Good soldering flux?

[shapirus]

Time to repeat the experiment. Here goes.

16 days later. Hi-res photos are attached.

Seven fluxes, left to right:

2) MG Chemicals 8341
3) Relife RL-422-IM
4) Mechanic 225
5) Mechanic 223
6) F-2000 (local stuff)
7) Lukey L2010
“Amtech” NC-559-ASM from Aliexpress

Of the above list, numbers 6 and 7 are not usable as fluxes at all -- same result as using no flux at all.

The list starts at number two, because the famous libreoffice calc does not allow to (or I don't know how to) unassign the number 1 from the first row used as header and start numeration in the actual data section.

I also did some quick soldering and conductance (before heating) testing, but the results require conversion from libreoffice spreadsheet to bbcode, so I'll maybe post them later. In a nutshell, 6 and 7 are useless (same as no flux at all).
With 6 and 7 excluded: all are good for soldering experience, except for 3, which among them has worse wetting (yet works quite fine in some scenarios), but is better in terms of the amount of fumes.
4 and 5 are the worst for fumes and cleaning, but are the best for wetting.
2 is the best for cleaning, 3 is slightly worse, the rest are significantly worse.
4 is the worst for conductance (8.5 GOhm in a controlled test that I'll describe later), followed by 5 (18 GOhm), 8 (21 GOhm), 2 (50 GOhm), and 3 (120 GOhm). 6 and 7 are outside of the megohmmeter's range (>200 GOhm).

Corrosive activity -- before heating -- can be seen in the photos. MG chemicals 8341 is the worst gel flux in this regard that I have ever used. "No clean", my ass.
Relife RL-422-IM has only a slight hint of green, F-2000 is corrosive (but it's not a flux anyway), and the rest are non-corrosive, except that the copper under number 8 became noticeably darker.

[Smokey]

Time to repeat the experiment. Here goes.

16 days later. Hi-res photos are attached.
...

...But you left off my favorite flux, TF5000, so I'm going to have to ask you to do that over...  hah!

Epic experiment!  Thanks for doing that.  If you post csv file, or contents here it's trivial to paste that back into any spreadsheet. 

About the flux attacking bare copper... how big of an issue is that really?  I'm only ever using flux over board area that is either soldered or has soldermask. 

[wraper]

...But you left off my favorite flux, TF5000, so I'm going to have to ask you to do that over...  hah!

I tried it quite a while ago. Cannot call it anything better than useable. Nothing exceptional in soldering properties and IIRC it was slightly conductive before reflow.

[shapirus]

...But you left off my favorite flux, TF5000, so I'm going to have to ask you to do that over...  hah!

Feel free to mail me some, as it's not sold here :).

Epic experiment!  Thanks for doing that.

More is coming: I ordered two more Relife fluxes, they'll be here in a week or two.

If you post csv file, or contents here it's trivial to paste that back into any spreadsheet. 

I think of creating a github repo with a markdown-formatted plain text file. Will use a proper yaml file instead of the spreadsheet junk for the raw data and then generate whatever representation I need from that.

Maintaining it on the forum manually will be tedious, especially as more data are added.

I think I'll do that when those new Relife fluxes arrive and I test them.

About the flux attacking bare copper... how big of an issue is that really?  I'm only ever using flux over board area that is either soldered or has soldermask.

Yeah that's a good question. Not much of a problem in practice, I think: most of the residue will be washed away (so much for "no-clean") during cleaning, and then the copper will be protected with either soldermask, or solder. But it's going to be worse for homemade boards without any coating, and worse yet for soldering insulated wires.

[shapirus]

More is coming: I ordered two more Relife fluxes, they'll be here in a week or two.

Both are crap.

I'm not gonna be bothered testing them for any other parameters after I measured their resistance:

1) RL-420S-UV: 83 MOhm at 260 V; <20 MOhm at 2.5 kV
2) RL-421S-OR: 23 MOhm at 260 V; arcs at 2.5 kV

Test setup: a layer of flux ~0.6 mm thick applied over two parallel strips of tinned copper on FR-4, each 20 mm x 2.5 mm, spaced 2.0 mm apart, long sides facing each other (resembling a capacitor symbol).

Other fluxes measured 8.5 to 120 GOhms in the same test.

