SMD resistor socket (2512,1206,0805,0603)

[OM222O]

Hello
I have designed a milliohm meter which works great. I wanted to have an easy way of measuring resistors without using probes, so I made these tracks directly on the front panel which go between the force and sense terminals:


However after getting the PCBs, I found out that they are very fiddly and practically slower than just using the probes, as I have to put a lot pressure at the right angle to get a good connection which in hindsight is pretty obvious 

I searched a lot for a socket that can fit resistor packages, but I can't find anything. there exist modules for SOIC or TQFP packages, but nothing for the standard resistors. If I can't readily buy the sockets, is there a trick I can use on the PCB instead?
I thought about having spring contacts that are really close together, but for the larger sizes it would again be really difficult to fit them, unless I have 3 sets of springs contacts which I don't have enough space for 

[ataradov]

You will have much better results if you remove the solder mask from underneath the components. The solder mask is pretty thick and the component will just rock on it. Scrape it from the boards you have and see if there is a difference.

[OM222O]

I will give that a shot, but I'd still like to have a better solution. thanks!

[DaJMasta]

Will definitely do better, you could try to solder on sort of tiny bus bars so that the only contact point is on the pads not the chip, but to get a good representation of it you'd still need pressure from the top.


Maybe some sort of clamp?  A board like this on the bottom, no solder mask, a thin board or other bit that just has suitably sized cutouts for SMD parts in line with the pads arrangement (to keep the parts contained), then a second board that comes down on top, sandwiching the part.  Spring load the thing in a jig and you could probably even use the two contacts on either board for four wire measurements of reasonable quality.  You can't really use your hands for the pressure if you want it to be accurate into the 100s of k Ohm range, but it probably does the trick fine when in very low resistance ranges.  I would be weary that finger gunk could eventually make the contacts on the board a bit gummed up which could effect measurements if you were holding the parts down each time.


I think the usual approach for component measurement is precision machined contacts in a clamp from either side and then a low leakage dielectric holder piece, there are LCR and impedance measurement test fixtures that can give you a good idea of what the test equipment companies think work well.  This approach with no solder mask is a cheap and simple alternative that's used in a lot of component testers for SMD parts.

[OM222O]

very interesting idea about using a PCB as a jig with suitable cutouts! I will try to do that but obviously I need to change my layout for that.
I had a look at the clamps, and they are seriously expensive! about 100$ while the board costs about 20$ ... so that's a no go 
What if I want to get some sort of plastic clamp manufactured with some copper inserts embedded in them? this is a product that I plan on selling so I can get custom parts made if they're not 5x the price of the product itself 

[sleemanj]

instead of footprints, just make two "bus tracks" then you can just slide the resistor down the bus which should help to remove oxidation and make a good contact wihtout needing precise alignment.  Remove solder mask from the whole lot.

[OM222O]

I tried manually scraping off the solder mask which didn't end up being very clean and I still had the same issue. is it really all that there is to the problem? I don't wanna spin out a batch of new boards just to have them be exactly the same.
Anyone knows a manufacturer where I can get custom plastic probes made? I think they must be injection molded over the copper pads?

[Brutte]

A "flat" resistor on a "flat" surface does the contact only in three points, worst case.
It might also be that if the center of its gravity is low enough and the resistor is "rounded" it will actually sit stable with contact in two points only.

If you expect it to do contact in more than two points reliably then I think some kind of spring mechanism is needed.
Tweezers are not that bad. Or you could put two PCBs one sliding over the other + some spring to apply constant pressure on a resistor while it is being measured.

[OM222O]

A "flat" resistor on a "flat" surface does the contact only in three points, worst case.
It might also be that if the center of its gravity is low enough and the resistor is "rounded" it will actually sit stable with contact in two points only.

If you expect it to do contact in more than two points reliably then I think some kind of spring mechanism is needed.
Tweezers are not that bad. Or you could put two PCBs one sliding over the other + some spring to apply constant pressure on a resistor while it is being measured.

