Test Equipment Anonymous (TEA) group therapy thread

[tggzzz]

Heads up, in the UK at least it is a bank holiday weekend - so fleabay has max sellers fee of £1 until the end of the month.

So competing for my time is:

• getting effing Win10 installed
• finding out about swift nesting boxes that I could fit under the eaves (work is restarting on the outside of my house, so the scaffolding won't be up forever, I hope)
• cutting grass, brambles, etc
• repairing my newsfox rss reader, which has just blown up
• swapping a battery holder for my Tek 1502, ready for fleabaying
• getting a description of my 2465dms for fleabay, and working out shipping costs
• inevitable family crap
• daring to poke around in a faulty 1kV PSU; no schematics, no indication of how to connect it to external equipment (other than the way it is currently connected)

Just as well I'm part of the idle retired.

[Specmaster]

... If you want a British designed and built car, try the Aston Martin DBS Superleggera, £300,007 https://en.wikipedia.org/wiki/Aston_Martin_DBS_Superleggera

Hehe. Clever pricing, Aston: £300,007

That was deliberate, the car is themed on the original car used in Gold Finger and hence the 007 on the price and also very limited production run.

[tggzzz]

@tggzzz:

I've never had this myself but I have seen it. Always just works for me. Which chipset and vendor is the board out of interest?

Asus TUF Gaming x570 with a Ryzen 3700x.

Shoving it in another hole is probably the correct answer. You can blame the USB standard for this not necessarily the PC or Windows. Check which ports your motherboard has and don't stick it in one of the USB3/3.1 holes and it'll be fine. If the cable to the front panel ports was a thick dual one that terminated on a pin header then it's probably USB 3. The back of mine has two clearly marked USB 2.0 ports (ASRock B450M Pro4 board)

Back panel:

• 3 x USB 3.2 Gen 2 ports (2 x Type-A ports; 1 x Type C port)
• 4 x USB 3.2 Gen 1 ports

Connectors for flying leads on motherboard:

• 1 x USB 3.2 Gen 1 connector supports additional 2 USB ports - I left these unconnected
• 2 x USB 2.0/1.1 connectors support additional 4 USB ports - I've connected these to the front panel

Yes, I've tried using the front panel USB 2.0 ports, with the same end result.

Edit: I rarely have hardware problems TBH. Most computers are scared of me and spring into action the moment they are presented to me. Apart from a Sony AIO PC which I beat the shit out of with a chair leg in my back garden and then left in a bin full of water for 6 months just to make sure the pile of shit was dead. Picture:

I've just put my last-but-one homebrew in the recycling bin. It was a dual processor (not core) Athlon server motherboard from ~2005. Worked beautifully as an XP/ubuntu dual boot.
I'm typing this on my current machine, an intel core 2 duo from ~2011. Works beautifully as an XP/ubuntu dual boot. Went to buy Win7 when XP support ended, Microsoft said "shan't sell you that".
My fallback machine is my daughter's round-Oz Asus eeePC from 2010. Worked beautifully until a disk died, tried to reinstall XP, Microsoft said "shan't, Samsung's fault", stuck ubuntu on it and it still works beautifully.
Now Microsoft says "shan't let you instally Win10, and you're on your own".

Once is happenstance, twice is coincidence, third time is enemy action.

Ok did some research. This is probably disparity between a UEFI and BIOS boot. Make sure you have written it using a UEFI capable tool. Instructions here: https://www.linuxbabe.com/ubuntu/easily-create-windows-10-bootable-usb-ubuntu

Make sure the board is set to true UEFI boot. Basically it expects to load the drivers from a known location during the boot process but it can’t see that because it’s in legacy BIOS mode. Just do a full config reset in the BIOS first before you do this. It should be set up right by default.

Mine probably works because I usually use a windows machine to write the stick using the MSFT image creation tool.

I thought I had done that kind of USB creation in the past. There is a wrinkle, of course: there's a >4GB file so fat32 can't be used and gparted won't create an exfat partition. No doubt the Win tool will, but that's back to bootstrapping again.

