Waiting for parts to test, time will tell.

Hi Sensacell,

So while you are waiting for parts, could you please a schematic of what you are planning to prototype?

Thanks,

Neko

 

If you have not already done so, you might want to stop by Power Integration's website. As far as I know they are one of the leaders in compact safety compliant power supplies. There you will find some helpful application notes, including one that gives advice on winding compliant transformers.

https://www.power.com/

 

Hi Sensacell,

So while you are waiting for parts, could you please a schematic of what you are planning to prototype?

Thanks,

Neko

Here is what I am testing- soon.

230 VAC / 50 Hz. input.

Output 5V and 3.3 V around 100 milliwatts total.

The first stage creates a 22 Volt input for the DC-DC module.
The DC-DC converter supplies only functional isolation, not SAFETY isolation.
The DC-DC converter output is really poorly regulated, I might need to add a wasteful ghost load resistor to the output to keep the output in check.

 

If you have not already done so, you might want to stop by Power Integration's website. As far as I know they are one of the leaders in compact safety compliant power supplies. There you will find some helpful application notes, including one that gives advice on winding compliant transformers.

https://www.power.com/

Hi DickCappels,

Thanks for the tip on Power Integrations.
They do have a several offline AC/DC converter chips. This one was of particular interest:



The chip is available in an SMD-8 package- nice and compact. The circuit in Figure 1 is similar to Danko's post of Viper06 in a simplified form. The question that I still have though is whether these simplified circuits are viable as is since no component values are shown.

Link to the datasheet: https://ac-dc.power.com/sites/defau.../data_sheet/topswitch-hx_family_datasheet.pdf

Any thoughts or comments?

Neko

 

Here is what I am testing- soon.

230 VAC / 50 Hz. input.

Output 5V and 3.3 V around 100 milliwatts total.

The first stage creates a 22 Volt input for the DC-DC module.
The DC-DC converter supplies only functional isolation, not SAFETY isolation.
The DC-DC converter output is really poorly regulated, I might need to add a wasteful ghost load resistor to the output to keep the output in check.

Hi Sensacell,

Here is what I am testing- soon.

230 VAC / 50 Hz. input.

Output 5V and 3.3 V around 100 milliwatts total.

The first stage creates a 22 Volt input for the DC-DC module.
The DC-DC converter supplies only functional isolation, not SAFETY isolation.
The DC-DC converter output is really poorly regulated, I might need to add a wasteful ghost load resistor to the output to keep the output in check.

Hi Sensacell,

Thanks for posting the additional info and schematic. Most DC-DC converter provide upwards of 1kV isolation. Wouldn't that be considered safety isolation? My experience with these tiny DC-DC converters is that they are poorly regulated and require a ghost resistor to keep the output from rising too high. Depending on the max input for your 3.3V regulator, you could always put a zener clamp on the DC-DC converter output.

Neko

 

Hi Sensacell,

Hi Sensacell,

Thanks for posting the additional info and schematic. Most DC-DC converter provide upwards of 1kV isolation. Wouldn't that be considered safety isolation? My experience with these tiny DC-DC converters is that they are poorly regulated and require a ghost resistor to keep the output from rising too high. Depending on the max input for your 3.3V regulator, you could always put a zener clamp on the DC-DC converter output.

Neko

It's a regulatory thing.

It's gotta meet certain creepage, clearance, and internal design criteria for it to be considered "Safety Isolation"
For example, the input and output pins on the CRE1S2405SC package are only 0.1" apart- that disqualifies it right there.

My design is also about achieving the minimum size, the since the SR086 concept uses the IGFET to switch-out the high voltage portion of the AC waveform, it avoids needing to rectify and filter the mains- and that giant 400V filter capacitor. All the downstream parts are low voltage, hopefully, I can condense this thing down smaller than a switchmode/transformer design.

 

The circuits in the Power Integrations data sheets work but the component values depend on the application. You can get these values using a calculator or following the examples in their application notes.

 

It's a regulatory thing.

It's gotta meet certain creepage, clearance, and internal design criteria for it to be considered "Safety Isolation"
For example, the input and output pins on the CRE1S2405SC package are only 0.1" apart- that disqualifies it right there.

My design is also about achieving the minimum size, the since the SR086 concept uses the IGFET to switch-out the high voltage portion of the AC waveform, it avoids needing to rectify and filter the mains- and that giant 400V filter capacitor. All the downstream parts are low voltage, hopefully, I can condense this thing down smaller than a switchmode/transformer design.

Got it. I worked in the medical device industry my whole career until I retired. Went through many rounds with UL and TUV. The STGD5NB also has a 0.1" spacing in the D-Pak- what am I missing? Neko

 

My application requires only functional isolation, which the DC-DC module provides. The IGFET plays no role in isolating anything, it just acts as the power switch for the regulator.

Here are some photos of my quickie test circuit.
It draws current only around the zero crossings- then switches off when the voltage rises over the Vout threshold. (terrible Power Factor of course)

One terribly annoying thing about this part is the fact that you cannot power it through an isolation transformer- the inductance of the transformer causes massive voltage transients that will kill the input rectifiers or the IGFET quickly. This is pointed out in the datasheet - and verified by yours truly as I blew my first test board up trying it.

After replacing EVERY PART on the board, it works fine on direct mains voltage, but I dare not try much more than using my isolated DMM and current probe on it to measure anything.

