Test Equipment Anonymous (TEA) group therapy thread

[mnementh]

Hmmmmm... initial inspection shows no evidence of fire or magic smoke, and the rails are steady on at +/- 25.5V from a good old-fashioned linear power supply buried in the sub enclosure. After an hour powered up, all finals feel slightly warm but none warmer than the others.

Based on specs, I expect this should deliver right around 25W continuous per TDA7296 in AB operation, and this thing has 8 of the little bastards with two in BTL mode for the sub at probably 75W continuous or so. Fukkin-a-diddy! That's $50 worth just in the finals!!!

Now I need to figure out how to do turn-on/un-mute to inject signal & see if they're all healthy... I have an eval circuit in the datasheet, so should be able.

mnem
 

Suggest putting the Toronto FD on alert. 

Wise idea, actually. Got it to turn on/unmute by pulling pin 11 & 12 LO, just as suggested in the pinout here. Start probing around with a jackleg signal injector (plugged iPod into 3.5mm headphone cable with a random ceramic cap soldered to the tip on the other end) and determine that all the analog amplifier chips appear to be working. Notice while I'm doing this the smell of something getting a bit overheated, but not smoky.

Eventually trace it down to a 100Ω resistor trying to unsolder itself that was feeding a 7805. 189 sez "KAAAK! Pin 2/3 Short!" but I can't find ANYTHING ELSE on the board getting hot ANYWHERE.

Due to hassle factor of disassembling the heat-sink assembly, I cheat and use my flush-cutters to separate the VOUT pin from the board... DAMN. Short is on the load side, not internal in the 7805.  Now of course I'm wondering if this fault was preexisting or an emergent property related to my tinkery.

At this point I decide it's time to back slowly away from the bench and go rot my brain a little with some Doctor Who on Prime... and find THEY DON'T CARRY IT on this side of Niagara Falls.

FUCK. ME. 

mnem
*Kix The Beeb in the 'nads, just on GP*

[GreyWoolfe]

I chose to stay out of yesterday's discussion of scars and stitches because I don't want to rehash in my mind all the nasty injuries I had as a kid with glass, rocks, and deli meat slicers. 

The stories I could tell and the scars to prove it. 

Dayum you guys are clumsy...

 Only had 2 injuries needing stitches in my 49 and counting years, both to fingers and to the bone, 3 stitches in my left index finger (trimming lego caterpiller tracks down to cam chain at age 14), self applied butterfly stitch to my right pinky (received from the chipped rim of a glass cup I was washing out to have a cup of tea  ).

Never broken a bone.

Oh wait, I remember, I've had 3 trips to hospital for my eyes, 2 to have brass swarf tweezered out, and once for a scratched cornea. That last one was probably the most uncomfortable injury I've ever had, like a burning itch that was impossible to scratch...


EDIT: OT, I been at it again, got a Robin SmartPat 5000, and have a Fluke 1953A counter on the way, pics to follow   

3 injuries requiring stitches.  I had a scratched cornea once also.  I was standing up from looking on a bottom shelf in a supermarket and caught a tine of a 2 pronged fork across the cornea hanging  on a display hook attached to the shelf.  2 borked pinky toes, one on each foot and fractured a metatarsal/hyperextended a ligament going over on my ankle.  That was a walking cast for 6 weeks.  Separated ribs multiple times.  The last time was about 2 months ago.  At almost 63, I find I take longer and longer to heal from crap like that.  2 shoulder surgeries from playing sand lot football in college and I can't remember the cause of half the scars, that's probably a good thing.

You, sir, have obviously lead a quiet life.

[VK5RC]

A great quote said to me initially about water-skiing 'if you aren't falling over, you aren't trying hard enough'!

[med6753]

This NanoVNA thing is the nuts. Thanks again xrunner. And also TEA FTW.

Quick filter synthesis and test case.

So using normalised filter tables, I picked 0.1dB ripple Chebyshev parameters and put together a simple test (this is not going to be used!) 7 element low pass LC filter. HP35s was programmed to generate the L(n) and C(n) parameters from global variables for frequency of 7.5MHz and normalised impedance of 50 ohms. Tables are available on the Internet. Mine came from hardcopy of EMRFD. Used DaveCAD on back of an RS envelope for the design.



Much futzing later. Double sided board gouged and filter assembled. Used Peak LCR45 to check the toroids and they needed 1 turn less than toroids.info calculated.



So on to the NanoVNA it goes. Filter sweep looks promising!



So plugged it into the PC and got a segmented S21 sweep. That gives 500 points across 1-30MHz.



Pretty good!

Measurements from sweep:

Corner freq = 7.22MHz
3dB point = 7.8MHz (close enough)
40dB down = 17.22MHz
Insertion loss = 0dB
Max passband ripple = 0.2dB (close enough)

Bing-fucking-o!

I don't know what the freak you're talking about but if you're happy then I'm happy. Good show! 

[mnementh]

Now of course I'm wondering if this fault was preexisting or an emergent property related to my tinkery.

