Test Equipment Anonymous (TEA) group therapy thread

[med6753]

]I would be concerned aboot that connecting the two transistor cases both thermally and electrically in a manner that affects the deliberate isolation provided by that BeO disc.[/size][/b][/i][/color]

mnem
 

Gorilla glue is an insulator.    

Nevertheless  I should check they aren't somehow electrically connected.

I wouldn't trust Gorilla glue to be an insulator. In fact, my experience with Gorilla glue is I don't even trust it to be a decent glue. 

[Andrew_Debbie]

]I would be concerned aboot that connecting the two transistor cases both thermally and electrically in a manner that affects the deliberate isolation provided by that BeO disc.[/size][/b][/i][/color]

mnem
 

Gorilla glue is an insulator.    

Nevertheless  I should check they aren't somehow electrically connected.

I wouldn't trust Gorilla glue to be an insulator. In fact, my experience with Gorilla glue is I don't even trust it to be a decent glue. 

runs off to check  ....

[med6753]

The Type 547 intermittent sweep issue. The horizontal deflection plates are on the bottom of the CRT behind the vertical amp board and completely out of view. It was strictly by feel. And sure enough, checking the leads one of them fell right off the CRT pin. I was able to get it back on and seat it firmly. No more wonky sweep.

That was the last of the issues with this scope. I brought this scope home on March 17th. It has taken until today to fix the final fault. Longest project I've ever had and I'm glad it's finally done.



And I have to revisit this guy. The intensity changes all by itself and sometimes goes out of focus. I suspect a bum connection but we'll see. A technician over on the Facebook Old Tek Scope group worked for Tektronix in a field office and he said no one liked working on these mini scopes. A PITA to service and reassemble. And I certainly agree. But I have no choice but to pull it apart again. 

This is why I do extensive burn-in's after restoration/repair. To shake out all this crap.


   

[mnementh]

]I would be concerned aboot that connecting the two transistor cases both thermally and electrically in a manner that affects the deliberate isolation provided by that BeO disc.

mnem
 

Gorilla glue is an insulator.     Nevertheless  I should check they aren't somehow electrically connected.

I wouldn't trust Gorilla glue to be an insulator. In fact, my experience with Gorilla glue is I don't even trust it to be a decent glue. 

runs off to check  ....

Ummm... what are you going to use to test with...? Do you have another meter with 100Gohm input impedance...?

Yes, I get that is only supposed to be guaranteed on the highest voltage range, but it seems to me the entire frontend PCB has to be designed with that in mind. 

EDIT: I certainly wouldn't trust Gorilla Gloo or thermal epoxy to have the same insulative qualities as needed here... that BeO disc is a pretty exotic material and was chosen for a reason. Even just changing the mass of the junction will affect thermal propagation, and therefore also will affect thermal loss to the surrounding environment.

mnem
 

[Andrew_Debbie]

~35MΩ   case Q3 to case Q5.     

[Andrew_Debbie]

measured in-circuit   

Not going to desolder unobtainium transistors.

[Andrew_Debbie]

Yes, I get that is only supposed to be guaranteed on the highest voltage range, but it seems to me the entire frontend PCB has to be designed with that in mind. 

mnem
 

It is 100,000MΩ on .2, 2, 20 and 200V DC ranges.      Only .2 and 2 go directly to Q5.   20 and 200V connect to the high voltage buffer, which uses the +/- 300V supply rails.

[mnementh]

Uhhh-huh. And you're certain that this will not affect the 100Gohm-rated portions of the circuit?

If I'm understanding this correctly, We're talking electrometer-grade voltages here... where waving your hand too close can affect readings.

mnem

[med6753]

measured in-circuit   

Not going to desolder unobtainium transistors.

If they are standard designation "2N"  and not some "in house" number you'd be surprised. Someone on Ebay has got them. Every old 2N transistor I needed for the Type 547 project was out there. 

[nfmax]

]I would be concerned aboot that connecting the two transistor cases both thermally and electrically in a manner that affects the deliberate isolation provided by that BeO disc.[/size][/b][/i][/color]

mnem
 

Gorilla glue is an insulator.    

Nevertheless  I should check they aren't somehow electrically connected.

I wouldn't trust Gorilla glue to be an insulator. In fact, my experience with Gorilla glue is I don't even trust it to be a decent glue. 

