balrog changed the topic of #homecmos to: Homebrew CMOS and MEMS foundry design | Wiki: https://github.com/homecmos/homecmos-wiki | Repositories: https://github.com/homecmos/ | Logs: http://en.qi-hardware.com/homecmos-logs/
Degi has quit [Ping timeout: 264 seconds]
Degi_ has joined #homecmos
Degi_ is now known as Degi
Bike has quit [Quit: Lost terminal]
_whitelogger has joined #homecmos
_whitelogger has joined #homecmos
_whitelogger has joined #homecmos
<Degi> nmz787: Maybe you can build it yourself. For 1.5 kW at 27.12 MHz I saw one using a vacuum capacitor (100-500 pF I think) between RF in and the coil and an air capacitor between RF in and ground, I think they should cost maybe 300 $ total?
<Degi> Oh also that was for an ICP torch, not capacitively coupled plasma.
<Degi> (Variable capacitors of course. You can get 500 pF 5 kV vacuum variable for 50 bucks on ebay, though this one used 10 kV)
<Degi> I think you need an inductor and a capacitor to match capacitive loads. I think low side cap can be air variable and the inductor should be variable too
<Degi> How good can your generator deal with reflections? Does it reduce power when the SWR increases?
<Degi> I think the 850 $ for the whole ICP unit sounds fine I think. Depending on how much time you want to invest
<Degi> You can get a whole mass spectrometer with 27.12 MHz 1.5 kW ICP torch for 400 bucks more though https://www.ebay.com/itm/HP-AGILENT-4500-ICP-MS-YOKOGAWA-G1820A-SHIELD-TORCH-PARTS/273629090416
Bike has joined #homecmos
_whitelogger has joined #homecmos
<nmz787> Degi: hmm, I am still not 100% clear on what capacitively coupled means... like, is that just where the electrodes are in the plasma chamber, and there's sort of an arc between them? basically the same physical config of the electrodes that would support DC plasma arc?
<nmz787> and yeah, I also saw some air caps on ebay for pretty cheap, not sure what kV 1kW would produce though... such that I could get the right kV rating
<nmz787> being cheap enough, I suppose I could just get some and try, and if it arcs across the plates, get another with larger plate spacing and try again
<Degi> Well capacitively coupled is what you have when you dont inductively couple
<Degi> Hm if you use it on the low side, it should be determined by the 50 ohms at 223 V if its properly matched
<Degi> (223 V RMS, I'd get one rated for 1 kV)
<Degi> Capacitively coupled is like when you have a big plate capacitor and inbetween is your plasma (The plates are probably electrodes outside of the vacuum chamber to prevent contamination) though ICP is better since it doesnt touch any walls at all (like an induction heater but with plasma instead of metal inside)
<Degi> Which RF generator do you have?
<nmz787> my amp's datasheet is weird, it says regarding "load mismatch tolerance" that it can "Continuous duty into any passive load without failure or oscillation"
<nmz787> but then for "protection" it says
<nmz787> "Forward power limits on PA current or excessive reflected power ... 75 watts"
<nmz787> those are the bottom of page 11, top of page 12 (as they wrote, not the actual PDF page)
<nmz787> pdf page 18 and 19
<Degi> Oh maybe it measures the reflected power and turns down power
<nmz787> possibly, it's variable power with pretty good resolution
<Degi> The RF gen I have (yet to get it working; it was extracted from a bigger device) downregulates output power when reflected power is high (also the FETs are rated 50:1 VSWR so welp)
<Degi> Oh does it need an input signal?
<nmz787> hmm, I don't think so
<nmz787> it's 13.56 MHz
<nmz787> there may be a modulation input if desired, but I don't think it's required
<Degi> It says something about exciter input kinda wondered
<nmz787> probably much lower frequency input if anything
<nmz787> looks like cap coupled means electrodes in chamber, touching plasma
<Degi> Hm you can put the electrodes outside too
<Degi> Hm no idea what the protection is. Maybe it downregulates the power to that point. What kinda device does it use for amplification?
<Degi> Oh geez the RF30S is pretty pricy on ebay.
