Test Equipment Anonymous (TEA) group therapy thread

[mnementh]

Those 2 orange tants. One is a 1uf/35V, the other is a 1uf/50V. I'm not sure which one is which and there's a 50% chance of getting it wrong resulting in smoke. So I'll wait until assembled and powered up then measure voltage on each.

Once again in the interests of fairness, I seriously doubt that mnementh would have suggested using 50V rated tants in both locations if med hadn't said that would have to "wait until assembled and powered up then measure voltage on each."

Yes of course it is an obvious solution, but is only applicable if you have both values of tants, rated at 50V to hand.

As both tants are 1µF, fair bet he has the correct value in 50V. 

In all honesty, I'd be more worried aboot him pulling a mnem and forgetting that some, but not all, tants are marked for the (+) leg, where most other types are marked for (-).

I suspect the likelihood of pulling that boner is a nigh-universal threat. 

mnem

[25 CPS]

the Agilent did not like that one bit and freaked out with an A-Err message and beeping when measuring the battery connected to the Fluke at the same time the Fluke was measuring the battery connected to the Agilent.  The Fluke didn’t care but the U1252A didn’t like that cross connected arrangement at all so a bench power supply connected to both solved that problem.

Yes, I do not remember which one I was using in the past but I also had bad experiences with the DMM measuring his own 9V battery. I think some esoteric designer did not account the possibility of the DMM leads connected to the DMM battery. I remember something bad but I regret not to have further details on that. To me never use the DMM to measure his battery.
As silly it sounds I believe the DMM are not design to measure anything in their own connected battery.

There's nothing esoteric in the design. This behaviour is by design.The DMM has a floating supply (the battery) and the potential of the COM (black / negative) is tied to a potential usually somewhere between battery plus and minus. This way the DMM has some kind of +/- supply in respect to the input terminal. If you connect the COM to one end of the battery, you'd short some of its internal circuitry, no wonder it shows errors or erroneous results.
You could leave the COM open and measure each battery plus and minus with a single test lead, and add these two values - should work.

I had been suspecting a ground loop-esque issue based on I could use one DMM to measure the other's battery no problem but having them check eachother at the same time was where the trouble happened given how the commons and battery negative poles were tied together in this test.  I didn't realize they tried to take a single battery and form a split supply by shoving the potential of the common input away from the two terminals with a voltage divider.  The Fluke was totally undisturbed by this though, so they must have taken a different design approach on the 27.

[ch_scr]

Todays "Fun with TE": news on the distributed amplifier!
Half-way build picture, with capacitors to ground vertically inside the pcb 
Idea was to reduce parasitics and required space. Turns out an 0603 fits pretty nice inside a 1.1mm hole. Bugger to solder tho.

Finished build, with and without gain pot. Gate 2 decoupling and connection bodged on with kapton tape and wire (aka poor man's triple layer pcb):

Measured with the Nanovna V2+ 
red=gain
blue=input return loss:

red=reverse isolation
blue=output return loss:

~5.5dB gain and gain flatness is not too impressive
but P1db at about 10mW/10dBm and 25dB isolation worst case (courtesy of the cascode nature of dual gate mosfets) make me happy nonetheless 
Not sure though why the gain goes up towards higher frequencies 
Schematic for the one's playing at home:

[Cerebus]

The following is mostly offered here as a PSA for those playing along at home:

I have bought parts through AliEx four times shipped via OrangeConnex from IIRC, Thailand and Korea to the GWN. Actually received one of them, and that was a month overdue.

I'll try to temper my judgment here with the knowledge that this was during the beginnings of the COVID crisis, when nobody knew what the pluck was going on with all of Asia, much less what was going on in there...

...but still, that track record, coupled with absolutely zero follow-through or ability to contact the carrier, has put any sales with them as a carrier in the "Not only no, but hell no!" column for me.

In short, be prepared for at best a 2-3 month wait, and then to have to order parts again when they don't come.

