X-ray machine

John Miles, KE5FX <jmiles@gmail.com> wrote:
> On Saturday, October 12, 2019 at 10:06:00 AM UTC-7, edward...@gmail.com wrote:
>> I don't think standard dental x-ray will be able to see bonding wire.
>> They are designed to see cavity without giving too much radiation.
>> You might have to increase intensity for higher resolution.
>
> Interestingly enough, dental X-ray machines are not that wimpy.  60 kVp
> at (IIRC) 8 mA is what I usually see on the panel at my dentist's office.
> Even at 1% efficiency they are shooting several watts at me.
>
> For PCB inspection, the power of the X-ray source is a *long* way down on
> the list of important considerations.  Given the right setup, 35 kVp at
> 300 uA is quite usable for electronic inspection on 6- and 8-layer PCBs,
> down to and including the bond-wire level.  The sensor quality and
> exposure time, along with the tube's focus spot size, are what make the
> difference.
>
> This thread is what got me into this stuff:
> https://www.eevblog.com/forum/reviews/faxitron-mx-20-x-ray-system-teardown/
>
> It really sucks that it's so hard to get your hands on a usable microfocus
> X-ray system.  These things are about as hazardous as the color TVs that
> most of us grew up sitting in front of... and that's assuming you've jimmied
> the door interlock.  Just about everybody here would sell a kidney or two
> for an MX-20 if they knew what they were missing.
>
> As it stands, getting a good deal on a good machine is pure luck.  And then
> you get to build a digitizer and write a bunch of software,  unless you were
> *really* lucky and got one that was (probably illegally) sold with its
> original PC full of patient records.  Grumble...
>
> Dental sensors are better than nothing, and at least some of the Gendex
> units have been reverse-engineered for use with free software
> at https://github.com/JohnDMcMaster/gxs700 .  Not sure what kind of results
> can be expected without a microfocus tube, though.  It's possible that a
> Gendex sensor and an older film-based Faxitron would be a good combination.

FWIW, an quick way to verify a system in an operatory is to have the 
sensor take an X-ray of itself. The GXS-700 sensor is a classic. 
    Carestream offers a system that captures ~hundred [1] images during 
X-ray bursts, software stitches the images together, colorizes them with
flesh and bone tones, and then renders the result in 3D images that can 
be turned and viewed from any angle. It's great for patient education 
because it drives home the thinness of your oral bone structure.

Note:

[1] ~hundred not hundreds of thousands. 

Thank you, 73,

-- 
Don Kuenz KB7RPU
There was a young lady named Bright Whose speed was far faster than light;
She set out one day In a relative way And returned on the previous night.

On Saturday, October 12, 2019 at 5:19:29 PM UTC-7, Sergey Kubushyn wrote:
> As I've just finished full testing of recently acquired Hamamatsu L9181-02
> Microfocus Source that turned up to be 100% success, fully working in all
> modes, went without errors through its 2-hour "warmup" (or "aging" as it's
> been called in older models) cycle I will be probably putting my Midmark set
> for sale really soon...

That looks like a great tube, all right.  It picks up at 40 kVp, right where 
the one in the Faxitron leaves off.  Do you know what machine(s) it was
originally used in?

Also, any good PCB images with the Midmark?

-- john, KE5FX

On Saturday, October 12, 2019 at 11:15:11 AM UTC-4, jlarkin wrote:
> 
> Excuse me for wasting bytes. Some day we'll all run out and I'll feel
> terrible.


Don't worry about it! The trolls and spammers waste more than you could use in 1000 lifetimes. :)

John Miles, KE5FX <jmiles@gmail.com> wrote:
> On Saturday, October 12, 2019 at 5:19:29 PM UTC-7, Sergey Kubushyn wrote:
>> As I've just finished full testing of recently acquired Hamamatsu L9181-02
>> Microfocus Source that turned up to be 100% success, fully working in all
>> modes, went without errors through its 2-hour "warmup" (or "aging" as it's
>> been called in older models) cycle I will be probably putting my Midmark set
>> for sale really soon...
> 
> That looks like a great tube, all right.  It picks up at 40 kVp, right where 
> the one in the Faxitron leaves off.  Do you know what machine(s) it was
> originally used in?

