3
\$\begingroup\$

I have a project where I'm about to order the PCB's for a first prototype, and I'm staring at the option of HASL vs. ENIG on the website. I will hand solder it with an iron, the smallest pitch part is 0.5mm (which I've succesfully soldered before, though with some difficulty, on a board with HASL finish). My skills with the soldering iron are not especially good, although as mentioned, I have succesfully soldered similar parts before, and I'm learning to be better all the time.

Now based on this question, I suppose an ENIG finish might save me some headache at these pitches. On the other hand, based on this website, one of the disadvantages of ENIG is "not reworkable".

Since this is a prototype, I will certainly have to swap around some 0603 resistors and maybe capacitors, once I get to evaluate the device in action, to find the right values, and test if some noise-reducing RC -filters are even necessary in the first place.

The question is: is ENIG so bad for rework that, considering I know for sure I will have to replace at least some resistors, I should go for HASL and just deal with the HASLle of soldering the fine pitch parts there?

In terms of price, the difference in the board house I'm using is 25$ for 10 PCB's between HASL vs. ENIG, which is not prohibitive, although if it's a close match, the price might be a tie-breaker in favor of HASL.

\$\endgroup\$
  • 3
    \$\begingroup\$ I've never heard of ENIG being "not reworkable" before. I do rework on ENIG boards a lot, and I don't see why that would be the case. Granted, I use leaded solder for rework, and therefore ruin the RoHS compliance, but it's rework, so that's not an issue. \$\endgroup\$ – uint128_t Jan 25 '16 at 15:49
  • \$\begingroup\$ My guess is that they mean the finish itself is not reworkable, not the soldered parts. \$\endgroup\$ – scld Jan 25 '16 at 17:19
  • \$\begingroup\$ Agreed, I have built many ENIG boards and never had a problem with rework because of it. Industry reworkability standards are often very conservative, like allowing 10 rework cycles without visible damage or performance change. In the real world you hardly ever need that many cycles. If you want to have easy reworkability it's more important to use larger components (0603 or 0805 instead of 0402 and smaller), and use full IPC pad dimensions instead of "as small as your assembly shop will let you". \$\endgroup\$ – The Photon Jan 25 '16 at 17:19
  • \$\begingroup\$ I second @uint128_t: if possible, specify leaded solder if you anticipate rework. It makes life much simpler! Of course, if you are really concerned of your ability to hit the pads with the unevenness of HASL, you can always wick the pads clean, place the IC, and then solder it down. \$\endgroup\$ – M D Jan 25 '16 at 22:13
2
\$\begingroup\$

I haven't noticed any particular problems "reworking" ENIG. It's a right tools for the right job thing. To pick off 0603 passives, heated tweezer tips are best.

That said, I've rarely been in the situation of having to swap out R's and C's once I've gotten up to the point of printing a board. Perhaps mounting your fine pitches on a breakout like from proto-advantage and using some other combination of prototyping techniques will let you figure this out in advance.

If you're talking about gearing up for a major production run, prototype it both ways (with and without your RC filters), or spend the engineering time in advance to figure out the values up front. Perhaps you can leave the RC filters in the board, and a combination of Do Not Populate and zero ohm resistors can replace more expensive parts. Or, you can print up some extra boards and populate them with different RC values until you have it right.

\$\endgroup\$
  • \$\begingroup\$ Ok, good to hear. This is a hobby project, although with some ambition that if it works out well, I might later pursue it as a product too. Anyway, even as a hobby, I will eventually make several units, but as the BOM cost is pretty big for this, I'd like to test out the various combinations on a single prototype first. \$\endgroup\$ – Timo Jan 25 '16 at 15:58
  • \$\begingroup\$ I've prototyped a lot of the sections on a breadboard, and everything non-trivial in LTSpice, but as the final board is a mixed signal device with an MCU driving several DACs controlling audio generation, the final noise performance will really have to be tested out with the PCB. So the may-or-may-not-need filters are on lines going from the digital side to the analog, and I don't really see how I could breadboard that such that I could rely on the level of digital to analog crosstalk being more-or-less similar to the final PCB. \$\endgroup\$ – Timo Jan 25 '16 at 16:03
  • \$\begingroup\$ @Timo -- sounds good, but I'd still raise red flags about building a PCB with the intent of figuring out passives values as you go. I'd ask my students to do better, and suggest that if you're having trouble figuring out whether you need a filter or not or what some of the RC values need to be, such matters would make fine new questions here. Also, if this is a hobby thing, I'd say build yourself the circuit you MIGHT need instead of the circuit you might be able to get away with. The few extra dollars on the BOM is a fine safety play. \$\endgroup\$ – Scott Seidman Jan 25 '16 at 16:04
  • \$\begingroup\$ I think I'm doing exactly what you're saying: the board and BOM contain the RC -filters, and what I intend to do, as I assemble the board a section at a time, is to first not connect the lines in question at all, measure the noise performance, then connect with a 0 ohm resistor, measure again, and if there's any appreciable increase in noise, replace the 0 ohm with the values for the RC-filter (which will have a cutoff in the audio range, I've calculated the values for that case). Repeat for every section, so I'll know a feature at a time how it affects noise and if the RC's help keep it down. \$\endgroup\$ – Timo Jan 25 '16 at 16:11
  • \$\begingroup\$ This is a music synthesizer, and the few cases where I might need to fine tune resistor values are things such as oscillator waveform balance, and I've of course got values that match what I've prototyped a section at a time. However, as the final balance is, in the end, an artistic judgment, I expect that some of those may change once I hear everything playing at once for the first time. \$\endgroup\$ – Timo Jan 25 '16 at 16:14
1
\$\begingroup\$

Don't worry about the technical differences between the two for this project. The "reworkability" they mention is either referring to the actual board process (as in, if you mess up the plating/immersion, you scrap the board, in which case you don't care) or it's referring to the soldering of components (in which case they're wrong).

I prefer ENIG when hand soldering purely due to the flatness of the finish but I think you could manage the finer pitch stuff with either HASL with a little more effort. This will come down to a personal choice between cost and preference.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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