The problem is not how it sits on the pcb, it's making good contact with the PCB. that's why applying pressure fixes the issue but as I mentioned before, it's really fiddly

[DaJMasta]

I think you'll find some of that is caused by the thickness of the solder mask.  The parts may have slight expansion downward when placed normally on the pads to overcome this in a normal installation, but if you can remove the higher center thickness from the solder mask, then more of the pressure goes onto the pads if it was supporting any of the pressure on the part, which should increase contact reliability.

Having them actually manufactured with ENIG plating could be worthwhile too, especially with thicker plating settings like card edge connectors, to minimize the small variations in resistance you can get from contacts.  I still think the best option for reliability and removed fiddlyness aside from removing the solder mask is a clamping mechanism of some sort.  It could probably be as simple as a piece of FR4 that slides over the parts held down by a screw with a washer between it and the board.  Even applying force by hand to that PCB lever thing would be more even and repeatable than just pushing it on with your finger, and it reduces surface contamination from whatever's on your fingers.

[OM222O]

it is actually ENIG (sorry for bad picture quality  ).

I will try to make some sort of clamp using pogo pins since I have them on hand. this thread gave me a few good ideas. thanks everyone.

[artag]

I would separate the force and sense contacts as you have on the 1225 layout. I can't see the layouts where you've made a tiny sense point in a force surround working. It seems to me that you'll be getting a 3-point contact (per pad) before that will connect both sides. The force current is not large so doesn't need thick traces.

[NorthGuy]

Get another PCB, make holes where the components are supposed to be. Attach on the top of your PCB. Now simply drop the component into the hole and press on top of it. You'll get reasonably good contact.

Why do you want to measure the SMD resistors?

[OM222O]

the force current can actually be up to 1A 

I need to measure these so I can bin my resistors and buy cheaper ones in the upcoming projects since I can focus purely on tempco rather than actual tolerance. I sometimes pay up to 50 cent for resistors which is just insane!

[coromonadalix]

why dont u use  kelvin tweezers or maybe somekind of test fixture like this:
https://www.ebay.com/itm/TH26008A-SMD-Components-Test-Fixture-for-Precision-LCR-Digital-Bridge-Meter/332079349542?

[OM222O]

why dont u use  kelvin tweezers or maybe somekind of test fixture like this:
https://www.ebay.com/itm/TH26008A-SMD-Components-Test-Fixture-for-Precision-LCR-Digital-Bridge-Meter/332079349542?

I think this should answer you:

very interesting idea about using a PCB as a jig with suitable cutouts! I will try to do that but obviously I need to change my layout for that.
I had a look at the clamps, and they are seriously expensive! about 100$ while the board costs about 20$ ... so that's a no go 
What if I want to get some sort of plastic clamp manufactured with some copper inserts embedded in them? this is a product that I plan on selling so I can get custom parts made if they're not 5x the price of the product itself 

As I mentioned, I will be building a very similar device using pogo pins. I might post an update but it'll be a while.
I seriously have no idea why they are so damn expensive. I'm guessing they don't sell like hot cakes? which can be explained by their price? it's one of those circles ... with injection molding they can be made for less than 10$! there isn't anything inherently special about them.

[tautech]

They're not that dear from member Franky's shop:
https://www.ebay.com/itm/ETA3105-Kelvin-Clips-1-Pair/173337601080?hash=item285bb9e438:g:uC8AAOSwJ7dbCpHe

You can probably get them cheaper from some unknown source but Franky stands behind his products.
https://www.eevblog.com/forum/buysellwanted/franky_s-sales-thread/

I bought a pre-made Kelvin clip set from Pintek when I was getting some stuff from them and the build quality was excellent but the choice of cabling to the clips was not. I moaned that the short leads between the wings of the clips needed to be of quality flexible cable not some stiff PVC insulated muck which they took onboard to make changes in future production. I'd not buy any pre-made Kelvin clips again without handling them first to check cabling.