I could have just posted the final graph in a blog post and said a few sentences, more efficient than video again. But this is an informational video blog, so you get information in video form. The "bla-bla" shows you exactly what I did, the step's I took, and explains why. If you don't like the bla-bla, you should unsubscribe, you are in the wrong place.
Dave, the effort that you went through to demonstrate this absolutely brings me to my knees. I am very appreciative of the efforts that you go through to bring us these lessons. You could retire very wealthy as a professor...
Same reason they think solder plating the traces doesn't help, because they know solder has higher resistivity than copper, so they just assume it's going to be negligible instead of actually measuring it.
Very, very interesting data, Dave! I always intuitively thought that tinning vias like this would make a difference in their resistance, but definitely not a drop like the one that you demonstrated. I'd love to see a follow up with smaller hole diameters, however, I think that it'd be a lot more accurate representation of actual vias used in real projects. Great video! chris
The image is still in commons, but it was removed from the wiki page. I thought the software mention was essential to know where the data originated from. I can't see how it can infringe any copyright?
A bit of a side trail, but many years ago we had certain machines in the military that would go from cold ambient to warm operating and would pop the vias causing intermittent operation. When we figured out what was going on, we had to solder wires in all the vias to prevent it from happening. Solved the problem. PCB processes have come a long way since then, but it is interesting that in certain situations, they can give you problems. Eye opening to see that even soldering through the holes significantly improves performance. Thanks for the test.
Man I never would have thought much about the effects of via current If I hadnt seen this video. Great stuff! thanks. I love discovering stuff like this. Thanks! Pat
Hey Dave, what do you think of a fundamentals friday vid about ground loops, star grounds, ground planes, split analog/digital ground planes. I know it's a really dry subject but it seems to come up a lot in forum discussions so I think it might still be a popular video.
After seeing your video on coating the traces in solder, I naturally began filling the vias, just had to be the right thing to do. No more kapton tape for wave process. Really interesting, I wish my job allowed me to dick around like this. Just wonderful, Thanks Dave!
Interesting post Dave. Hadn't seen this until now but it's a great vid. I suspect there's more at work here but overall the premise is sound. Solder, as we know, isn't as conductive as copper, and thus plugging a via will not give a 1:1 uplift in current carrying capacity vs. increased dimension. If you plug it with a component pin or a wire, this is certainly going to help versus solder but solder only performs about 10% as well as copper in terms of conductivity. An effect that may not be as obvious is the effect the additional solder will have on cooling...thus lowering the temperature rise thus resistivity, thus increasing the current carrying capacity. This tends go overlooked by the PCB greenhorn as it's much harder to model, but can have a substantial impact on the via's ability to carry currents. Nice video and great experiment!
Awesome video!!! I'll have to do a video response to this one. We fight board terminal thermal failure almost daily with these portable hot tub control boards. The heater circuit is switched and handled directly on the board. 5.5 kw heater on 240 volt mains has 22-24 amps running on vias! Not reliable, I'll have to show you the failure modes.
Hi @John Eastmond and @Dave. I am considering to (try) to fill vias with solder paste to decrease via thermal resistance. After Internet searches on vias filled with conductive epoxy (actually a mixture of metal particules and epoxy) such as the DuPont CB100, I am concerned with thermal expansion of solder, which would increase delamination risk. Would you be concerned? Can you share some of your experience on thermal failure, John? Thanks in advance!
Silver Solder? I just ran across an advert for "Dayton Audio 4% Silver solder" (can be found on Amazon US) in which it says "higher electrical conductivity than standard solder." I wonder if that would have better results on your test rig?
Metcal recommends tinning the tip with a little solder before leaving it in the stand, and cleaning it before making a joint if it's been in the stand for any length of time. It makes sense to me - the tip is likely to oxidise and so need cleaning before you make the next joint anyway, and IMHO it's better to have a thin layer of cheap solder oxidise than the plating of the expensive tip. Doesn't need a big glop either - just a thin coating.