That's odd because my experience with orangeconnex has been completely the opposite. They've always been reliable and the one and only time anything that was passing through their hands went astray it was when they used Royal Mail as the last mile carrier and I can prove conclusively that it was Royal Mail's fault on that occasion (Stupid postwoman shoved it through next door's postbox while next door was unoccupied and scanned it as delivered. Builders came to renovate last week and found it and passed it on.).

In particular I've always found tracking info from orangeconnex to be accurate, sometimes more accurate than publicly available tracking info from the carriers they subcontract to. Also they seem to be significantly faster (2-3 days) on those semi-express services from China than other carriers - they seem to be faster at getting stuff onto a plane out of China than others (e.g. 1-2 days from "ready for collection" to "on a plane" compared to 3-5 days for other carriers) leading me to believe that they run their own linehaul operation in China where others rely on third party linehaul.

The wildly differing experience leads me to suspect that whether they are good or not is down at least in part to the country of destination. There's a lot of competition for goods passing through Heathrow and a canny buyer can probably negotiate better prices and service for something passing through London than perhaps they can for other destinations. Orangeconnex appear to have their own depot in London (Pump Lane, Hayes), which then also gives them flexibility with onward carriers; who they use and for what, to get from Hayes to here on the other side of London seems to vary from package to package. Some places they may be faced with an incumbent with a monopoly/near-monopoly and have to take what they can get - I have no idea whether this applies to the GWN.

[Cerebus]

You are a day late and a dollar short. 

My mantra.....ALWAYS read ahead before responding. 

I read your post about those components earlier and I wouldn't insult your intelligence by even recommending that ... 

I'm not insulted nor offended. I knew the Captain Obvious's would come out of the woodwork and I would have to explain my motivation.

Ya still haven't learned how to use an apostrophe though, have you Mr. Smartypants? 

[Cerebus]

Todays "Fun with TE": news on the distributed amplifier!
Half-way build picture, with capacitors to ground vertically inside the pcb 
Idea was to reduce parasitics and required space. Turns out an 0603 fits pretty nice inside a 1.1mm hole. Bugger to solder tho.

I can't decide whether that's a brilliant idea or an abomination. 

[mnementh]

The following is mostly offered here as a PSA for those playing along at home:

I have bought parts through AliEx four times shipped via OrangeConnex from IIRC, Thailand and Korea to the GWN. Actually received one of them, and that was a month overdue.

That's odd because my experience with orangeconnex has been completely the opposite. They've always been reliable and the one and only time anything that was passing through their hands went astray it was when they used Royal Mail as the last mile carrier and I can prove conclusively that it was Royal Mail's fault on that occasion... (SNIP) Some places they may be faced with an incumbent with a monopoly/near-monopoly and have to take what they can get - I have no idea whether this applies to the GWN.

Yeah, there is some confusion here... but in this case, pretty sure it never left the country of origin, as the tracking trail stopped with a particular single-instance air freight flight number, which further investigation showed as having landed before the item was actually purchased. And further, one of my subsequent purchases ended exactly the same way, with the same exact flight number.

Again... I've chalked it up to COVID; it only cost me assache in the form of an eBay dispute, after all.

But what I do still consider a big "no way in 'ell" is the fact they have no customer service portal or service ticketing system in place so you can ask someone at the carrier anything aboot a parcel. As these are still supposedly a tracked shipment, that's a large caveat.

mnem
 

[med6753]

Yeaaaah... that post was more aboot delivering a terrible pun than it was aboot advising papa smurf on how to fix a Tek... That would be like telling your octogen-stripper how to suck... eggs. 

mnem
And now I get to repeat said terrible pun...   

You keep mentioning that stripper. Do you require a refresher re-post? 

You keep threatening all of us with that stripper. Do you need a refresher repost of Kerm-goatse...? 

mnem
I have a pink frilly tutu and I'm not afraid to use it... 

I've done nothing of the sort.......recently. 

[med6753]

You are a day late and a dollar short. 

My mantra.....ALWAYS read ahead before responding. 

I read your post about those components earlier and I wouldn't insult your intelligence by even recommending that ... 

