Hi, welcome to tear down! Tuesday We' got ourselves a power analyzer today. Haven't had a look inside a power analyzer before, so should be a little bit interesting. This is a Voltec PM 300 and if you like me, you haven't heard of Voltech before. They were sold off to Uh Tetronics some time ago and Tectronics still do, um, sell some of the Voltech power analyzers, the PA 1000 and the PA 4000.

You can still buy them. but anyway, this is a nice little bit of Kit which I picked up at an auction and it is a three-phase power analyzer. There's the voltage and current inputs and it's got a nice big graphical display on it as well, which shows all sorts of stuff, including all three channels at once. It even has a waveform mode that can actually show the waveform.

and of course it measures a whole range of uh, you know, power stuff, not only for Mains although you know it goes up to like 500,000 volts or something like that. um, but it can also Al do uh, smaller voltages and currents as well. multiple ranges and it's a really nice bit of Kit And of course it can measure everything. You know we can measure voltage, current, uh, phase angle, real power, apparent power.

can even do inrush current, which is really handy. It can Uh has a balast mode where it can do you know, high frequency balast outputs, and ah, all sorts of stuff Peak and averages and all sorts of things like that. So really, very handy Power analyzer for really, uh, you know, measuring the Power consumption of products cuz it's not that easy. You know you get one of these cheap ones.

You buy an eBay for 10 15 bucks. uh, gives you a rough indication, but you know these things really go to town. Uh, pretty accurate too. 0.1% on pretty much all the ranges.

So um, and all the uh functions. So pretty awesome. Should be interesting inside. let's take a look.

Not hugely complex, but hey, you never know. So you know what we say here on the E blog. don't turn it on, Take it apart now. I'm not exactly sure how old this unit is, but uh, it has a calibration date on the back of 2004 so it's at least a 10-year-old uh design, but still pretty kickass.

By modern stand is to get a .1% three-phase power analyzer with all the bells and whistles. And the good thing about learning about new brands like this Voltec one that I haven't heard of before, you add them to your eBay watch list I've a list of you know of the more obscure name uh products out there and you just keep that eBay watch list and an alert pops up if you know something on the market comes up. Now this is actually an exmilitary unit same as the oscilloscopes. I uh, scored recently, but it came in a nice padded Uh custom Pelican uh case genuine Pelican case and it looks like it's hardly ever been used.

Got some writing on the front here, but yeah I think I Just sat in a warehouse for like 10 years there. it is calibration due 2004 so I expect this have been manufactured maybe in the early uh, 20000s or something like that. We probably only got like calibrated once or something like that when they originally got it and so it could be like 23 be Ming Guess it's got what looks like an optional module here with the printer and Rs 232 interface, so you can actually get the data out of this thing. Standard: I means input switchable of course, which is really handy.

It's got a voltage ction, uh, switch there and three separate Uh channels for your three-phase uh measurement. I've got no need for three-phase measurement here, of course I don't even have three- face power com to the lab. Oh no, no. actually technically I do I think I have it into the switchboard, but I certainly don't have any uh, three-phase Outlets that's for sure.

So I'll just use it as a single Channel and you can actually hook up an external Uh current, use an external current shut and you can actually program the in. but it recommends of course that you use the internal uh current shunts for these things and they can be I'll link in the Manu the user manual for this thing down below and it shows how it can be uh configured in Del y configurations for the various phases and all sorts of stuff. So very flexible input so it looks like you know we're likely going to see inside construction, probably three separate cards for this, a huge baseboard in there, something like that, um, and maybe a separate display board on the front and front panel board. So I would expect these to be separate isolated.

they are galvanically uh, isolated. So the obvious? Well, there are a couple of ways to do that either you try and uh, do it analog or digital. For me, if I was designing this thing, I would do it digital. so I'd have the because these are very, you know, very precise.

They're going to have a, you know, like a 16bit converter in there or something like that. you know it needs to be really precise. um, ad conversion in these things. and uh, front end amplifiers and Rain switching and stuff like that.

So I expect all that, including the adcs to be on each separate card there and then like a Serial data interface to that because this thing I think uh s like has a 50 khz bandwidth or 100 khz or something like that. It's not, you know, huge. So you can just get the data over a digital serial interface. and it's much easier to galvanically isolate a digital interface just using some Opto cupers than it is to do it analog wise.

