-
#75 Reply
Posted by
German_EE
on 10 Dec, 2015 11:21
-
I was always taught that there should be at least two layers of insulation between a 230V conductor and the outside world, a length of heatshrink or spiral tubing would fix the problem here but has this particular requirement now fallen into disuse?
As for the rust, coating the steel after it has been cut and bent to shape would cure this issue. However, the shoddy wiring, the problems with the output connections and the low quality components mean that there is no way I would recommend this product to anybody. Spacing decent quality connectors 19mm apart is not a difficult task.
-
#76 Reply
Posted by
madires
on 10 Dec, 2015 12:07
-
And the shorted turn of the toroidal transformer. The mounting screw doesn't seem to be isolated. Rust, again. The bodged ethernet. Build quality is the same like my Mastech clone. I wouldn't pay more than €300 for that Lab PSU.
if the turn was really shorted you would know, a 300W transformer with a ~0.5V turn shorted, something would be glowing
The magnetic field is mostly whithin the toroidal core and the screw is in the middle of the toroid. My guess is that the magnetic field lines would cancel out each other in the middle of a perfect toroidal transformer. And the field strength in/around the middle should be less than directly around the core anyway. Anyone with a clamp meter and a toroidal transformer laying around? It would be really interesting to know how much power is lost in a shorted turn in the middle of the toroid.
-
#77 Reply
Posted by
FrankBuss
on 10 Dec, 2015 12:23
-
FrankBuss, by thew way, did you see rust and burr when you reviewed the SPD3303D inside or you remember the brand of the main filter capacitors? If somewhat different in metalwork and main components choices those clues could indicate a shift of assembly resources for the X series.
See the images I linked in my first post at the end here:
https://www.eevblog.com/forum/testgear/siglent-spd3303d-review/I think all capacitors were Rubycon. And yes, there was some rust, I commented one image where you can see it clearly.
-
#78 Reply
Posted by
VK3DRB
on 10 Dec, 2015 13:03
-
Good tear down.
Rubycon caps aren't that great. I have seen quite a few dried up Rubycons over the years, especially in TV sets.
One thing you did not pick up was the terrible smudged top overlay printing on the front panel board. Their PCB makers have crap QC. For that sort of money, you would expect a quality PCB, not one where you cannot read the component identifiers because they are smudged.
The screen saver might be just to do what it is supposed to do - prevent an image burning into the screen. It happens on LCD's and especially on OLEDs.
The cheap-arse terminal posts are a big put off. Nothing worse that wires that don't fit or slip off cheapo terminal posts and short elsewhere. I did a linear supply back in the 80's that could provide 4.5 to 15V at 25A continuous. I used real terminal posts... fine threaded brass bolts with suitable wing nuts and a large hole in the bolt side to hold any decent wire on one axis and a small hole at right angles. Worked a treat for any wire or group of wires.
I think the power supply is a little over priced for what you get.
-
#79 Reply
Posted by
Salas
on 10 Dec, 2015 13:07
-
FrankBuss, by thew way, did you see rust and burr when you reviewed the SPD3303D inside or you remember the brand of the main filter capacitors? If somewhat different in metalwork and main components choices those clues could indicate a shift of assembly resources for the X series.
See the images I linked in my first post at the end here:
https://www.eevblog.com/forum/testgear/siglent-spd3303d-review/
I think all capacitors were Rubycon. And yes, there was some rust, I commented one image where you can see it clearly.
Thanks! Detailed pics. I had seen your video before and the circuit mods but somehow I had missed the pics.
-
#80 Reply
Posted by
AF6LJ
on 10 Dec, 2015 13:15
-
I would repair that stuff, and debur the sheet metal.
I would also do a video on it and send a link to siglent as well as post it here.
As far as the wiring goes, putting plastic tubing over the wires isn't the answer, cleaning up the edges is.
There is no excuse for that.
Fair enough, but I'd totally flunk it in a review (yeah, I know, this was a teardown, not a review, but it gave opinions) if you had to open it up, voiding the warranty, and modify it yourself to make it safe and usable. It's supposed to be a power supply, not a power supply kit!
I totally agree, I am working from the position I just bought this and took it apart to inspect it. Something I do quite often when I buy something, almost always if the item is used.