At this point I decide it's time to back slowly away from the bench and go rot my brain a little with some Doctor Who on Prime... and find THEY DON'T CARRY IT on this side of Niagara Falls.

FUCK. ME. 

mnem
*Kix The Beeb in the 'nads, just on GP*

Emergent property gets the kewpie doll. I'm too embarrassed to say just how I caused it; just glad I stepped away before I did any permanent damage.

mnem
I blame bd139; he made me say "penis fingers" a third time and they came for me.

[Specmaster]

You have to be very careful saying that, they can come and get you at any moment when you least expect them fingers to be around...W..H..A..M they got you 

[VK5RC]

Come on, be brave and 'fess up how you stuffed up!
My fav is how I can connect the red to the black wire. WTF

[Brumby]

Hey @Brumby want to raise the blood pressure of SHMBO 'some more' 

Needs some love but if it was closer I would find a place for it in the toy collection  eBay auction: #193188410543

Weight 17kg
30cm x 30cm x 80cm

This would have to become a piece of furniture!

.... and just about double the travel time for the 8658Bs he led me to.  It's next door to Lucas Heights which is famous for being the home of Sydney's only nuclear reactor (Used for nuclear medicine, sterilisation, etc. not power generation)


If Bean keeps this up, he'll have me travelling to Victoria!

[Brumby]

(Added to my Watch list.   )



   

[Neomys Sapiens]

@Brumby: Victoria? So he is still keeping the knowledge of the electronics scrapyard in Alice Springs from you?

[mnementh]

Come on, be brave and 'fess up how you stuffed up! My fav is how I can connect the red to the black wire. WTF

Ummm... Okay... 

To turn the unit on and unmute it, you need to pull pin 11 (STBY) and pin 12 (UNMUTE) LO on the controller interface. I accidentally tried to pull pin 6 (+5V) LO instead of (UNMUTE). And since I have zero assorted resistors and zero DSUB15 connectors, I did it with a twist-n-soldered "Y" of single 0.100" pigtails, because they were handy.

I absolutely deserved every little nip & blister that thing gave me... 

mnem

[beanflying]

(Added to my Watch list.   )

   

I didn't even have to twist your arm much 'this time' 

[VK5RC]

Come on, be brave and 'fess up how you stuffed up! My fav is how I can connect the red to the black wire. WTF

Ummm... Okay... 

To turn the unit on and unmute it, you need to pull pin 11 (STBY) and pin 12 (UNMUTE) LO on the controller interface. I accidentally tried to pull pin 6 (+5V) LO instead of (UNMUTE). And since I have zero assorted resistors and zero DSUB15 connectors, I did it with a twist-n-soldered "Y" of single 0.100" pigtails, because they were handy.

I absolutely deserved every little nip & blister that thing gave me... 

mnem

Mixing up pin 11 and 12 - that is NOT embarrassing,
I was thinking of something much more in my line of stupidity. 😜

[Specmaster]

Come on, be brave and 'fess up how you stuffed up! My fav is how I can connect the red to the black wire. WTF

Ummm... Okay... 

To turn the unit on and unmute it, you need to pull pin 11 (STBY) and pin 12 (UNMUTE) LO on the controller interface. I accidentally tried to pull pin 6 (+5V) LO instead of (UNMUTE). And since I have zero assorted resistors and zero DSUB15 connectors, I did it with a twist-n-soldered "Y" of single 0.100" pigtails, because they were handy.

I absolutely deserved every little nip & blister that thing gave me... 

mnem

Mixing up pin 11 and 12 - that is NOT embarrassing,
I was thinking of something much more in my line of stupidity. 😜

No no, he actually tried to pull pin 6 and then it bit him 

[Brumby]

(Added to my Watch list.   )

   

I didn't even have to twist your arm much 'this time' 

There is an unsettling trend developing here.....

Where's the next HP score...?!!



(Wait ... Did I just say that?)

[med6753]

An 8810A ongoing update. It took just under 2 hours to reach yesterday's cal point. Not unreasonable but the similar 8800A's reach that point in under .5 hours. But what IS encouraging is that it's been powered up for 5 hours now and once it reached that cal point it has NOT drifted any lower. 

I think I know why the change from last week's miserable total drift from 9.9974V down to 9.9964V but I'll wait and see once it has close to 10 hours on it.

My Solartron 7081 has a procedure for when it has been in storage for a while. You wrap it in a blanket, turn it on, and let it cook at 40C for a day. Apparently that evaporates any volatiles on the PCBs.

Maybe something similar is happening here.

I suspect part of it is exactly that since it had not been used in quite a while but I don't think it explains all of it.

Another 8810A update. I decided to leave it on all night into this morning. So it accumulated 25 total hours. After settling to set point at 2 hours into the run there was no further drift. Measured a steady 9.9969V with blips to 68 and up to 70. So why is it now stable and last week had unacceptable drift? I have a few assumptions with some suspected “smoking guns”.