It is one the Hamamatsu general purpose Microfocus X-Ray sources:

https://www.hamamatsu.com/us/en/product/type/L9181-02/index.html

and here is their full current line-up:

https://www.hamamatsu.com/us/en/product/light-and-radiation-sources/microfocus-x-ray-source/index.html

Those are very expensive -- I've been offered that exact L9181-02 with a 50%
discount by their sales people, for "just $55K" -- but they are very good
and de-facto standard. There are many other models (which are quite often
the same units but just assigned different part numbers for using them in
customer machines) but this lineup is their off-the-shelf general purpose
ones sold as standalone units.

The last in lineup is L12161-07 which is their current (improved?) version
of those L8121-01/03 units that I had two damaged ones. It looks exactly the
same, just with different control box made for remote control only. The old
one that I still have has full-featured control panel with all manual
controls and displays and it can also be remote controlled so that new one
is actually a downgrade. That last set is their crown jewel, most powerful
of them all and going up to 150kV, higher than anything else. My current
aquisition, L9181-02 is lower power (39W vs 75W) and only goes up to 130kV
but it is very good unit nevertheless, second best of their general purpose
sources. There are 50W units there but they have 15 micron minimal focal
spot size while L9181-02 has 5 micron. And none of those go above 110kV.

The very similar one, L9181-05 is the same as L9181-02 but with wider beam
angle (100 vs 45 degree) that is better for microscopy but not as good for
general purpose use.

Almost all of those except the last two have all control electronics
built-in so they don't need any external control box -- they are controlled
with text commands over RS-232 port.

I would've happily exchanged my L9181-02 for a working 150kV L8121-01 unit
in good condition, even without control box (as I already have one) but it
is almost absolutely unlikely to find one so I'm very happy to have their
second best :)

> Also, any good PCB images with the Midmark?

They are decent enough to find some soldering problems under BGA chips but
not all that great. Then, that size 1 sensor is rather small so you can only
catch some part of interest so it requires several shots to inspect the
entire board. It is kinda useful for some other things though -- e.g. to
find out what holds that @#$#@ plastic case after you seem to remove the
last screw or discover what's inside some potted assembly or something like
this but you won't see bonding wires unless you move the tube far away and
make a maximum possible exposure and even then it would be not all that
great.

BTW, I don't know what could be the use of anything below 30kV (maybe some
soft tissues or something very transparent, low density?) Those low energy
X-Rays are always filtered out on medical machines as they are useless for
imaging while at the same time readily absorbed by the body thus increasing
the radiation doze significantly. It is what they call "Aluminum equivalent"
or so. Most of BGA chips are completely opaque at such low energies; you
need something like at least 50kV to see through them.

However there is a use of low energy X rays but it is even lower than 30kV.
Hamamatsu make what they call PhotoIonizer units that are used as
electrostatic removers:

https://www.hamamatsu.com/us/en/product/type/L12645/index.html

Those run at something like 9.5kV and are of very low power. I have a couple
of their L9873 units that are like that L12645 but have all control
electronics built-in and don't require an external control box. Those are
really nice and they do work. And as it is X rays they are effective at
relatively long distances unlike e.g. "Nuclespots" which use alpha radiation
so they are very short range. And unlike Nuclespots they don't have any
decay with time (VERY SHORT time for nuscespot as it is Po210 source) so
they last forever. On the other hand nuclespots don't require external power
supply but that is a questionable advantage...

---
******************************************************************
*  KSI@home    KOI8 Net  < >  The impossible we do immediately.  *
*  Las Vegas   NV, USA   < >  Miracles require 24-hour notice.   *
******************************************************************

jlarkin@highlandsniptechnology.com wrote:
> On Thu, 10 Oct 2019 20:39:04 -0700 (PDT), tabbypurr@gmail.com wrote:
> 
>> On Friday, 11 October 2019 04:17:44 UTC+1, jla...@highlandsniptechnology.com  wrote:
>>> This counts parts, on reels, trays, or in tubes.
>>>
>>> https://www.dropbox.com/s/7dxshzvpv53y5c3/VisiCon.JPG?raw=1
>>>
>>> Sadly, its resolution is mediocre imaging a PC board. I wish I had
>>> something for small parts, wire-bond sort of resolution, but I
>>> wouldn't use it often enough to justify buying one.
>>
>> Maybe your next project is making one.
>> I gather handling the high EHT is a significant cost. Perhaps the secondary could be in the tube somehow, eliminating all external EHT wiring - I doubt it really.
>>
>>
>> NT
> 
> ebay has dental x-ray units starting around $500, but I think they
> need wet film developing. A digital imager is probably the hard part.
> 
> 
> 
   Don't some x-ray machines use a silicon array for the imager?
   And aren't they more sensitive requiring less exposure time?
   Can't be too hard with all of those 'phones around for homebrew 
prototype.

