Test Equipment Anonymous (TEA) group therapy thread

[med6753]

Finally got some time to start the re-cap of the Type 547. Started with C802, which is part of the decoupling network to keep the HV oscillator from feeding back into the PSU. Original was a 40uf/400V. Could have used either a 39uf or 47uf. Went with the 47uf/400V. The replacement cap was small enough that I could mount it inside the HV assembly and leave the original can in place.

Hm. I'd have probably used a 39µF electrolytic capacitor and parallel to it a 820nF foil capacitor and a 100nF ceramic capacitor for better RF blocking.

Why? The original circuit didn't have that. Also, the HV Oscillator is complete enclosed in a metal case although I have not seen any degraded performance or noise with the cover off. But your point is well taken because if you run a Type 535A with the case open you DO see noise on the trace. I've seen it myself on my Type 535A but not on any of my three Type 547's. 

I've always been told that it is better to increase the capacitance rather than drop its value, so with that in mind, I'd have done exactly the same as you and go with the 47uF 

Yes, within reason it's always better to go higher rather than lower but given tolerances the difference between 40uf and 39uf is a nit.

[Cubdriver]

Finally got some time to start the re-cap of the Type 547. Started with C802, which is part of the decoupling network to keep the HV oscillator from feeding back into the PSU. Original was a 40uf/400V. Could have used either a 39uf or 47uf. Went with the 47uf/400V. The replacement cap was small enough that I could mount it inside the HV assembly and leave the original can in place.

Hm. I'd have probably used a 39µF electrolytic capacitor and parallel to it a 820nF foil capacitor and a 100nF ceramic capacitor for better RF blocking.

Why? The original circuit didn't have that. Also, the HV Oscillator is complete enclosed in a metal case although I have not seen any degraded performance or noise with the cover off. But your point is well taken because if you run a Type 535A with the case open you DO see noise on the trace. I've seen it myself on my Type 535A but not on any of my three Type 547's. 

I've always been told that it is better to increase the capacitance rather than drop its value, so with that in mind, I'd have done exactly the same as you and go with the 47uF 

The only time it's a real concern is if you're replacing the first filter cap after certain tube rectifiers - for instance, a 5U4G is limited to 40µF for the first filter cap - too large a value can draw too much current on start up and damage the tube.  Beyond that, going a bit higher in value is normally kosher.

-Pat

[med6753]

Finally got some time to start the re-cap of the Type 547. Started with C802, which is part of the decoupling network to keep the HV oscillator from feeding back into the PSU. Original was a 40uf/400V. Could have used either a 39uf or 47uf. Went with the 47uf/400V. The replacement cap was small enough that I could mount it inside the HV assembly and leave the original can in place.

Hm. I'd have probably used a 39µF electrolytic capacitor and parallel to it a 820nF foil capacitor and a 100nF ceramic capacitor for better RF blocking.

Why? The original circuit didn't have that. Also, the HV Oscillator is complete enclosed in a metal case although I have not seen any degraded performance or noise with the cover off. But your point is well taken because if you run a Type 535A with the case open you DO see noise on the trace. I've seen it myself on my Type 535A but not on any of my three Type 547's. 

I've always been told that it is better to increase the capacitance rather than drop its value, so with that in mind, I'd have done exactly the same as you and go with the 47uF 

The only time it's a real concern is if you're replacing the first filter cap after certain tube rectifiers - for instance, a 5U4G is limited to 40µF for the first filter cap - too large a value can draw too much current on start up and damage the tube.  Beyond that, going a bit higher in value is normally kosher.

-Pat

I was just going to add that and you posted it first. Tube rectifiers have very specific max capacitance first stage or as you said can get a little violent.  After that first stage and with some surge or voltage dividers you can add more capacitance as needed....within reason.

That's one reason why the Type 105 Square Wave Generator had two rectifiers in the PSU. There's a ridiculous amount of capacitance down stream to achieve a clean PSU. 

[Specmaster]

Finally got some time to start the re-cap of the Type 547. Started with C802, which is part of the decoupling network to keep the HV oscillator from feeding back into the PSU. Original was a 40uf/400V. Could have used either a 39uf or 47uf. Went with the 47uf/400V. The replacement cap was small enough that I could mount it inside the HV assembly and leave the original can in place.