First, as tggzzz posted the drift seen last week might have been the result of an extended “off time”. That's possible. This 8810A had been unused for several months. However, the nearly identical 8800A that I fixed last week had been sitting in the TEA closet for over a year and did not exhibit any appreciable drift once repaired. The 8810A and 8800A share the exact same PSU and reference circuit. In fact, I've been unable to determine so far where they are different. But I have noticed a difference in how they warm up from a cold start. Both 8800A's from initial power on read one or two minor digits low and then after about 20 minutes or so come up to set point and remain there. The 8810A does the opposite. From initial power on it reads one or two minor digits high then settles down after about 2 hours to set point then remains stable. I can't explain why the difference.

I also discovered that these old Flukes are extremely sensitive to proper and accurate adjustments. If they are the least bit off you'll have an instrument that's inaccurate and will apparently drift. 2 critical adjustments: Master Off-set and Bias seem to be required to be dead nuts or you'll have issues. On the 8810A they were only one or two minor digits off but seems to have made all the difference. And at the same time it apparently has validated my procedures and methods for calibrating these Flukes as sound and accurate.

So before I consider this one a wrap I installed the 8810A back into it case and going to let it sit or a day or so unpowered. Then bring it up again and do another burn-in and see what happens.

     

[mnementh]

   

MNEM

[med6753]

Are you suggesting sun spots and solar storms are the cause? 

[mnementh]

Oh, I would never, ever do that. Not unless it was absolutely, positively AMUSING.

mnem
 

[mnementh]

To turn the unit on and unmute it, you need to pull pin 11 (STBY) and pin 12 (UNMUTE) LO on the controller interface. I accidentally tried to pull pin 6 (+5V) LO instead of (UNMUTE). And since I have zero assorted resistors and zero DSUB15 connectors, I did it with a twist-n-soldered "Y" of single 0.100" pigtails, because they were handy.
I absolutely deserved every little nip & blister that thing gave me... 

mnem

Mixing up pin 11 and 12 - that is NOT embarrassing, I was thinking of something much more in my line of stupidity. 😜

No no, he actually tried to pull pin 6 and then it bit him 

Looking back on the incident over my morning , the funny part is... my little "Y" adapter was made of cutoff servo wires, so WRB color coding. And clearly playing a subtle joke on both of us, while I made sure the Black wire was plugged into pin 9 (GND), Murphy made sure the one I plugged into pin 6 (+5V) was the Red wire. The White wire on pin 11 (STBY) wisely stayed out of the shenanigans... 

So in the most literal sense I did, in fact, hook the Red wire up to the Black.

What I find interesting is that I was able to get output on all channels even with (UNMUTE) not pulled LO. I guess the digital tone/gain/compensation control circuitry isn't going to do much except sit there looking stupid with its (Vcc) somewhere around (+0.8V) instead of (+5V), though. Or that (+0.8V) is little enough that the analog TDA7296 ICs saw it as LO across both pins, so turned on as usual. 

mnem
*Raises a mental toast to the hardy little LM7805 and 100Ω SMD resistor that proved "tolerant" of such abuse as dropping approx. 24.5V across them for several minutes while I  poked & prodded the amplifier's innards*

[mansaxel]

Ok, made an aggressively low offer on a Philips PM 6662 counter, in a pretty sorry, albeit claimed operating, state. That did not bite, but scored a counter-offer. Down to 30% of asking. I had offered 5% of asking. I countered with 7,5% and that was accepted. Just paid. Shipping inside EU so could be fast.

[AVGresponding]

You, sir, have obviously lead a quiet life.

Au contraire meine kamerade, I just have exceptionally good spatial awareness, a good eye for details, and razor sharp reactions.

Which is not to say I haven't had a plethora of minor nicks and scrapes, which I just let bleed to wash the crap out. They usually clot within a few minutes, which reminds me of this classic comedy clip starring Dirk Bogarde and James Robertson Justice, "What's the bleeding time, boy?".

Also had more than a few hair-raising near misses, but that's another story.


Some pics now:

Made in Murica!

This is a complete unit that I stripped those mercury switches from.

Sadly, this is microfarads, not millifarads   


What you don't want to see when you arrive to test a fire alarm system:

This one's a real glitchy sum-bish, the system has been configured incorrectly imo, but it's NMP except insofar as it's prone to crashing at the end of a full test. Luckily today was only a partial test, so I didn't have to spend an extra half-hour resurrecting it at the end...
 

[med6753]

Oh, I would never, ever do that. Not unless it was absolutely, positively AMUSING.

mnem
 

Then you would have to include quarks, kryptonite, and moonbeams. 

[mnementh]

Oh, I would never, ever do that. Not unless it was absolutely, positively AMUSING.

mnem
 

Then you would have to include quarks, kryptonite, and moonbeams. 

Don't forget random      incursions... 

mnem
Oh, and Trump Hair:   

[mnementh]

Another Installment in the Saga of the $9 Process Meter...



After waiting since Sept 27, and absolutely ZERO activity since Oct 22, suddenly SOMETHING is in Canadian Customs... 

mnem