-
#81 Reply
Posted by
AF6LJ
on 10 Dec, 2015 13:46
-
a simple zip tie would have fixed it, just zip tie it to the bundle right next to it and it never get near that sharp edge
That was my thought too - tie it to all that stuff two inches above it, away from the metal.
As far as the wiring goes, putting plastic tubing over the wires isn't the answer, cleaning up the edges is.
There is no excuse for that.
Nope. You could still scissor the wires if you ever took the lid off (or it could even happen on the assembly line...)
Better to move the wires completely away from the metal edges.
My issue with that is dealing with the problem that way is a true band-aid fix. It doesn't address the actual problem; that being the unfinished sheet metal.
A second issue I have with that part of the overall design is the mains leads running along side where screws are used to secure the cover.
A third issue I have with that aspect of the design is the lack of a cable clamp to keep the mains wiring from flapping in the breeze.
The above are what really needs to be done in order to bring that instroment up to what I would consider worthy of it's asking price.
For extra credit; placing the mains wiring in a piece of fire retardant plastic tubing would be nice, but not really necessary, but nice.
The crappy binding posts are a blatant attempt to cut costs and hope nobody will notice.
As for the UI; ;it is obvious that whole UI was borrowed from somewhere else and retrofitted for use on a power supply.
If they let me design it...
No screen saver...
I agree with Dave the unused items on the UI have to go until they are put into use.
A simple color change would suffice for attempting to set the voltage outside of the operating range.
One last issue;
What is the point of having a precision power supply that has adjustable output voltage in one millivolt steps while having one millivolt RMS ripple and noise? This is stupid. How is the resolution meaningful if it is obfuscated by noise??
-
#82 Reply
Posted by
Deathwish
on 10 Dec, 2015 14:53
-
As far as I am concerned the description " fit for purpose " is what matters.
If it is dangerous or potentially so in any way then it is not.
-
#83 Reply
Posted by
Kalvin
on 10 Dec, 2015 15:05
-
One last issue;
What is the point of having a precision power supply that has adjustable output voltage in one millivolt steps while having one millivolt RMS ripple and noise? This is stupid. How is the resolution meaningful if it is obfuscated by noise??
You can change the DC-level of that noise in one millivolt steps.
Edit: Then you just state the measurement voltage was N volts +/- 1 millivolts. For example: The voltage applied was 100 millivolts +/- 1 millivolts.
-
#84 Reply
Posted by
Tek_TDS220
on 10 Dec, 2015 16:27
-
I'm in the market for a power supply, but this one is unacceptable. The cheesy power connectors, the flimsy voltage setting slide switch, the hazardous mains line with the potential to short against the chassis, bad capacitors, hot bridge diode and pass transistor.
The price (~$500 US) is simply crazy for this level of quality. There are so many other options, I don't know why anyone would choose Siglent.
-
#85 Reply
Posted by
Kleinstein
on 10 Dec, 2015 17:49
-
Even if the field in the center if a torroidal transformer is near zero, a turn through the center will give a voltage and possibly large current. The voltage does not depend on the field at the wire, but on the flux flowing through the loop. There a strange kind of non lokal effect in inductance, it's not the magentic field at the wire that cause the voltage.
There are EU regulations for loadfactor for supplies that take more than 75 VA (If I remeber right) - but this applies to the currend drawn under usuall operation. So for audio amplifiers and lab power supplies this takes the no (or low) load current. So usually lab supplies don't have a problem with that. To meet the requirements of the higher power level, full PFC is likely needed. So just a simple linear supply will not work - even with the rather large passive PFC it gets difficult to meat the limits.
Using SCR a preregulator may cause EMI problems and midium frequency (e.g. kHz range) interference which are rather difficult to filter. So a good switched mode supply may be lower noise, as filtering in the 100 kHz range is easier, but still needs some effort. There is also the option to use a kind of class H output stage - so use a BJT of FET to do the tap switching, but this needs a more complicated circuit and more filter caps and diodes.
The trouble I have with the relay tap switching is, that it causes quite some current spikes. So this constant up/down switching with a tricky load can wear out the relays and possibly overload the transformer. So just a 1 V hysteresis looks like rather small. A good supply should even turn off with an error in case of constant tap switching.