I'm not insulted nor offended. I knew the Captain Obvious's would come out of the woodwork and I would have to explain my motivation.

Ya still haven't learned how to use an apostrophe though, have you Mr. Smartypants? 

Oh go sit of your fist and lean back on your thumb. 

[mnementh]

Todays "Fun with TE": news on the distributed amplifier!
Half-way build picture, with capacitors to ground vertically inside the pcb 
Idea was to reduce parasitics and required space. Turns out an 0603 fits pretty nice inside a 1.1mm hole. Bugger to solder tho.

I can't decide whether that's a brilliant idea or an abomination. 

Given that the MMLCC almost certainly has a different thermal expansion coefficient from the FR4 & trace and this assembly technique puts the entire body of the capacitor in tensile stress, I'm going to go with the latter. You know, to err on the side of caution.

Now whether it's better or worse than tombstoning SMD caps/resistors... 

mnem
before you say it... yes... *guilty as charged*

[med6753]

212 mini, round 5.

Completed tant replacement on prior mentioned board. Before diving into the main board decided to perform a quick functional test. It works, on both batteries and mains.  Going to let it cook for a while.

[mnementh]

Just adorbs.     

mnem
*toddles off to do something... constructive.

[Cerebus]

Todays "Fun with TE": news on the distributed amplifier!
Half-way build picture, with capacitors to ground vertically inside the pcb 
Idea was to reduce parasitics and required space. Turns out an 0603 fits pretty nice inside a 1.1mm hole. Bugger to solder tho.

I can't decide whether that's a brilliant idea or an abomination. 

Given that the MMLCC almost certainly has a different thermal expansion coefficient from the FR4 & trace and this assembly technique puts the entire body of the capacitor in tensile stress, I'm going to go with the latter. You know, to err on the side of caution.

Now whether it's better or worse than tombstoning SMD caps/resistors... 

mnem
before you say it... yes... *guilty as charged*

Yeah, but the same applies if it's soldered on the top normally, so in that respect it's no better or worse than 'conventional' SMD construction.

Mechanically it's going to be supported by bulk solder, rather than a thin layer of solder over copper - again I'm not in a position to assess whether flex on the board is going to affect a vertical mounting like this more or less than it would if it was conventionally mounted. Anybody got a stress/strain curve for solder alone versus solder between a pad and a component?

Edit: I think my discomfort here is that it's a situation where normal mechanic's instincts fail. Instinct says "there's something hinky about this" but I can't offer a rational description to back up instinct, which is usually an indication that instinct may be off in the particular case, or at least unreliable.

[mnementh]

Thinking aboot the thermal stress that way, it does seem to boil down mostly to the fact of soldering one end then the other... which resultant contraction from cooling is one of the surest ways of fracturing a MMLCC or ceramic substrate and is the reason we take such care to minimize dwell time when hand-soldering.

This property is of course the whole reason most SMD assembly process control ensures the board is preheated, soldered, then cooled slowly.

My gut reaction is that in this case, the surface mount technique has some "give" in the tiny bit of elasticity of the trace/solder fillet, where being butted against the outer sides of the PCB takes at least the elasticity of the trace away, while concentrating more heat on the component rather than the trace/PCB, so increasing potential for expansion.

No idea if any of that "gut reaction" is any more correct than yours, though.

Random thought... what is the comparative expansion coefficent of the FR4 in each relevant axis? Any idea if there is any material difference due to the layers of matting?

mnem
 

[Saskia]

Yeaaaah... that post was more aboot delivering a terrible pun than it was aboot advising papa smurf on how to fix a Tek... That would be like telling your octogen-stripper how to suck... eggs. 

mnem
And now I get to repeat said terrible pun...   

You keep mentioning that stripper. Do you require a refresher re-post? 

I can assist with a smurf-licking dragon. 

Should I choke you with my tits ?
(You were asking for this, sorry)

[med6753]

Yeaaaah... that post was more aboot delivering a terrible pun than it was aboot advising papa smurf on how to fix a Tek... That would be like telling your octogen-stripper how to suck... eggs. 

mnem
And now I get to repeat said terrible pun...   

