supposed to leave a more resilient surface than shearing, but, I wonder about the risk of burning away the protective coating.

The problem is probably crap protective coating (cutting the proportion of zinc because it is expensive and it doesn't look any different would be very typical Chinese cost saving mentality) and poor finishing where they manually ground off burrs from blunt punches or plasma recast and ground off more of the crap protective coating.

Bottom line is it shouldn't rust in the future and it sure as hell shouldn't be rusty out of the box. Siglent quality control is crap.   

[Rasz]

From what I have seen and read it is common to punch and form steel sheet with no further processing provided the sheet has some form of coating offering cathodic protection (galvanising or zinc loaded paint).

Yes, most if not all cheap computer cases are like that, but they ARE NOT SHIPPED PRE RUSTED, whats more they dont even rust after few years.

[EEVblog]

UPDATE:
I have been informed that Siglent took the rusting issue seriously and have retooled their process to fix it. They are sending me the new production model so I can show the differences. 

[IvoS]

Nice. My SDG5082 is on its way. I will have a peek inside as soon as I get it to see how the chassis looks like these days.

[Smokey]

TheSignalPath did a tear down of a current production unit and they really cleaned up the internals since the EEVBlog review teardown.

[IvoS]

A few weeks old unit SDG5082 I purchased from Tequipment.  Rust, rust.....,  left me with rusty taste in my mouth. Another thing I noticed was pinched wire against the chassis. So, I am kind of glad I opened it up and caught it. This pinched wire could have the insulation deteriorating with the heat and one day it would leave me with blank screen.
I don't think I would be buying anything from Siglent any time soon.