If a suppy is oscillation with an active load, this might be the fault of the active load. Especially the constand resistance mode seems to be done in a kind of slow software mode - so here it is likely the load that is not well behaving. Still a valid worst case test for the supply robustness, but not for the quality of regulation.
-
#86 Reply
Posted by
stranger
on 10 Dec, 2015 21:11
-
This is clearly a fail for basic reasons, another most basic problem to add to the pile: The 1mV resolution is pointless before wondering about accuracy of the readings because Siglent have ignored the most basic calculations when they made the pigs breakfast of the terminal design.
This is an incompetent product.
Remember the 1mV resolution, look at the small holes in the binding posts, 12AWG may be the largest wire that goes through the hole. Assume a typical 1m wire length so a voltage drop of about 8mV per amp in each lead, so for 3A a total of 48mV dropped in the supply & ground wires. There is no point in the 1mV resolution of what the PSU thinks it is generating then delivering something VERY different to the customers application. Remote sensing capability is compulsory for accurate PSU, this PSU doesn’t do it , if Siglent want to join the serious players they need to find out what the game is.
The terminal design and spacing are typical of companies that just don’t get it, the lack of remote sensing when providing high resolution as a marketing and pricing feature is hilariously revealing.
-
#87 Reply
Posted by
0b01010011
on 10 Dec, 2015 21:50
-
Yet another issue - the UI.
Timer option that no-one cares about taking 20% of the screen
White on Cyan for editing - are you shitting me
Waveform mode on the stock picture that Dave used on the web item only takes a postage stamp size of the screen, potentially one of the most useful features??
They get almost all the way there on hardware and then piss all over it with crap software.
-
#88 Reply
Posted by
Tek_TDS220
on 10 Dec, 2015 22:00
-
^^You speak the truth, stranger.
-
#89 Reply
Posted by
bitwelder
on 10 Dec, 2015 22:04
-
What's that connector hanging in the breeze towards the back panel (ending a pair of orange wires) ?
Something that is used only in a different PSU model?
-
#90 Reply
Posted by
Macbeth
on 10 Dec, 2015 22:25
-
Yet another issue - the UI.
Timer option that no-one cares about taking 20% of the screen
White on Cyan for editing - are you shitting me
Waveform mode on the stock picture that Dave used on the web item only takes a postage stamp size of the screen, potentially one of the most useful features??
They get almost all the way there on hardware and then piss all over it with crap software.
I wondered why Dave didn't try out the waveform / history modes. Then again this is just a tear down and not a review. So I guess we will get the more in depth try out of it later? Like Dave actually forces himself to use this for a while and reports back?
-
#91 Reply
Posted by
tautech
on 10 Dec, 2015 22:48
-
Yet another issue - the UI.
Timer option that no-one cares about taking 20% of the screen
White on Cyan for editing - are you shitting me
Waveform mode on the stock picture that Dave used on the web item only takes a postage stamp size of the screen, potentially one of the most useful features??
They get almost all the way there on hardware and then piss all over it with crap software.
I wondered why Dave didn't try out the waveform / history modes. Then again this is just a tear down and not a review. So I guess we will get the more in depth try out of it later? Like Dave actually forces himself to use this for a while and reports back?
Dave should, it's incomplete until he does.
-
-
Thanks Dave. Well, I'm not to be buying any Siglent power suppies any time soon, actually probably never.
Those binding posts are crap, real crap, I bet they cost less than 1 Yuan each, or maybe 1 Yuan for the whole lot.
The only nice features were the Analog Devices parts. It's a power supply with precise measurement built in and that's it PERIOD.
As an analogue and RF engineer trying to make low noise measurements I don't need precision in a power supply only good isolation and low noise, same thing really, good common mode isolation is importent when making low noise measurements. It's Kirchhoffs law.
For an analogue bench supply it's a big thumbs down, you lost the plot guys, it's a disgrace. There is no common mode isolation from the mains and the front panel probably has high near field emissions, go wave a magnetic probe at it if you can. Noisy power supply with multimeter built in. Big thumbs down from the anlogue guys.
