Test Equipment Anonymous (TEA) group therapy thread

[grizewald]

Thank you, now if all of our meters were analogue, we would each and every one of us argue that they were all bob on, am I right? Even with Tins new toy, they would all be differant.

Yup.

When you can measure to 6 1/2 to 7 1/2 digits, the only way out is to get your own Josephson junction reference, something to measure it with, and to throw away anything which won't agree!



I added some more TE to the collection since I was here last. Since I took the photo, I also added a GPSDO and another OCXO which will go in the 5334A. Time is much easier to get that "agreement to n zeroes" feeling with.

EDIT: I notice that the GPSDO in question is actually sitting on top of the stack there. The photo was taken before I got a GPS antenna which could actually receive enough satellites in my urban canyon to get a lock.

[grizewald]

Hi Grize,

What's the meter on top of the Voltage standard? Looks like it's the boss

Dek.

It's a Prema 4000. An excellent but not so well known German multimeter based on the LM399 reference.

[grizewald]

Hi Grize,

What's the meter on top of the Voltage standard? Looks like it's the boss

Dek.

It's a Prema 4000. An excellent but not so well known German multimeter based on the LM399 reference.

It came with a dead insect of unknown variety and age on one of the main smoothing capacitors. Quite how the insect got in there, I have no idea as the case is pretty much sealed.



There were also some interesting flux residues left over from a previous repair:



It's a great piece of kit though. Lots of hand drawn tracks and guard traces:

[worsthorse]

Specmaster is right on the mark with his comments.



There's a maximum disagreement of 11 microvolts from the nominal 0.01V that the Time reference is set to. It bugs me much more than it should. What bugs me even more is how the disagreement changes as I change the reference output to step through the various ranges on the meters. I'm almost resigned to never knowing who is right and who is wrong.

The only way out is to fly to Taiwan and steal TiN's new toy!

(With apologies for the extended absence. I see mnem has turned into a Canadian while I was away.  )

I have a simple cure for all this fifth and six digit madness:



Simply apply over the offending digits one-by-one until all the meters match.   

_______
hillybilly? me? just the part from the holler...

[madao]

Actually , i repair  TDS540  with less capacitor plague.

It looks good, CPU Board no fail  despite empty batterie NVSRAM.
Acquisition board has minor fail,  shut up , because i must sleeping.

[grizewald]

That was a labour of love madao! Were all those capacitors really bad?

[Specmaster]

Actually , i repair  TDS540  with less capacitor plague.

It looks good, CPU Board no fail  despite empty batterie NVSRAM.
Acquisition board has minor fail,  shut up , because i must sleeping.

Whats your technique for removing these capacitors because those little bastards always give a bit of grief getting them out.

[Cubdriver]

Actually , i repair  TDS540  with less capacitor plague.

It looks good, CPU Board no fail  despite empty batterie NVSRAM.
Acquisition board has minor fail,  shut up , because i must sleeping.

Whats your technique for removing these capacitors because those little bastards always give a bit of grief getting them out.

I watched a video done by Mr. Carlson repairing a Tek scope with these surface mount can caps, and he was just grabbing them with a pliers and twisting them off their leads, then desoldering the lead stubs remaining on the board.  I've tried it a time or two since with good results, though it does still make me a tiny bit nervous about tearing up pads.  So far so good, though, so that concern is decreasing with experience.

FWIW.

-Pat

[grizewald]

I like that idea Pat. The bond between the track and PCB is never going to be stronger than it is then.

[WastelandTek]

I watched a video done by Mr. Carlson repairing a Tek scope with these surface mount can caps, and he was just grabbing them with a pliers and twisting them off their leads, then desoldering the lead stubs remaining on the board.  I've tried it a time or two since with good results, though it does still make me a tiny bit nervous about tearing up pads.  So far so good, though, so that concern is decreasing with experience.

FWIW.

-Pat

I saw that, seemed to work surprisingly well, and Paul hasn't steered me wrong yet.



sigh, the forum is modifying the link and I can't make it work for the life of me

[Specmaster]

Actually , i repair  TDS540  with less capacitor plague.

It looks good, CPU Board no fail  despite empty batterie NVSRAM.
Acquisition board has minor fail,  shut up , because i must sleeping.

Whats your technique for removing these capacitors because those little bastards always give a bit of grief getting them out.

I watched a video done by Mr. Carlson repairing a Tek scope with these surface mount can caps, and he was just grabbing them with a pliers and twisting them off their leads, then desoldering the lead stubs remaining on the board.  I've tried it a time or two since with good results, though it does still make me a tiny bit nervous about tearing up pads.  So far so good, though, so that concern is decreasing with experience.

