EEVblog 1698 - Awesome Ebay AC Power Source: GW Instek APS-7050

[EEVblog]

I scored this GW Instek APS-7050 Precision AC Power Source on Ebay, let's check it out.

[ballsystemlord]

So if the voltage coming in is 230v, and that's 230×1.414==325.22Vp-p how are the capacitors, which are in parallel and rated for 63v, not blowing up?!?
If they were in series, that would be 189v*2=378v.

[EEVblog]

So if the voltage coming in is 230v, and that's 230×1.414==325.22Vp-p how are the capacitors, which are in parallel and rated for 63v, not blowing up?!?
If they were in series, that would be 189v*2=378v.

The input transformer is a step down.

As an aside, with 2nd thought this is basically a big linear audio power amplifier rather than a traditional switching DC-AC inverter.

[ballsystemlord]

Does it not need a PFC circuit? Or perhaps the incoming transformer works well as a passive PFC?

[Kleinstein]

The output transformer is needed to bring up the voltage from the linear amplifier. The linear transistor amplifiers have a preferred voltage range - too high gets tricky with SOA of the transistors. I would expect the voltage steps and winding switching at the output side, not so much the input side.

There is likely some injection for the fast glitches after the transformer. At least there would be the current measurement and likely also some feedback for the voltage.

The 2nd PCB on top if the heat sink is just for mechanical stability with isolation (the heat sinks are likely at some +-40 V). It is a nice detail.

It does not look like they have PFC. Test instruments and similar may not need it as they are usually not used very often or in high numbers. The more common case is stand-by with low power. It may make sense to have PFC in such a construction as this usually comes with a crude regulation of the raw voltage and get away with a smaller input transformer.

[pdeal]

Very cool instrument Dave!

I know you didn't really reverse engineer it exactly but I guess the typical power supply just turns the transistors on and so they are full on or full off but this one must modulate the transistors in a linear mode to get the desired output? Wouldn't there have to be some sort of realtime feedback of the output? Any idea how that would be done? Seems like for an instrument like this it would need to be a many times per cycle measurement of the output.

I'd like to look at the output of a variable frequency motor drive and see what kind of sin wave output these generate. Seems like these would need to have to modulate the transistors in a partial on state too? Lots of heat I would think.

[ballsystemlord]

<snip>
I'd like to look at the output of a variable frequency motor drive and see what kind of sin wave output these generate. Seems like these would need to have to modulate the transistors in a partial on state too? Lots of heat I would think.

Actually, for a VFD, you'd almost always be using IGBTs.

[Kleinstein]

The APS7050 here very much looks like a classic linear amplifier (not that different from audio) and than the output side transformer to step up the voltage.
So the transistors are not swiched, but in linear mode. The linear amplifier part is what needs the large heat sinks.
I don't know if the do the feedback from before or after the transformer (likely a mix from both sides). The extra transients are quite steep and thus likely added after the transformer (the PCB below the control and thus hidden).

The is some SMPS part, but it looks like this is to power the digital part and output side modulation.

VFDs are usually switched mode and give a lot less clean output. These more correspond to a class D amplifier and the output may looks more like the square pulses with zero in between, but it depends on the model.

[Doctorandus_P]

Th factual errors in this video are really astounding. The big transistors are no FET's but BJT's, and the low voltage electrolytics, while dave is pointing at 115V and more transformer windings with the relays.

It would have been really nice to discuss a more detailed block diagram, and then also do some measurements, for example whether the fault simulations are injected at the "audio amplifier" stage, or in that block after the transformer. But dave yapping about making general assumptions and then 5 minutes later discovering they were wrong and saying something else is a bit annoying. The APS7050 itself is an interesting piece of kit though.

The APS7050 here very much looks like a classic linear amplifier (not that different from audio) and than the output side transformer to step up the voltage.

...

VFDs are usually switched mode and give a lot less clean output. These more correspond to a class D amplifier and the output may looks more like the square pulses with zero in between, but it depends on the model.