Not particularly Gorrilla-y either

[factory]

Small box of IC's & other bits, sadly I've worked out what the board with 9 readouts came from  , I'll see if anyone else can work it out.



And finally, a very dead Racal counter, readouts & some boards missing, but they left the crystal & the vacuum-filled-transistor-bulb* in the front end.

*Blame shango for this incorrect name.

David

A counter of some sort, beyond that I have no idea.

"vacuum filled"? How the fuck can anything be "filled" with a vacuum?   

It's not from a counter, maybe I should give a clue, it not from TE and I should probably mention they were 10 bulbs fitted originally.

Well Shango does like to give things daft names, this video at approx 3:50 minutes, he calls them vacuum-packed-transistor-bulbs  https://youtu.be/l63SJfQJ3O0?t=230 , would take too long to find the other one (if I got it right).

David

A 10 digit calaculator?

Correct, just need to guess which one....

David

Not a Sumlock ANITA by any chance?

Yes it's from an Anita calculator.

David

This picture would have given it away too easily, the board is marked Bell Punch Co Ltd, I can't confirm exactly which Anita is was from, but it's from one of the 1000 series. http://anita-calculators.info/html/transistors_-_integrated_circuits.html


And to make things worse, the Hanger where the Hack Green sale takes place has one of these calculators on display, it wasn't turned on this year, but here is a picture from the previous sale in 2019.


David

[mnementh]

measured in-circuit   

Not going to desolder unobtainium transistors.

If they are standard designation "2N"  and not some "in house" number you'd be surprised. Someone on Ebay has got them. Every old 2N transistor I needed for the Type 547 project was out there.

Yeah, in this case they are a "selected pair", so do have a in-house part number and will be calibrated after install.

I understand wanting to disturb this "calibrated system" as little as possible... which is a large part of why I don't understand the apparently cavalier attitude in choice of adhesives and is the main reason my brain goes at the thought of encapsulating it in epoxy.

mnem
 

[edfarrington93]

Talking about Tek scopes in 2021.

I got two quotes for my company at work.
One is $153K (6 Series) the other one (5 Series) is about $140K.

Problem is in both over 70% of the money are just SW options... go piss a rope Tek!

Jeezus! That's crazy, i didn't realize how pricey some of the new gear is. I checked the website and they have several options that are 10 grand extra. I hope there is some groundbreaking engineering inside those things for that kind of coin.

[mnementh]

]I would be concerned aboot that connecting the two transistor cases both thermally and electrically in a manner that affects the deliberate isolation provided by that BeO disc.[/size][/b][/i][/color]

mnem
 

Gorilla glue is an insulator.    

Nevertheless  I should check they aren't somehow electrically connected.

I wouldn't trust Gorilla glue to be an insulator. In fact, my experience with Gorilla glue is I don't even trust it to be a decent glue. 

Not particularly Gorrilla-y either

I dunno... I have some of their epoxy and it does sortof smell like some jungle animal... 

mnem
takes forever to set, tho; so I rarely use it.

[factory]

measured in-circuit   

Not going to desolder unobtainium transistors.

If they are standard designation "2N"  and not some "in house" number you'd be surprised. Someone on Ebay has got them. Every old 2N transistor I needed for the Type 547 project was out there.

Yeah, in this case they are a "selected pair", so do have a in-house part number and will be calibrated after install.

I understand wanting to disturb this "calibrated system" as little as possible... which is a large part of why I don't understand the apparently cavalier attitude in choice of adhesives and is the main reason my brain goes at the thought of encapsulating it in epoxy.

mnem
 

Q3 might be obtainable (2N1304), but Q5 the dual transistor with custom markings is probably unobtainium.

David

[mnementh]

Talking about Tek scopes in 2021.

I got two quotes for my company at work.
One is $153K (6 Series) the other one (5 Series) is about $140K.

Problem is in both over 70% of the money are just SW options... go piss a rope Tek!

Jeezus! That's crazy, i didn't realize how pricey some of the new gear is. I checked the website and they have several options that are 10 grand extra. I hope there is some groundbreaking engineering inside those things for that kind of coin.

We were talking in here a while aback aboot some used off-lease Infiniium scopes in the 1/4 to 1/2 million dollar range. 

mnem
For that kindof dosh it better make me pancakes, eggs, and bacon the morning after...

[Cerebus]

]I would be concerned aboot that connecting the two transistor cases both thermally and electrically in a manner that affects the deliberate isolation provided by that BeO disc.[/size][/b][/i][/color]

mnem
 

Gorilla glue is an insulator.    

