Test Equipment Anonymous (TEA) group therapy thread

[bd139]

Penis fingers strikes again. Breadboarding accident. Building an internal power supply for my Williams pulser because I can't find the other one I made in the junk boxes. However this is my last LT1073's switch pin...



So need to work out how to screw 90V or so out of something else I have lying around . Don't think I even have any switcher ICs left at all.

Edit: haven't even got any MOSFETs which I can bodge stuff with

Edit: I WILL NOT BE DEFEATED. How not to build a boost converter: on a breadboard with a 555 and a random MOSFET I found in the junk box an IRF620 ... but it works  . Feedback loop is via shorting the control pin to ground crudely with a 2n3904 and a voltage divider. Had to futz with it for a while to get the conduction cycle right for the pikey inductor. Doesn't need heat sink now due to thing bouncing an amp during conduction 

This thing produces 135V very efficiently:



Source:
http://www.nixiekits.eu/service.php

The .pdf for the Nixietherm Mk III, Assembly manual:
http://www.nixiekits.eu/Downloads/NT_MkIII_Aufbauanleitung_Bausatz.pdf

Yeah it's not that much different to that actually. Mine misses the frequency compensation and niceties though and the MOSFET rings like fuck

Edit: Problem with this class of design is that it's very supply voltage and temperature sensitive. I have a better design somewhere I did a few years back that used a couple of LM311's and an LM385 reference. One to control the switching duty cycle as the master oscillator and another to influence that to control the output cycle. That took me a month to get working but it was around 85% efficient!

Also there are some much cooler LM723 designs out there but I haven't tried them yet! One day.

Edit 2: I'm not sure why people use the 555 for this to be honest especially that implementation as a production item. It's totally shit for it. Only reason to do it is if you have nothing else lying around. Why not use an MC34063 or literally everything else that is not that crap. There's a ton of stuff from China in this space which costs nothing these days 

[bd139]

Penis fingers strikes again. Breadboarding accident....      Edit: I WILL NOT BE DEFEATED. How not to build a boost converter: on a breadboard with a 555 and a random MOSFET I found in the junk box an IRF620 ... but it works  . Feedback loop is via shorting the control pin to ground crudely with a 2n3904 and a voltage divider. Had to futz with it for a while to get the conduction cycle right for the pikey inductor. Doesn't need heat sink now due to thing bouncing an amp during conduction 

      

Looks at bd's circuit... looks at the MkII... Looks again...



mnem

That's basically the Williams reference design that is.

Edit: these always look interesting: https://www.analog.com/en/products/adcmp572.html ... 35ps rise time!

[mnementh]

Yup. The MkII incorporates a stripline output cap on the back of the PCB so the output isn't so peaky and generates a nice identifiable plateau. There's easily a hundred people on eevBlog who'd love to have one of SVG's PCBs to build one; but the guy has bigger fish to fry, what with being hired to design boards for Tesla. 

[EDIT] That ADCMP572 freaks me the fuck out man... operating at that kind of slew rate and generating usable reference is fucking voodoo. 

(SNIP)
Edit: Problem with this class of design is that it's very supply voltage and temperature sensitive. I have a better design somewhere I did a few years back that used a couple of LM311's and an LM385 reference. One to control the switching duty cycle as the master oscillator and another to influence that to control the output cycle. That took me a month to get working but it was around 85% efficient!

Also there are some much cooler LM723 designs out there but I haven't tried them yet! One day.

   Yeah, when I ordered a replacement LM723 to rebuild my ginormous PHASE III 25A Linear, I got a handful plus some other goodies to have them to putter around with. 

mnem
*hiding out from the suck again*

[tautech]

Yup. The MkII incorporates a stripline output cap on the back of the PCB so the output isn't so peaky and generates a nice identifiable plateau. There's easily a hundred people on eevBlog who'd love to have one of SVG's PCBs to build one; but the guy has bigger fish to fry, what with being hired to design boards for Tesla. 

mnem
*hiding out from the suck again*

Actually bigger stuff than just design, he manages one of Tesla's divisions.
Few years old but still worth a listen if you have the time:
https://www.eevblog.com/forum/amphour/our-3-hour-interview-with-free_electron/

[bitseeker]

I could totes be Captain Canuck-Dwagon:

"Now there, young man... put that firearm away..."