<nmz787> hrmm, not sure, haven't opened it up :P
<nmz787> got it for free though, so I can't complain too much about needing to spend some $$$ for matching
<Degi> Heh yes
<Degi> Got mine for free too
<fltrz> nmz787, internal electrodes will typically wear, and may cause x-rays
<fltrz> depending on vacuum pressure
<fltrz> you also want to make sure the RF stays in the chamber, but I've seen mostly glass ones too I think (I may be confusing that last part with sputter chambers)
<fltrz> nmz787, if you ignore the magnetic field and the spectroscopy facet, it should be possible to compute the impedance from the equations in https://warwick.ac.uk/fac/sci/chemistry/research/oconnor/oconnorgroup/teaching/ch908-copy/recommendedtexts/msr_1998_marshall_ftms_primer.pdf
<fltrz> energy is dissipated by ions colliding with other molecules
<fltrz> nmz787, it's nto clear from the manual (I just started reading it) what the RF output port of the device is like. is it 2 shielded conductors? or is it on big fat coax?
<fltrz> nmz787, you are correct about optionality of exciter input, this is for maintaining phase lock between multiple identical generators, the exciter should signal have the same frequency as the unit and is essential a lower power version of what the unit should produce
<nmz787> yeah def aware the internal electrodes will wear, x-rays usually not as worrisome from a process standpoint afaik
<nmz787> since the plasma will be so much more reactive
<nmz787> sputter chambers often have one electrode being the sputter "target" metal, and the other electrode is the chamber wall
<fltrz> nmz787, your output impedance is 50 Ohm
<nmz787> yeah, I saw that
<nmz787> just opening that PDF now
<fltrz> nmz787, I don't know if electrode material contamination can be an issue for you, so I guess that is a reason for exciting externally
<nmz787> and yeah I have seen the sections of that manual about coupling generators, but haven't dug too deep into the manual details
<fltrz> i.e. capacitively through the walls, what is your chamber made of? glass or metal?
<nmz787> mostly just checked basic specs, pinout of the analog connector, reading up on the digital (rs232/485) connections
<nmz787> chamber is yet to be decided
<nmz787> I know ICP needs it to be dielectric
<nmz787> at least the plasma generation portion
<nmz787> I've seen some stuff like test tube with coil, which can be connected to the side of a metal chamber
<nmz787> so you can reduce the amount of glass I guess
<fltrz> nmz787, also make sure you verify the coax cable voltage and power limits, and the absorption spectrum per meter, so you can calculate the heat generated in the first meter of cable
<fltrz> what does ICP stand for btw?
<nmz787> inductively coupled plasma
<fltrz> oh yeah
<fltrz> I wonder if you need to spark a small plasma to get it started, or if the generator will shut down if some small amount of background ionization doesn't ionize other molecules fast enough...
<fltrz> nmz787, if you have a scope and a signal generator you could check what the reflection on the coil/electrode setup is with a small signal, but I expect the impedance to change once the plasma lights up
balrog has quit [Ping timeout: 240 seconds]
<nmz787> apparently SWR meters are usually used for efficiency check and adjustment feedback
<nmz787> Standing Wave Reflection
<nmz787> the plasma cleaner I've used before doesn't seem to have a sparker inside
<nmz787> just a coil around a small section of glass tube
balrog has joined #homecmos
<fltrz> nmz787, the generator kind of has its own SWR built in, it measures reflected power, so consider that fraction of the forward waves with same amplitude as the reflected power waves, together they form the standing waves
<nmz787> but I haven't heard back from the mfg sooo I can't see the pinout or other specs to see if I could interface it to my generator
<nmz787> and it's a lot more money to gamble with vs a few $50 on variable caps an inductors
<nmz787> not too sure what the chamber is for, it doesn't look like the bottom is open, since the one pic of the bottom has some reflections visible
<nmz787> but hard to say
<fltrz> nmz787, I didn't fully read the manual yet, but it looks like they claim theres a matcher built in?