I've tried various combinations of BIOS "compatibility module:disabled" and "os type: windows uefi only", to no avail.

I haven't touched the BIOS's key management tweaks; I know what I don't know and might bugger up.

Anyway, copying the files onto an NTFS formatted partition; maybe that will work?

[bd139]

That's a possibility. I've done it before when I accidentally installed Linux on all my PCs  . Roughly:

1. install ntfsprogs package
2. fdisk it to NTFS partition type (can't remember which ID that is)
3. Mount the iso using loopback
4. Copy all files to it.
5. (optionally) use ms-sys to install an MBR (do this on the disk device not the partition): http://ms-sys.sourceforge.net/

This is much easier on windows believe it or not:

1. diskpart
2. ...list disk
2. ...sel disk X
2. ...clean
3. ...create par pri
4. ...active
5. ...format fs=ntfs quick
6. copy files to stick with windows explorer
7. reboot

[McBryce]

Right, the Xtal tester and cheap arse frequency counter is finished and has finally passed testing. I say finally because it was apparently working fine but the only crystal that I had to test it with, although it being a new one, was dead 

Anyway the kit was well thought out apart on area, the tester socket was mounted flat on the PCB, which is fine if you didn't to put it into some form of a case. I purchased it with an acrylic case which means that unless the crystal under test has long leads, then your out of luck. I'll look at ways of bringing this upwards so it finishes pretty flush with the case. The supplied instructions are pretty crap to be honest. Apparently, when used as a frequency counter, the decimal point serves a dual function, firstly as as decimal point and secondly according to other online video I came across, if it steady, it denotes the display is in MHz but if it flashes then it means the display is in kHz, none of this is explained in the instructions.

Anyway here are some photos of the unit in test and finshed in its case (less the future modification with its test socket).


The completed board.


The solder side.


The board has passed the post test, 4th digit showing a zero means it has passed power on test.


Testing with a 2.4MHz duff crystal 


Testing with a 54MHz crystal (max is 50MHz) this shows the third overtone (thanks bd139)


Testing a 20MHz crystal


And in its case with the duff crystal

All in, its not a bad piece of TE given the low cost £8.86 plus shipping, while not a precision device it is a very handy thing to have to provide a quick no nonsense test of a crystal.

It's NOT finished until you've removed the protective sticker from that 7-Seg display!!

McBryce.

[Ero-Shan]

Here we go again with another round of measurements (I think BU508 was asking for something like that).

But before I present them, I have to briefly go back to the capacitance measurements: Ceramic caps with more than about a nF showed a pronounced change in capacitance over time. Some started about 10 % low and needed up to one minute to settle on the final value. Others started high and came down. The time was much shorter with the HP.

Back to the subject. I've made a couple (voltage only) of measurements with my newly acquired AC Voltage/Current Standard YEW 2558. All instruments were given sufficient time to find their thermal equilibrium. The frequency was set to 50 Hz.

Without any further ado, here's the beef:

YEW 2558     HP3458A             HP34401 (A)        HP34401 (B)      HP 3478A         Solartron 7150  Fluke 8040A     Fluke 8060
Nom. output  min-max             min-max            min-max          min-max          min-max         min-max         min-max
-----------------------------------------------------------------------------------------------------------------------------------
120.00 mV    120.587-120.613     120.106-120.128    120.083-120.112  120.037-120.053  120.22-120.29   119.74-120.56   120.38
1.2000 V     1.19872-1.19901     1.19800-1.19811    1.19793-1.19794  1.19773-1.19783  1.19905-1.19910 1.1989-1.1990   1.2015-1.2009
12.000 V     11.9924-11.9942     11.9888-11.9899    11.9851-11.9858  11.9810-11.9824  11.9248-11.9266 12.000-12.002   12.024-12.025
120.00 V     119.9711-119.9889   119.8603-119.8664  119.837-119.844  119.847-119.862  119.611-119.618 120.02-120.04   120.24-120.25
750.00 V     749.691-749.802     749.034-749.103    749.401-749.434  n/a, 300 V max   746.18-746.22   750.1-750.2     750.7-750.9

For reference: The YEW claims an accuracy of ±(0.08 % of setting + 0.015 % of range) for all voltage outputs above 20 % of range (range is 1.2 * 10ⁿ V).