"What the fuck are you talking, ya crazy lizard...?!?" *KA-BLAM!!!*

*Ptoooie* *Ping* "Enough with all the noise, ya hoser... and would ya lay off the foul language, ehh?"

FUCK YOOOOOUUUU!!!" *KA-BLAM-BLAM-BLAM-BLAM-BLAM-!!!* click-click-click* 
*FWOOOOOOOOOM!!* *Sweeps up the ashes, because he is a civilized dwagon*

Now that would be a major deterrent of crime (and avoid clogging up the courts and jails).

[bitseeker]

but then i discovered that the MSD thumbwheel did not work. 

tracked it to a dead 1820-0732 on the option board.  what is an 1820-0732?  since i don't have a manual for the k20 and can't  find the pin outs for the IC on the interweb.....who the fuck knows what it really is?

did some tracing on the board and snooped around with a scope and decided it is a preset counter (set to be a down counter it turns out....well yeah....that is only logical after you noodle it out).  ordered a down counter and worked up a three IC replacement for the one mystery chip.  (ugly as a dog turd, but it works)

success!  here you see it counting a 1 Mhz signal from an 8640b with the gate time set for 456789 uSec.

Nicely done! Added to POI.

[nixiefreqq]

just gotta' get this off my chest.

took the pledge a long time ago to never ever become a tekaholic.

and then along comes med and points out that the guys in new jersey still have some 485's.

leave it to damn new yorker to tempt me with blue boxes from the garden state.

if i cave in to temptation "EXCELSIOR" will become my new favorite curse.



ps  always thought the official motto of the empire state was "wadda' you lookin' at".  turns out it is "excelsior". 

pps  just realized that "excelsior you fathead" was one of sheps lines.  if you are gonna steal....steal from the best.

[0culus]

Well, after just having a 2465B, I have two 7k mainframes now. Really nice stuff.  Once you get it you can't stop. 

[worsthorse]

I powered up the 7603 yesterday, just to see if it worked. What a nice scope it is! I am not sure it if will make it to my bench given my space restrictions but I can see why people like them so much.

[mnementh]

$74 on Amazon...

mnem
 

[worsthorse]

(Attachment Link)

$74 on Amazon...

mnem
 

if you don't behave yourself, i am going to start drop shipping stuff from my shelves to your place in texas.   

[BU508A]

Yeah it's not that much different to that actually. Mine misses the frequency compensation and niceties though and the MOSFET rings like fuck

Edit: Problem with this class of design is that it's very supply voltage and temperature sensitive. I have a better design somewhere I did a few years back that used a couple of LM311's and an LM385 reference. One to control the switching duty cycle as the master oscillator and another to influence that to control the output cycle. That took me a month to get working but it was around 85% efficient!

Also there are some much cooler LM723 designs out there but I haven't tried them yet! One day.

Edit 2: I'm not sure why people use the 555 for this to be honest especially that implementation as a production item. It's totally shit for it. Only reason to do it is if you have nothing else lying around. Why not use an MC34063 or literally everything else that is not that crap. There's a ton of stuff from China in this space which costs nothing these days 

Does it really matter in this case, if this circuit is prone to temperature changes and/or supply voltage as long as you get 90V+?
All you want is to trigger the avalanche.

I have here for comparison my Jim Williams Pulse Gen and the one from Leo Bodnar:

Jim Williams:




Leo Bodnar:

[Specmaster]

@mnem
"I eat my peas with honey;
I've done it all my life.
I know it may sound funny,
but it keeps them on my knife." 