You keep mentioning that stripper. Do you require a refresher re-post? 

I can assist with a smurf-licking dragon. 

Should I choke you with my tits ?
(You were asking for this, sorry)

Death by motorboating? Sounds.....nevermind. 

[ch_scr]

[...] capacitors to ground vertically inside the pcb  [...]

I can't decide whether that's a brilliant idea or an abomination. 

Thanks, I'm not so sure myself about that 

[...]
Random thought... what is the comparative expansion coefficent of the FR4 in each relevant axis? Any idea if there is any material difference due to the layers of matting?

mnem
 

Yeah, and it's obvious  (If you already know, as always...) The matting keeps it pretty stable in X and Y, but in Z there is almost no restriction - the layers just sit atop each other. So Z just moves with the thermal expansion of the resin. Soldering with a hot air gun on this pcb is probably a bad idea.

[BU508A]

Yeaaaah... that post was more aboot delivering a terrible pun than it was aboot advising papa smurf on how to fix a Tek... That would be like telling your octogen-stripper how to suck... eggs. 

mnem
And now I get to repeat said terrible pun...   

You keep mentioning that stripper. Do you require a refresher re-post? 

I can assist with a smurf-licking dragon. 

Should I choke you with my tits ?
(You were asking for this, sorry)

That's really very thoughtful of you but I'm afraid I'll have to decline.

But if Papa Smurf is interested, well ...   

[Cerebus]

Yeaaaah... that post was more aboot delivering a terrible pun than it was aboot advising papa smurf on how to fix a Tek... That would be like telling your octogen-stripper how to suck... eggs. 

mnem
And now I get to repeat said terrible pun...   

You keep mentioning that stripper. Do you require a refresher re-post? 

I can assist with a smurf-licking dragon. 

Should I choke you with my tits ?
(You were asking for this, sorry)

Death by motorboating? Sounds.....nevermind. 

[Wolfgang]

The upper gate has floating DC. not a good idea, IMHO.

[Saskia]

... I'm no longer willing to put other people's repair jobs ahead of my own projects on the bench once those three committed once are finished up and out the door.

My final red line was crossed when a remote acquaintance asked if I would be so nice to have a look at a minor problem his girlfriend was having with her stereo amp. Surely something I could fix in a few minutes. I agreed to have a look.

On the day, she arrived with an amp of some cheap chain store house brand, the thing was a total wreck and looked as if it had been buried under a dungheap which wasn't far from reality - she owned a pub. Switches had lost their click stops, the volume pot was broken and there were a number of black spots on the PCB.

When I suggested we'd better put this thing out of its misery she got all upset. "Well, thanks for nothing, Mister Electronics Wizard!"

My workbench hasn't seen anything that wasn't my own, ever since.

Yeah I have kinda already come to the same conclusion too...

Yesterday spent WHOLE day, 9 Hours straight, working on the two electric fence motors. Managed to get one of them going, jut dirty micro switches that detect the passing of the fence. Opened them up, some Deoxit, exercise... good as new. electronic board two happy and driving the motor.
Other one, not such luck. The circuit board itself was sick, very sick. Supplies look good and CPU is running, but doing strange stuff and fails to drive the front panel board (LCD  Keypad). Obviously no service manual so can't go any further, and even if I could the guy would not want to pay for the time it would take to figure it out, never mind pay for the parts if need be. So I called it quit on that second motor.

So I e-mail my friend saying cool I managed to save one of the motors, taht will be 50 Euros, cheap for 9 hours of work and considering the motor still sells brand new, for 500+ Euros. I have not gotten a reply yet...

