Ceramic pcb evaluation project brainstorming

[coppercone2]

I am getting close to being about to make a silver glass sintered pcb.

what would be a simple circuit used to test the performance of my work against a fr4 card?

 I wanted something that might highlight a quality ceramic pcb over a regular board from the stand point of a rf or microwave measurement.

Does anyone have any ideas? Like circuit topologies that are fussy about substrate.

I know I can try to measure a strip line or something, but I was curious if there are some amplifiers or filters that are unstable on normal substrates but might behave on ceramic ones.

[uer166]

An easy one is a 1/4 wave stub transmission line notch filter: the notch will have sensitivity to Dk and length variation. You can make it on both kinds of boards and make a notch frequency vs temperature graph from -40 to 85C to compare stability.

[coppercone2]

I can actually meet that temperature range, its exactly what I was looking for

I expect my trace quality will be low, but it should be stable. the temperature sweep sounds like it will be the 'proof' of this concept working.

[coppercone2]

well my test trace works. I need to figure out the screen printing or whatever, but the chemical mixture works. I would love to know what palladium does to the silver/glass mix though. I see they add 10% - 20% palladium.

[T3sl4co1l]

Note that you can control sensitivity to Q by varying the coupling factor to the stub; for example, a parallel-coupled microstrip might have say -30dB coupling (give or take spacing and electrical length), which will match nicely to a Q ~ 30 resonator, which could be a one-end-grounded TL stub.  Coupling should be towards the GND end of course.

The stub can be a spiral or zig-zag if you need more length, to drop the resonant frequency in the range of whatever instruments.  Structures of some inches length will place in the low GHz, and so on.

Might be worth quantifying if you can make a solid ground plane as well, i.e. minimal porosity.  A three-layer capacitor, with one side driven (say, a smaller polygon within the bounds of the PCB), shield/GND in the middle, and a sense electrode on the far side (also pulled in from the edges -- that way the driven and sense electrodes have minimal fringing between each other), which can be just insulated with plastic film and held with adhesive, and then measure the voltage on one by driving the other.  Impedance will be high and signal level low, so consider a high-impedance (JFET etc.) amp and averaging to reduce noise.

Next step up from that, magnetic shielding, simply do the same thing but with a pair of induction coils, and low impedance amplifier; it'll be transparent at low frequencies (not that the coils themselves will induce much, but you know, calibrate with coils in position vs. with PCB inserted between them), but the interesting question is how much isolation is had at high frequencies (100s kHz), and where the break point is.

Trace resistance of course also an important figure.

Preferably, many of these with production statistics, to capture how consistent/reliable it is, but that probably comes later, heh.

Tim

[T3sl4co1l]

well my test trace works. I need to figure out the screen printing or whatever, but the chemical mixture works. I would love to know what palladium does to the silver/glass mix though. I see they add 10% - 20% palladium.

Sintering aid, or perhaps something about corrosion resistance?  Phase diagram, for what it's worth:



So, MP lowered a little bit, mushy range means a lower-melting mixture and ready diffusion while solid phase can kind of retain its position, give or take exact composition and firing temperature of course.

Tim

[coppercone2]

Hmm I think maybe the key to making it 'solid' might be to crush the wet paste with rollers. They seem to use rollers like for dealing with silicone to make the paste in some documents I read. Ball milling the powder might also make it better if it beats the glass into the silver.

I see another part of the game is going to be getting high quality alumina, of course the low porosity one is more expensive

I will get palladium first to see how it does on the other side of the razor blade

I have a feeling the best procedure would be to ball mill the mixture together for a while, then mix up a sizable amount with the PVA binder using rollers.


I am glad this works well, unlike that total bullshit with oxidized tin chloride 

[coppercone2]

it also does have a white coating of some kind that is shown in applied science video and appears in my silver too. Maybe the palladium prevents that from growing.

[coppercone2]

lol, what if I put a tad of silver brazing black flux into this mix?

[twospoons]

Seems like a very expensive way to put a conductive trace on alumina (Pd is currently USD 30 per gram).  Did you consider electroless copper, then electroplate up to required thickness, then use standard pcb etch techniques? Only needs catalytic quantities of Pd (mg).

https://www.researchgate.net/publication/321292890_ELECTROLESS_COPPER_ON_ALUMINA_AND_OTHER_SUBSTRATES  Though the pre-etch for the alumina in that paper is a somewhat terrifying mix of sulfuric acid, hydrofluoric acid and chromic oxide  .

[coppercone2]

that electroless stuff. I tried it, I bought high purity tin, turns out its oxidized. getting tin chloride pisses me off more then paying for the precious metals lol

its supposed to dissolve clear, I end up with a glass of milk

 My opinion on electroless plating is similar to my opinion on the ITER

I expect a circus trying to get tin chloride that dissolve clear, even in strong HCl.