So anyway, um, it could be interesting. Let's crack it open. You bet your ass we're going to avoid the warranty. Let's do it.

Ah, one of the most satisfying things you could do look at that. Brilliant. There's a tiny little fan on the base of the unit here. It doesn't really need much ventilation at all because, well, yeah, there's no huge power dissipation inside this thing.

It might be drawn I don't know, You know, 10 watts or something just to power the electronics. Not even. and uh, the current shuns aren't going to be dissipating much power Anyway, So there you go. Um, I Expect some through hole technology in here, of course.

Probably a mix of uh, uh, through hole and surface mount. probably predominantly surface mount the processor. you know, processor inside this thing. They're not going to have a huge arm processor or anything like that, just some sort of 8 or 16bit micro.

uh, possibly something like that. so looks like we need to get the back off here. and uh, anything else. No, don't know.

Okay, let's try and slide this forward. H Well, there you go. I Just took out the serial parallel module and yeah, it's all through hole. So this thing could be maybe entirely through hole or mostly, uh, through hole.

Anyway, obviously got some sort of micro over there. huge big um through hole package. very traditional style card. Edge Connector routed into the board out in here, so that that is quite a nice design if somewhat ancient.

Uh yeah. date code as I suspected 2003 There you go got ourselves an Hitachi H8. so if we got one on here, eh, very likely to find an Hitachi micro also inside. uhuh I was wondering where the other screws were.

There we go. That's a bit sneaky. No. I can't get get this rear bezel off cuz it's got I can see under here.

It's got another probably three screws up under there, but it's under the case so I can't slide this case off forward or backward. so I'm probably just going to have to leave her off one side like that. Tricky almost in and uh yeah, it did have those screws on top. so I had Buckley's chance of sliding the thing back out so it's got to come off over like so oh man, that's one tricky little mongr.

But Tada we're in. Well it looks like I was wrong on the three board. uh, construction there for the three different channels. No, we can see right through there.

hello. Looks like um, what they've done is they've split it into a uh, single phase. so maybe they sold a single phase unit of this and you only got the bottom? Bard cuz we can see some shunts down there and the ranges for uh, phase number one. You know the puts for phase number one down here and the top board is just two additional Uh phase boards.

And yes, the single phase model is the PM 100 the manual uh shares between those I forgot about that and uh, it's really interesting kind of uh construction like this. Check out the board up in there like that. There we go. Actually a a connector sandwich between the top and bottom boards.

here. it's got card edges is on both sides. that's really quite fascinating and I've done the screws off there and it looks like it's just going to pop off. Oops, here we go.

Tada And apart from the the wiring we're in, there, we go. So if you just got the Pm100 model, you just get that baseboard down the bottom. And what's kind of annoying is that they haven't really put uh, servicing thought into this. I mean they haven't left enough lead length here for me to swing this board out completely.

But I guess even then, of course, to actually plug this board in and get it working, you have to plug it in to this top board up in here like this. It's got to plug into there so it's got a sandwich construction. You can't get access to the damn board to even probe it while it's operational or it's very difficult. Anyway, yeah, real pain in the ass to service something like this.

One of the first things I noticed was this little wire bodge over here going down. well it's powering the fan. They basically got the fan. directly as a red wire connected there and a black one going all the way over to the other side of the cap there.

and they're just bodgy soldered onto the leads of these axial uh, main output filter caps from the Transformer here. Almost as if the fan is an afterthought, because if you're designing a fan in from the start, you would have put a two pin connector on there to plug your fan into, right? So ah, beats me. And they've put a cut out in the board here and they've mounted what looks like I can't get the number on that yet. It's probably a, you know, a linear rig or something 78.

Yeah, I think it is a 7805 and they're using that big block down there as a heat sink. and of course, yeah, there's nothing uh special with the Transformer there. They've used a uh, just a PCB Mount Mains Transformer there. and well, there's a fullwave bridge rectifier on the output there.

There's for uh, 1 N, Uh 41s there, plus some beefy output filter caps, but that's it. You know it doesn't need much power to power all the circuitry as I said. And there you go. as we guessed, exactly the same processor used on the inside here.