Observations:

• All of the DMM with more than 4½ digits are terrible at showing a stable value. Whether this is due to The YEW or the DMM, I can't say for sure. But remembering the readings of other measurements, I tend to blame  the DMMs for the most part.
• 120 mV looks a little high with all DMMs, which hints at the YEW.
• Both Flukes show a little higher than the rest of the bunch.
• The Solartron apparently has problems at higher voltages; this is emphasized by the fact that it made a reset (!) when I plugged the 750 volts in.
• With most of these instruments, I should be able to measure the line voltage with sufficient accuracy.

[Specmaster]

@McBryce, spotted that and removed, just never updated the photo 

[BU508A]

Here we go again with another round of measurements (I think BU508 was asking for something like that).

But before I present them, I have to briefly go back to the capacitance measurements: Ceramic caps with more than about a nF showed a pronounced change in capacitance over time. Some started about 10 % low and needed up to one minute to settle on the final value. Others started high and came down. The time was much shorter with the HP.

Back to the subject. I've made a couple (voltage only) of measurements with my newly acquired AC Voltage/Current Standard YEW 2558. All instruments were given sufficient time to find their thermal equilibrium. The frequency was set to 50 Hz.

Without any further ado, here's the beef:

YEW 2558     HP3458A             HP34401 (A)        HP34401 (B)      HP 3478A         Solartron 7150  Fluke 8040A     Fluke 8060
Nom. output  min-max             min-max            min-max          min-max          min-max         min-max         min-max
-----------------------------------------------------------------------------------------------------------------------------------
120.00 mV    120.587-120.613     120.106-120.128    120.083-120.112  120.037-120.053  120.22-120.29   119.74-120.56   120.38
1.2000 V     1.19872-1.19901     1.19800-1.19811    1.19793-1.19794  1.19773-1.19783  1.19905-1.19910 1.1989-1.1990   1.2015-1.2009
12.000 V     11.9924-11.9942     11.9888-11.9899    11.9851-11.9858  11.9810-11.9824  11.9248-11.9266 12.000-12.002   12.024-12.025
120.00 V     119.9711-119.9889   119.8603-119.8664  119.837-119.844  119.847-119.862  119.611-119.618 120.02-120.04   120.24-120.25
750.00 V     749.691-749.802     749.034-749.103    749.401-749.434  n/a, 300 V max   746.18-746.22   750.1-750.2     750.7-750.9

For reference: The YEW claims an accuracy of ±(0.08 % of setting + 0.015 % of range) for all voltage outputs above 20 % of range (range is 1.2 * 10ⁿ V).

Observations:

• All of the DMM with more than 4½ digits are terrible at showing a stable value. Whether this is due to The YEW or the DMM, I can't say for sure. But remembering the readings of other measurements, I tend to blame  the DMMs for the most part.
• 120 mV looks a little high with all DMMs, which hints at the YEW.
• Both Flukes show a little higher than the rest of the bunch.
• The Solartron apparently has problems at higher voltages; this is emphasized by the fact that it made a reset (!) when I plugged the 750 volts in.
• With most of these instruments, I should be able to measure the line voltage with sufficient accuracy.

"With most of these instruments, I should be able to measure the line voltage with sufficient accuracy. "

 

https://www.eevblog.com/forum/projects/show-us-your-mains-waveform!/?all

[tggzzz]

Heads up, in the UK at least it is a bank holiday weekend - so fleabay has max sellers fee of £1 until the end of the month.

