I've been assembling low numbers of reasonably complicated PCBs professionally for a few years now, and while I can get good results eventually, one thing that's dragging the process out is the need for post-reflow cleaning. I always find quite an excessive amount of dark, hard flux, which takes a lot of effort to remove, particularly from in-between fine-pitch components. I find the flux itself makes inspection difficult to do with confidence so I feel the need to remove it. Usually, I have to scratch at it in a few places, in addition to the river of IPA and brushing.

My process is -

  1. Wipe down the board with IPA
  2. Apply Solder paste using 0.12mm Stencil
  3. Place components
  4. Reflow
  5. Spend a long time cleaning and inspecting.

The paste used in the examples below is SMD291AX from Chipquick, it's a leaded, no-clean flux solder paste. In this example, it was a fresh tube that arrived that morning. My reflow oven is homemade, but the profile tracks the recommended profile pretty well, within a few degrees.

While I can get good results I'm proud of, I'm entirely self-taught and have just figured this out from trial and error, so I could well be missing something. Some of my googling suggests that no clean fluxes are pretty tough to remove, so perhaps I should look at changing paste.
The images below are all how the PCB came out of the oven
enter image description here In the centre is a 0402 resistor, below an SOIC.
enter image description here SOT-23 next to 0402s
enter image description here Unpopulated 0.5mm pitch header

Is this a typical amount of flux and hence cleanup? Is there a way to improve my methods to reduce it?
To be clear, in this question, I'm not looking for advice on how to remove this flux, but on how to reduce the buildup in the first place.

  • 2
    \$\begingroup\$ Can the boards be stacked in a bin full of 91% isopropyl alcohol and left for an hour? Might make cleaning a non-issue. No-clean flux can be conductive so should be removed. Ironically. \$\endgroup\$
    – rdtsc
    Commented May 11, 2023 at 11:46
  • 1
    \$\begingroup\$ @LordTeddy Do you have any ethyl acetate you could add to the isopropanol? Maybe at 5%. Just to see if it makes it a better (faster) solvent for that flux? For some fluxes, using commercial flux remover will have you laughing at trying to use only isopropanol, in my very limited hobbyist experience. \$\endgroup\$ Commented May 11, 2023 at 11:52
  • 1
    \$\begingroup\$ @glen_geek For the avoidance of doubt, I was suggesting adding EA to the IPA for cleaning, exactly as LordTeddy interpreted it. I mistakenly thought that the question included possibly making the cleanup faster. \$\endgroup\$ Commented May 11, 2023 at 14:13
  • 1
    \$\begingroup\$ A no-clean flux is used when you can't or don't want to clean the board after assembly. For example, you have a component like a moisture sensor that can get damaged during cleaning, We found for some applications a no-clean flux messes with our signals (for example, it lowers the output of high impedance piezoelectric transducers). If you want to reduce the brown mess, reduce the stencil thickness. An ultrasonic cleaner automates the scrubbing process, but it usually takes a few tries. If you're going to go the cleaning route, it's better to use a flux that isn't no clean. \$\endgroup\$
    – C. Dunn
    Commented May 11, 2023 at 14:32
  • 1
    \$\begingroup\$ To me it looks like something is wrong with the reflow, either bad paste or temperature profile. Thermocouple costs a few dollars online, maybe worth checking that you are hitting the temperatures you think you are? \$\endgroup\$ Commented May 11, 2023 at 16:23

3 Answers 3


The simple answer is no, this is not a typical amount of flux and flux cleanup, whether no-clean or not.

Diagnosing the cause may be more complex. There may be several possible factors, and you may have more than one operating here.

I think the most likely cause is excessive solder paste. I see solder wicking up on top of the SMT parts and making too large a bubble on the unmounted pads. The stencil may be too thick, have rough edges at the apertures, or your squeegee technique may not be removing enough. (I like stainless stencils from OSHStencils.)

Second possibility is defective/degraded solder paste. Did you buy from an authorized distributor, or from a less-reliable source? I store the syringe tip-down so if there's separation, it will be solder-rich rather than flux-rich.

Third possibility is the board is not reaching a high enough temperature, so the solder does not become fully liquid and the flux does not boil off. Do a test run, open the door (or pull the PCB out) at the peak temperature and quickly measure the board surface temperature with an infrared non-contact thermometer. I shoot for 220-240 deg C.

I've been making boards with a strictly amateur $10 hot plate for years and rarely have to clean any flux except where I do any manual touch up.