They can also be easily measured with BM869s (low voltage) in the nanosiemens range.

Requested a full refund for my Aliexpress order on the grounds of false description (which specifically mentions "high insulation" and "no conductivity", albeit the latter is stated along with "no resistance"). Will see what they will have to say lol.

They were packed nicely, though, and had needles included. But who cares if they are useless for what they are advertised for.

[jonpaul]

Geuine rosin, Kester 44

We use only 63/37 Eutectic solder with those flux cores.

Avoid commie china junk solder or wire



j

[tooki]

Geuine rosin, Kester 44

We use only 63/37 Eutectic solder with those flux cores.

Avoid commie china junk solder or wire

Do you even bother reading threads before replying? It’s clear from the original post onward that this thread is about external flux, not flux cores.

I love Kester 44 flux-core solder, but unfortunately Kester stopped making the 44 paste flux many years ago. The liquid version, 1544, still exists but is only sold in large jugs. (There are some resellers that break those down into small bottles, at least in USA.)

Of course, both 44 paste and 1544 would fail shapirus’s requirement to be non-conductive before heating. I don’t think he understands that flux specs refer to the conductivity of the post-soldering residues.

[shapirus]

Of course, both 44 paste and 1544 would fail shapirus’s requirement to be non-conductive before heating. I don’t think he understands that flux specs refer to the conductivity of the post-soldering residues.

Not that I don't understand, but I don't care. I test the fluxes I get my hands on for my specific requirements, which is hand soldering of both THT and SMD parts. And I believe this specific use case is of interest to many electronics hobbyists that can read this, too. Another way of looking at it is worst case scenario testing.

Besides, the two fluxes from my last post are specifically advertised for BGA rework and general repair, which does not guarantee full heating, either (I believe even when using a hot air gun). And hey, they are 500-1000 times (!) more conductive than the other fluxes I tested.

It would also be interesting to test conductivity after full heating, but that would require having a reflow oven, which I don't have, and the motivation to do it, which I also don't have, since I don't do that kind of soldering.

...but I might try something, just for a quick test: for example, set the hot air gun to what, say, 280 °C, heat the PCB with the flux on top from below (to avoid direct blowing on the flux so that more of it stays in place) for, say, 30 seconds, then let it cool and measure what remains. Just to see if the "it must be heated to stop conducting" claims are substantiated, and to what extent, if they are.

[forrestc]

I realize I'm late to the party here but....

Indium TACFlux 020B-RC is pretty much the only flux I use for hand soldering/rework anymore, for the reasons everyone mentions in this thread.   

From the datasheet:

"TACFlux® 020B-RC passes SIR testing in the unreflowed state. Oftentimes during hand soldering and rework, the flux may not be heated properly due to spot heating methods, thus the flux is not adequately activated. TACFlux® 020B-RC is designed to solve this problem and avoid reliability concerns such as dendritic growth over time. "

The 020B doesn't seem to be as active/aggressive as the TACFLUX-089HF, but 089HF needs to be fully heated to pass a SIR test.   Note that there is also a slightly different formulation of 020B which is labeled as tacflux-020 or tacflux-020A or something similar and is either thicker or thinner (can't remember which).

Note that SIR is the test the industry uses to verify a flux isn't going to corrode/cause dendritic growth over a long period, even in high humidity situations.   

[shapirus]

...but I might try something, just for a quick test: for example, set the hot air gun to what, say, 280 °C, heat the PCB with the flux on top from below (to avoid direct blowing on the flux so that more of it stays in place) for, say, 30 seconds, then let it cool and measure what remains. Just to see if the "it must be heated to stop conducting" claims are substantiated, and to what extent, if they are.

So I did exactly that (but 50 seconds and 290 degrees) with RL-421S-OR, the one with higher conductance.

Yes, heating the flux reduces its conductivity, and, in this particular case, it did that to a great extent. That's kind of expected, because what remains looks like plain rosin. No numbers, however, since this test setup can't be considered controlled and repeatable with the flux melting and flowing all across the board.

Still, my primary objective is not to test whether fluxes meet the datasheet specs (and I'm not sure that any of mine, except MG Chemicals 8341, even have datasheets), but to test which of them are more applicable than others for applications where neither full heating, nor full removal of residues can be guaranteed, and find a good go-to flux that I'll be happy with. I've had a device malfunction due to a conductive flux, and I don't want to make that mistake again.

