250g-500g 1/0.5mm Lötzinn Lötdraht Löt mit Flussmittel Lötkolben Lötstation DHL (WHL #68)
(Updated, see note at the bottom)
A fancy German item description for today, as I got it second-hand after even a soldering novice noticed this is total BS. “250g-500g 1/0.5mm Solder Solder Wire Solder with Flux Soldering Iron Soldering Station DHL” would be the closest translation. It’s solder – fake Wan Hung Lo leaded solder. Sold on eBay (probably all over the world), not even super cheap at 14€ a roll (250g, large diameter).
It is labelled “JINHU SOLDER CORE WIRE”, Type CF-10 Model B-1, comes in diameters of 0.3/0.4/0.5/0.6/0.8/1.0/1.2/1.5/2.0/2.3 (not necessarily true) with 2.0% flux (probably also not true), and proudly states “MADE IN CHINA”. No label on the back. Now even without the history of this basically being discarded as crap, the mention of “60/40” makes this stuff suspicious. Every manufacturer should be able to describe his product. Not in a colloquial manner like one does for the most basic Sn60Pb40 solder, but in the proper way – there’s clearly enough space on the label for it. If they don’t, it’s probably to avoid liability. Same goes for the diameter that is not given in millimeters.
So, since I like solving mysteries, I did run a couple of tests on that stuff. For the sake of simplicity, let’s call this China product C0.5 for the 0.5mm version and C1.0 for the 1.0mm roll. I tested it against Felder ISO Core RA Sn60Pb39Cu1 with 2.5% flux, which I use both at home and at work. Most tests were run with F0.5 for direct comparision, so 0.5mm diameter, but I also have F0.7 ready (actually the preferred diameter for around 90% of things).
Let’s test density first. If it is 60-40 solder, it should weigh about the same for the same length of wire, right? That one percent copper and the slight difference in advertised flux amount shouldn’t account for huge discrepancies.
C0.5: 0.60 to 0.65mm
(C0.5, F0.7, C1.0 from left to right – the China solder looks slightly more shiny in the pictures, but I cannot spot a difference just looking at it)
I have two issues with that.
1) Usually smaller diameter solder is more expensive due to tighter production tolerances, more time consuming production processes for the thinner strands, typically more flux with higher requirements in even distribution along the solder, as well as smaller roll sizes. For example, I use a slightly different Felder 0.5mm solder with 3.5% flux at home, it’s a 100g roll instead of the 250g used for bulk solder joints. If you pay for 0.5mm and they deliberately send you 0.6mm or 0.65mm, that’s a scam right there.
2) In case 1) doesn’t apply and this was a production mishap with people or machines at fault: Well, glad your QA guys caught that stuff before it was sent out.
Anyway, since the C0.5 now doesn’t have a direct comparison partner, I will need to correct for the greater diameter than expected. Let’s assume this is 55% larger than it should be, so weight will be adjusted by this factor.
Weight of a 100cm piece – measured on a folding ruler, I’d guess no more accurate than 1% due to the bendy nature of thin solder wire.
C0.5: 2655mg (first piece) – corrected for diameter: 1.7g (1713mg raw)
C0.5: 2682mg (second piece – I really doubted myself when the Chinese solder was heavier than expected) – corrected: 1.7g (1730mg)
F0.7: 2.8g (weighed at home – I don’t have scales with µg resolution in my kitchen, sorry )
With the generous assumption that solder wire is roughly cylindrical, that yields the following densities:
F0.7: 7.28g/cm³ (I’m ashamed of using the last digit given the weight is only known to two significant digits…)
Here’s the snag: Those are the densities of the solder wire (well yeah, duuh). This is not the density of the final solder alloy, since the raw solder wire includes flux, which is not present in the metal density calculation. We can go two ways from here: Recalibrate those measurements with the known density from the brand name product to eliminate flux influence, and use the relative deviation for further calculations, OR calculate an estimated density including flux and start from there.
Of course we will do both and compare the results, that’s how you catch your own mistakes.
Route one: Felder published density of this very solder to be 8.5g/cm³. With 7.2g/cm³ present in our data (weighted heavily on the result from the better scale), we need to add around 18% to eliminate the lower density flux. Note that a) the China stuff has 2.0% flux instead of 2.5% so the effect isn’t as pronounced, b) it is highly likely that different flux with different density is used, and c) the 2.0% is just a marketing thing and I do not know if that is true until I melt down an entire roll to get a chance of calculating density the old-fashioned way. Let’s eyeball 15% as a correction factor.
Average density: 9.9g/cm³.
Route two: I googled flux densities (not the nuclear ones!) on the interwebs and found something like 1.0 to 1.1g/cm³ (Kester publishes 1.013g/cm³ for use in SMD paste), which seems reasonable when I pick up the stainless shot glass filled with boiled off flux from expired SMD solder paste. Also assuming that flux and solder do not chemically mix (no water-alcohol intermixing density shenanigans!), one can crudely estimate the solder wire density with a volume consideration: 2.5% of weight (rarely clarified this is not volume!) will be flux at 1.013g/cm³, 97.5% will be metal at 8.5g/cm³. Total density will be – and no, this is not doctored to fit my values, but of course I will use a disputable third significant digit to overemphasize my precision – 7.17g/cm³. I got 7.16g/cm³ in the above calculation. Bazinga.
Doing the very same with 2.0% flux content yields 7.41g/cm³, a considerable change for such minute decrease in flux amount. The raw results for the China solder wire however averaged at 8.62g/cm³, which even exceeds the density of the pure 60-40 compound with zero flux. Reversing the calculation to get density at fixed 2.0% flux yields 10.18g/cm³ for the raw China solder alloy. Which is not too dissimilar to the estimated 9.9g/cm³ with the other method shown above, I have to say.