The argument that sifterjoe is putting is that the lower temperature in itself is responsible for the reduced resistance. This occured to me as well. However, on reflection I don't think this is the case - it will alter the slope of the curves in your final plot as the difference will be greater at high currents, but they are all virtually flat anyway. And the % difference between 299.8K and 302K is so small that the change in resistance due to the heatsink effect is simply lost in the noise.
Fair enough, I was taught that working as a solder jockey on an assembly floor where they treated the tips like gold and my personal irons have been various cheap wood burners that do run to hot for most things.
I remember seeing on hackaday a while back a guy made a via punch - for diy boards where you can't plate your own holes - you insert a wire plug then crush it with a press so it rivets through the hole, no solder needed, just mechanical connection.
are there any potential downsides to filling a via with solder, for example if you have a PCB where a via is noticeably getting warmer, would it help to fill it with solder, or will it cause problems since it came from the factory like that?
Very interesting video, Dave I am curious (this would probably not work for the smaller via's but for the larger ones if you took some stranded wire and soldered that in the via would it be able to flow more the a say solid wire or just filling the solder.
iMovie does not do circuit simulation, but Ltspice, TINA-TI, and even the software Dave used in the video all install and run just fine on my 7 year old MacBook Pro via wine/winebottler. =P
Had a small board with really nearly no place for any extra traces. A lot of power-connections between boardsides were placed onto solder-pads of other SMD-components - the pad is already there so the via does not take up any more board-space and on the pads we often had space for 2 vias. Well, hand-soldered boards. With Reflow i think this would be a nightmare and ruin the connections. Sadly many programs do not have an option for filled vias, and the same for cheap manufacturing. Vias filled with copper from the manufacturing have such low thermal and electrical resistance - well, of course, they have far thicker copper than anything else on a normal board.
Interesting, the walls of the via are much thinner = more resistance, what if you just put a copper rivet in the hole? that would work excellent I reckon
Paul Avesaath That's what I thought - why bother with the test?. If you add ANY conductive material, (even carbon - though typically used for its resistive properties, it's still conductive), you add conductance and you therefore lower resistance due to their inverse relationship (G = 1/R). It could be my cynicism, but I imagine the myth was started by PCB manufacturers who wanted to make a few more cents out of each PCB. Either that, or it started during the early days of pick-and-place machines, when soldering through a via may have been less reliable and no cheaper than just whacking a few more vias on a link, making it less of a myth and more a piece of history that's no longer relevant.
It should be same for vias as traces. The resistance decreases with wire diameter. Would expect it to be same if you add some solder to hollow via. I could be totally wrong tho
I use a trick for connecting traces top and bottom using a pad I just feel the hole with solder and connection will be established also other best way is to place a piece of leftover wire from a resistor or diode into the hole and refill with solder that is the best choice
I disagree in that if the via was 2.2C at 1.8A, then you stuck a big wire in there acting as a heat sink, if you check the calc for let's say 1C you jumped down to a different curve. The low power dissipation is even more argument for why it matters, it is easier for that resistor lead to be effective heat sink. If you side cutter that wire flush and redo measurement, it will be much closer to solder filled hole numbers.
9:28 Was Dave actually afraid to call this a logarithm? Don't get me wrong, I love a good comedy trope, even if it is unintentional, but if the curve fits...and by the way, that curve is Gabriel's Horn. It's only natural. Looks to me that the kink in current capacity is at the point where the via diameter equals the board thickness. This fascinates me; perhaps studying it could make for a good thesis paper for an EE graduate student.
It was kind of obvious, but i didn't expect that much of a difference. How about inductance. I really want to see a similar research on that damned parasitic..
Hai Dave, For prototyping DIY pcb's I use these: "www.bungard.de/index.php/en/products/through-hole-plating-line/through-hole-plating". Does your software Altium supports through hole rivets automatic dimensioning when placing set in via-rules.
I could have just posted the final graph in a blog post and said a few sentences, more efficient than video again. But this is an informational video blog, so you get information in video form. The "bla-bla" shows you exactly what I did, the step's I took, and explains why. If you don't like the bla-bla, you should unsubscribe, you are in the wrong place.