Nevertheless  I should check they aren't somehow electrically connected.

For some value of "insulator". We're talking about a 200G input here, so what counts as an insulator by normal experience may not count as an insulator in this particular application. However, if memory serves this is a long tailed pair and if the cases are going to be connected to anything internal to the transistors it's going to be the 'common' side of the long tailed pair where the circuit is only going to have something between a few ohms and a few k ohms between the two devices so the leakage from a 1G 'short' between them isn't going to matter more than a fart in a hurricane.

[Andrew_Debbie]

]I would be concerned aboot that connecting the two transistor cases both thermally and electrically in a manner that affects the deliberate isolation provided by that BeO disc.

mnem
 

Gorilla glue is an insulator.     Nevertheless  I should check they aren't somehow electrically connected.

I wouldn't trust Gorilla glue to be an insulator. In fact, my experience with Gorilla glue is I don't even trust it to be a decent glue. 

runs off to check  ....

Ummm... what are you going to use to test with...? Do you have another meter with 100Gohm input impedance...?

Yes, I get that is only supposed to be guaranteed on the highest voltage range, but it seems to me the entire frontend PCB has to be designed with that in mind. 

EDIT: I certainly wouldn't trust Gorilla Gloo or thermal epoxy to have the same insulative qualities as needed here... that BeO disc is a pretty exotic material and was chosen for a reason. Even just changing the mass of the junction will affect thermal propagation, and therefore also will affect thermal loss to the surrounding environment.

mnem
 

I was  at a loss what to do.   Leaving it as is was not an option.  An air gap was wrong and there was a possibility of small BeO fragments coming loose.

There was some sort of adhesive residue on Q5, mostly in the center of the washer.    It looked like cyanoacrylate but could have been anything.


Unfortunately  I haven't come across anyone who worked on the ground floor at Dana.   I was over there once or twice in 1979 but that hardly counts.

   I'd love to sit down and have a chat with the design team but I doubt that is remotely possible.



 

[Specmaster]

measured in-circuit   

Not going to desolder unobtainium transistors.

If they are standard designation "2N"  and not some "in house" number you'd be surprised. Someone on Ebay has got them. Every old 2N transistor I needed for the Type 547 project was out there.

Yeah, in this case they are a "selected pair", so do have a in-house part number and will be calibrated after install.

I understand wanting to disturb this "calibrated system" as little as possible... which is a large part of why I don't understand the apparently cavalier attitude in choice of adhesives and is the main reason my brain goes at the thought of encapsulating it in epoxy.

mnem
 

I think if the glue has had an effect here, it is more likely to do with the sheer amount of glue and the fact that be examining the photo at 700% scale, it appears is if some has not only gone down around the transistor case, affecting its thermal characteristics, but also it looks like it has crept up the side of the BeO disc, which will if the glue is not a good isolator, affect the electrical isolation between the 2 transistors.

Personally, I'd been inclined to have just a tiny spot of superglue between them, if it crept out around the edge of the disc then it would be way too much.

[mnementh]

measured in-circuit   

Not going to desolder unobtainium transistors.

If they are standard designation "2N"  and not some "in house" number you'd be surprised. Someone on Ebay has got them. Every old 2N transistor I needed for the Type 547 project was out there.

Yeah, in this case they are a "selected pair", so do have a in-house part number and will be calibrated after install.

I understand wanting to disturb this "calibrated system" as little as possible... which is a large part of why I don't understand the apparently cavalier attitude in choice of adhesives and is the main reason my brain goes at the thought of encapsulating it in epoxy.

mnem
 

Q3 might be obtainable (2N1304), but Q5 the dual transistor with custom markings is probably unobtainium.

David

I get that... but then we come back around to the whole "calibration" of the circuit issue again. Just changing the adhesive used probably affected the thermal transfer/thermal loss of the assembly. I still cringe at that pic of the gorilla snot all over that transistor. 

OTOH, "the damage is done" now... undoing it is as likely to mess things up even worse as it is to fix anything... so I think I'll go find something else to worry aboot. Thankfully, this is all Andrew's problem, not mine.

mnem
 

[Andrew_Debbie]

Here is the circuit again --  Q5 is printed MPS200201  on the can.   Hand selected or custom part.     Even if I could find one,  that would mean figuring out the FSV resistors, which is beyond what I can do at home.