Yes.
A brother of mine bought a Chinese Class D audio amplifier and added a mains transformer, and built something very similar to this thing. First he tried it with a 230Vac transformer, but he had saturation problems and the Class-D amp triggered it's overcurrent protection. After that he bought another transformer made for 400Vac, and that apparently worked nicely.

And further. If you want to build something like this. it seems logical to put in an LISN too.

[picburner]

From reading the user manual it appears that the output voltage can be brought up to 600V and the frequency up to 1KHz only via two USB sticks (Opt. APS-003 and APS-004) so the hardware is already set up to operate at these values.

[thm_w]

From reading the user manual it appears that the output voltage can be brought up to 600V and the frequency up to 1KHz only via two USB sticks (Opt. APS-003 and APS-004) so the hardware is already set up to operate at these values.

Question is if the APS-7050E has the hardware built into it as well? I'm guessing not.

[picburner]

I've read the APS-7050E user manual too, and it doesn't mention the possibility of adding these two options.
This is understandable, given that the 7050E is a cut down and more affordable version of the 7050.
Who knows if the reduction is just in the software or if they've also modified (reduced) the hardware.
In particular, the number of windings in the output toroidal transformer.

[pope]

I mostly work on audio gear and I was wondering whether an AC source like the APS-7050E would be suitable as a solution for clean mains supply for the audio equipment.

I’ve used a much cheaper AC source in the past, but it was really only useful for testing the minimum and maximum AC voltage that audio devices could operate at. When it came to providing a clean AC signal though, it was a nightmare. Full of harmonics and essentially unusable as a clean AC source.

[Kleinstein]

The efficiency of the supply is not very good and there is also quite some no load power consumption. So I would consider this not suitable for long term powering anything. It is more something for testing different voltages / frequencies.

Audio equipment should be OK with the normal mains, maybe with some extra passive filtering if there is an inverter near by that causes EMI outside the normal range.

[pdenisowski]

The efficiency of the supply is not very good and there is also quite some no load power consumption. So I would consider this not suitable for long term powering anything. It is more something for testing different voltages / frequencies..

They are very useful as a "clean" source of "standard" AC mains power when running some types of compliance tests, in particular the IEC 61000 series tests.  Most AC mains power (at least where I live in the United States) has an unacceptably high THD* and cannot be used to make valid measurements of the DUT.

When you're doing tests like this (and using analyzers that costs 10s of thousands of USD), efficiency of the supply is not a significant concern 

* "unacceptable" THD means anywhere from 2% down to less than 1%, depending on the test.

[pope]

OK, so it's not a totally stupid idea...

[EEVblog]

I mostly work on audio gear and I was wondering whether an AC source like the APS-7050E would be suitable as a solution for clean mains supply for the audio equipment.

I’ve used a much cheaper AC source in the past, but it was really only useful for testing the minimum and maximum AC voltage that audio devices could operate at. When it came to providing a clean AC signal though, it was a nightmare. Full of harmonics and essentially unusable as a clean AC source.

This one is specced at 0.5% THD to 500Hz into a resistive load and is a linear design which is what you want here. That's about as low as you'll get.
Certainly not something you'd usually use for full time power, but they are very suitable as a "clean" source AC supply.

You can of course get the finest in audio-wankery:



https://stromtank.com/wp-content/uploads/2020/12/STROMTANK_S-1000.pdf
Only $27,000 
https://indiehifi.com.au/products/stromtank-s1000-battery-power-unit

I'd bet that it's not any better than that 0.5% THD in the GW-Instek

[EEVblog]

You can of course get the finest in audio-wankery:
I'd bet that it's not any better than that 0.5% THD in the GW-Instek

Yep, I was right, the $27,000 only gets you 2% THD 
https://www.stereophile.com/content/stromtank-s-1000-computer-controlled-battery-power-source-specifications

[pope]

 

[woody]

As an aside and a bit back to the subject, there is a FW update (3.18) for the 7050/7100 that solves a funny bug I encountered in earlier FW versions where, if I saved and then recalled a setting that had its range set 155V, this recalling blew the 16A breaker in my wiring closet. Repeatable.

GWInstek took a real interest in this problem and solved it, which is nice support in my book.