I have a good feeling that silver won't expire and rage me either. I am telling you, its a circus, I got like 10 beakers and lab bottles, a whole bunch of different chemicals. Then I find the process online that everyone keep recommending is freaking dubious as hell (the patents describe using 10x more palladium chloride) and the reagents are downright unstable. IMO that guy that wrote that article about plating is lucky as hell that it worked so smoothly.


I made a list of hopeless technologies that are way under estimated in their ability to piss you off and cause difficulties
1) aluminum anything
2) plating
3) literarlyany-FET technologies
4) plastic films

[uer166]

Note that you can control sensitivity to Q by varying the coupling factor to the stub; for example, a parallel-coupled microstrip might have say -30dB coupling (give or take spacing and electrical length), which will match nicely to a Q ~ 30 resonator, which could be a one-end-grounded TL stub.  Coupling should be towards the GND end of course.

Can you do a quick drawing? In my mind I would like to see a directly connected stub to a transmission line for very narrowband notch, with maximum sensitivity to test this out. If you reduce the coupling to stub, presumably that will also reduce the rejection at the notch? Would it also increase bandwidth?

[coppercone2]

it kind of sounds like a directional coupler

[T3sl4co1l]

Note that you can control sensitivity to Q by varying the coupling factor to the stub; for example, a parallel-coupled microstrip might have say -30dB coupling (give or take spacing and electrical length), which will match nicely to a Q ~ 30 resonator, which could be a one-end-grounded TL stub.  Coupling should be towards the GND end of course.

Can you do a quick drawing? In my mind I would like to see a directly connected stub to a transmission line for very narrowband notch, with maximum sensitivity to test this out. If you reduce the coupling to stub, presumably that will also reduce the rejection at the notch? Would it also increase bandwidth?

Decrease actually.  So maybe that works against sensitivity, depending on what instruments one has to test it.

Something like this:



I would guess ballpark, the coupled length should be maybe 1/4 the total length of the stub?  But again, depends on coupling factor and stuff.  If the build ends up really high Q, little coupling will be required to still get a prominent (if extremely narrow!) notch.

I am kind of assuming you have the means to make vias.  If not, just curling ground around the edges should suffice, for geometry like this.

The far end terminator can be a thru connection of course, in which case you can measure s12 = s21 and the transmission dip is a reflection peak.  As a one-port, reflection (probably use a reflectance bridge) of course will do.

Note that it'll have significant radiation resistance, depending on how it's zig-zagged/coiled up.  A shield over top turns that into more of a hybrid-dielectric stripline geometry that reflects radiation back, keeping Q high (while introducing cavity modes of course, but don't worry about those, just concentrate on the microstrip mode).

Having just a regular transmission line isn't a bad idea either, microstrip with two edge-launch footprints at either end.  Make it zig-zag around a bit to make the most of the board area, and characterize that with whatever means are available (VNA, TDR..).  Printing a range of trace widths is probably a good idea.

Tim

[David Hess]

Compensated high impedance dividers are sensitive to the substrate's dielectric constant being consistent at low frequencies.  They will not function properly with all FR4 substrates, so sometimes an exotic substrate is used.

[coppercone2]

I see the neutron star is missing from the middle of the circuit


I think one of my filter programs does microstrip filter generation for band stop/notch (distributed), I wonder what that will come up with. I just need to find it lol


Brazing fast-on 250 connectors is no fun, I had to silver the thermocouple wire and solder it twice... the fast on really like wicking solder into the tab slot. I guess I should have bought some machined fast-on connectors with a solder cup instead of being a cheap skate.. if they even exist. or packed it with something to prevent wicking in there... god damn omega for not putting screw terminals on this shit. who puts fucking blades for thermocouples!?!! almost every aspect of this kiln project is extremely whack and annoying

I guess they expect you to put another 50$ interconnect thats going to go intermittent open next to the PID just so you can crimp the stranded thermocouple wire and connect that into solid thermocouple wire. Those TC blade connectors are such shit, but even that would be preferable to using 250 fast ons. Clearly they expect you to use both to interface with their circuit, or get some kinda solid high gauge thermocouple wire crimping system, which I think I would rather literately die then buy.


I hate the thermocouple so god damn much. seriously why the FUCK dont they just add some screws there for thermocouple wire instead of making me play this god forsaken games with these fossilized vines


But yeah if you silver it with 56% braze then clean it you can solder it into fast-on connectors OK. Its probobly gonna give some kinda error but fuck it, at least I think it might not start giving my open circuit bullshit like the ceramic thermocouple blade connector I put in my reflow oven. yeah that turned into a percussive maintence requirement. rated to 700c my smelly ASS, it won't last a year with 350 

I know no one likes them, at work we found a big drawer full of old ass fiberglass stiff wire thermocouples, they just kept shoving it in there to make a giant tumble weed over a decade. the crazy technician was gonna try to sort it out but I took it to the fucking dumpster in a hurry. it was like a 2x2 x1 foot cube of thermocouple wire, you could pour cement in there and it would be good for a nuclear bunker

[coppercone2]

How tall can I make a trace before it starts causing problems?