Hey, you know you've already got the development tools for it. It's already in your bill of materials. Why not reuse it on the serial one as well? even though It's overkill for the serial interface card, it's probably you know it's good enough for something like this. Anyway, to drive the graph IAL LCD display and clearly it speaks French H I don't that's for sure.

And the only surface mount stuff we' got in this whole design is on as you'd expect on the front panel. LCD And yeah, this is not a Voltec design. They've just used an off-the-shelf module. or they've got an Oem.

You know, to supply them a module and that's uh, pretty much it. The interesting part about it is that the ribbon cable goes to this vertical I Think they call it an interface board like, you know? It's just crazy why you'll notice all the uh pins on one side. There are uh, shorted out, especially on that top one there. But you like why And look behind it.

they've got a an unused IDC header connector there. just ah. it's bizarre. and in case you're wondering, no, that one can't just be plugged into there.

it's actually two pin short. So that raises the bizarre question that, well, what do they do in the PM 100 the one Chant one phase module, They would have to have this vertical Riser board. Still, even though it says it's the PM uh 300 interface board there, like huh? what the and on this interface board? Well, just a whole bunch of 8bit shift registers 74 HC 4094 and what are they? 74 HC 597 Go figur in a just a miscellaneous uh Hz4 inverter over this side. so they're obviously ly Shifting the data in from the other two channels on the board on top.

but it's just it's bizarre. Anyway, well, we do have an 8bit micro here, so we got to interface everything to that 8bit micro bus. but jeez, and I'm never a big fan of crystals just free standing like that, just flapping in the breeze. You get the right vibrational mode and eh, off it's going to come.

We' got our serial interface board. uh, being plugged into this card. Edge connector on the front panel. so there's obviously a front panel connector board in there.

I'm not going to bother taking this whole thing apart just to get it that to look at. Uh, you know, the front panel user interface and stuff like that. Eh, boring as bat poo. It does look like though.

they've just got the all the buttons in a matrix of course. Uh, the front panel. Coming back, there's actually two of these, uh, multi-way ribbon connectors going down to the front panel board behind that shield in the front. Now, let's take a look at the interesting stuff and we're just going to have a look at one channel of this.

uh, top two channel board. Yes, I have looked the other IDE identical. The other channel on the other board. uh, down at the base there for the PM 100.

It's just another duplicate of all this channel. So we're only con and these look like two identical channels so we'll just concentrate on one and no surprises for guessing exactly what I thought it would be. Uh, they're all galvanically isolated. Check this out.

You can see the split in the ground plane, totally separate ground planes between all the channels. Looks like we got some input attenuation here. We'll take over that very interesting looking current shunt on the input. there.

we've got some protection. We've got some interesting looking uh Transformers here. we'll have a look at those in a second. and there's our uh, galvanic isolation using three opto couplers there.

So we're just getting serial data out of this thing and that's exactly what you'd expect. Now, if you were paying attention, you would have realized that there's no power. There's no other wires com to this board ordinarily in an isolated design like this for Simplicity sake. The main Transformer Main's Transformer on the input would have separate winding isolated winding.

Taps for each of these and you you know you'd have a connector over here or something. or you know, plugs in to each module coming from each isolated tap on the Transformer but they haven't bothered to do that. All of the Power is coming in Via this card Ed connector here and so aha, how are they getting the power across? It's not coming across. The Opto cou is bingo, It's coming across.

these two Transformers up here little hand hand wound jobs. Interesting. and this is very, very interesting. You can tell it's power transfer by the big beefy tracks on one side.

Looks like they might have a switching transistor here. so they're obviously, uh, driving this at, you know, switching the thing and then they've just got. You know, this one's actually got two turns on it. This one's only got a single little turn.

and look, they're actually using two in series like that. Look, this wind in here is just directly connected to that wind and you can see the traces going going like that. So look at that. It's a two, well, a effectively a two-stage Transformer there going to the other side and of course it's a dead giveaway.