  • \$\begingroup\$ Excessive paste has been suggested and certainly seems sensible. What thickness stencil are you normally using? My technique is not something I'd thought that much about and definitely worth looking at. TY. The solder paste was from RS and used on the day of purchase, so probably fine. Temperature could be an issue, but it's more likely that it's running too hot IMO. The flux is brittle, like when you've left an iron on a joint for far too long. I can't see the solder being under temperature causing this. \$\endgroup\$
    – LordTeddy
    Commented May 11, 2023 at 17:37
  • 3
    \$\begingroup\$ @LordTeddy I use 4 mil (0.1 mm) stainless or 5 mil polyimide stencils. I squeeze a bead of paste at one end, squeegee it across with a credit card which leaves the stencil mostly clean except for the apertures, then use a clean card to strip off any paste that sticks up. After flowing, there should be enough solder to create concave fillets at the edges of the SMT leads, but no convex bubble on top. \$\endgroup\$ Commented May 11, 2023 at 20:30
  • 2
    \$\begingroup\$ @LordTeddy there are countless things you could accidentally do wrong when using the stencil, leading to too much paste applied, even tiny gaps between the stencil and the pcb can cause this. rough edges, warping due to not being flush with the pcb, thus bending down there, and thus up in the middle. If I was faced with this problem I would experiment a bit with hand dispensed amounts of paste on some unpopulated pads. I have seen these amounts of flux in boards of products I have repaired too, so you are not alone ;) \$\endgroup\$
    – PlasmaHH
    Commented May 11, 2023 at 21:19

Thanks for everyone's input. The answer is two-fold.

There is an issue with too much solder paste. I ran a number of test pieces, and I found often that there was quite a variation in the thickness of the solder applied, and it was usually thicker than the stencil. This is probably part technique, but also part stencil quality. The stencils I get are never particularly flat, and getting them to sit flat on the PCB is tremendously difficult. This explains the quantity of flux left over

With my test pieces, just pasting up the 0.5mm connector used in my question, I ran the oven at different temperatures. Just up to temperature and off again, no preheat. Starting at 235C, decreasing in 15C steps to 160C. Have a look at the difference between 235 (top) and 160 (bottom).test peices
You can clearly see the difference between 160C, where the solder is clear and still fluid (it can be wiped off with a cloth), and 235C where it's opaque and crusty.
I actually also found that the position of my thermocouple made a big difference, moving it from 1cm away, to touching the board edge, was the difference between the solder flowing and not.

I checked the datasheet and SMD291AX has a melting point of 183C, so it seems pretty unlikely that the thermocouple is reading well, and it's clear that my little hacked toaster oven doesn't produce a very even temperature, but that's hardly surprising.
Looking at this, I figured that my reflow profile is around 15 or so degrees higher than it should be, and the flux is basically charing on the board instead of boiling off. In order to test this, I heated the board to what my oven called 160C, held it there for a minute or so, and then increased it to 205 before shutting it off. Something that begins to approximate a reflow cycle (ie preheat and flash) The result was much better. It's pictured below (bottom) with just the straight-to-205C board above for comparison. enter image description here
This is now much better. There is a small amount of flux left over, but it's clear, and I wouldn't feel the need to clean it (it's totally imperceivable to the naked eye). Once I get a sensible amount of paste down, there'll be even less flux. Compare this to the version that went just straight to 205C, which hasn't had the chance for the flux to boil off, so you can see it's starting to char on the board.

  • \$\begingroup\$ If you aren't using one already, there are devices for holding stencils in place with the proper amount of force. Maybe look into getting one of those in the future; solder paste stencils aren't intended to be (though often can be) held by hand. \$\endgroup\$
    – Hearth
    Commented May 11, 2023 at 23:21
  • \$\begingroup\$ I found this video very useful \$\endgroup\$ Commented May 12, 2023 at 14:19
  • \$\begingroup\$ That looks a lot better. There looks to be some shorting, and perhaps it's because this was a test. If it's still a problem, try reducing your aperture size (make the holes in the stencil smaller than the pads) or try switching to a 0.051 mm (2 thou) stencil (less metal to short). \$\endgroup\$
    – C. Dunn
    Commented May 12, 2023 at 16:07
  • \$\begingroup\$ @C.Dunn Yes, shorting is definitely due to excess paste, from poor technique rather than the stencil thickness I think \$\endgroup\$
    – LordTeddy
    Commented May 12, 2023 at 16:19
  • 1
    \$\begingroup\$ If you are using a converted toaster oven, it is probably giving too much infrared radiation and too little convective heating. This causes the excess flux to heat too much, burning it into a dark residue. \$\endgroup\$
    – jpa
    Commented May 12, 2023 at 16:19

Others have already pointed out the too-much-paste problem you are having. I will take my answer in a different direction.

No-clean flux is not meant to be removed easily, so it isn't. Meanwhile, water-based fluxes are very easy to remove after reflow. Often, a quick rinse is all it takes. No scrubbing or chemicals needed.

Where I work, we use a solder paste with water-soluble flux (Kester R520A) and after reflow the boards go into a dishwasher and run through a rinse cycle. They come out perfectly clean, beautiful shiny boards. Of course they are allowed to dry overnight before powering up.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.