EEVblog #589 - Voltech PM300 Power Analyser Teardown

[JackOfVA]

The A/D data sheet says the sample rate that can be realized is slew rate limited, up to the 200 KS/s maximum based on the recommended clock speed and number of clock cycles for conversion.

Opportunity for some games here by reducing the voltage applied to the A/D's input S/H stage (thereby reducing the slew rate) and also reducing the A/D voltage reference to maintain the dynamic range?

[vanarebane]

So this product helps to test other mains products?

I have problem with my DIY reflow oven that, when operable, makes all the lights in the house flicker. I suspect that switching Solid State Relay in PWM mode, that it makes the mains voltage somehow off the standards.. With this tool, I would be able to find the problem of this flicker?

Maybe there is a tech university nearby that can give a test with some similar tool to this...

Great vid, Dave!
Thanks

Go in the same room and turn on another large load (vacuum cleaner, toaster...) while watching the lights.

The lights only got a bit less darker. What does this tell you?

[delmadord]

Dave,
I would also like to see that puppy in action. At this point, I can only imagine it is doing the dishes :-D Will there be quick review please?

[c4757p]

Go in the same room and turn on another large load (vacuum cleaner, toaster...) while watching the lights.

The lights only got a bit less darker. What does this tell you?

I was hoping it would tell you something. The lights will always dim when loads are switched, you're not going to get around that. It has more to do with your wiring than with the device connected to it.

Try slowing down your PWM a bit to see if it makes it less annoying. Or speed it up so you can't see the flicker, but then power dissipation in the switch will increase.

[vanarebane]

Go in the same room and turn on another large load (vacuum cleaner, toaster...) while watching the lights.

The lights only got a bit less darker. What does this tell you?

I was hoping it would tell you something. The lights will always dim when loads are switched, you're not going to get around that. It has more to do with your wiring than with the device connected to it.

Try slowing down your PWM a bit to see if it makes it less annoying. Or speed it up so you can't see the flicker, but then power dissipation in the switch will increase.

Yes I was already editing my first post when you already answered.
I followed the DIY reflow oven instructions here:
It uses PWM to turn on and off the power to the heater and has no soft-start or such, so yes, this makes the power jump up and down causing the flicker effect. This was my initial though also.
Changing the PWM would change the heating properties yes, but the oven is doing the reflow nice now and I'm afraid that changing PWM would make it not work on the reflow part. The flickering is not that annoying, I can turn some lights off or leave the room while reflowing, but I'm afraid of damaging other nearby more sensitive mains devices. Two laptop chargers already have started to act weird (laptops run in the lowest CPU speed when plugged and full power when on battery, 2 chargers, 2 laptops with same responses)
Would it help if the mains voltage is stored in some sort of cap or inductor?
I guess I could start a new thread on that instead of sidetracking it here.

[nitro2k01]

I was hoping it would tell you something. The lights will always dim when loads are switched, you're not going to get around that. It has more to do with your wiring than with the device connected to it.

I would fully expect that in a 120V country (*recalls Dave's story about the copying machine he tried to use in the US and sniggers and giggles*) but not so much in a 230V country. I live in Sweden, a 230V country where the wiring is typically properly installed, and I never really notice any flicker when turning on heavy appliances like a vacuum cleaner, microwave oven or a toaster, even when the lamp is connected to the same phase as the appliance.

I would guess vanarebane's mains wiring is simply underdimensioned. As a short term solution, I would recommend getting an extension cable that you connect to the power outlet with the shortest path to the fusebox, and route that to the room with the oven and use that for the oven. Make sure the extension cable is rated for the current the oven is using, of course. As a long-term solution, I would get the mains wiring checked out by an electrician and maybe replaced, if I were you.

[hikariuk]

I would fully expect that in a 120V country (*recalls Dave's story about the copying machine he tried to use in the US and sniggers and giggles*) but not so much in a 230V country. I live in Sweden, a 230V country where the wiring is typically properly installed, and I never really notice any flicker when turning on heavy appliances like a vacuum cleaner, microwave oven or a toaster, even when the lamp is connected to the same phase as the appliance.

Assuming his flag is accurate, vanarebane is in Estonia, which is also 230V.