Dave, the effort that you went through to demonstrate this absolutely brings me to my knees. I am very appreciative of the efforts that you go through to bring us these lessons. You could retire very wealthy as a professor...
Some people do that, but it's only for more exotic applications. Generally, solder fill is much easier and cheaper.
Yes, filling will help. Less dissipation = less heat. It will also make your board a smidgen heavier.
Same reason they think solder plating the traces doesn't help, because they know solder has higher resistivity than copper, so they just assume it's going to be negligible instead of actually measuring it.
Very, very interesting data, Dave!
I always intuitively thought that tinning vias like this would make a difference in their resistance, but definitely not a drop like the one that you demonstrated. I'd love to see a follow up with smaller hole diameters, however, I think that it'd be a lot more accurate representation of actual vias used in real projects.
Great video!
chris
The image is still in commons, but it was removed from the wiki page. I thought the software mention was essential to know where the data originated from. I can't see how it can infringe any copyright?
I love that 'I only give negative feedback' tshirt. Amazing.
Very interesting - I expected it to make a difference, but the actual difference is quite significant. Thanks Dave.
Dave made over 540 Videos and is still possible to increase the level of Quality and Nerdyness. :D
The mat is not conductive, it is static dissipative.
This is probably the best tutorial that I ever found online about vias !
A bit of a side trail, but many years ago we had certain machines in the military that would go from cold ambient to warm operating and would pop the vias causing intermittent operation. When we figured out what was going on, we had to solder wires in all the vias to prevent it from happening. Solved the problem. PCB processes have come a long way since then, but it is interesting that in certain situations, they can give you problems. Eye opening to see that even soldering through the holes significantly improves performance. Thanks for the test.
Man I never would have thought much about the effects of via current If I hadnt seen this video. Great stuff! thanks. I love discovering stuff like this.
Thanks!
Pat
On big power stuff, it is not uncommon to use vias this size.
Hey Dave, what do you think of a fundamentals friday vid about ground loops, star grounds, ground planes, split analog/digital ground planes. I know it's a really dry subject but it seems to come up a lot in forum discussions so I think it might still be a popular video.
Your channel is great Dave, I like learning new stuff from your vids
After seeing your video on coating the traces in solder, I naturally began filling the vias, just had to be the right thing to do. No more kapton tape for wave process. Really interesting, I wish my job allowed me to dick around like this. Just wonderful, Thanks Dave!
Kind of, they are called embedded resistors or resistor in via technology.
I added the graph to Wikipedia, but it was removed.
Interesting post Dave. Hadn't seen this until now but it's a great vid. I suspect there's more at work here but overall the premise is sound. Solder, as we know, isn't as conductive as copper, and thus plugging a via will not give a 1:1 uplift in current carrying capacity vs. increased dimension. If you plug it with a component pin or a wire, this is certainly going to help versus solder but solder only performs about 10% as well as copper in terms of conductivity. An effect that may not be as obvious is the effect the additional solder will have on cooling...thus lowering the temperature rise thus resistivity, thus increasing the current carrying capacity. This tends go overlooked by the PCB greenhorn as it's much harder to model, but can have a substantial impact on the via's ability to carry currents. Nice video and great experiment!
Excellent video Dave, every day I learn something new... Keep them coming mate!
Someone removed it for some other reason, I greatly doubt it was copyright.
At 800uW power dissipation at 2A, I don't think it's going to affect the measurements too much :->
It was removed because the person didn't think the data was reliable. Plus it wasn't formatted to their liking.
Fantastic!!
Awesome video!!! I'll have to do a video response to this one. We fight board terminal thermal failure almost daily with these portable hot tub control boards. The heater circuit is switched and handled directly on the board. 5.5 kw heater on 240 volt mains has 22-24 amps running on vias! Not reliable, I'll have to show you the failure modes.