So competing for my time is:

• getting effing Win10 installed
• finding out about swift nesting boxes that I could fit under the eaves (work is restarting on the outside of my house, so the scaffolding won't be up forever, I hope)
• cutting grass, brambles, etc
• repairing my newsfox rss reader, which has just blown up
• swapping a battery holder for my Tek 1502, ready for fleabaying
• getting a description of my 2465dms for fleabay, and working out shipping costs
• inevitable family crap
• daring to poke around in a faulty 1kV PSU; no schematics, no indication of how to connect it to external equipment (other than the way it is currently connected)

Just as well I'm part of the idle retired.

Well, that's "1" sorted, thanks to bd, ignoring all the crap on all the guides on the net, and using NTFS on the usb stick. "1" now mutates into "installing usable operating systems alongside Win10". I wonder how long Win10 will operate; it thinks I don't have an internet connection

Sorting "4" was delightfully easy. Locate a likely looking directory in a backup, and cp -r that to the normal place. Now how easy wouldn't that be in Windows? Which bits of That Big Mistake (the registry) have to be twiddled, and how do you resurrect it if you screw up?

Can't put off measuring the house eaves for "2" much longer I'll have a my third cup of coffee first, then I'll have to go to the post office to get milk and maybe other things the local supermarket doesn't have anymore (e.g. eggs!).

[bd139]

Sorting "4" was delightfully easy. Locate a likely looking directory in a backup, and cp -r that to the normal place. Now how easy wouldn't that be in Windows? Which bits of That Big Mistake (the registry) have to be twiddled, and how do you resurrect it if you screw up?

It's exactly the same but your config directory is c:\users\yourusername\AppData\Roaming\

cp -r works as well 

Things have come a loooong way. Also there is registry on Linux but it's called gconf 

As for no internet connection, it works fine airgapped. It'll just not prompt you to create an online account (which you can skip anyway)

[tggzzz]

Sorting "4" was delightfully easy. Locate a likely looking directory in a backup, and cp -r that to the normal place. Now how easy wouldn't that be in Windows? Which bits of That Big Mistake (the registry) have to be twiddled, and how do you resurrect it if you screw up?

It's exactly the same but your config directory is c:\users\yourusername\AppData\Roaming\

cp -r works as well 

Things have come a loooong way. Also there is registry on Linux but it's called gconf 

As for no internet connection, it works fine airgapped. It'll just not prompt you to create an online account (which you can skip anyway)

Yeah, I worry about systemd and gconf

Seems like MS is re-learning what it knew in the early 80s (e.g. with Xenix and everything personal in /users), and forgetting the reason Motif was such an advance in the late 80s (flat contextless GUI widgets, ugh).

And it seems like the Linux mob is re-learning all the mistakes Microsoft made in the early 90s, e.g. kitchen sink prograns and/or data structures

[ch_scr]

Look what the postman dragged in. Works fine , and was really well packaged (thankfully, bugger is plenty heavy).

[bd139]

Sorting "4" was delightfully easy. Locate a likely looking directory in a backup, and cp -r that to the normal place. Now how easy wouldn't that be in Windows? Which bits of That Big Mistake (the registry) have to be twiddled, and how do you resurrect it if you screw up?

It's exactly the same but your config directory is c:\users\yourusername\AppData\Roaming\

cp -r works as well 

Things have come a loooong way. Also there is registry on Linux but it's called gconf 

As for no internet connection, it works fine airgapped. It'll just not prompt you to create an online account (which you can skip anyway)

Yeah, I worry about systemd and gconf

Seems like MS is re-learning what it knew in the early 80s (e.g. with Xenix and everything personal in /users), and forgetting the reason Motif was such an advance in the late 80s (flat contextless GUI widgets, ugh).

And it seems like the Linux mob is re-learning all the mistakes Microsoft made in the early 90s, e.g. kitchen sink prograns and/or data structures

Yeah. You can thank Freedesktop for that. Microsoft has had plenty of time to fix all the problems now. Once it's going I find it 100% hassle free.

I'm still fighting to keep it going in that direction where I can.