[nitro2k01]

Assuming his flag is accurate, vanarebane is in Estonia, which is also 230V.

Indeed. The first paragraph is for c4757p and the second one is for vanarebane .

[kriebz]

The transformer isolation on the power bus is fascinating. Best guess on the need for twin transformers is that a single transformer would need two windings with lots of turns and this would have a risk of flashover.

My thoughts exactly.

Though I think I'd have used at least 2 turns on each former (the transformer interconnection), the one only has just over 3/4 turn on it. But, it obviously works well enough,

Ooh, I've actually seen setups like this only much larger, for total isolation.  I'm not so good at theory, but as I understand, the low-turns inner segment is at a lower voltage, which means higher current.  The inductance of the windings resists current change, i.e. in the case of transients, giving a much cleaner output, at the cost of a fair bit of loss.  There might also be a fusible link in the inner circuit.  I'll see what else I can find out about this type of circuit.

[tom66]

It looks like Tek discontinued production though.... not sure what technology Voltech would have contributed unless it was some obscure patent.

A couple of complete production ready products, with all the firmware, hardware and circuit design etc done. Do those all-in-one power chips do 0.01%?

Yes. We use similar ADE-chips at work and they go through an extensive calibration rig which can get accuracy down to about 0.1%. I'm sure if we tried harder and used a more accurate cal rig, we could do better.

[vanarebane]

I would fully expect that in a 120V country (*recalls Dave's story about the copying machine he tried to use in the US and sniggers and giggles*) but not so much in a 230V country. I live in Sweden, a 230V country where the wiring is typically properly installed, and I never really notice any flicker when turning on heavy appliances like a vacuum cleaner, microwave oven or a toaster, even when the lamp is connected to the same phase as the appliance.

Assuming his flag is accurate, vanarebane is in Estonia, which is also 230V.

I was hoping it would tell you something. The lights will always dim when loads are switched, you're not going to get around that. It has more to do with your wiring than with the device connected to it.

I would fully expect that in a 120V country (*recalls Dave's story about the copying machine he tried to use in the US and sniggers and giggles*) but not so much in a 230V country. I live in Sweden, a 230V country where the wiring is typically properly installed, and I never really notice any flicker when turning on heavy appliances like a vacuum cleaner, microwave oven or a toaster, even when the lamp is connected to the same phase as the appliance.

I would guess vanarebane's mains wiring is simply underdimensioned. As a short term solution, I would recommend getting an extension cable that you connect to the power outlet with the shortest path to the fusebox, and route that to the room with the oven and use that for the oven. Make sure the extension cable is rated for the current the oven is using, of course. As a long-term solution, I would get the mains wiring checked out by an electrician and maybe replaced, if I were you.

Yup, both right, 230V mains here and you can probably see from that far that house wiring is a bad crows nest
All thanks to my dad, who built this house and did the wiring himself. The oven is quite far from the fusebox and has many shitty crowsnest in between.
I went to the fusebox today and the oven performed without flickering other lights. So your hunch was right.
Is there any component that I can work into the oven to compensate the flicker?

Thanks!

[calexanian]

Not that it would matter but the devices across the current shunt, D13 and D14 kinda smell of TVS's Just a gut feeling looking at the partial part numbers and either an impending stroke or my spidey sense.

[Pillager]

Yup, both right, 230V mains here and you can probably see from that far that house wiring is a bad crows nest
All thanks to my dad, who built this house and did the wiring himself. The oven is quite far from the fusebox and has many shitty crowsnest in between.
I went to the fusebox today and the oven performed without flickering other lights. So your hunch was right.
Is there any component that I can work into the oven to compensate the flicker?

Thanks!

This really could have made it's own thread 

You can install a soft-start-circuit, but that will likely throw off your oven's performance a bit. You would have to test it and compensate for the changes.

Using a separate extension cord from the fuse box is another way, but it shouldn't be permanent. You could put in a new power line from the fuse box to your oven, or let an electrician do it. Do it yourself ONLY IF YOU KNOW WHAT YOU'RE DOING! Mains power is not to be underestimated! Make shure the power is OFF, and you should have another person there, just in case something does happen.