Hm. I'd have probably used a 39µF electrolytic capacitor and parallel to it a 820nF foil capacitor and a 100nF ceramic capacitor for better RF blocking.

Why? The original circuit didn't have that. Also, the HV Oscillator is complete enclosed in a metal case although I have not seen any degraded performance or noise with the cover off. But your point is well taken because if you run a Type 535A with the case open you DO see noise on the trace. I've seen it myself on my Type 535A but not on any of my three Type 547's. 

I've always been told that it is better to increase the capacitance rather than drop its value, so with that in mind, I'd have done exactly the same as you and go with the 47uF 

Yes, within reason it's always better to go higher rather than lower but given tolerances the difference between 40uf and 39uf is a nit.

True it is only a nit, but its a nit that you also decided it was better to go higher.

[Brumby]

... but given tolerances the difference between 40uf and 39uf is a nit.

Tolerances with electrolytics has always fascinated me.  The actual difference in capacitance replacing a nominal 39uF with a nominal 40uF could go down, rather than the expected up.  Usually not a problem, but occasionally I will check - just out of curiosity.

I have found there is a tendency for actual values to be higher than marked values, but the variation is not consistent - so a marked 39uF could actually be 50uF and a marked 40uF could be 47uF.  All within spec, but the actual value with such a replacement has gone down.

[ch_scr]

Following up on the discussion about scope probe storage, an interesting idea popped up:
VHS Tape boxes with custom labels
http://www.pa3cor.nl/electronics/scope-probe-storage/
He even has some labels to grab for inspiration 

[Peter_O]

That's a nice idea.

But for a real TEA the concept stores way to much air.   
Would like to have the room for that a solution.

I put the probes I don't use regularly into zip bags and put all of them into one box.

The others are hanging on a quite regular cable holder, ready to grab.

[VK5RC]

Here's my Chinesian contribution, still better that that POS:

eBay auction: #https://www.ebay.co.uk/itm/284443493043

Link is borked, not that I want that one.

Is there any reason why these couldn't be made from a square of regular copper PCB, or even a square of copper plate?

David

This one is working: https://www.ebay.com/itm/284443493043

UK version: https://www.ebay.co.uk/itm/284443493043

Maybe because of thermal mass. A copper plate will take much longer to reach equilibrium than a PCB with some copper traces.

+1 to what BU508A just said. To avoid thermal EMF, you want a short with a low thermal mass to reach the temperature of the connectors on the multimeter quickly.

Also notice the low gauge wire used on the rustic Agilent short.

I was absolutely certain that someone would comment on the shorting block used in this video (fast forward to 8:10):

Marco Reps got a black 3458B 3458A last year from Keysight:

I am pretty sure that Keysight's 3458A zeroing procedure specifies a copper wire bent into an omega shape as the 'proper' shorting tool for use.

[Cubdriver]

Marco Reps got a black 3458B 3458A last year from Keysight:

Am I the only one who expected to see a fuschia feline following behind with a more shocking color of paint for the front panel?



-Pat

[tggzzz]

Finally got some time to start the re-cap of the Type 547. Started with C802, which is part of the decoupling network to keep the HV oscillator from feeding back into the PSU. Original was a 40uf/400V. Could have used either a 39uf or 47uf. Went with the 47uf/400V. The replacement cap was small enough that I could mount it inside the HV assembly and leave the original can in place.

Hm. I'd have probably used a 39µF electrolytic capacitor and parallel to it a 820nF foil capacitor and a 100nF ceramic capacitor for better RF blocking.

Why? The original circuit didn't have that. Also, the HV Oscillator is complete enclosed in a metal case although I have not seen any degraded performance or noise with the cover off. But your point is well taken because if you run a Type 535A with the case open you DO see noise on the trace. I've seen it myself on my Type 535A but not on any of my three Type 547's. 

I've always been told that it is better to increase the capacitance rather than drop its value, so with that in mind, I'd have done exactly the same as you and go with the 47uF 

Presuming this is for PSU smoothing capacitors, there can be some downsides.

Increased turn-on current draw can also be a problem for Si rectifiers, hence the current limiting thermistors in some PSUs.

Some modern voltage rectifiers behave badly with too much load capacitance coupled with too low ESR. Replacing tants with polymer electrolytics can reveal that, since the latter have lower ESR.

