EEVblog #880 - Mailbag

[EEVblog]

Mailbag Monday
Teardown bonanza

SPOILERS:
LCD's and CCD sensor under the Tagarno microscope
The Aoyue3D Ritocco 3D shaping , cutting and finishing tool http://www.aoyue3d.com/
amazon
1980's multimeters (Fluke ripoff)
Wandel & Goltermann optical attenuator
Cheap mains energy meter teardown
DTV TV tuner teardown
Wayback Wednesday material, a collection of the Cyclopedia Of Applied Electricity
Introducing Teirnan
Radio Shack calulator and diary teardowns
Dodgy ebay industrial light controllers

[iamdarkyoshi]

Those old multimeters? Yeah I have some under yet another label

[Razor512]

I would love to see various generation of CCD and CMOS sensors under a good amount of magnification.

I took apart a few cameras a while back, mainly for some capacitors, card slots, and the backlights.



Full size:

[n3vti]

Holy crap! I remember the RadioShack shopping calculator, and I used to own the black organizer! Ah, the quirkiness that is was RadioShack back then...

Speaking of that, sorry if I haven't been on in a while. I just got promoted to a manager of a store, so now I just might be able to get some (decent) equipment! Because of the long hours (and longer drive), I won't be on as often. 

[bktemp]

The lamp ballasts actually look pretty decent. They are Osram PT VIP 03 MID.

[Brumby]

What's wrong with this picture?

[bktemp]

What's wrong with this picture?

There is no Vol 3, but Vol 8?

[dexters_lab]

hehe, i'll hold my hands up to electronic organisers

had a few over the years, Casio SF-4000, PalmPilot Pro 

i also had an Apple MessagePad 110 

[Muttley Snickers]

What's wrong with this picture?

There is no Vol 3, but Vol 8?

It was on one of the pages shown in the video declaring 8 volumes and I sent Dave a link earlier to a US Ebay listing specifically for volume 3 but the listing states they may not ship to AUS and he may not want it anyway. I realised afterwards being slightly blinded by what was missing rather than the realisation of what he did have.

[Fungus]

What's wrong with this picture?

There is no Vol 3, but Vol 8?

Volume III was borrowed by the missus to press some flowers....?

[nitro2k01]

Oh look, yet another bodge! (On the PSU of the set top box.) Guess they didn't need a negative rail and wanted to make assembly simpler.

[nitro2k01]

I think what happened on the lighting PSU was that the transistor touched the plastic film capacitor and melted it and shorted internally. A temperature that might be fine for the transistor could be deadly for the cap, and the exact alignment (whether they touch or not) could be what makes the PSU work or fail.

[Fungus]

I think what happened on the lighting PSU was that the transistor touched the plastic film capacitor and melted it and shorted internally. A temperature that might be fine for the transistor could be deadly for the cap, and the exact alignment (whether they touch or not) could be what makes the PSU work or fail.

I'm thinking it's more like:
a) The PSU designer did a decent job of making a PSU for some combination of price and ambient conditions.

b) The PSUs were then bought by a disco light designer who doesn't read datasheets and was all, "Airflow? We don't need no stinkin' airflow!"

c) ...this video.

[sean0118]

I think what happened on the lighting PSU was that the transistor touched the plastic film capacitor and melted it and shorted internally. A temperature that might be fine for the transistor could be deadly for the cap, and the exact alignment (whether they touch or not) could be what makes the PSU work or fail.

Might be possible? But it's weird that they added such a large heatsink to the output transistor/diode which is probably just a boost converter.

I think the four 19N20C mosfets are being used for input rectification (see link), I thought that would still require a fairly large heatsink?

http://www.thetaeng.com/FETBridge.htm

[Fungus]

I think the four 19N20C mosfets are being used for input rectification (see link),

Maybe.

I thought that would still require a fairly large heatsink?

Nope. A fully-turned-on MOSFET has very little voltage drop so hardly any Watts are dissipated.

[bktemp]

I think the four 19N20C mosfets are being used for input rectification (see link), I thought that would still require a fairly large heatsink?

http://www.thetaeng.com/FETBridge.htm

I think Dave swapped the input and output connectors.
Osram PT VIP 03 MID seems to be a HID ballast optimized for projection lamps. It even has a UART connection on the control board for controlling the lamp! The transformer next to the output is most likely the ignition transformer (insulated wire, multiple sections on the bobbin). The H-bridge probably generates a low frequency squarewave for the lamp. So it has almost no switching losses and only moderate conduction losses. The mosfet on the heatsink instead is probably used as buck converter for generating a constant current. Because it converts the rectified mains voltage into a lower voltage, it has quite high switching losses.