[AVGresponding]

TEA battery gurus, here's one for you. I am working on an HP 208A test oscillator. This is a cute little battery powered sine oscillator mainly for audio frequencies (but covers up to about 560 kHz) that is a bit of a unicorn. Pat (Cubdriver) was able to track down a manual for it, and it's now on the usual manual sites. Basically, the only function of the mains power in this instrument is to trickle charge the batteries when plugged in (charging current is adjustable with a pot on the power supply board). The instrument does not function without batteries, and mine has had the batteries removed. It calls for 4 6.5 volt NiCads. What would be a suitable modern replacement?

In series, series parallel, or parallel?

I'd look at using SLA batteries, if there's space. If you do, don't use cheap fleabay cells, buy them from a reputable seller or you'll end up with rubbish cells that fail within weeks, and/or a tiny fraction of the advertised capacity.

SLAs generally can be made to work as a substitute for NiCd on a 3:4 cell ratio, provided care is taken with understanding boost/float charge rates (SLAs are slightly less tolerant of higher V) as though they can handle the current, some NiCd chargers will output a bit too much V looking for a dv/dt EOC slope.
My guess is an old charger like that will not be so complex, so I'd just measure the o/c V and see it's not above about 2.3V/cell.

Since you can adjust the trickle charging current, you can probably substitute NiMH batteries. Usually, NiMH aren't a direct replacement for NiCd because NiMH require a lower trickle current.

My UPS/EL-CBU background is making me biased   

One thing I would say though, is I've only ever seen SLAs spew electrolyte (and sometimes plates!) due to serious and long term overcharging abuse. Normally they fail by just losing all capacity either due to sulphated plates or drying of the electrolyte.

[Ice-Tea]

Since you can adjust the trickle charging current, you can probably substitute NiMH batteries. Usually, NiMH aren't a direct replacement for NiCd because NiMH require a lower trickle current.

Been wondering about that lately. I read NiCd trickle is set at 0.1C, NiMh at 0.05C. But as NiMh has higher density, wouldn't that mean that if you replace a 1Ah NiCd with a 2Ah NiMh you're fine?

[tggzzz]

Since you can adjust the trickle charging current, you can probably substitute NiMH batteries. Usually, NiMH aren't a direct replacement for NiCd because NiMH require a lower trickle current.

Been wondering about that lately. I read NiCd trickle is set at 0.1C, NiMh at 0.05C. But as NiMh has higher density, wouldn't that mean that if you replace a 1Ah NiCd with a 2Ah NiMh you're fine?

My limited understanding is that 0.02C is better, but even then there can be problems with continuous trickle charging. The other issue is that the actual capacity of an "XAh" cell may only be 0.7XAh

[jxjbsd]

Three of the four modules are damaged and under repair.

[BU508A]

Here is one of these rare instruments which combines an analog instrument with a digital meter:

metrix MX 570

https://www.ebay.com/itm/402270553414



No affiliations with the seller.

[Cerebus]

<much snipery>

Without any further ado, here's the beef:

YEW 2558     HP3458A             HP34401 (A)        HP34401 (B)      HP 3478A         Solartron 7150  Fluke 8040A     Fluke 8060
Nom. output  min-max             min-max            min-max          min-max          min-max         min-max         min-max
-----------------------------------------------------------------------------------------------------------------------------------
120.00 mV    120.587-120.613     120.106-120.128    120.083-120.112  120.037-120.053  120.22-120.29   119.74-120.56   120.38

Observations:

• All of the DMM with more than 4½ digits are terrible at showing a stable value. Whether this is due to The YEW or the DMM, I can't say for sure. But remembering the readings of other measurements, I tend to blame  the DMMs for the most part.

I think what you should blame is physics. Let's look just at that 120mV into the HP3458A. You're showing a span between min and max readings of 26uV. That's just noise, an unavoidable consequence of physics and you are not going to see stable values on a sensitive enough instrument. That's 26 uV peak-to-peak, so 4.3uV rms (using the 6:1 rule of thumb). You haven't stated what your measurement bandwidths are, but let's assume that you're using the analog (as opposed to sampling) path  of the HP3458A at it's full 2MHz bandwidth. That's a noise density of 3nV/sqrt(Hz) which is an excellent figure and you can't expect not to get the "instability" you're seeing. For comparison a 50 ohm resistor sitting all on its own at room temperature is going to generate a Johnson noise of 0.9 nV/sqrt(Hz), 1k ohms 4nv/sqrt(Hz).