[Saskia]

on my way to my doctor prescribed walking to get the shit out of my system (the allergens and toxins that caused so much headache) I just *had* to stop by my dealer to get my fix.
UT8802e bench dmm (rather unappealing specs) for silly money, ti 30 eco for 5€, raspi 4b4gb incl. touch screen / case, lt 50€.
the ut is for flipping, that's just what I'll do ...

I hate this autofuckuping feature they call autocorrection. may Google go to hel. Along with usoft and apple ...

[Cerebus]

Marco Reps got a black 3458B 3458A last year from Keysight:

Am I the only one who expected to see a fuschia feline following behind with a more shocking color of paint for the front panel?



-Pat

That's only for the models being delivered to the Pink Panther's alma mater, Durham Durham University.

[mnementh]

...
How can I do that with the bare chips in my hand? I have a bunch of programmers and chip clips etc but I'm not up with writing and reading directly to I2C chips...

Some time(~2months) ago I've tuned up my AN8008's: power off @ one hour, back light tied to button etc... I've done it on TL866-II Plus, which allows editing the programmable content in the buffer.
You can try the TL866 software in demo mode, without the device, to see the capabilities.
Oh, and the EEPROM's in case are 24C02(I²c).

I've been meaning to get a TL866 for a while now, this might be the excuse...

Currently at my disposal, I have an AVR Dragon, USBASP clones, Atmel AVRISP STK500 clone, Silicon Labs MCU ToolStick, Bus Pirate, and a couple other bits and pieces.
Oh, I also have a CH341A on order too.

Did you see this CH341A bundle? https://www.amazon.com/Organizer-EEPROM-CH341A-Adapter-Programmer/dp/B07V2M5MVH/

Looks like a lot of flexibility and convenience for $14 via Prime so I got it coming; I suspect it's prolly available a bit cheaper on your side of the world.

mnem

[TERRA Operative]

...
How can I do that with the bare chips in my hand? I have a bunch of programmers and chip clips etc but I'm not up with writing and reading directly to I2C chips...

Some time(~2months) ago I've tuned up my AN8008's: power off @ one hour, back light tied to button etc... I've done it on TL866-II Plus, which allows editing the programmable content in the buffer.
You can try the TL866 software in demo mode, without the device, to see the capabilities.
Oh, and the EEPROM's in case are 24C02(I²c).

I've been meaning to get a TL866 for a while now, this might be the excuse...

Currently at my disposal, I have an AVR Dragon, USBASP clones, Atmel AVRISP STK500 clone, Silicon Labs MCU ToolStick, Bus Pirate, and a couple other bits and pieces.
Oh, I also have a CH341A on order too.

Did you see this CH341A bundle? https://www.amazon.com/Organizer-EEPROM-CH341A-Adapter-Programmer/dp/B07V2M5MVH/

Looks like a lot of flexibility and convenience for $14 via Prime so I got it coming; I suspect it's prolly available a bit cheaper on your side of the world.

mnem

Yeah, I saw that set. I've got one coming from ebay, eventually....

[mnementh]

Yeah, I just got it to add to my collection of programming dongles... 

mnem
There's no such thing as too many dongles...

[med6753]

Tempting but I think $95 USD is a bit much. Heath's answer to the fancy hp AC voltmeters. A quick check of the specs states an accuracy of 4% full scale. Doesn't sound very impressive to me. How does that stack up against the hp's?

According to the listing this one was factory built and not a kit so it should be unmolested.

https://hudsonvalley.craigslist.org/ele/d/florida-ac-voltmeter-heath-im-5238/7478330110.html

[Vince]

Yeah I hope that's it.  I have just ordered x3 NOS 6AQ5 from Ebay France, so hopefully I shall receive them late next week all going well.

30 Euros shipped. 10 Euros per tube. Not cheap by my wallet standards, but not crazy expensive either, for NOS tubes I mean.

Best deal was 4 NOS ones for 18 Euros shipped, but they come from Ukraine so delivery time, or delivery at all really, is unknown. I don't want to waste 18 Euros down the drain nor wait 3 months to get them 

So hoping to get my tubes late next week then. In the meantime I will resume work on the Rochar Nixie DMM  I think

How fast are your peek hold meters, can't you expect some results from them?