FWIW.

-Pat

Yeah, I saw video and I must say that I tried it, it did rip a pad right of board, so clearly some boards are more equal than others and I've not had the courage to try it again.

[Cubdriver]

I watched a video done by Mr. Carlson repairing a Tek scope with these surface mount can caps, and he was just grabbing them with a pliers and twisting them off their leads, then desoldering the lead stubs remaining on the board.  I've tried it a time or two since with good results, though it does still make me a tiny bit nervous about tearing up pads.  So far so good, though, so that concern is decreasing with experience.

FWIW.

-Pat

I saw that, seemed to work surprisingly well, and Paul hasn't steered me wrong yet.



sigh, the forum is modifying the link and I can't make it work for the life of me

It's working for me.  Ha, I was just going to edit my post to add the link.  You beat me to it.

-Pat

[bd139]

It’s all ok until a pad comes off or you break a .4mm trace next to the pad. This might not show up until you get some weird instability because the decoupling network is shagged. And on cheap PCB laminate it’ll fuck it instantly.

The correct way is soldering tweezers. Or just use two irons (this is what I do, the second being an antex C15).  Also don’t use hot air as the pads run under the capacitor too far and you can’t heat it up too much or it’ll blow up.

Honestly it’s such shit bodger advice that he discredits himself.

[bsudbrink]

It came with a dead insect of unknown variety and age on one of the main smoothing capacitors. Quite how the insect got in there, I have no idea as the case is pretty much sealed.

Don't worry, there was just a little bug in the design.

[Zucca]

Got the Convoy UV Light, after one month on the china boat DOA.

Never saw something like this before, when the current in the LED is > 50mA is start to flash with about 1.3Hz frequency. 

Full refunded immediately 5 sek after with aliexpress.

[VK5RC]

For a while I headed down the 'volt-mutt' rabbit hole, and still have a leg in there to tell the truth.
I ended up with a 3458 and about a dozen ltz10000 references of different construction. I am still doing a long term experiment with the references.
Reliable and reasonably accurate and repeatable voltage readings down to the 10s of uV (in the 10V range) are pretty achievable and 6 1/2 dmm can get there, eg my 34461A. Those meters are I think very good value for money.
Below that starts getting crazy,
1 Most LTZs do drift esp early on (month or so) and I suspect so do all the other components. Of my voltage references, two look pretty good, ie not much drift, lowish temp co-efficient.
2 Room/lab/leads etc starts getting crazy, a EMI noisy laptop can blow readings in my set up by many uV.
Volt nuttery is and was for me a learning experiment - in all seriousness I can't recall ever having to measure a voltage better than a few 10s of mV.
The other issue with 'calibration' is that it is not a static or one-off event - everything drifts and that needs to be a 'known', hence most reference meters have a 'calibration cycle'. Part of my reasoning with so many volt references was to be able to reduce the cost of repeated formal (expensive - and risky - shipping a 3458 800km) calibration.
Rob

[grizewald]

So while I was away, I became afflicted with a severe case of TN syndrome. This led to extensive ESI and ultimately, the first steps along the road to more serious problems.



Here's how my addiction started....

First I found an HP 3325A (one of five) in apparent good condition at a reasonable price. When it turned up, it looked like it spent its life in a rack in a clean environment. The rear dust filter was immediately replaced and I must find a replacement for the fan which sounds like a small airport has moved in to my apartment when the machine is powered on. It looks great though!



The insides were equally as clean but also revealed that it had the OCXO option fitted. Win!



The frequency synthesiser deserved a partner and my unremitting ESI found me one...

(continued in next post so as not to break the max images per post limit)

[grizewald]

The logical partner was the HP 5334A.

This machine had a harder life then the first. Not only was it filthy, it had accumulated a third of a century's worth of calibration stickers!
(seller's image)



It looks like the people at the calibration labs had tried removing the old ones at first, but had given up in the end. The adjustment points at the back of the case and every screw for removing the case had those white, ultra fragile "Calibration void if broken" stickers on them. 

Nothing withstands Label-off, iso and a man with a serious case of TEA!!



Inside, it's perfect. My worst fear was that the backup battery had leaked, but apart from a dried mark on the top of the completely dead battery pack's case, the PCB was pristine. Love the star grounding.





This is the utterly crap TCXO which it gets as standard:



It wanders around 10MHz like someone who's drunk eight pints of beer. It doesn't matter how long you have the counter turned on; three of the least significant digits are about as useful as a ..... notusefulthing  I leave the analogy to the reader.

A counter with a crap reference being kind of pointless deepened the TN which had gripped me and let to extensive bouts of both google flu and ESI. The end of the story (so far) in the next post...