They're just doing some filtering and then, yeah, some rectification and then some. uh, filtering. There probably got some uh, voltage local voltage regulation in there. sort of to 92, you know, 785s or something like that happening.

but that's rather interesting. Look at that. They've kind of quite a bit of trouble there to C spec and wind these Transformers and they've used two of them. So why they've gone with two like that in series? you know your guess is as good as mine.

really. I What did they not get the voltage specification? They were looking for the isolation specification with just the one, Huh? And there's our switching transistor on the primary side. it's a 2sk 940. so they're just doing crude switching of the primary side there and and well, they just rectify the output of the second Tre and Bob's your uncle.

So it's very crude and very inefficient. but you know, hey, it gets the job done, whatever floats your boat. And no surprises for guessing for finding some pretty beefy uh Opto couplers here from Aigo. love that name.

Ah, have a go you mug love it. Um, in these case, these are uh, 10 megit so pretty darn quick. Uh I think 3500 volts. RMS Opto couplers with five uh, kilovolts per millisecond transient voltage on them.

so you know, really pretty beefy Opto couplers. Well, I was way off the mark on the ADC Look at this pretty old school. Two of them as you'd expect, one for the voltage, one for the uh current on its Channel but very old school ad 7575 Adc's gez I Used the These Donkeys years ago. 8 bit successive approximation converters.

Really nothing fancy it all. although I guess all they need is8 bit resolution on this thing cuz it does have various ranges. It's got like eight voltage in current ranges or something like that. Speaking of which, the range switching.

nothing fancy going on here at all. Just some 74 HC 4053 and log Max's uh, Tl71 is it? Yes, that's a Tl071 Op amp and another 74 you know, 4,000 series. Marx I mean jeez, Nothing happened in here at all. so this is pretty interesting.

I mean obviously got two pathes here. One on this side here is for the voltage the voltage comes in. We'll take a look at that in a second. The other side here is for the current.

We can see the big current shunt down the bottom. basically identical uh paths like this. and they just got that um, you know, 4,000 Umti series, analog switch in and various opamp gain ranges. So clearly what? they? this is an auto ranging unit.

So clearly what they're doing is just at the higher uh gain levels. They're just, um, letting the other opamps uh, you know, overload and uh, saturate and then switch in between uh, those as it goes down in ranges. so they're obviously getting away with that. Nothing fancy here at all.

It's not like they got, you know, a relay switching or anything like that. really? You know, basic configuration stuff and the input here. Very basic attenuation. They just got four dropper resistors on each leg here, basically attenuating the input signal here.

Down to bugger. all. uh to go into this. uh MX here.

so you know really, no matter how high the transient voltage, it's not going to damage the input there. That's why we basically don't see any input protection if we have a look around here, you just don't need it because you're just attenuating the input so much. I Mean there's nothing on the input protecting that at all. And as far as our current Shunk goes here, well look at this.

This is a fancy pancy. It's an SMD current shun I Guess you could call it because it is actually um, you know, solded although surface mount onto large pads over here and that's super duper Whyde? what is that like? you know, 30 mm? uh, wide. That's going to be a low Tempco metal of course. and they looks like they've made no attempt at all to to actually trim.

that trimming would be done in software afterwards. So yeah, they're not going to dick around and laser cut that or you know, sold a little bits on there to, sort of, you know, change the resistance a minute amount. something like that. Yes, that is a cutout down in the boards there so that can extend down below the board.

And of course there's no fusing on this puppy at all. It's just yeah. Straight into there look big beefy blade terminals there solded directly into our current shant. And if you're wondering what this uh, circuitry here, this is the external current shunt.

uh, input. So they just got added some protection there. There's a themister on the input there, dead giveaway. th1 looks like they got a bodged uh ceramic cap on there just for some, uh, extra noise suppression there.

Hey, if you're going to bodge it, do it. Nice. They've put heat shink tube in over the leads there and a couple of back-to-back diodes in a series resistor there. so that's just a feed in an external shunt voltage if you want to.

So anytime you're feeding in external voltages like that, yeah, have to add some input protection and some noise suppression. and on every backpedal connection there for the voltage and current inputs, they got noise suppression to ground high voltage ceramic cap there 6 kts Uh 22 Narad. By the looks of it, look at those ceramic noise suppression caps down to earth as far as the eye can see all the way through there every back panel connector except for the external Uh current shunt input as we saw which is uh, taken care of on the main board as we just saw before and right next to our ADC there we have ourselves our voltage reference. There you go: little Uh, two pin shunt and the voltage Regulators on the output of the voltage isolation that we uh, saw before for the power coming in.