Obviously the full working out of noise figures is more complex than this, needing to know bandwidth, instrument mode and so on before you can properly calculate them, but as a rough calculation the above would suggest that what you're calling "unstable readings" is just basic unavoidable measurement error due to noise.

[mnementh]

Heads up, in the UK at least it is a bank holiday weekend - so fleabay has max sellers fee of £1 until the end of the month.

So competing for my time is:

• getting effing Win10 installed
• finding out about swift nesting boxes that I could fit under the eaves (work is restarting on the outside of my house, so the scaffolding won't be up forever, I hope)
• cutting grass, brambles, etc
• repairing my newsfox rss reader, which has just blown up
• swapping a battery holder for my Tek 1502, ready for fleabaying
• getting a description of my 2465dms for fleabay, and working out shipping costs
• inevitable family crap
• daring to poke around in a faulty 1kV PSU; no schematics, no indication of how to connect it to external equipment (other than the way it is currently connected)

Just as well I'm part of the idle retired.

Well, that's "1" sorted, thanks to bd, ignoring all the crap on all the guides on the net, and using NTFS on the usb stick. "1" now mutates into "installing usable operating systems alongside Win10". I wonder how long Win10 will operate; it thinks I don't have an internet connection

Sorting "4" was delightfully easy. Locate a likely looking directory in a backup, and cp -r that to the normal place. Now how easy wouldn't that be in Windows? Which bits of That Big Mistake (the registry) have to be twiddled, and how do you resurrect it if you screw up?

Can't put off measuring the house eaves for "2" much longer I'll have a my third cup of coffee first, then I'll have to go to the post office to get milk and maybe other things the local supermarket doesn't have anymore (e.g. eggs!).

"4" is easy, provided you DIDN'T disable Shadow Copy. Just find a suitable (as in it makes sense to YOU how it organizes things) Shadow Copy browser/viewer on the internet. Sorted.

"2" sounds like a job for some disaffected teenager in your life. A $10-30 (depending on just how disaffected and just how many minutes out of his/her oh-so-busy day) STEAM or NINTENDO ONLINE gift certificate should provide adequate motivation.

mnem

[tonyalbus]

Right, the Xtal tester and cheap arse frequency counter is finished and has finally passed testing. I say finally because it was apparently working fine but the only crystal that I had to test it with, although it being a new one, was dead 

Anyway the kit was well thought out apart on area, the tester socket was mounted flat on the PCB, which is fine if you didn't to put it into some form of a case. I purchased it with an acrylic case which means that unless the crystal under test has long leads, then your out of luck. I'll look at ways of bringing this upwards so it finishes pretty flush with the case. The supplied instructions are pretty crap to be honest. Apparently, when used as a frequency counter, the decimal point serves a dual function, firstly as as decimal point and secondly according to other online video I came across, if it steady, it denotes the display is in MHz but if it flashes then it means the display is in kHz, none of this is explained in the instructions.

Anyway here are some photos of the unit in test and finshed in its case (less the future modification with its test socket).


The completed board.


The solder side.


The board has passed the post test, 4th digit showing a zero means it has passed power on test.


Testing with a 2.4MHz duff crystal 


Testing with a 54MHz crystal (max is 50MHz) this shows the third overtone (thanks bd139)


Testing a 20MHz crystal


And in its case with the duff crystal

All in, its not a bad piece of TE given the low cost £8.86 plus shipping, while not a precision device it is a very handy thing to have to provide a quick no nonsense test of a crystal.

Indeed very nice ...

If you like the original design and schematics and background.. i made a review with some links to the documents.



Tony

[mnementh]

Sorting "4" was delightfully easy. Locate a likely looking directory in a backup, and cp -r that to the normal place. Now how easy wouldn't that be in Windows? Which bits of That Big Mistake (the registry) have to be twiddled, and how do you resurrect it if you screw up?