Hi @John Eastmond and @Dave. I am considering to (try) to fill vias with solder paste to decrease via thermal resistance. After Internet searches on vias filled with conductive epoxy (actually a mixture of metal particules and epoxy) such as the DuPont CB100, I am concerned with thermal expansion of solder, which would increase delamination risk. Would you be concerned? Can you share some of your experience on thermal failure, John? Thanks in advance!
They don't exist yet.
Really enjoyed this one. Well explained, and a simple topic that is accessible to many people. Hope you get lots of views :)
Silver Solder? I just ran across an advert for "Dayton Audio 4% Silver solder" (can be found on Amazon US) in which it says "higher electrical conductivity than standard solder." I wonder if that would have better results on your test rig?
But I also lowered the resistance by 60-70% at the same time.
Metcal recommends tinning the tip with a little solder before leaving it in the stand, and cleaning it before making a joint if it's been in the stand for any length of time.
It makes sense to me - the tip is likely to oxidise and so need cleaning before you make the next joint anyway, and IMHO it's better to have a thin layer of cheap solder oxidise than the plating of the expensive tip.
Doesn't need a big glop either - just a thin coating.
This is very interesting. Nice experiment.
Great thanks Dave
The argument that sifterjoe is putting is that the lower temperature in itself is responsible for the reduced resistance. This occured to me as well.
However, on reflection I don't think this is the case - it will alter the slope of the curves in your final plot as the difference will be greater at high currents, but they are all virtually flat anyway. And the % difference between 299.8K and 302K is so small that the change in resistance due to the heatsink effect is simply lost in the noise.
Thanks for making this video; nice job
Fair enough, I was taught that working as a solder jockey on an assembly floor where they treated the tips like gold and my personal irons have been various cheap wood burners that do run to hot for most things.
I remember seeing on hackaday a while back a guy made a via punch - for diy boards where you can't plate your own holes - you insert a wire plug then crush it with a press so it rivets through the hole, no solder needed, just mechanical connection.
Dave, are you selling the µRuler? I would like to get my hands on one.
beauuutiful video. Thank you Dave !
Great video, Dave. I may have missed it, but were you using leaded or lead-free solder?
Dual, not Joule
Dave mentioned that it was standard 60/40 tin/lead solder.
But that via is made of a proper resistance material, not just proper thicked copper?
14:03 What do I need to type into Amazon/ebay to find a ruler like that, with all those useful things on it?
Just a thought is that a lot of resistor leads I have met are actually tinned steel wire, not as conductive as copper wire.
are there any potential downsides to filling a via with solder, for example if you have a PCB where a via is noticeably getting warmer, would it help to fill it with solder, or will it cause problems since it came from the factory like that?
Very interesting video, Dave I am curious (this would probably not work for the smaller via's but for the larger ones if you took some stranded wire and soldered that in the via would it be able to flow more the a say solid wire or just filling the solder.
Needed to clip that resistor wire to prevent any heat sinking effects. Otherwise, very interesting video. Good data. Like!
iMovie does not do circuit simulation, but Ltspice, TINA-TI, and even the software Dave used in the video all install and run just fine on my 7 year old MacBook Pro via wine/winebottler. =P
This might be a stupid question, but could there be designs that would purposely design resistance in vias?
good to know for sure now :D thanks !
Great video.
Torrent search engine in the bookmarks :P
I am sure it for the very big datasheets :D
You should put more thinner cables in to the same hole and test it too
All the way through I wondered if he'd blow it up to test the current limit? Disappointment. :)
Had a small board with really nearly no place for any extra traces. A lot of power-connections between boardsides were placed onto solder-pads of other SMD-components - the pad is already there so the via does not take up any more board-space and on the pads we often had space for 2 vias. Well, hand-soldered boards. With Reflow i think this would be a nightmare and ruin the connections.
Sadly many programs do not have an option for filled vias, and the same for cheap manufacturing. Vias filled with copper from the manufacturing have such low thermal and electrical resistance - well, of course, they have far thicker copper than anything else on a normal board.