[shapirus]

I realize I'm late to the party here but....

Indium TACFlux 020B-RC is pretty much the only flux I use for hand soldering/rework anymore, for the reasons everyone mentions in this thread.

Nah it's never too late to add new information, especially personal experience (ideally accompanied by numbers), be it good or bad. It'll be useful to someone some day.

[tooki]

Of course, both 44 paste and 1544 would fail shapirus’s requirement to be non-conductive before heating. I don’t think he understands that flux specs refer to the conductivity of the post-soldering residues.

Not that I don't understand, but I don't care. I test the fluxes I get my hands on for my specific requirements, which is hand soldering of both THT and SMD parts. And I believe this specific use case is of interest to many electronics hobbyists that can read this, too. Another way of looking at it is worst case scenario testing.

That’s fair.

My main concern was you trying to get a refund for flux because of it not meeting the claimed conductivity specs, even though you hadn’t reflowed it. That’s not fair unless they specifically said it was nonconductive in a non-reflowed state.

Of course, both 44 paste and 1544 would fail shapirus’s requirement to be non-conductive before heating. I don’t think he understands that flux specs refer to the conductivity of the post-soldering residues.

Not that I don't understand, but I don't care. I test the fluxes I get my hands on for my specific requirements, which is hand soldering of both THT and SMD parts. And I believe this specific use case is of interest to many electronics hobbyists that can read this, too. Another way of looking at it is worst case scenario testing.

Besides, the two fluxes from my last post are specifically advertised for BGA rework and general repair, which does not guarantee full heating, either (I believe even when using a hot air gun). And hey, they are 500-1000 times (!) more conductive than the other fluxes I tested.

Hot air alone is not adequate to achieve full heating. It heats locally, but also blows barely-molten flux all over the place. However, BGA rework is usually done with both preheating from below and hot air (or infrared) from above. And above all, if the balls reach reflow temperature, then the flux between them will also have reached that temperature. It’s only whatever flux has flowed far away that is of concern.

It would also be interesting to test conductivity after full heating, but that would require having a reflow oven, which I don't have, and the motivation to do it, which I also don't have, since I don't do that kind of soldering.

You should reconsider. Reflow in an oven is exceedingly efficient. (And practically mandatory for many modern packages.) A simple $50 toaster oven is enough for basic use.

...but I might try something, just for a quick test: for example, set the hot air gun to what, say, 280 °C, heat the PCB with the flux on top from below (to avoid direct blowing on the flux so that more of it stays in place) for, say, 30 seconds, then let it cool and measure what remains. Just to see if the "it must be heated to stop conducting" claims are substantiated, and to what extent, if they are.

You make it sound as though that’s some crackpot theory… No, it’s how no-clean fluxes are designed to work.

However, I would be leery of your proposed methodology, because it’s quite possible that it wouldn’t heat it evenly enough to draw any conclusions. If you do this, you’d want to do it with a custom test PCB that has solder pads that you can apply solder paste to, so that you can actuate whether it reached reflow temperatures.

Remember: there are TWO temperature thresholds to reach. The first (lower) one activates the flux, the second (higher) one neutralizes it again. If you heat to a temperature that reaches the activation temperature, but not the neutralization temperature, it may be left in a more conductive, more corrosive condition than the unheated state! So the only fair test of this “claim” is to actually get it hot enough to melt solder, since that’s what it was designed for.

[shapirus]

My main concern was you trying to get a refund for flux because of it not meeting the claimed conductivity specs, even though you hadn’t reflowed it. That’s not fair unless they specifically said it was nonconductive in a non-reflowed state.

That's 50/50, I would say. They never said that a full reflow profile was required for the loss of conductivity, either, and they advertised the flux as purposed for general repair work. Many other fluxes with similar claims have resistance on the order of gigaohms before heating, and that sounds fair and reasonable to me. Obviously I will prefer those over the more conductive ones.
These two, however, conduct in megaohm range, and that's a completely different story.

It’s only whatever flux has flowed far away that is of concern.

That's the point. Unless you heat entire board, you can't guarantee that the still conducting residues do not get trapped in some high impedance circuitry where it can cause trouble. I agree that it's not very likely, especially with almost exclusively digital circuits like laptop or smartphone boards, but it may be an issue with sensitive analog devices.