It's exactly the same but your config directory is c:\users\yourusername\AppData\Roaming\

cp -r works as well 

Things have come a loooong way. Also there is registry on Linux but it's called gconf 

As for no internet connection, it works fine airgapped. It'll just not prompt you to create an online account (which you can skip anyway)

This is actually the recommended way to force Windoze to install with only a local account. I'm going to copy/paste from older posts for the rest of my win10 install advice; use what you need, disregard the rest:

Windoze10 will activate based on a serial number in the BIOS; they've actually been doing this with some OEMs since the Win7 daze. I have one USB of Win7 Pro that I used to use on Dells... it would install and activate on pretty much anything Dell made in the last 10 years. Win10 will ALSO run quite happily with no activation; it just won't allow you to personalize the system or get Defender updates. After 14 or 21 days it will revert to the "Not Genuine" nag desktop, which, since you can't personalize, you can't make go away.

Other than that, and the fact that when MS rolls out the next major update it will break all your Windoze Automatic update ability, you CAN run Win10 unlicensed... they do this on purpose as "Trial" mode.

Of course, if you can find a legit non-OEM license from Win7/8/8.1, you can ALSO just use THAT key to activate Win10; then that license will remain YOURS in principle "til the end of time". I'm currently running on a legit Win7 .edu license from 10 years ago.  ...They want you away from everything non-10 enough to give it to you for free so they don't have to support archaic shit like that.

At setup, you have to complete the install with no network connection. Then you will be given the option to install a local user account only. You'll be given a couple pages of security/privacy settings; select the most "NO!" answer you can on all of them. After that, follow the instructions in the above video; it's pretty good.

I ALWAYS uninstall OneDrive, and all the crapware games in the Start menu. Then, once you have that done, download & run O&O ShutUp10 with "Recommended Settings" and toggle off whatever other invasive processes you feel a need to kill as well. It's really good at shutting Cortana up and hogtying her pretty good.  You CAN leave it TSR and it will fix anything Winbloze tries to change; I just run it every other week or so as maintenance.

You can also make Win10 totally your local/domain user bitch by editing Account policies as outlined here: https://www.isunshare.com/windows-10/2-ways-to-disable-or-block-microsoft-account-in-windows-10.html

Good luck and welcome to the age of Winbloze As A Service...

mnem
*makes sign of the cross* Dominy, Dominy, Dominy...

[tggzzz]

Heads up, in the UK at least it is a bank holiday weekend - so fleabay has max sellers fee of £1 until the end of the month.

So competing for my time is:

• getting effing Win10 installed
• finding out about swift nesting boxes that I could fit under the eaves (work is restarting on the outside of my house, so the scaffolding won't be up forever, I hope)
• cutting grass, brambles, etc
• repairing my newsfox rss reader, which has just blown up
• swapping a battery holder for my Tek 1502, ready for fleabaying
• getting a description of my 2465dms for fleabay, and working out shipping costs
• inevitable family crap
• daring to poke around in a faulty 1kV PSU; no schematics, no indication of how to connect it to external equipment (other than the way it is currently connected)

Just as well I'm part of the idle retired.

Well, that's "1" sorted, thanks to bd, ignoring all the crap on all the guides on the net, and using NTFS on the usb stick. "1" now mutates into "installing usable operating systems alongside Win10". I wonder how long Win10 will operate; it thinks I don't have an internet connection

Sorting "4" was delightfully easy. Locate a likely looking directory in a backup, and cp -r that to the normal place. Now how easy wouldn't that be in Windows? Which bits of That Big Mistake (the registry) have to be twiddled, and how do you resurrect it if you screw up?

Can't put off measuring the house eaves for "2" much longer I'll have a my third cup of coffee first, then I'll have to go to the post office to get milk and maybe other things the local supermarket doesn't have anymore (e.g. eggs!).