On Saturday, October 12, 2019 at 7:29:37 PM UTC-7, Sergey Kubushyn wrote:
> I would've happily exchanged my L9181-02 for a working 150kV L8121-01 unit
> in good condition, even without control box (as I already have one) but it
> is almost absolutely unlikely to find one so I'm very happy to have their
> second best :)

Yeah, I can't see wanting more than 130 kVp. I wonder if that would be enough
to start damaging parts.  Not sure where the limits are.  

> BTW, I don't know what could be the use of anything below 30kV (maybe some
> soft tissues or something very transparent, low density?) 

This machine was used for bone density studies on rodents at a big pharma
company.  It came with documentation showing annual service by the Faxitron
tech, right up to the point where they lost the password to the PC.  I don't 
think they used it much, certainly not after that.  Wish I knew how many 
hours were on the tube.

Obviously there were no HIPAA concerns, but they still didn't sell the PC
with it. :(

> Most of BGA chips are completely opaque at such low energies; you
> need something like at least 50kV to see through them.

Fortunately that's not the case.  Some favorites:

http://www.ke5fx.com/XEM3005_MX20_58098.175292_35kVp_15s.png
(XEM3005 FPGA board with Spartan3E and Cypress FX2LP, 3x mag IIRC)

http://www.ke5fx.com/MCP23018_edge_37kVp_15s.png
(MCP23018 I2C extender QFN, standing on edge)

http://www.ke5fx.com/LT8650S_MX20_58230.228355_37kVp_7s.png
LT8650S buck regulator, 5x mag.  Package size is 6 mm x 4 mm.

http://www.ke5fx.com/BB_37KV_325uA_30s.jpg
BeagleBone Black, 1x mag

http://www.ke5fx.com/BB_BGA_small.jpg
Closeup of SoC on BeagleBone, 5x mag

Note that these required exposures up to 30 seconds.  If the machine were
limited to typical dental exposure durations, it wouldn't be very useful.

-- john, KE5FX

On Sunday, 13 October 2019 09:09:26 UTC+1, John Miles, KE5FX  wrote:

> Yeah, I can't see wanting more than 130 kVp. I wonder if that would be enough
> to start damaging parts.  Not sure where the limits are.  

A lot lower than 130kV.  I once used a lot of one-time programmable
microcontrollers (Hitachi 63P01) and wondered whether I could erase
them with X-rays.  I had access to a small microfocal X-ray machine,
so I experimented.
The result was disappointing.  I used the lowest energy I could, which
was probably 30kV or 35kV.  I exposed a quartz window CMOS eprom for
initial testing as this was more convenient - I could read back the
contents at regular intervals during breaks in the exposure.  After
something like 10 minutes exposure at a range of maybe 10cm from the
target the eprom was erased.  I was able to reprogram it and it verified correctly, but only for a short time.  Unfortunately, it discharged
within minutes, so there was permanent damage.  I didn't try things
like annealing it which I only thought of much later.  In any case,
if even short-term damage was being caused by an exposure only just
sufficient to erase the eprom then this was clearly not going to be
a sensible way of treating devices that would be used in a critical
application.
John
 

John Miles, KE5FX <jmiles@gmail.com> wrote:
> On Saturday, October 12, 2019 at 7:29:37 PM UTC-7, Sergey Kubushyn wrote:
>> I would've happily exchanged my L9181-02 for a working 150kV L8121-01 unit
>> in good condition, even without control box (as I already have one) but it
>> is almost absolutely unlikely to find one so I'm very happy to have their
>> second best :)
> 
> Yeah, I can't see wanting more than 130 kVp. I wonder if that would be enough
> to start damaging parts.  Not sure where the limits are.  

It is nice to have and it doesn't damage anything. Also it has almost twice
the power (Large Spot only -- they have same limit with small spot) that
might be useful at times for either reducing exposure time or being able to
see through at all.