That could be a chorus for your Canadian Dwagon song [emoji1787][emoji1787][emoji1787]

Sent from my POT-LX1 using Tapatalk

[Specmaster]

@mnem you don't hear anything like that happening in Canada, are guns not allowed there or are they just more civilised[emoji4]

Sent from my POT-LX1 using Tapatalk

Lots of restriction on guns here. The only thing you can get easily (need to do a 2 days training session) is hunting rifle and shotgun. Side arms are heavily restricted (police screening, can only be used at a shooting range). Automatic weapons are forbidden.

That must be the British influence coming through. Are people searched on the border with America for weapons then?

Sent from my POT-LX1 using Tapatalk

[med6753]

just gotta' get this off my chest.

took the pledge a long time ago to never ever become a tekaholic.

and then along comes med and points out that the guys in new jersey still have some 485's.

leave it to damn new yorker to tempt me with blue boxes from the garden state.

if i cave in to temptation "EXCELSIOR" will become my new favorite curse.



ps  always thought the official motto of the empire state was "wadda' you lookin' at".  turns out it is "excelsior". 

pps  just realized that "excelsior you fathead" was one of sheps lines.  if you are gonna steal....steal from the best.

If you're trying to make me feel guilty it ain't working.  You better get one before addicted2analogtek grabs them all. 

[med6753]

(Attachment Link)

$74 on Amazon...

mnem
 

if you don't behave yourself, i am going to start drop shipping stuff from my shelves to your place in texas.   

I've always been, and will continue to be, misbehaving. Do you want my address? 

[capt bullshot]

Edit: I WILL NOT BE DEFEATED. How not to build a boost converter: on a breadboard with a 555 and a random MOSFET I found in the junk box an IRF620 ... but it works  . Feedback loop is via shorting the control pin to ground crudely with a 2n3904 and a voltage divider. Had to futz with it for a while to get the conduction cycle right for the pikey inductor. Doesn't need heat sink now due to thing bouncing an amp during conduction 

Don't need no stinkin' switcher IC, nor 555, nor MOSFET to generate somewhat higher voltage

[med6753]

@mnem you don't hear anything like that happening in Canada, are guns not allowed there or are they just more civilised[emoji4]

Sent from my POT-LX1 using Tapatalk

Lots of restriction on guns here. The only thing you can get easily (need to do a 2 days training session) is hunting rifle and shotgun. Side arms are heavily restricted (police screening, can only be used at a shooting range). Automatic weapons are forbidden.

That must be the British influence coming through. Are people searched on the border with America for weapons then?

Sent from my POT-LX1 using Tapatalk

The one and only time I've been to the Great White North in 2006 I was not searched for anything. And this was at a very busy border crossing at Niagara Falls. But I will tell you that the border guard who questioned us (Her) was not very friendly and it seemed like she wanted to tell us "Yankee Go Home". 

[bd139]

Does it really matter in this case, if this circuit is prone to temperature changes and/or supply voltage as long as you get 90V+?
All you want is to trigger the avalanche.

Yeah it doesn't matter too much for sure but certainly no good for a production design IMHO. The thing you do tend to find is that sometimes after it has warmed up a bit, the MOSFET's conduction changes for the worse so the fixed duty cycle can chop off a chunk of the saturation cycle of the inductor. Also because I'm using a shitty MOSFET, if you don't run enough Vcc into it the thing runs in the linear region and just gets warm. Plus the 555 tends not to be as good at sinking current as sourcing it. Also the control divider's switching transistor's action is rather soggy. All situations lead to low output voltage and the avalanche to fail randomly unless you run it hard and that sometimes leads to a "face full of MOSFET" 

The LM311 one I built was pretty sneaky. It used a simple LM311 oscillator with charge steering diodes but the high to low switching point was controlled by comparing the MOSFET drain voltage against Vcc. So literally when the inductor hit near saturation (Vd = ~0.9*Vd), it'd switch state and dump that into the cap. The voltage control was done by changing the off cycle using another LM311 and a voltage reference controlling a BJT current source into the timing cap. I'm sure the 555 could be bodged to do the same with some thought unconventionally as it is basically two comparators in an IC.