"4" is easy, provided you DIDN'T disable Shadow Copy. Just find a suitable (as in it makes sense to YOU how it organizes things) Shadow Copy browser/viewer on the internet. Sorted.

"2" sounds like a job for some disaffected teenager in your life. A $10-30 (depending on just how disaffected and just how many minutes out of his/her oh-so-busy day) STEAM or NINTENDO ONLINE gift certificate should be adequate motivation.

I have no idea what a shadoow/copy browser might be, but I doubt it is relevant to newfox. That reader has the best set of ergonomics I've seen, but I don't think it works with modern firefox browsers. When I finally migrate to my new machine, I'm seriously thinking of keeping an old firefox around solely for newsfox; but we'll see.

If only finding the right nesting box was so simple. There are quite a few on the market, but most won't fit my house, and swifts are picky about where they nest. I suspect I'll end up making a few to fit my eaves, but I need to discuss that with a local expert

As for teenagers, no longer available: sproglet is working hard and surviving despite falling through the gaps between all the UK government handouts. Key phrases: newly self-employed with savings.

[mnementh]

Ok did some research. This is probably disparity between a UEFI and BIOS boot. Make sure you have written it using a UEFI capable tool. Instructions here: https://www.linuxbabe.com/ubuntu/easily-create-windows-10-bootable-usb-ubuntu

Make sure the board is set to true UEFI boot. Basically it expects to load the drivers from a known location during the boot process but it can’t see that because it’s in legacy BIOS mode. Just do a full config reset in the BIOS first before you do this. It should be set up right by default.

Mine probably works because I usually use a windows machine to write the stick using the MSFT image creation tool.

Or he COULD just pop a blank DVD into a USB Optical drive and burn the .iso to that from ANY OS, even Chrome or Win XP. I've done it; it works.

mnem
 

[xrunner]

Heathkit IG-18 continued ...

Now that the power supply is working well I replaced the obvious - the 100 uF cap with corrosion that was across the meter. Then I reviewed the alignment procedures in the assembly manual and proceeded to investigate the distorted sine wave. When I attempted to turn the "feedback" and "bias" pots I could see the distortion vanish at points, but the pots were so dirty that I couldn't keep the adjustment stable.

I applied Deoxit to both of the pots and turned them back and forth several times. After that  - presto - I could set the adjustments properly and was rewarded with a good sine wave on the output. I also tried to adjust the other two pots on the main board - "meter cal" and "symmetry" (for the square wave) but were as dirty as the others, so same treatment and again, worked fine after cleaned.

Anyone want to take up smoking now?

Facing the last major issue, I turned to the sine amplitude - the meter amplitude indication only went halfway and the owner stated this had not been the case until lately. He stated that this problem had appeared slowly over some time. After probing the generator circuit I found that the input amplitude to the output stage (fine adj. pot and attenuators) was good. However, we have a serious problem.

At some point in time the original fine adj. pot was replaced by another one - why I have no idea. But ... the job is an abomination (read: Gorilla Joe got to it before me).

As I stated before, by observation, the pot is mounted cock-eyed and it's coming apart at the seams. The original shaft was cut off and a slot filed into to mate with the wire link, it but it's not even centered. The pot is supposed to be 5k. The thing had to be removed.

After checking, it was presenting 4.7 k max - and here's the main thing - only 1.3 k at minimum on the center or variable pin. This is why the amplitude is not going full scale any longer. I took it apart to see what's inside. The shaft stuck at a certain point so that explains the reason it won't go to zero. The rear of the shaft is not on-center and there is nothing at the back inside it to keep it aligned there. A Gorilla Joe mod again?

I lashed up a test with a new 5k pot I had on hand, and with that setup, the output - min to max - worked perfectly. However, it's not the same shaft size as the old one, the shaft is too long to fit, and too small for a good tight fit in the plate hole. It's connected to the variable shaft by an odd "U" shaped wire link that fits in a hole on the main shaft and then in the cut-out of the pot shaft.

So, we're looking at some cutting, filing, Fusion 360 / 3D printing for a customized adapter or bushing.

Be back soon ...