>> BTW, I don't know what could be the use of anything below 30kV (maybe some
>> soft tissues or something very transparent, low density?) 
> 
> This machine was used for bone density studies on rodents at a big pharma
> company.  It came with documentation showing annual service by the Faxitron
> tech, right up to the point where they lost the password to the PC.  I don't 
> think they used it much, certainly not after that.  Wish I knew how many 
> hours were on the tube.
> 
> Obviously there were no HIPAA concerns, but they still didn't sell the PC
> with it. :(

Eh, they almost never do and those Faxitrons sold with PCs (I saw a couple
on Ebay) had ridiculously high asking prices.

>> Most of BGA chips are completely opaque at such low energies; you
>> need something like at least 50kV to see through them.
> 
> Fortunately that's not the case.  Some favorites:
> 
> http://www.ke5fx.com/XEM3005_MX20_58098.175292_35kVp_15s.png
> (XEM3005 FPGA board with Spartan3E and Cypress FX2LP, 3x mag IIRC)
> 
> http://www.ke5fx.com/MCP23018_edge_37kVp_15s.png
> (MCP23018 I2C extender QFN, standing on edge)
> 
> http://www.ke5fx.com/LT8650S_MX20_58230.228355_37kVp_7s.png
> LT8650S buck regulator, 5x mag.  Package size is 6 mm x 4 mm.
> 
> http://www.ke5fx.com/BB_37KV_325uA_30s.jpg
> BeagleBone Black, 1x mag
> 
> http://www.ke5fx.com/BB_BGA_small.jpg
> Closeup of SoC on BeagleBone, 5x mag
> 
> Note that these required exposures up to 30 seconds.  If the machine were
> limited to typical dental exposure durations, it wouldn't be very useful.

It is all rather significantly above 30kV. And long exposure times. For PCB
inspection you need real-time video and moving object fixture that allows to
move and rotate the inspected object. This allows to do optical zooming and
do everything in real time. You turn the voltage knob and see in real time
how different parts become transparent. You move the object between the tube
and sensor and magnification changes in real time so you can look at
whatever part you need at the angle and magnification you need and you can
chose optimal voltage/current so you can see what you want to see. It is not
like higher is always better -- any material has it's range when it is
opaque, totally transparent (i.e. disappearing from the picture) and
anything in between.

---
******************************************************************
*  KSI@home    KOI8 Net  < >  The impossible we do immediately.  *
*  Las Vegas   NV, USA   < >  Miracles require 24-hour notice.   *
******************************************************************

On Sunday, October 13, 2019 at 11:57:10 AM UTC-7, Sergey Kubushyn wrote:
> It is all rather significantly above 30kV. 

True, but my point is, you don't need 50 kV.  You'd *like* 50 kV+, given the
option, but 35 is much better than nothing.

This thing goes down to 15 kV if I remember correctly, but I've never 
seen an application for that.  Occasionally when shooting through plastic 
or something similarly transparent I can get better contrast by going down 
to 25 kV or so, but the difference isn't dramatic.  I overclocked it to 37,
and that's usually where I run it.

> And long exposure times. For PCB inspection you need real-time video and 
> moving object fixture that allows to move and rotate the inspected object. 
> This allows to do optical zooming and do everything in real time. You turn 
> the voltage knob and see in real time how different parts become 
> transparent. You move the object between the tube and sensor and
> magnification changes in real time so you can look at whatever part you 
> need at the angle and magnification you need and you can chose optimal 
> voltage/current so you can see what you want to see. It is not
> like higher is always better -- any material has it's range when it is
> opaque, totally transparent (i.e. disappearing from the picture) and
> anything in between.

Yep, all very true.  This is the kind of thing you want near your 
workbench, not on the factory floor.  On a production line a low-voltage
cabinet machine wouldn't be ideal.  

BGA inspection in particular is a lot more involved than just shooting 
X-rays at the board and calling it good or bad.  JL has some nifty optical 
toys for that job already.

-- john, KE5FX

On Sunday, October 13, 2019 at 3:26:26 AM UTC-7, jrwal...@gmail.com wrote:
>I was able to reprogram it and it verified correctly, but only for a 
> short time.  Unfortunately, it discharged within minutes, so there 
> was permanent damage.  I didn't try things like annealing it which 
> I only thought of much later. 

Interesting.  How would you anneal it?  Heat and/or high voltage?

I haven't tried erasing EPROMs with mine, but it does make sense that
a 10-minute exposure to X rays in the PHz rangee would be comparable to 
a much longer exposure (days?) to UV in the EHz range.  I've definitely 
ruined EPROMs with excessive UV erase time.

I should try nuking a 24LC64 or something and see how that goes.

-- john, KE5FX