All the parts on the board are quite small, I can bake this down to a nice small board, but I have concerns about how finicky it is about the input source characteristics.

 

My application requires only functional isolation, which the DC-DC module provides. The IGFET plays no role in isolating anything, it just acts as the power switch for the regulator.

Here are some photos of my quickie test circuit.
It draws current only around the zero crossings- then switches off when the voltage rises over the Vout threshold. (terrible Power Factor of course)

One terribly annoying thing about this part is the fact that you cannot power it through an isolation transformer- the inductance of the transformer causes massive voltage transients that will kill the input rectifiers or the IGFET quickly. This is pointed out in the datasheet - and verified by yours truly as I blew my first test board up trying it.

After replacing EVERY PART on the board, it works fine on direct mains voltage, but I dare not try much more than using my isolated DMM and current probe on it to measure anything.

All the parts on the board are quite small, I can bake this down to a nice small board, but I have concerns about how finicky it is about the input source characteristics.

Sensacell,

Lots of copper on your protoboard! If you want to bring the line up slowly, you might consider a VARIAC. Staco makes some smaller units that can be found on eBay.

Am not sure what you scope photos are showing? When I do line measurements with a scope, I don't use the ground clip since the scope is already earthed.

Neko

 

Ahh my bad- the waveforms show the AC line current as measured with my Tek TCP202A current probe.
It draws rather nasty spikes of current over a minuscule conduction angle, peaking at 2A. (measured with no load- the idle current of the DC to DC module)

I might try softening these with some series resistance, but this will reduce efficiency.

I have a Variac and an Isolation transformer, I can use neither on this circuit. The inductance of these devices banged with these current pulses produces destructive transient voltages.
The datasheet for the SR087 admonishes to avoid this pitfall in the fine print.


I use the techniques shown in this video to make my protos:

 

Ahh my bad- the waveforms show the AC line current as measured with my Tek TCP202A current probe.
It draws rather nasty spikes of current over a minuscule conduction angle, peaking at 2A. (measured with no load- the idle current of the DC to DC module)

I might try softening these with some series resistance, but this will reduce efficiency.

I have a Variac and an Isolation transformer, I can use neither on this circuit. The inductance of these devices banged with these current pulses produces destructive transient voltages.
The datasheet for the SR087 admonishes to avoid this pitfall in the fine print.


I use the techniques shown in this video to make my protos:

OK, so those were current probe photos- got it. That video on prototyping is great. Over the years, I have used all those techniques and then some. Thanks, Neko

 

OK, so those were current probe photos- got it. That video on prototyping is great. Over the years, I have used all those techniques and then some. Thanks, Neko

Just received these today- smallest AC/DC converter that I've seen yet, 5V, 700mA. Will test today.

https://www.ebay.com/itm/192612896363

 

Just received these today- smallest AC/DC converter that I've seen yet, 5V, 700mA. Will test today.

https://www.ebay.com/itm/192612896363

Did a load regulation test today. This little beast is looking good. Although rated at 700mA, it got past 800mA until the foldback current limit was reached at over 800mA. Data from the testing are shown in the table below:

AC/DC unit 1	Line 122.9 VAC
Load (mA)	Vout (DC)
0	4.859
10	4.802
50	4.773
100	4.788
200	4.79
300	4.783
400	4.789
500	4.801
600	4.801
700	4.82
800	4.81
> 800	3.5

 

Wow! and at $2.00 each, I got nothing to say!
Wonder how EMI nasty it is?

Does it seem to have any filtering?

 

Wow! and at $2.00 each, I got nothing to say!
Wonder how EMI nasty it is?

Does it seem to have any filtering?

Hi Sensacell.

Yah, crazy at two bucks a pop! I don't have a HiPot tester but > 40M Ω on ohm meter between L/N and DC output. I have a new scope for which I haven't explored the FFT function yet. If this is functional on this scope, I'll see what I can find out about EMI. Are you needing to meet IEC60601 standards? CISPR11?

Can always add a LC lowpass filter on the Line side to help EMI. Honestly don't even know if this is a buck converter?? So few components.

I'd like to reverse engineer this petite module. It may be relatively easy (or not) to identify components. The PCB looks 4 layer, I may be able to get an x-ray of the board to see any inner traces. Done this before.

The tough part is the magnetics. I could dissect the transformer and get (maybe) turns and construction info from the bobbin. A former research colleague claims he can characterize the core and at very least identify the μi of the core.

Combined with a schematic, it would get us on track to a copy-cat of this micro design.

What do you think?

Neko

 

It looks pretty nifty. If I was going to use it I would take a 1K resistor (parallel .15 uf parallel capacitor optional) from the - out pin to earth ground and measure the voltage across it to determine the leakage current, the repeat with the AC plug released. I would much rather see any leakage currents more than a couple hundred microamps than feel it.

 

It looks pretty nifty. If I was going to use it I would take a 1K resistor (parallel .15 uf parallel capacitor optional) from the - out pin to earth ground and measure the voltage across it to determine the leakage current, the repeat with the AC plug released. I would much rather see any leakage currents more than a couple hundred microamps than feel it.

I measured leakage current at 60 Hz and it's impressive. With load applied, I measured 8.9uA between Earth and the - out pin. With the AC plug released, I measured 3.1uA . Looks pretty good for $2.00.