Interesting, the walls of the via are much thinner = more resistance, what if you just put a copper rivet in the hole? that would work excellent I reckon
I mentioned in the registration text (which of course people never read!), if that happens, send me an email with your username and email.
if the VIA is fat - does that mean a change in thermodynamics - does it act as a mass - and holds energy (heat)?
.. ah.. always finish the video . .
would the board sitting on a conductive mat make a difference?
We've had an arguments about this me and my friends years ago. I won . I've always filled my via
isn't it pretty obvious it would make a difference.. (just common sense?) any way nice vid...!
Paul Avesaath
That's what I thought - why bother with the test?. If you add ANY conductive material, (even carbon - though typically used for its resistive properties, it's still conductive), you add conductance and you therefore lower resistance due to their inverse relationship (G = 1/R).
It could be my cynicism, but I imagine the myth was started by PCB manufacturers who wanted to make a few more cents out of each PCB. Either that, or it started during the early days of pick-and-place machines, when soldering through a via may have been less reliable and no cheaper than just whacking a few more vias on a link, making it less of a myth and more a piece of history that's no longer relevant.
Interesting what it does to inductance, especially in RF applications
I think the equation [ Resistance = Resistivity x length / section ] should describe your experiment close enough....Anyway Thumb Up, great work.
It should be same for vias as traces. The resistance decreases with wire diameter. Would expect it to be same if you add some solder to hollow via. I could be totally wrong tho
I use a trick for connecting traces top and bottom using a pad I just feel the hole with solder and connection will be established also other best way is to place a piece of leftover wire from a resistor or diode into the hole and refill with solder that is the best choice
If you design the pcb with over size pads the solder will fill the holes all the way through even if there is a component leg going through the pcb.
I disagree in that if the via was 2.2C at 1.8A, then you stuck a big wire in there acting as a heat sink, if you check the calc for let's say 1C you jumped down to a different curve. The low power dissipation is even more argument for why it matters, it is easier for that resistor lead to be effective heat sink. If you side cutter that wire flush and redo measurement, it will be much closer to solder filled hole numbers.
Using torent search engine. I wander. :)) nice post although. Keep up good work!
Where are the answers from Q and A?
The wire should be clipped off, top and bottom ... wouldn't the extra wire and the mass of the resistor modify the results?
Just a thought, but sticking in that resistor like could act as a heat sink.
9:28 Was Dave actually afraid to call this a logarithm? Don't get me wrong, I love a good comedy trope, even if it is unintentional, but if the curve fits...and by the way, that curve is Gabriel's Horn. It's only natural.
Looks to me that the kink in current capacity is at the point where the via diameter equals the board thickness. This fascinates me; perhaps studying it could make for a good thesis paper for an EE graduate student.
but that is the thing about these videos, he is like our showman!.
Nice video! What's up with the torrent search engine ;)
Yes I see that I did a little research and my Idea isn't original XD, its also a good way to do a via on a home-made pcb according to what I read
It probably could effect the data somewhat
Hi Dave
Does not tinning your tip make a difference, watching you put that iron back in the stand with a dry tip made me all o.O
It was kind of obvious, but i didn't expect that much of a difference.
How about inductance. I really want to see a similar research on that damned parasitic..
Silver plate on vias, is definitely higher!
8 people have never etched a board...
How to design 100amps pcb
multi layer board with RF is unusual.
Awesome
Only if you wrapped the mat around so it touched both sides of the via!!!!
Ceramic RF worlde is different!
Or Linux distros :)
>TorrentZ
oh noes, it will cool down to room temperature faster!
option 0 on a mac º =P
Hai Dave, For prototyping DIY pcb's I use these: "www.bungard.de/index.php/en/products/through-hole-plating-line/through-hole-plating".
Does your software Altium supports through hole rivets automatic dimensioning when placing set in via-rules.
Notepad... Barepad...
DaveCAD...
Torrent Search engine ;)
If you leave the whole "bla-bla" away, the video will be just 5 min long, and as much informative as it is now.
bla-bla
MS EXCEL