Test Equipment Anonymous (TEA) group therapy thread

[Specmaster]

Hopefully, the notification service of postings being made will be back up and running next?. Judging by the lack of postings here now, I guess that there was only a handful of us making regular checks to see if the server was back on line?

[AVGresponding]

Is this enough fine control of wattage...? They make these for putting lots of cheap LED floodlights in the garage; I've used one with appliance bulbs to fine-adjust max current on my dim-bulb tester. Costs $10-15 on Amazon: https://www.amazon.com/s?k=bulb+splitter+7+in+1

Interesting ! 

However no it still does not cut it. It makes it more practical to adapt the wattage to the DUT, but still is an on/off safety device. I was talking about fine control as far as ramping up the voltage slowly. A dim bulb tester is no help for this, no matter how fancy one makes it.

Yeah, I'd still want one of these inline with a variac, myself. A lot of the time you'll find that once things start turning on in a SMPS, especially the older designs, that's when a short appears. The scariac doesn't stop the magic smoke from coming out then, but rather encourages the switcher transistor to try and sink more current to make up for the lower voltage.

mnem
 

I don't think I'd try to soft start an SMPS with a variac; an SMPS is a constant power device, it'll just draw more current at lower voltages...

[AVGresponding]

Hopefully, the notification service of postings being made will be back up and running next?. Judging by the lack of postings here now, I guess that there was only a handful of us making regular checks to see if the server was back on line?

Personally I have notifications switched off, otherwise I'd be bombarded with pings and bongs.

Hopefully from now on they remember not to turn on the kettle at the same time as the toaster, down at the server farm.

[psykok]

An author kind of TEA arrived in my lab.


Audio Analyzer are like DMM, you always need one more  but this one is special

This audio Precision System One is quite old (1983) but still a reference in the domain.

[Robert763]

What was I saying about using "the cloud" for everthing last week....

And "the cloud" is immune to fires in the server building?

McBryce.

My point was it is a bad thing, it was in relation to having a micro SD card in a phone for local storage and backup.

[nixiefreqq]

for a dso always thought my hp 54601a did a really good job with xy mode.


my just arrived 54622 is much better.


(damn firemen in utah are responsible for the boredom which caused me to buy another redundant dso.  really like it however.)


edit   have been in server rooms after a fire.  never saw any gear actually ruined by fire or smoke......but the water damage was always pretty bad.  stuff built to mil stds you could usually pull and reuse (like crypto).  we usually got there just a couple of days after the fire so the mold was not too bad yet (but there was always a stink)

[bd139]

What was I saying about using "the cloud" for everthing last week....

The cloud would have survived this. This would have been an "availability zone failure" which is easy to architect around. In fact I spent a week about 2 weeks ago actually just pulling plugs metaphorically speaking with no side effects at this level.

What happened here is "lowest bidding hoster had a massive shit show" not "the cloud".

And just remember that the guys running the "lowest bidding hoster" know more than the average person.

Your stuff is safer in the cloud. If you do it properly.

And the data security argument is void. What you post on a public forum is public. What you post in a private S3 bucket in AWS is not.

Edit: I notice the forum is running like shit now as well. Another problem easily resolved by upping the instance size transparently rather than buying another physical machine 

[med6753]

Edit: I notice the forum is running like shit now as well. Another problem easily resolved by upping the instance size transparently rather than buying another physical machine 

It sure is. Herky Jerky and sometimes bad Gateways.

[McBryce]

What was I saying about using "the cloud" for everthing last week....

And "the cloud" is immune to fires in the server building?

McBryce.

My point was it is a bad thing, it was in relation to having a micro SD card in a phone for local storage and backup.

Ah, ok. I've (the company I worked for) already had massive problems that cost millions due to lost cloud data.

McBryce.

[bd139]

ALB + Auto scale group + baseline EC2 instances across different AZ's + RDS w/ AZ failover = self healing win.

Anyway, fun time...

5384A PCBs. OCXO module actually fits the footprint I designed properly (thank feck). Will populate this evening.



Crystals for SA 2nd IF bandwidth filter and second IF downconverter, plus GPIB fun coming (to match all the crystals)

[bd139]

Ah, ok. I've (the company I worked for) already had massive problems that cost millions due to lost cloud data.

I've read about 50 case studies where that happens and it's always someone doing something stupid, a low ball provider, or PaaS, or Atlassian 

IaaS cloud = win.

[Cerebus]

What was I saying about using "the cloud" for everthing last week....

[beanflying]

Tektronix TEA Slide night anyone  eBay auction: #124596338613 or a  fest?

[xrunner]

Well, we're still running!

Nice pic, looks cleaner for sure, brand new. Looks like you trained your camera well, it knows its stuff now : it automagically focused the picture on the trouble maker : that yellow dipped tantalum cap ! Good camera, good camera ! 