[Towger]

What's wrong with this

They are in the wrong order and volume 3 is missing.

So much for my thoughs that Dave was the product if the Christian Brother education system or similar organisation.

[gregallenwarner]

I've noticed when tearing down multimeters, Dave always jokingly mocks the "crappy" transistor testers found on the cheap ones. Forgive my ignorance, maybe I haven't been around long enough to know why this is, but can anyone tell me for my own edification, why does Dave regard the transistor testers so poorly?

[iamdarkyoshi]

I've noticed when tearing down multimeters, Dave always jokingly mocks the "crappy" transistor testers found on the cheap ones. Forgive my ignorance, maybe I haven't been around long enough to know why this is, but can anyone tell me for my own edification, why does Dave regard the transistor testers so poorly?

It seems to just be an unneeded feature added to cheap meters to make them look fancy and more feature packed.

[bitwelder]

At 23:05 the name of Nicola Tesla was listed among the 'authorities consulted' (which may have given his input on the foreword paragraph on 'transmitting electricity without wires' read by Dave)

[HKJ]

why does Dave regard the transistor testers so poorly?

One point may be that they are not needed, but more important is isolation distance. If you place a finger over the transistor socket the distance to the metal inside is less than a mm, compare that to the isolation at the test probes. Due to this a transistor tester will usual prevent any CAT rating, except the fake ones.

[sean0118]

I thought that would still require a fairly large heatsink?

Nope. A fully-turned-on MOSFET has very little voltage drop so hardly any Watts are dissipated.

But you don't know how many amps are running through them? Even at 3A these would be ~90C above ambient, unless I'm mistaken?

I think the four 19N20C mosfets are being used for input rectification (see link), I thought that would still require a fairly large heatsink?

http://www.thetaeng.com/FETBridge.htm

I think Dave swapped the input and output connectors.
Osram PT VIP 03 MID seems to be a HID ballast optimized for projection lamps. It even has a UART connection on the control board for controlling the lamp! The transformer next to the output is most likely the ignition transformer (insulated wire, multiple sections on the bobbin). The H-bridge probably generates a low frequency squarewave for the lamp. So it has almost no switching losses and only moderate conduction losses. The mosfet on the heatsink instead is probably used as buck converter for generating a constant current. Because it converts the rectified mains voltage into a lower voltage, it has quite high switching losses.

Yeah that sort of makes sense, I was going off the letter which said 60VAC in and 300VDC out (Are there lights that require 300VDC?).

I thought that transformer with the split bobbin was actually a common mode choke to filter the input AC.

If the side with the heatsink is the input, where is the rectifier? Unless the boards DC input AC output?


Anyway, still not closer to why the mosfets died, maybe the lamps just pull too much current? 

[Fungus]

why does Dave regard the transistor testers so poorly?

a) They're useless

but mainly:

b) They're a sure sign of a crap multimeter. When you see a transistor tester you know you're in for a laugh if you open up the meter and look inside.

[Fungus]

Nope. A fully-turned-on MOSFET has very little voltage drop so hardly any Watts are dissipated.

But you don't know how many amps are running through them? Even at 3A these would be ~90C above ambient, unless I'm mistaken?

Let's see:

http://datasheet.octopart.com/FQP19N20-Fairchild-datasheet-87852.pdf

Rds(on) is 0.12 Ohms - higher than I expected for a MOSFET. 

Let's do the math then: At 3A the voltage drop across them is 0.36V. 0.36V at 3A is a bit more than 1W of power dissipation.

Wikipedia says "When a TO-220 package is used without a heatsink the heatsink-to-ambient thermal resistance in air for a TO-220 package is approximately 70 °C/W", so ... 90°C is in the ballpark, yes.

The datasheet says their operating temperature is up to 150°C so they'll be OK with that dissipation in open air. I guess it's all down to the airflow inside the lamp but you're probably right; those particular MOSFETs need heatsinks.

[bktemp]

Let's do the math then: At 3A the voltage drop across them is 0.36V. 0.36V at 3A is a bit more than 1W of power dissipation.

In a h-bridge only half of the mosfets are active at the same time, so the avarage power power dissipation is 0.5W per mosfet. That's ok without a heatsink.

I can not find any datasheet or other informating for the ballast or the lamps.
Here is the current model:
http://www.osram.com/osram_com/products/electronics/electronic-control-gears-for-specialty-lighting/ecg-for-projection/index.jsp
The ignition transformer is completely potted with the output wires attached.
PT VIP 03 MID seems to be mainly being used in video projectors. Typically there is a PFC on the main power supply with a 350-400Vdc output going to the lamp ballast.
Maybe the ballast requires forced air cooling, beause is it anyway necessary for the lamp.