As for his Asus tablet, I wanted to be good and check the charging board inside the thing, as it failed to power up the tablet in the absence of the battery, and even when the battery was present and the USB charging cable... it still was draining the battery.
So I soldered a couple wires to the charging board so I could feed it 5Volts via my lab power supply, to monitor the current.... Wanted to the take the board out so I could see the underside, see where the easiest place would be to solder wires to. Removing the board invites unplugging 3 large very high density flat flex cables. First time I see this kind. Connectors too were not the usual flat flex connector we see, so had to figure out how to undo them... but they were so fragile that I fucked one of them. One of the two that handles the touch screeen. SO now the touch screen does not work anymore.
Of course I can't tell him that... so I need to buy a new board, 15 bucks or so, and swallow the cost. So won't male any money on this Asus, well I will even Lose money for sure.

9 hours on these motors, and he won't even pay 50 euros for it by the looks of it... A whole day WASTED when I have so many projects of my own to work on, and a house to work on as well.... all that to either lose money or make 10 bucks for fixing a 500 Euros motor, and the guy not even being grateful ?!

FUCK THAT !!!

I am done... I am waiting for my friend to get back to me, and I will tell him I don't do repairs any more, not worth my time and people just expect miracles for pretty much free, as if electronic skills were totally worthless. If it's worthless then go fix your own stuff, have fun !!!   

Similar thing: fixing whitestar 2 CPU board for a Harley Davidson pinball which was damaged because they did not bother to change the batteries.
Board would not boot.
Took 6 hours to repair it. parts cost irrelevant, but a replacement board is unobtainable. Needed reconstruction of the vias and eyelets. Y'all know how much a bungard press costs and how much work it is to fix this kind of damage.

"100 € ? that much "

my reply: well, next time you need a repair go talk to the parking meter ... just don't ring again.

[mnementh]

[...] capacitors to ground vertically inside the pcb  [...]

I can't decide whether that's a brilliant idea or an abomination. 

Thanks, I'm not so sure myself about that 

[...]
Random thought... what is the comparative expansion coefficent of the FR4 in each relevant axis? Any idea if there is any material difference due to the layers of matting?

mnem
 

Yeah, and it's obvious  (If you already know, as always...) The matting keeps it pretty stable in X and Y, but in Z there is almost no restriction - the layers just sit atop each other. So Z just moves with the thermal expansion of the resin. Soldering with a hot air gun on this pcb is probably a bad idea.

Yeah, that's the theoretical difference, which I already knew was the case. Just not sure what the actual thermal coefficient would be, and even if I could be arsed to look it up ATM, not sure how to characterize the way it would react to a more concentrated point source of heat, like a soldering iron. 

I was actually hoping it might bug someone else enough to go down that rabbit-hole, as I'm a bit overwhelmed IRL with the honey-do list and mounting time pressure.

mnem
*currently gusseting a bedframe for my son*

[Saskia]

Yeaaaah... that post was more aboot delivering a terrible pun than it was aboot advising papa smurf on how to fix a Tek... That would be like telling your octogen-stripper how to suck... eggs. 

mnem
And now I get to repeat said terrible pun...   

You keep mentioning that stripper. Do you require a refresher re-post? 

I can assist with a smurf-licking dragon. 

Should I choke you with my tits ?
(You were asking for this, sorry)

Death by motorboating? Sounds.....nevermind. 

I do indeed accept credit card for my services which are conveniently priced at 360€/h ...

[ch_scr]

The upper gate has floating DC. not a good idea, IMHO.

The gate 2 goes to the potentiometer to set gain. The gate 1 is internally biased on BF1005 
And while connecting it directly (as in DC-wise) to something (say between the "parallel" fet's) is not recommended by the datasheet, I hoped the part-to-part variation (on the internal biasing) would be small enough for it to work (and it seems to).
The alternative - dc-decoupling each gate 1 individually from the input line - took too much space and parts for my liking.
So rest easy, it is not floating. In fact, I also used this kind of fet on the active probe - too bad none of these are in production anymore  ! The only dual gate fet I could find (on a distributor) was also NRND and without the internal bias - maybe this kind is more expensive to manufacture, and DG-MOS out of (mass production) use in general?

[mnementh]

In this context... I'd be willing to risk it. 

mnem
*toddles off to do something... even if it's wrong*