I am actually not familiar with the copper thickness for RF boards and striplines. Can you go thicker on these like 3oz equivalent ?

I need to make a decision on how thick to make my stencil. If I go thin maybe I can save silver but if I go thick I have the option of maybe lapping / polishing it to get it nice

I think most RF stuff is between 0.5 - 1 oz?

[T3sl4co1l]

Top surface finish doesn't matter, nor does thickness much. Most of the current flows on the bottom face, and at that, towards the edges.  The surface roughness and porosity of the substrate matter much more.

Tim

[coppercone2]

Hmm that does make sense that it travels close to the negative. You still need to scrape it, unless the palladium gets rid of the white residue, so I prefer to have it on the thicker side. At least it means I don’t need to polish it

[coppercone2]

Alright I did a 20:1 ratio of Ag/Palladium


Now I am getting suspicious that it might work with other stuff to make other things.

How about some bullshit cheap materials mixed with glass

Ideas:
carbon + glass
nichrome powder + glass


Does anyone know if you can make some resistors? Is silver the only thing that is low temperature enough to sinter? Carbon is just weird because it sublimes

How about bootleg printed resistors made with maybe iron, manganese, other elements? I mean for like stenciled components, not big hardware parts or slugs etc


Is it really like that, we can't make anything but traces without getting crazy?

I need a array of very very small mixing cups

I almost feel like rust HAS to sinter easily. maybe rust-glass. It just seems like it does


My intuition tells me that it would be unlikely for there not to be bootleg resistors you can make, but it might be unlikely to get good dielectrics that you can make. All that metalish stuff is sticky, it all welds. I can understand that pottery is difficult but metal freaking welds. usually the problem is that its brittle. I am thinking that as usual, magnesium and aluminum are gonna require NASA, everything with aluminum is always bullshit, regardless of the form. I think the only thing might be to avoid titanium, aluminum, magnesium and toxic metals like lead and ruthenium. I am thinking there is alot of different powders that are relatively safe to try to make resistors with.

[twospoons]

The thing about silver is it will cold-weld at room temperature. Its one of the reasons the old wirewrap stuff lasted so well: silver plated pins and silver plated wire would bond better over time.  Heating simply accelerates the process.
I'd bet bronze powder would work. Sintered bronze bearings are common.

I'd stay away from metals that react readily with oxygen at temperature: titanium, aluminium and magnesium all fall into that category.  Unless you want to start messing with inert atmospheres using argon as a shielding gas.

Thick-film resistors are made with things like ruthenium and rhenium oxides.  If you want low tech and low temp, just use graphite with a suitable binder to make a conductive ink.

[coppercone2]

Ehh I just wanted to maybe try to mix small amounts of different cheap common oxides like a small dot and measure it, I doubt I will get any high end materials. It is just fascinating

[coppercone2]

But I do have a question. They say these weird metal oxides.


How about silver oxide? is it just too conductive?


edit: finally nailed that fucking fly that has been harassing me for an hour, I think the room temperature just dropped 30 degrees.   

[twospoons]

You could try Antimony doped Tin oxide - thats conductive. So is lead dioxide, but you said you don't want to play with lead. Theres also the titanium suboxides - Ti4O7 can be made from titania (TiO2) by reduction with carbon, hydrogen or titanium metal at 900C to 1000C.
Pretty sure magnetite is conductive too, and so is manganese dioxide.

The trick is picking something that will sinter readily at temperatures you can attain.

[coppercone2]

Actually the ruthenium is not that bad because you would need to heat it to above 1100C in air to get anything to happen. At glass temperatures its pretty much like anything else. I got scared because the next oxidation state makes something potentially unpleasant, but after reading chemistry book it looks like you need to have some intent to 'form' oxides, they have some specific apparatus for this with gasses and quartz etc

I could only see that as a problem for someone using doing something really stupid

For lead, I was always taught that getting it too hot is when the trouble starts. Like welding lead with a OA torch.


I think I can try ruthenium-glass. it seems like its well known and defined in various patents etc.


And no, making your own Rhenium is not easy either, to get from the powder to O2 it looks like a pretty serious process. 24 hours of heating, at two temperatures, then some quartz flow reactor.. thats a damp and drunk squid . And iridium is literally too expensive, like seriously.


Its expensive but I see enough about ruthenium that it looks like its pretty much a good bet like the silver glass is. And it looks like you can add copper or nickel to tune the TCR. And it looks like some of these cheap oxides might be the stuff used to make PTC/NTC.



It also begs the question, can you make a MOV while your at it. I probobly should take 2 weeks off before buying anything. this may be a bad financial decision. 


There is another interesting 'integration' feature I noticed with this silver too. They have those little ceramic heat sinks. You could probobly use them as a substrate to make a "power" substrate. You could solder a transistor tab right to a metalized region on a ceramic heat sink that is acting like a PCB. This means you can make like a pretty beefy metal oxide power resistor. I imagine its very stable considering there would be no thermal grease on this type of circuit.