You should reconsider. Reflow in an oven is exceedingly efficient. (And practically mandatory for many modern packages.) A simple $50 toaster oven is enough for basic use.

I don't doubt that. I just won't have any use for it. I don't do repairs, I don't assemble complex boards (or any boards for that matter) on a regular basis. I mostly make one-off devices to try one idea or another, implement something that is of interest to me at a particular time, basically just what a hobbyist does.

You make it sound as though that’s some crackpot theory… No, it’s how no-clean fluxes are designed to work.

However, I would be leery of your proposed methodology, because it’s quite possible that it wouldn’t heat it evenly enough to draw any conclusions.

As I mentioned, it was an idea of a quick and dirty test to get a first rough estimate. It actually worked: I confirmed that heating at least one particular flux greatly reduces its conductivity. It was reduced in all areas that were heated to different temperatures.

Of course, proper testing would require a more complex setup with a corresponding reproducible temperature profile and a board (or whatever) designed to retain a given initial amount of flux in place as it's being heated to be able to measure it in a controlled and reproducible way.

And considering this:

Remember: there are TWO temperature thresholds to reach. The first (lower) one activates the flux, the second (higher) one neutralizes it again. If you heat to a temperature that reaches the activation temperature, but not the neutralization temperature, it may be left in a more conductive, more corrosive condition than the unheated state! So the only fair test of this “claim” is to actually get it hot enough to melt solder, since that’s what it was designed for.

...I don't think I'm gonna be bothered. It's calling for a much more complex test setup that I'm willing to use, and that, in addition, would serve no purpose for my practical application of fluxes.

Yes there is a certain probability that a comparison of unheated fluxes cannot be extrapolated to the same fluxes heated partially or fully. This is a disclaimer that has to be made.

Empirically, however, I've been observing, so far, that partially heated or fully heated fluxes are not worse than unheated ones, and if that is true, then testing unheated fluxes represents the worst case scenario, which I am happy with.

If someone wants to extend testing to partially and fully heated fluxes, it will be wonderful. But not me... at least for the time being :).

[tooki]

It’s only whatever flux has flowed far away that is of concern.

That's the point. Unless you heat entire board, you can't guarantee that the still conducting residues do not get trapped in some high impedance circuitry where it can cause trouble. I agree that it's not very likely, especially with almost exclusively digital circuits like laptop or smartphone boards, but it may be an issue with sensitive analog devices.

I know that. I said that in my original reply up above. As I said there: that’s why I always tell people to always clean no-clean flux (except from flux-core solder) when using it for rework.

[shapirus]

As I said there: that’s why I always tell people to always clean no-clean flux (except from flux-core solder) when using it for rework.

Correct. But cleaning also does not guarantee that all of the residue will be removed. I'm not sure just how much of an issue will the trapped flux be, but even when washing entire boards in ultrasonic bath, some flux easily remains under SMT parts and generally in hard to reach spots.

And it's the combination of this and the fact that we can't, with enough certainty, heat all the flux to where it loses conductivity, that motivates my filtering for fluxes that will be reasonably low-conductive before heating. Once a flux passes this check (many do!), I can look into other parameters.

[tooki]

As I said there: that’s why I always tell people to always clean no-clean flux (except from flux-core solder) when using it for rework.

Correct. But cleaning also does not guarantee that all of the residue will be removed. I'm not sure just how much of an issue will the trapped flux be, but even when washing entire boards in ultrasonic bath, some flux easily remains under SMT parts and generally in hard to reach spots.

As I said above, which you didn’t respond to then, either:

Well, by “cleaning” I mean “actually cause to be clean”, not just “perform a cleaning procedure regardless of outcome.”

I do recognize that this can be hard to do with some components. I’m very glad I have compressed air at work, since that’s super useful for getting stuff out from under SMD parts.

The emphasis is “cause to be clean”. That is, perform cleaning until it is actually clean. Not just “do” a cleaning once, regardless of outcome, and consider it good.