OK Vince, now with your keen eye you pointed out a tantalum cap in one of my pics of the hp E3611A - a trouble maker as you stated. That happens to be the only one in the parts list - a 6.8 uF 35V.

Now, that happens to be a 3-legged tantalum. Works the same as a 2-legged. They made them so that they couldn't be installed backwards. Although there are no polarity markings on it, I checked the schematic and the center pin should be positive. Also indicated by a voltage check too.

I'll let you decide, since you are appreciative of vintage test equipment - do you say it's OK to put in a 3-legged part in it's place, just to keep it original.

Thanks.

 

[Cerebus]

Is this enough fine control of wattage...? They make these for putting lots of cheap LED floodlights in the garage; I've used one with appliance bulbs to fine-adjust max current on my dim-bulb tester. Costs $10-15 on Amazon: https://www.amazon.com/s?k=bulb+splitter+7+in+1

Interesting ! 

However no it still does not cut it. It makes it more practical to adapt the wattage to the DUT, but still is an on/off safety device. I was talking about fine control as far as ramping up the voltage slowly. A dim bulb tester is no help for this, no matter how fancy one makes it.

Yeah, I'd still want one of these inline with a variac, myself. A lot of the time you'll find that once things start turning on in a SMPS, especially the older designs, that's when a short appears. The scariac doesn't stop the magic smoke from coming out then, but rather encourages the switcher transistor to try and sink more current to make up for the lower voltage.

mnem
 

I don't think I'd try to soft start an SMPS with a variac; an SMPS is a constant power device, it'll just draw more current at lower voltages...

Once it's up and running, yes. But most of the inrush current at start-up is charging the DC rail capacitor(s), that is going to be voltage dependant (a simple lumped component model would be an RC circuit and some diodes). Somewhere there's a crossover point where the steady state current exceeds the inrush current.

It's all theoretical though, because for most off-line SMPS the minimum voltage that it will start up at is going to be pretty close to the rated minimum input voltage. If you've got an off-line SMPS rated for 90-265 V rms input the wiggle room at the bottom end is going to be pretty small because at some point you'll hit the duty cycle limit of the buck regulator - despite the usage of every bloody sports commentator and every overblown drill sergeant in the world, you can't "give it 200%".

[Cerebus]

And the data security argument is void. What you post on a public forum is public. What you post in a private S3 bucket in AWS is not.

Unless you you leave your S3 bucket credentials on Github. 

[GreyWoolfe]

Yeah I'd buy one but not for that. Main advantage is it has RS232 which means no PITA GPIB adapter required.

So you're saying the fact of GPIB-only is the reason the 3478A gets no respect...?

Seems way overpriced to me, considering a working 3478A can be had for ~100 quid almost any day; less if you can be patient and shop carefully.    Shameful, actually; for some reason the 3478A just gets no respect.

Is it just the 3478A, or any HP gear of that generation, ie with that horrible tiny hard to read LCD that almost nobody likes ? LED or VFD gears are so much easier to read, from a distance, at any angle, whatever the lighting conditions dark or bright.

Maybe if they had made a much larger LCD with digits 3 times as tall, fatter segments, and a bright and even backlight... people would value them more ? Still,  you would still have the viewing angle problem.which neither LED or VFD suffer from.

At any rate this LCD is a killer for me personally. I would rather get an older gear with LED displays (still small but easy to read and no viewing angle problem), or save some more money and get a newer generation with VFD. 

Having lived with one everyday for a while... the transflective LCD on the 3478A is very good for the breed. Viewing angle is quite wide, and unless you like to work on a dark bench, the fact of passive display is a total non-issue.

I think the problem is

a) not sexy; everything aboot it screams "dumptruck" design ethos.
2) general ignorance aboot B/L vs non-B/L display; folks pre-judge it and just say "pass..." because they've never actually had one.

Bottom line is the 3478A is one of the most straightforward UI, generally easy-to-use meters ever made. If not for the fact they're proportioned like a Dachshund (dog and a half-long; half a dog wide) they'd be the perfect all-around bench meter. 

mnem
 

I have 2 of them.  I have more than enuf light on the bench that viewing is never an issue.  Also paid less than 100 US bux for each of them and checking them on a DMM Check + shows good enuf for me.  I like them.  I also have 2 3466's sitting next to them.

[Robert763]

Well, we're still running!

Nice pic, looks cleaner for sure, brand new. Looks like you trained your camera well, it knows its stuff now : it automagically focused the picture on the trouble maker : that yellow dipped tantalum cap ! Good camera, good camera ! 

OK Vince, now with your keen eye you pointed out a tantalum cap in one of my pics of the hp E3611A - a trouble maker as you stated. That happens to be the only one in the parts list - a 6.8 uF 35V.

Now, that happens to be a 3-legged tantalum. Works the same as a 2-legged. They made them so that they couldn't be installed backwards. Although there are no polarity markings on it, I checked the schematic and the center pin should be positive. Also indicated by a voltage check too.