Time for a density overview.
β-Sn: 7.27g/cm³ (notable for lead-free solder types such as Sn99Cu0.7 – this rules out lead-free solder entirely)
(Felder) Sn60Pb40 and Sn60Pb39Cu1: 8.5g/cm³
Cu: 8.92g/cm³ (see, traces of copper for tip protection do not make a huge difference)
Ag: 10.49g/cm³ (let me know once Chinese sellers try fucking you over by selling Ag35Cu36Zn27.5Sn2.5 silver solder instead of leaded one…)
Pb: 11.34g/cm³ (lead, baby!)
Ruling out both heavily silver-based hard solder that would have a completely different melting point and would cost a fortune to scam buyers of leaded solder, as well as heavily bismut-based solder for having again a very different melting point (lower this time), the only way to create a solder wire with density pushing north of 10g/cm³ is using tons of lead. For simplicity let’s use the table from efunda (can be calculated as well) and mix in melting point/range data from Kester:
Sn0Pb100: 11.34 g/cm³ (pure lead), 327°C
Sn05Pb95: 10.80g/cm³, 309°C
Sn10Pb90: 10.50g/cm³, 268-302°C
Sn25Pb75: ?, 183-268°C
Sn30Pb70: 9.66g/cm³, 183-258°C
Sn40Pb60: 9.28g/cm³, 183-238°C
Sn50Pb50: 8.90g/cm³, 183-216°C
Sn60Pb40: 8.52g/cm³ (matching the value from Felder), 183-190°C
Sn63Pb37: ?, 183°C (eutectic)
From this data, I’d estimate this to be 25-75 solder (give or take 5%), not 60-40. Same can be eyeballed from a phase diagram of the binary Sn-Pb system, although I very much suspect there’s a significant amount of contaminants present.
The seller’s description by the way is very helpful here: Of course they don’t specify the 60-40 thing, but reassuringly they put “Material: Zinn und Blei. ” in there. “Material: Tin and Lead”. Yes, I would have assumed that, but anything more detailed would have given the buyer the opportunity to open a dispute…
Now, what about the melting points?
Using the Xytronic LF-8800 soldering station, the temperature can be adjusted in 1K steps. That cannot be translated into absolute accuracy, so I tested this with the Felder solder first. Melt quite a bit, shake off from tip, and repeat twice. After that, I consider the stuff on the tip to be the very same stuff that is in the solder wire, so no alloys that would skew melting points. With the datasheet value of 183-190°C (slightly non-eutectic!) I got it to melt a tiny piece at 193°C displayed value. As a second data point I do have pure bismuth which should melt at 271°C but only did so at 305°C (it was a substantially larger piece, though). Rough estimate: Real temperature = Temperature reading * 0.79 + 31K
Yes, that tip looks a bit crusty, but the bismuth is fantastic!
C0.5 and C1.0 behave alike, just to put that first. Without those cleaning steps, the Chinese solder starts melting at around 205°C (193°C corrected), but once there’s a pure alloy on the top, that gets way up – consistent with the idea of 60-40 where more tin allows for a lower melting point, and the estimated low tin content in our sample solder.
Once we’re in the pure C0.5 alloy system, it takes at least 230°C (corrected 213°C) to melt. The strange thing however is that it isn’t really melting like regular solder, it’s getting very soft. It’s getting so soft that strands of the solder can be placed, even layered on the iron tip, and it disconnects from the spool as shown above. The piece of solder however does not turn into a blob of molten solder that wets the entire tip, no no, it stays there. The picture below has been taken at 325°C (corrected 288°C) and it stayed that way for a couple of minutes until disturbed. If not moved at all, I managed to do that up to 395°C which is absolutely crazy. Regular solder, even lead-free one, would move upwards due to surface tension and quickly ball up way before that temperature, but this stuff just softens and needs to be agitated to leave the wire form. Never seen that before.
Of course that means the surface tension is a complete joke on the Chinese solder, which is why even the inexperienced guy I got it from knew that the solder joints were bad and it is difficult to process. He tried again with known-good solder and it’s as different as chalk and cheese. It’s not wetting properly, and there’s very little tension to keep the solder in place instead of spreading out in an uncontrolled matter. Not being an eutectic alloy with wide temperature gap between solidus and liquidus temperature makes it even worse. This can be tested another way – flinging it off onto the table.
Large blobs, somewhat difficult to even remove from the tip unless huge amounts are present. There’s even blobs of the flux nearby.
On the contrary the C0.5 has a splash pattern like this:
Many more balls, some elongated or deformed, most of significantly smaller size, very easy to fling off and the solder will split up into several pieces on impact on the table. It’s a bit like water on a hot stove except for the fact that boiling water still has a lot of surface tension that keeps the droplets together. That stuff is terrible for any solder joint, no doubt about it.
Alright, time to summarize: Don’t buy Wan Hung Lo solder unless you have known-good solder to test it against, and the time and leisure to do so in order to open a dispute before any buyer protection runs out. This solder should not be used in any electronic product, DIY or commercial setting (even without the RoHS crap *cough*), and anyone looking for this very specific alloy will have a trusted source and a reason for it. It is not cheap enough to justify any experiments, and it has much worse properties than any available properly engineered solder, leaded 60-40 or lead-free. Do. Not. Buy.
For me, however, this is a very nice base material for another project. The bismuth shown above is a big hint…
Update: I have added test data of flux content and a visual comparison of leaded, unleaded and this solder in post WHL #68F1