[mnementh]

]I would be concerned aboot that connecting the two transistor cases both thermally and electrically in a manner that affects the deliberate isolation provided by that BeO disc.

mnem
 

Gorilla glue is an insulator.     Nevertheless  I should check they aren't somehow electrically connected.

For some value of "insulator". We're talking about a 200G input here, so what counts as an insulator by normal experience may not count as an insulator in this particular application. However, if memory serves this is a long tailed pair and if the cases are going to be connected to anything internal to the transistors it's going to be the 'common' side of the long tailed pair where the circuit is only going to have something between a few ohms and a few k ohms between the two devices so the leakage from a 1G 'short' between them isn't going to matter more than a fart in a hurricane.

What is your opinion aboot the change in mass of the junction? It seems to me the BeO disc was chosen for both thermal conductivity and electrical insulation properties...

mnem
 

[mnementh]

]I would be concerned aboot that connecting the two transistor cases both thermally and electrically in a manner that affects the deliberate isolation provided by that BeO disc.

mnem
 

Gorilla glue is an insulator.     Nevertheless  I should check they aren't somehow electrically connected.

I wouldn't trust Gorilla glue to be an insulator. In fact, my experience with Gorilla glue is I don't even trust it to be a decent glue. 

runs off to check  ....

Ummm... what are you going to use to test with...? Do you have another meter with 100Gohm input impedance...?

Yes, I get that is only supposed to be guaranteed on the highest voltage range, but it seems to me the entire frontend PCB has to be designed with that in mind. 

EDIT: I certainly wouldn't trust Gorilla Gloo or thermal epoxy to have the same insulative qualities as needed here... that BeO disc is a pretty exotic material and was chosen for a reason. Even just changing the mass of the junction will affect thermal propagation, and therefore also will affect thermal loss to the surrounding environment.

mnem
 

I was  at a loss what to do.   Leaving it as is was not an option.  An air gap was wrong and there was a possibility of small BeO fragments coming loose.

There was some sort of adhesive residue on Q5, mostly in the center of the washer.    It looked like cyanoacrylate but could have been anything.


Unfortunately  I haven't come across anyone who worked on the ground floor at Dana.   I was over there once or twice in 1979 but that hardly counts.

   I'd love to sit down and have a chat with the design team but I doubt that is remotely possible.

Yeah, I kindof figured it was like that after we discussed it at length. Just because my hackles rise at the thought doesn't mean I'm right aboot this; I do in fact hope I'm just being overly paranoid.

Again... I'm glad this one is on someone else's bench. If I found myself in such a situation with no good source of technical reference, I would have probably gone with a little dot of CA myself. However, there's no way of knowing what other atrocities I might have committed while I was in there.

mnem
 

[Cerebus]

]I would be concerned aboot that connecting the two transistor cases both thermally and electrically in a manner that affects the deliberate isolation provided by that BeO disc.

mnem
 

Gorilla glue is an insulator.     Nevertheless  I should check they aren't somehow electrically connected.

For some value of "insulator". We're talking about a 200G input here, so what counts as an insulator by normal experience may not count as an insulator in this particular application. However, if memory serves this is a long tailed pair and if the cases are going to be connected to anything internal to the transistors it's going to be the 'common' side of the long tailed pair where the circuit is only going to have something between a few ohms and a few k ohms between the two devices so the leakage from a 1G 'short' between them isn't going to matter more than a fart in a hurricane.

What is your opinion aboot the change in mass of the junction? It seems to me the BeO disc was chosen for both thermal conductivity and electrical insulation properties...

mnem
 

The whole point of the assemblage is to keep the two transistors as near the same temperature as possible to keep input offset voltage from the temperature sensitivity of V_BE as close to fixed as possible. Any changes here are going to affect offset voltage and offset voltage drift with temperature. The offset voltage can be trimmed with the front panel offset pot, drift is minimised by maximising the thermal coupling of the two transistors. That this comes with a front panel offset adjustment suggests that the whole thing was not Mr. Stable in the first place so there's clearly no long term 'baked in' stability to worry about. With that in mind, going to the expense and trouble of a BeO washer attached to the tops of the cans in hoping to replicate a monolithic matched pair (that's how it's drawn so we must presume that at some point in its production history there really was a monolithic matched pair there) with just a matched pair smacks of magical thinking.

[Neper]

Just what exactly are the properties of BeO that make it suitable for this application?