(I once had a roommate who washed dishes like that: he’d run the soapy sponge across a pan or dish and declare it “washed” because he had done the motion, without even looking whether the food had actually been removed!  Super frustrating for me to grab a pan or dish from the cupboard, only to find it still covered in visible, smellable residue from last week’s dinner. Once I didn’t pay attention and started cooking a dessert, then noticed it smelled like chicken…)

[shapirus]

As I said above, which you didn’t respond to then, either:

Well, by “cleaning” I mean “actually cause to be clean”, not just “perform a cleaning procedure regardless of outcome.”

I do recognize that this can be hard to do with some components. I’m very glad I have compressed air at work, since that’s super useful for getting stuff out from under SMD parts.

The emphasis is “cause to be clean”. That is, perform cleaning until it is actually clean. Not just “do” a cleaning once, regardless of outcome, and consider it good.

Sometimes it's easier said than done. Even with ultrasonic and compressed air (I do both, plus brushing) some flux still remains under SMT parts. Easy to check: just heat the washed and dried board, and the remaining flux begins smoking and flowing from under components. Or desolder some components and see what's under them.

I wash my boards with a mix of IPA and ethanol (ethanol is there just to save the more expensive IPA, the mix works about the same as pure IPA). Sometimes, in more difficult cases, I add some purified gasoline. Maybe some chemicals can penetrate and dilute the flux better. I can't justify using (or bying to test) them for home use. Alcohol plus low-conductive flux plus reasonably thorough washing work well enough for me, or I want to believe they do.

[tooki]

In my experience, proper flux cleaners work very noticeably better than pure IPA/ethanol. I wish people would stop relying on those, especially for modern no-clean fluxes which plain and simply are not fully soluble in alcohol. It’s no wonder you haven’t had good luck getting them cleaned out. Modern no-clean fluxes contain ingredients that swell up in alcohol, but don’t dissolve into it, requiring significant mechanical scrubbing to more or less remove.

In contrast, they do dissolve in proper flux cleaners, especially ones designed for lead-free fluxes.

Try a proper, modern flux remover like Electrolube LFFR (my favorite solvent-based flux remover), or for your ultrasonic bath, splurge the whopping $60 for a 5-liter canister of Electrolube Safewash Super (SWAS), which you then rinse off with tap water, followed by a final rinse with DI water or alcohol.

[T3sl4co1l]

A friend said a -- drat, now I forget which brand exactly, but probably any will do? -- window cleaner is effective on flux residues.

Seems like it would be slow, to me, but maybe a solvent-rich variety, or boosted with a little alcohol or ethyl acetate, would cover that.

There are a few different flux chemistries and it stands to reason flux cleaners need similar variation; alcohol works fine for your basic rosin based ones, but other solvents, surfactants, chelators, and acids or bases may prove more helpful for some.  Which is basically to say... try a lot of things and see what works best, but also, just buy the stuff that's recommended and be done with it?

Tim

[tooki]

A friend said a -- drat, now I forget which brand exactly, but probably any will do? -- window cleaner is effective on flux residues.

Seems like it would be slow, to me, but maybe a solvent-rich variety, or boosted with a little alcohol or ethyl acetate, would cover that.

The brand likely doesn’t matter, since the formulations are basically all the same, at least for the standard ammonia-based window cleaner. And my suspicion is that the ammonia is the key ingredient, insofar as aqueous ammonia is basic, and a base is what is used to saponify rosin. (Commercial aqueous flux remover, like the SWAS mentioned in my previous reply, uses plain old sodium hydroxide for that purpose! I assume the remaining important ingredients of the formula are corrosion inhibitors, surfactants, and redeposition inhibitors.) I’m just leery to use cleaners not intended for electronics, because of the risk of corrosion.

I just got a small amount of SWAS from another electronics lab at work, I’ll try it out soon. I’ve used the old Safewash in spray cans before, but they discontinued that recently. It worked pretty well, but not as fast as solvent cleaner. You have to give it time to soak. But since it’s water based, it doesn’t evaporate quickly like solvents.

Aside from the LFFR I also mentioned in my previous reply, I do have bottles of ethanol and acetone at my bench. The ethanol is mostly for cleaning of things that aren’t flux. I do use acetone to remove flux — it’s extremely effective, more than anything else I’ve used — when the materials allow. (For example, I used it yesterday to clean flux from silicone-insulated wire soldered to an all-metal banana socket.) But since acetone damages many plastics, and wipes the markings off many components, I rarely use it on PCBs and the like.