Yeah, no surprises as I guess 78 lo5s and of course make that a split Supply with the 79 lo5. Negative: So you might still be wondering. the input configuration here. voltage and current they're are obviously not galvanically isolated, so how does that work? Well, of course, the entire channel here is galvanically isolated by the Uh by the Transformer over here this switching transforma that provides the positive and negative 5 Vol rails for all the logic.

But basically what they're doing here is uh. they're referencing everything to the current shunt here. so the current shunt would effectively be the reference voltage. And because we've got such a large input attenuation here, on the voltage side, you can pretty much do whatever you want on the voltage inputs here.

and uh, pretty much not blow anything on the input stage. That's why you don't find any protection in there as I said before, and of course, you wouldn't be able to do that if you only had single side attenuation here. If you were say, referencing one of the voltage uh terminals, it would have to be common referenced to the current shunt down here. But because they've got two separate Uh high voltage attenuators on the input effectively, and this voltage, you can hook it up wherever you want over the entire operational range of this thing and you're not going to blow anything.

and then the data output side of our ADC As I said, it's an 8bit parallel output, so they're just, uh, whacking that into some shift registers. and of course, we don't have enough uh bits there. so they're just uh, feeding that over the opto coupler and well, that's it. Nothing fancy whatsoever.

Oh, I Forgot to mention there is a trim pot down here. I'm not sure exactly what they're uh, trimming and that's only in the voltage path they don't have a similar trimmer pot in in the current path, so not exactly sure what they're doing there. Um, you can calibrate this thing in software anyway to um, you know there's a whole procedure in the manual to do that. So there you go I Hope you found that relatively interesting.

That was the Voltec PM 300 power analyzer. Well, and it was a little bit different than I Expected is clearly with all through Ho technology and the ancient Hatachi processor stuff like that, you know it's based on a much older design and they've just carried that over into future designs cuz you know this thing well. it was manufactured around uh, that 2003, uh date or something, early 2000s. they've only manufactured uh, this is serial number 2,000 or something so hasn't manufactured too many before that.

So you know it's like late 90s design. uh at best. So I'm surprised we didn't see any surface mounting there. but that's what you get in these products.

The designers of these uh products. If they're familiar with all the stuff before, then they going to change everything when just for a new model. Oh let's all go surface. Mountain Well, Why? you know it's worked before all our Footprints all our pack in our CAD package.

We'll just run the damn thing again. Heck, even our lay can probably reuse a layout from an existing product or something like that. But yeah, anyway. um, there's very crude sort of uh ADC front end just 8bit adcs that fixed uh or fix gain ranges with the you 4,000 series marks on there to switch.

They're obviously overloading the other ranges when you switch low, but eh, it doesn't matter. it's just internally clamped and it's all just fine. and you can manually or Auto uh, switch between those Uh ranges seamlessly. CU This thing just continuously samples all the time.

Of course, at F you know, like at 100 khz or some fixed Uh sample rate or something like that. it does that to be able to do the uh, transient capture and all that sort of stuff. so actually collect the data and then dump it or display the waveform at to date so it's got to be able to do that. Auto Range switching and that's just a crude easy way to do it.

and with the voltage input, they don't even need any uh, you know, any protection or clamping on the input just attenuate the buggery out of the voltage input. That's it. I Don't know I Expected a little bit more complexity from all that, but eh, it is what it is. There you go, It obviously quite.

It works. It's really quite accurate and professional. This is a professional level product. It's not something that you'd buy on eBay It's from a company that specializes in these sorts of power analyzers and there probably been designing them since before I was born I Don't know.

Anyway, now it's owned by Tetronics and the Dan her group grown anyway. I Hope you enjoyed that. If you did, please give it a big thumbs up. and if you want to discuss it, the Forum links are down below.

That's the best place to do it and as always I will have highres photos of this tear down available on E B Blog.com and the link for that will be down below as well. So go there and check out the photos. Hope you enjoyed it. Catch you next time.