-
#93 Reply
Posted by
alho
on 11 Dec, 2015 01:25
-
Don't shove the wires between pcb and casing, problem solved. No reason for them being in there. Better to zip tie them to the loom going to input voltage selection board.
Will the rust really case any problems during life span of power supply? Will it rust trough the casing? Anyway I'd assume it would take decades. Not living in humid are so I don't have experience in this.
I bet most of you have electric appliance with sharp sheet edges even on outside surfaces. So don't go look behind your stoves, washing machines, microwave ovens or refrigerators or you might find that you need to replace these poor quality products with more costly ones. I'd like to hear how the sales people laugh in breakroom after one of you refused to buy a stove because it has sharp edges that can get you an owie.
The binding post tough thats just horrible.
-
#94 Reply
Posted by
alter Ratz
on 11 Dec, 2015 01:28
-
As for the rust, coating the steel after it has been cut and bent to shape would cure this issue. However, the shoddy wiring, the problems with the output connections and the low quality components mean that there is no way I would recommend this product to anybody. Spacing decent quality connectors 19mm apart is not a difficult task.
This is not rust, its metal oxide. Must have fallen out from some leaky MOSFET.
-
#95 Reply
Posted by
stranger
on 11 Dec, 2015 10:56
-
Regarding the connector pitch on the PSU, Siglent and others might want to read this. Most instruments with dual banana plug connections follow the well established convention of using a 0.75inch mounting pitch.
Why? As has been suggested it allows dual plugs to be used such as:
http://www.abbatron.com/products/28-dual-instrument-test-adaptersor
http://www.digikey.co.uk/product-search/en/connectors-interconnects/banana-and-tip-connectors-jacks-plugs/1441927?k=dual%20banana%20plugI use these for both multimeter leads and PSU connections when I need to make frequent connections. For meters when doing 4 wire Kelvin measurements it speeds connections and cuts errors. My standard low voltage multimeter leads long ago had their plugs chopped and were permanently wired into a dual plug.
All my PSU and instruments of various USA, UK and Chinese brands meet the 0.75inch pitch standard which may go back 50 years, it becomes very noticeable when something does not, Japanese expression "A nail that sticks out will be hammered".
-
#96 Reply
Posted by
Fungus
on 11 Dec, 2015 11:59
-
Regarding the connector pitch on the PSU, Siglent and others might want to read this. Most instruments with dual banana plug connections follow the well established convention of using a 0.75inch mounting pitch.
All my PSU and instruments of various USA, UK and Chinese brands meet the 0.75inch pitch standard which may go back 50 years, it becomes very noticeable when something does not, Japanese expression "A nail that sticks out will be hammered".
Question is: How can somebody in charge of designing a PSU
NOT know that.
-
#97 Reply
Posted by
Macbeth
on 11 Dec, 2015 12:24
-
The worst of it is all the references to Master and Slave plastered on the front. How on earth do they think they could get any western universities or schools labs kitted out with this blatant racist triggering? Not safe space friendly at all. Even the hard disk manufacturers stopped with that racist stuff back in the IDE days when they renamed Master and Slave to Primary and Secondary. (Where's mojo to back me up on this one?)
-
#98 Reply
Posted by
Rasz
on 11 Dec, 2015 13:42
-
Regarding the connector pitch on the PSU, Siglent and others might want to read this. Most instruments with dual banana plug connections follow the well established convention of using a 0.75inch mounting pitch.
All my PSU and instruments of various USA, UK and Chinese brands meet the 0.75inch pitch standard which may go back 50 years, it becomes very noticeable when something does not, Japanese expression "A nail that sticks out will be hammered".
Question is: How can somebody in charge of designing a PSU NOT know that.
its female interns fault!
-
#99 Reply
Posted by
Kleinstein
on 11 Dec, 2015 15:36
-
At least the supply is consistant: ther terminals are a little to large and a little to far appart.
So maybe there are dual 4.2 mm plugs somewhere in china.
At least for europe the terminals have an additonal problem: they are not compatible with safty plugs (those with a fixed tube). So this supply does seem to comply with EU standards for lab equipment. If we are lucky there is a EU versions with better terminals.