I have here for comparison my Jim Williams Pulse Gen and the one from Leo Bodnar:

Jim Williams:




Leo Bodnar:

Thanks for this. Great results. I have a Leo Bodnar GPSDO already and the engineering is excellent. I may have to make another purchase soon 

Edit: I WILL NOT BE DEFEATED. How not to build a boost converter: on a breadboard with a 555 and a random MOSFET I found in the junk box an IRF620 ... but it works  . Feedback loop is via shorting the control pin to ground crudely with a 2n3904 and a voltage divider. Had to futz with it for a while to get the conduction cycle right for the pikey inductor. Doesn't need heat sink now due to thing bouncing an amp during conduction 

Don't need no stinkin' switcher IC, nor 555, nor MOSFET to generate somewhat higher voltage

Yes aware of that. I've got a few HT transformers here which are probably easier now I think about it  ... but not as much fun.

[BU508A]

Don't need no stinkin' switcher IC, nor 555, nor MOSFET to generate somewhat higher voltage

What is that HARTU 1300.401.1?
Google didn't find anything. 

[capt bullshot]

Don't need no stinkin' switcher IC, nor 555, nor MOSFET to generate somewhat higher voltage

What is that HARTU 1300.401.1?
Google didn't find anything. 

You won't find anything about it. It's a custom transformer that happened to sit in my junk box, intended for some other application. I just reuse all kind of scrap (here: http://wunderkis.de/poweronov2/ - it's labeled unfinished stuff, indeed the hw/sw is finished long time ago, but I didn't write up anything about it. It's a power on/off box to control my lab power distribution).
Anyway, the circuit is rather general, will work with many transformers, just needs some windings on a core, roughly matched to the input to output voltage ratio.

[bd139]

Ah yes, a BJT flyback driver with feedback loop. That makes sense. I built one a number of years ago which used a TIP41C and a gapped E core. I discovered pretty quickly how loud one of them can get if you make your own shitty magnetics like I did and it ends up firing up in the 10KHz range . Worked really well though for a bit. It was an attempt to build a converter to generate +/-2Kv for CRT bias from 12V DC but the secondary HT winding flashed over a couple of times and the core former was getting a bit melted from all the repairs so I shelved the project.

Also Royer converters work here pretty well. Here's one I built a few weeks back:



Building switching converters has always interested me.

[capt bullshot]

Ah yes, a BJT flyback driver with feedback loop. That makes sense. I built one a number of years ago which used a TIP41C and a gapped E core. I discovered pretty quickly how loud one of them can get if you make your own shitty magnetics like I did and it ends up firing up in the 10KHz range . Worked really well though for a bit. It was an attempt to build a converter to generate +/-2Kv for CRT bias from 12V DC but the secondary HT winding flashed over a couple of times and the core former was getting a bit melted from all the repairs so I shelved the project.

Remember electronic photo flashes? One could literally hear them charging the flash capacitor - same circuit, often even single transistor without feedback, output voltage limited by the circuit itself - and a neon indicator to show "flash is ready"

[bd139]

Yep that's actually what got me interested in switching converters. I used to dismantle disposable cameras for the flash board and reverse engineered one (zapping myself in the process a couple of times) with the intention of using it as a portable HT supply for a radio but alas it didn't knock out enough current to sustain a high voltage.

Edit: I did find that if you built a naive relay oscillator, you could rectify the spike off the coil and use that to charge a flash capacitor. Relay didn't last more than a couple of minutes though 

[tautech]

Switchers when done properly are bloody marvelous, I had a Uni student do one for me out of 4000 IC's to get a proper VCO controlled one for a piss ant 6V 100mA PV.
It could charge a 2200uf cap to 150V in seconds in bright sunlight or a few minutes under a twin fluorescent light at night.

Dumb sucker didn't know much about HV clearances though as he did the PCB with a solid pour and 20 mil clearances ! 
Charge...charge.... CRACK....rinse and repeat !