I am really sorry but I don't understand what you mean ?! 
My English is not up to sniff it looks like 

Could you rephrase that please ?

Sorry, my bad.

The detailed message was for your earlier oscillator tube circuit capacitor measurement.
My thought was that your scope didn't work or the capacitor is bad.
Your earlier low voltage says that there's no short to ground but reason for later 0V has few options.

Maybe you should do a special peak detector.
My guess is that you'll see similar stuff also in the future.
Yes, I don't trust that scope is (not)triggering if I don't see it.

Cap is good, I replaced it as I said, and tested it of course before soldering it in. A string of 4 film 1.5nF 630V caps to get me close enough to the 390pF of the old mica cap.

I guess my scope is just unable to trigger on it when single shooting, though I don't see why. It's an easy signal. Not sure what's going on. I will try other scopes see if they do any better.

Making a peak detector to capture that pulse ? Well I am not going to go that far... I have other things to do. For now I am waiting for my replacement 6AQ5 see if that gets it going. If not yeah, more troubleshooting will be on order I guess....

[mnementh]

Boring Mundane Everyday Print #307: Bed Leg Riser   

      AutoDesk Online 3D Viewer

Simple problem; easy fix with 3DP. We have these beds that are designed to be used without boxspring; they're awesome in that they're just tall enough (~12.5"/320mm clearance) to slide a full-size 57L Sterilite storage bin under. Except when they're not... like on carpet where the bed sinks in a little.

So I designed these polygon block risers to go on the 30.5mm square tube legs; they'll lift the whole bed up 25mm to make up the difference, and will also allow use of some other bins we have that are ~6mm taller than the Sterilite ones. These are a pretty high-density print and 65mm tall, so I broke the job up between two printers. Even with both machines running simultaneously, ~15 hours and 625 grams filament.

As always, full details and 3DP nerd stats can be seen over on the 3DP thread:

https://www.eevblog.com/forum/3d-printing/3d-printer-yet/msg4152823/#msg4152823

mnem

[Vince]

Had a quick google. Someone on some forum says it's just the gettering process.
Colour can be grey or silver but sometimes black.  So I guess it all depends what material, and how much of it, a given manufacturer, for a given lot of tubes, decided to use...

OK so I will keep these black 6AQ5 then, precious rare little things that I have only 3 of...

If it covers the ENTIRE inside of the tube, it’s likely not gettering.  Look closely at the coating, especially near the top mica insulator.  Look where the points of the insulator that bear against the envelope are, and see if you see drag marks in the coating below them towards the base of the tube.  If so, this will clearly indicate that the coating was there when the tue was assembled and the 'guts' of it slid into the envelope, with the mica points scraping the coating away.

-Pat

Hmmm... you are correct. I have just looked at a black one outside in teh sun light, and one can easily see all the regularly spaced vertical scrape marks going all the way top to bottom.
I can even see a little bit through the coating and see the internals of the tube, a bit.

So it's still a mystery then. If you ever remember what that is, don't hesitate to post ! 

[Robert763]

The graphite coating is to prevent the accumulation of charge on inside of the glass. This charge can affect the operation of the valve. Some valves use other coatings or doped glass to the same end.

[Vince]

Wonderful Robert, thank you very much ! But where have you been for all that time, you left us wondering for soooo  long, at least 48Hours if not more, that's cruel of you !!! 

[AVGresponding]

I am pretty sure that Keysight's 3458A zeroing procedure specifies a copper wire bent into an omega shape as the 'proper' shorting tool for use.

That's fine, if like the 3458 your meter has binding posts, but the vast majority these days have 4mm safety banana sockets, and good luck getting one of those to fit in a Fischer or Lemo socket as you might find on an old Solartron!

[m k]

An ADG368, an ADG368, my kingdom for an ADG368......


So, dealing with a recalcitrant TDS754D here that apparently lost its cal constants from the EEPROMS (U1052 and U1055) on the acquisition board.
I've had luck replacing them in the past, but this time I can't read or write anything to or from them using the EEPROM/NVRAM read/write utility stuff from somewhere on the forum that I put on floppy disk (It works on another working scope so my disks are good).

I replaced the EEPROMS with brand new ones and then tried writing the cal data from another TDS754D to at least get something force-fed in there but no dice...