[Richard Head]

That's a damn nice instrument.
I've got a PM100 (single phase) and it basically tells you everything you'll ever need to know.
I also scored a 100:1 current clamp and a 1000:1 shunt all for $50!
Hang onto it Dave. I certainly would.

Dick

[EEVblog]

Hang onto it Dave. I certainly would.

I plan to. I need a good meter like that for the lab. Not easy to come by, or come by cheaply.

[Richard Head]

Meant to say 1000:1 CT.

[rsjsouza]

That was a very good teardown; growing in the eighties I have a weak spot for discrete logic, despite knowing that one or a few CPLDs/FPGAs could replace the entire set.

[NiHaoMike]

Yup, both right, 230V mains here and you can probably see from that far that house wiring is a bad crows nest
All thanks to my dad, who built this house and did the wiring himself. The oven is quite far from the fusebox and has many shitty crowsnest in between.
I went to the fusebox today and the oven performed without flickering other lights. So your hunch was right.
Is there any component that I can work into the oven to compensate the flicker?

Thanks!

Swap the lamp ballasts for a good quality inverter type with active PFC. (PFC doesn't really do good for home use, but active PFC is also a voltage regulator.)

[Pilot3514]

I was wondering why the design uses a shunt instead of a current transformer.

My only guess is that the power meter could be used to measure DC power as well.

I am interested in building a device to measure and log the voltage, current, and frequency of a two phase generator.  I may also want to raise an alarm if the values go outside of set parameters. 

Any advice to "young players"?

[SeanB]

Shunt likely is cheaper to get trimmed to good accuracy over a wide frequency range than a CT. Ct will have difficulty getting that wide range from 50Hz to 50kHz with sub 1% error without complex compensation or frequency dependant calibration.

[turbo!]

This looks like a reasonable low cost design. This one is the same concept as the plug-in type cheap but very accurate kWh meter. They're cheaply built to be within ballpark figure range and calibration is pushed into the Flash ROM. This relies on calibration constants set-in and the difference is in how good the calibration constant stays constant over time and temperature change.
 
The ranging circuits are like your every day DMMs. I think the most expensive component of this instrument is the R&D expense for the program software algorithm in that thing labeled U40. I'm pretty sure Tek bought them, because they wanted the intellectual property behind the software algorithm.

The two channels are integrated mathematically in the CPU and accumulated digitally. Big problem with this type of integration is that pulses that slips out of sampling or display refresh will slap past you.

[turbo!]

That's a damn nice instrument.
I've got a PM100 (single phase) and it basically tells you everything you'll ever need to know.
I also scored a 100:1 current clamp and a 1000:1 shunt all for $50!
Hang onto it Dave. I certainly would.

Dick

What's the software version in yours?  Does yours let you change ranges manually or use long integration?  I could use new firmware.bin.

[Richard Head]

Turbo
I haven't used it in a little while so I can't answer you now.
I'll power it up later and check the software version.
I also have some PM300 units available to me at work. They are housed in the same enclosure as the PM100 but are 3-phase units. I think it's the one Dave did the teardown on.

Dick

[PTR_1275]

I know I am digging up an old thread, but how easy are the service manuals to find for the Voltech gear? I am trying to find a copy for one now, but keep going around in circles.

-Chris

[bittumbler]

Hi,

sorry for replying to such an old thread, but i just now watched the video.

Regarding the "dual ring core transformer" setup, i think this is a form of (electrostatic) shielding.

A normal isolating transformer has a capacitance between primary and secondary windings of a few nF.
This dual transformer setup reduces the capacitance to a few pF.
Both "middle" windings are designed for low voltage/high current (low number of turns) to minimize electrostatic transfer between windings (low voltage means low electrostatic fields).
However typically one of the middle windings should have a center tap to ground. That seems to be missing.

Another approach to solve the same shielding problem is using coax cables as transformer windings. Then you need only one ring core.
The center of the coax will be your normal conducting winding wire. The shield of the coax will be your electrostatic shield to minimize capacitance between primary and secondary.

For this device such a shielding seems to be important, because it has multiple separate low impedance ground connections:
the power supply that powers it, and the measurement circuits. You want them isolated as much as possible.

Best Regards

Matthias