I'll let you decide, since you are appreciative of vintage test equipment - do you say it's OK to put in a 3-legged part in it's place, just to keep it original.

Thanks.

 

The 3 leg tant's are made for low impedance, not being reversable. Unless it has more then 25V DC on it I'd let it stay where it is.

[bd139]

And the data security argument is void. What you post on a public forum is public. What you post in a private S3 bucket in AWS is not.

Unless you you leave your S3 bucket credentials on Github. 

You've probably done something wrong if that matters ...

1. Your access credentials shouldn't survive long enough for that to be an issue (temporary creds always)
2. IAM role should have a trust policy saying where you can access the bucket from (least priv)
3. Your account shouldn't have the ability to do that anyway or make the bucket public (least priv)
4. Require MFA for console access so if you lose your login, meh (mfa)
5. Actually look at the complaining Trusted Advisor does (audit/analyse)

If you skipped all that then it's your own fecking fault 

[xrunner]

The 3 leg tant's are made for low impedance, not being reversable. Unless it has more then 25V DC on it I'd let it stay where it is.

Oh OK, I never ran across that explanation, thanks.

I'm going to check the max voltage it can have (it varies according to the output voltage setting.)

[Robert763]

Just snagged a tidy Fluke 8060A on ebay for £42.20 inc post 
The faulty "display fades" does not bother me. either a dirty zebra strip or dodgy capacitor. Should not be too hard to fix and everybody needs at least 3 8060As
www.ebay.co.uk/itm/Fluke-8060A/154397971430

[Cerebus]

Now, what was I saying before I was interrupted by someone sneeking off behind the generator shack for a quick fag? Oh yes, OCXOs.

So, my batch from what turned into a distributed group purchase took 9 days door to door. Ordered two and like BD, got twice as many for my money.

Desoldering turned out to be relatively easy for a part on a 10 layer board - just desoldered each pin with Chem-Wik (that stuff's good) and went around the board with a small screwdriver for a lever and gave each pin a lift while heating it with the old Metcal. No application of 60/40, chip-qwik or anything similar needed.

All four survived the ordeal of having the board next to them violently cut and being desoldered. Here's how they all performed:



Sorry for the laziness of a screenshot, but converting to SMF's table notation is too much of a fag. Readings taken with my TTi TF930 that was last adjusted against GPS a year ago.

All with enough range left to be steerable to 10.000 000 00 MHz at an appropriate EFC voltage with room to spare for future ageing (i.e. GPSDO fodder). None good enough to be used "out of the box" as a 10 MHz reference without trimming, unless you were more interested in stability than accuracy.

I tried a few random selections for stability, using another known good uncalibrated 10 MHz OCXO as reference input to the TF930, and they stayed stable at the limit of the counter's resolution (10 mHz or 1 ppb) for as long as I was prepared to wait - several hours in one case.

Conclusion: £13.90 well spent.

[Brumby]

Ah....  We're back!

That was a long few days.

Unfortunately, SWMBO noticed my lack of focus and I fell into a strange world activity which I had not experienced for such an extended period of time.  Damn I got a lot of stuff done around the house!

I fear this may have repercussions....

[bd139]

Just snagged a tidy Fluke 8060A on ebay for £42.20 inc post 
The faulty "display fades" does not bother me. either a dirty zebra strip or dodgy capacitor. Should not be too hard to fix and everybody needs at least 3 8060As
www.ebay.co.uk/itm/Fluke-8060A/154397971430

Very nice  . Definitely my favourite meter so far those are.

[bd139]

Now, what was I saying before I was interrupted by someone sneeking off behind the generator shack for a quick fag? Oh yes, OCXOs.

So, my batch from what turned into a distributed group purchase took 9 days door to door. Ordered two and like BD, got twice as many for my money.

Desoldering turned out to be relatively easy for a part on a 10 layer board - just desoldered each pin with Chem-Wik (that stuff's good) and went around the board with a small screwdriver for a lever and gave each pin a lift while heating it with the old Metcal. No application of 60/40, chip-qwik or anything similar needed.

All four survived the ordeal of having the board next to them violently cut and being desoldered. Here's how they all performed:



Sorry for the laziness of a screenshot, but converting to SMF's table notation is too much of a fag. Readings taken with my TTi TF930 that was last adjusted against GPS a year ago.

All with enough range left to be steerable to 10.000 000 00 MHz at an appropriate EFC voltage with room to spare for future ageing (i.e. GPSDO fodder). None good enough to be used "out of the box" as a 10 MHz reference without trimming, unless you were more interested in stability than accuracy.

I tried a few random selections for stability, using another known good uncalibrated 10 MHz OCXO as reference input to the TF930, and they stayed stable at the limit of the counter's resolution (10 mHz or 1 ppb) for as long as I was prepared to wait - several hours in one case.

Conclusion: £13.90 well spent.

Excellent post