So taking a look at the schematic, the EEPROM write control pins are tied to a 74HC86, which receives 5v from the famous 'Write Enable' switch on the processor board, via a 15V to 12V vreg and voltage divider.
All voltages test ok, so I replaced the 74HC86 because it was easy, still no dice. hmmmm.

I have also double checked all traces for continuity, all perfect like the day it was born.

So, the I2C data comes from a 160-pin chip, ADG368. Hmmm, all my spare boards use ADG308 chips. Well, lets get the hot air gun out and swap to see... No boot. Shiet, incompatible.
Swap back and we are booting again but errors are back to square one. So I think the ADG368 chip is most likely the bad bit at this stage (or whatever is feeding the AGD368..) I'll have to scope the I2C lines to the EEPROM chips too, to see if there is activity there.

Now, about finding another ADG368......... 

Last resort, pull good EEPROMS full of cal data from a scrapped TDS754D, stick them in and hope it works, if so, call it good. 
I also want to rig something up to directly read the data from the old EEPROMs to see if they might actually still be good, maybe something is just stopping the disk utility (and hence the scope) from seeing the data.
How can I do that with the bare chips in my hand? I have a bunch of programmers and chip clips etc but I'm not up with writing and reading directly to I2C chips...

Any idea what that 12V does?
My guess is that some version of NVRAM needed it for writing.

The 12V is brought down to about 5V with a voltage divider to provide 5V to the Write Protect pin on the EEPROMs to put them into write protect mode.
It also goes through some other bits and pieces of logic and ends up connected to the write protect pins on the NVRAM's.

Do you mean the 12V line from top of R2009?

How many NVRAMs there are?

Ah, my bad. That arrowhead on the 'NVWR_EN/I' is actually the line going to the NVRAM's, there are two of them, one containing the RTC.

Ok.

I'm a bit lost with names, types and sizes.

X24C02 is a pure serial E2PROM of 256x8 bytes.
Somebody extracted stuff from DS1650Y partitionable NV SRAM of 4M bits
and DS1486 RAMified Watchdog Timekeeper of 128k bytes.

TDS700x FAS
has low and high NVRAM addresses starting from 0x10 instead of 0x00.
tex = LOW_NV_RAM_ADDR("67108880")
tex = HIGH_NV_RAM_ADDR("67633168")

Relative starting addresses being 262144(256k) and around 131072(128k).

Different stuff being around 327680(256k + 64k).
VAR = hwAInstrSerialNumberDigit(UNIT "" VALUE 327685)
VAR = hwAOption1M(UNIT "" VALUE 327686)
VAR = hwAOptionMathPak(UNIT "" VALUE 327689)

So some models have clearly had two DS1486 chips since RTC and Watchdog are using first bytes
and write protected and NVRAM marked X24C02 chips are very small.

A11 board schematics is naming write protect switch for flash and NVRAM.
Then TDS700x FAS, it seems, is using name NVRAM for everything.

If those not X24C02 MVRAMs are stuff of swapped boards then CRC error is clearly for X24C02, otherwise not.

Can the U1050 be so dumb that X24C02s are read through it?
Pretty far fetched, I know.

U1050 has SDOUT/SDATA pin27 that is going nowhere, like are WNR and ALE, but they are also test points.
Maybe you can get some status data out from that pin.

[bd139]

Good morning, evening, afternoon or whatever the hell it is where you are. Decided my mid-life-crisis is now thoroughly over so back to the TEA thread and TE it is, my true home on this planet.

On the journey of self-discovery and self-enlightenment I discovered many things, all of which you know anyway I'm sure:

1. People are really really weird.
2. Some of them are incredibly stupid.
3. Even more are maliciously incompetent.
4. Alcohol is really really bad for you in large quantities but terribly fun 

However got some time to catch up with the kids, improve fitness, clean up some personal traumas and messes, shelf old ideas, reconnect with old friends and find new friends too, the latter two of which I'm now going to ignore and doom scroll through ebay 24/7 and shitpost on here

Currently still working on moving house (it's been 3 fucking years now on this treadmill) so involvement will be small and not boat anchor sized for a bit. Excuse my lack of trying to catch up with the thread because quite frankly I'd be dead by the time I managed to read back to my last post and work out what the hell is going on.

Anyway, hello again

[capt bullshot]

Welcome back!