[grizewald]

The obvious fix for the 5334A was an OCXO. Finding one took a while, particularly one where I could find some clues as to whether it would match with the voltages available on the OCXO socket on the counter's PCB.

The search culminated in this:



and how to power it via the completely anonymous pins on the back.

Of course though, I could not have any idea how accurate the counter or the new OCXO or even the new synthesiser was.

When it comes to the VN side of the TEA addiction, it's those poor guys who have the most expensive addiction. I realised that mine could be solved by the US military! I found a BG7TBL PLL GPSDO and bought that as my reference.

Unfortunately, the ceramic antenna they provided gave me horrible S/N ratios from the satellites it could see. Even though the apartments where I live are only three stories, they and the trees create a really bad urban canyon when it comes to receiving GPS signals. Putting an antenna on the roof is out of the question. So I bought a small quadrifilar antenna from Aliexpress.

After much waiting and paying Postnord's ransom to receive my import package, the antenna finally arrived. I hung it out of the window in totally the wrong orientation - on a north facing window - and got GPS lock from cold in around one minute!

So I printed a mount for it, used a proper window pass through and stuck it on a pole on my east facing balcony. The GPSDO has been locked and stable since then with no alarms. The final configuration was to use the GPSDO as the reference for the counter and calibrate my OCXO against that. The OCXO in the 3325A is .3Hz slow and stays that way. The Ovenaire OCXO is 0.3Hz high since I last nudged it to agree with the GPSDO, but by tomorrow, after being on since this evening, I expect it will show exactly 10MHz tomorrow morning.

Here is the 5334A measuring a 10MHz 250mv P-P sine wave from the 3325A:



and now measuring the Ovenaire OCXO (powered by the PL320 supplies):



It's dropped to ...0.2 while I was writing these posts.

So finally, I can have time to at least the accuracy of my GPSDO which also has an OCXO of unknown virtue. It's good enough for most things I need to do!

There's a problem with both machines though - humming transformers. Am I right in saying that the best thing to do with them is take them out, dump then in a bucket of resin, apply as much vacuum as I can generate, then remove and allow the resin to cure?

I've not tried it, but I can imagine that dribbling superglue into my laminations will only get me a transformer as a borg-like addition to my hand?

[tautech]

.....................
There's a problem with both machines though - humming transformers. Am I right in saying that the best thing to do with them is take them out, dump then in a bucket of resin, apply as much vacuum as I can generate, then remove and allow the resin to cure?

50 vs 60 Hz transformers ?
Also I'd be checking the main smoothing caps haven't dried out or are just old and tired. 

[bd139]

Mine hummed. This improved a bit when I took it completely to bits to do the RIFAs in it. You did replace them didn’t you? 

Basically tighten the big screws holding it down and it’ll compress the laminations some more.

Edit: and oh yes the built in TCXO is total crap. Very disappointing. My 5316 which has the basic crystal oscillator is more stable!

[WastelandTek]

Is that a Bodnar pulser I see there on your bench grize?

[Cerebus]

It’s all ok until a pad comes off or you break a .4mm trace next to the pad. This might not show up until you get some weird instability because the decoupling network is shagged. And on cheap PCB laminate it’ll fuck it instantly.

The correct way is soldering tweezers. Or just use two irons (this is what I do, the second being an antex C15).  Also don’t use hot air as the pads run under the capacitor too far and you can’t heat it up too much or it’ll blow up.

Honestly it’s such shit bodger advice that he discredits himself.

And there was me expecting you to say "Rubbish! What you need is a really small shaped charge."

[WastelandTek]

what do you guys think about this one?  $167 delivered

https://www.ebay.com/itm/HP-3325A-Synthesier-Function-Generator-w-Power-Cable/254467450725?epid=1505466257&hash=item3b3f712765:g:VysAAOSwzwReCkQH:sc:FedExHomeDelivery!98408!US!-1

[Kosmic]

I watched a video done by Mr. Carlson repairing a Tek scope with these surface mount can caps, and he was just grabbing them with a pliers and twisting them off their leads, then desoldering the lead stubs remaining on the board.  I've tried it a time or two since with good results, though it does still make me a tiny bit nervous about tearing up pads.  So far so good, though, so that concern is decreasing with experience.

FWIW.

-Pat

I saw that, seemed to work surprisingly well, and Paul hasn't steered me wrong yet.



sigh, the forum is modifying the link and I can't make it work for the life of me

Please, never, never, never try to do that. You are going to break a board. Can't believe he did a video recommending this technique. You could also try to throw the board on the ground until the cap come off ?

Like bd139 recommended, use tweezers, 2 irons. Also hot air (with controlled temp) was always working fine for me.