<fltrz> also given the claim about any load
balrog has quit [Ping timeout: 250 seconds]
<nmz787> there are some connection diagrams that show a separate matcher box tho
<fltrz> oh didnt get that far yet
<nmz787> which isn't there when they show a dummy load box
<nmz787> pdf pg 27, vs 29
<nmz787> then after 29 it shows the match network hooked back up to the supply, presumably for the automatch function
<nmz787> I guess maybe using the supply's integrated SWR to feedback to the matcher
<nmz787> or to lower the power in case of mismatch or arcing detected
balrog has joined #homecmos
<Degi> Hm so do you have ICP or capacitive coupling?
<Degi> Lol you sure that RF matcher is for only 1 kW?
<nmz787> no idea, it says "10" in the model, so kinda assuming since my generator is also "10" and means 1000 watts, that it's similar
<nmz787> but that could be completely wrong
<nmz787> thus why I have held off making any money decisions until hearing back from the manufacturer, who didn't respond all week
<nmz787> found a forum post online about a microwave generator, and the poster said the company didn't respond after 3 months
<nmz787> (same company)
<Degi> It could be for 10 kW
<nmz787> and I don't have capacitive or ICP yet... still undecided on which regime to use, different stuff might call for different processes is my current thought
<Degi> Also if you want microwave coupled plasma, just use a microwave magnetron yet
<nmz787> nah I don't think I want microwave!
<nmz787> just that search came up when I was googling for info on the company
<Degi> I think it has not so good penetration
<nmz787> like, maybe this company just sucks and I want to avoid
<Degi> Heh why no microwave
<nmz787> I just haven't heard of it being used, so far
<nmz787> 13.56 MHz is quite common
<nmz787> I did however make some microwave oven plasma a week ago
<Degi> My DIY microwave oven sometimes does that
<fltrz> Degi, how do you make a DIY microwave oven??
<nmz787> just took a beaker turned upside down, lit a match under it, and quickly turned on the oven
<fltrz> I think 13.56 MHz will be in ISM band so unregulated, wouldn't try that with microwave at high power though...
<nmz787> a piece of very charred wood also worked quite well
<fltrz> PA is Power Amplifer?
<nmz787> did it after watching this video https://www.youtube.com/watch?v=l0u8Vtf2GoQ
<nmz787> fltrz: that's my guess
<Degi> Lol just stick a magnetron onto two cut-out acetone cans?
<Degi> Idk 2.54 GHz is ISM
<Degi> Hm beaker didnt melt?
<Degi> You can easily melt basically everything with a microwave
<Degi> With enough time and thermal insulation
<nmz787> it got hot
<nmz787> but I ran things for several minutes
<nmz787> threw some plastic in there too
<nmz787> it was not the best smelling idea
<nmz787> there's the whole microwave sintering too
<Degi> Hm idk I've reached above 2000 C within a minute or so. But the beaker has probably high enough therma conductivity
* nmz787 steps out
balrog has quit [Ping timeout: 256 seconds]
<SpeedEvil> 2000C? Not F?
<Degi> Lol yeh
<Degi> You can probably melt thoria with it
balrog has joined #homecmos
<fltrz> matching network: 2 variable capacitors on coax
<fltrz> presumably just shafted variable air capacitors?
<Degi> Hm
<Degi> In the ICP matcher I had it used a vacuum variable and an air variable cap
<fltrz> so the coax has ground shield, and an inner conductor, so they put one in series with the inner conductor and one in parallel
<Degi> Apparently theres an inductor to
<fltrz> i.e. "short" coax with a cap, and put one in series with either shield or inner conductor I guess
<Degi> Series with the shield?
<Degi> The ICP matcher had an air variable parallel to the coax input, then one in series to that with the coil in series too. Basically a resonant tank with a split capacitor to introduce energy
<fltrz> Degi, I read that super simple "schematic" as put RF---coax---innerconductor---var cap----Chamber "+" and RF----coax----shield---var cap----[inner conductor---chamber "-" or Chamber "-"]]
<fltrz> Degi, ah I see
<fltrz> p 12 has a more "detailed" schematic
<fltrz> nmz787, p 15 shows they use the electrodes internally indeed, for capacitive wafers on a wafer holdr, and both a ceiling plate and wafer holder are electrodes
<Degi> p20 shows ICP version
<fltrz> ok my education for today has finished