Single Sided Measurement, Kickstater Coming to Canada

Just a reminder, the above board made by APCircuits has worked out well. They are a 7/7 mil service however the board was designed for 6/6 mil process of OSHpark. The problem was the turn around. APCircuits had the order on a Wednesday night and it was delivered Monday afternoon. Total cost is $88 dollars shipped with Fedex Priority which was $30 for 4 boards. OSHpark is 17 shipped for 6 boards.

APCircuits are very good, they reminded me that there were “errors” in the board, but as previously mentioned I told them to go ahead. I haven’t found a single problem with the boards and I’ve tested all the small traces. I have tested the RTD sensor lines but they are larger to decrease resistance so I’m not concerned.

Above is the board connected to the debugger, battery and a single full bridge setup which is connected to my 4 setup test arm used in thermal testing from here. This means it’s pretty much equivalent to the Stages One powermeter. Single sided, but that’s just for testing out, coding and debugging.

Below are a few more pictures.

I expect that I’m going to have to re-adjust timelines. I’m away beginning of next month for a wedding and waiting until I’m back to order the strain gauges needed. I’m going to look into re-adjusting my proposed schedule and see what I can do. The “to do” list is still extensive but nothing impossible. The hardest thing really was the reflow. While still not perfect it’s fine. Again, the solution was to ditch the BGA package. That’s advice for any hobbyist with a DIY reflow oven. BGA is too much of a headache, QFN is hard but not impossible. Still haven’t received any import bill for mouser yet.

1. Finish coding that couldn’t proceed until the boards were completed and working
• I only coded initially for one side
• I’m cheating and only outputting raw reads to the power, this is mainly for debugging
• Sort out UART connection to help “debug”
• Look at implementing TE/PS measurements – This is low priority as no head units can read it
• I won’t code thermal until later and I get a sample RTD from omega
2. Enclosure design – I decided to hold back a bit on this due to adjustments that might need to be made for the board design.
3. Instrument another crank (or V3 to conserve cranks, have to rework it by removing old gauges and installing the new ones in the right locations (slight movement and different gauges)) and connect up powermeter.
4. Mount battery, measure current consumption.

My rate of work on this has to slow down anyway from both a financial position and a time position. The board was the main hurdle at this point, right now getting it together. However I am seriously looking at Kickstarter when it comes to Canada in order to offer the Board (~$40), Dev units ($500 – 600), and final units on Hollow Forged Cranks ($700 FCC Approved). At least that’s the goal. I’ll know more when I get circuit board manufacturing quotes and find out minimum order quantities in a few weeks.

Continually jumping off cliffs…

A friend of mine tweeted today:
“We have to continually be jumping off cliffs and developing our wings on the way down.” - Kurt Vonnegut
Since Friday (July 5th) I’ve jumped off of a lot of cliffs. About 9/10 times my “wings” didn’t work, leaving me to ask why. Okay, enough with the metaphors. On to ramblings about solder reflow and where my prototype boards are.

PCBs Arrive! Also, what is a Beta test?

The PCB’s arrived today. Six of them! So I have five I can ruin. I generally have good luck with PCB’s but I suspect this time I will not be so fortunate. I’ll mess up a few reflows before I get the IC’s soldered. Basically my first attempt will be just the Balun, nRF51422, and the crystals. I can sort the rest by soldering iron. It’ll be a lot of work, but I don’t want to ruin so many parts unnecessarily.

Here is something that has blown my mind. I’m basically still on track. I’ve got about 2 weeks to assemble and tune the PCB trace antenna and finish the enclosure design. . I am actually getting better estimating timelines it seems.

Really this is going to spill over to the build 5 – 10. Why? Well it’s simple -- I don’t have the strain gauges. I have three for the left sensor, but zero for the right (and I need 2 strain gauge rossettes for the right arm). I have lots of test beams to work with though, so continuing development won’t be a problem.

So here is a picture of where it will sit on the arm.

And now with chain rings.

There was a comment in my previous post asking what is my goal of all this. I’m going to explain this in a few days. The question asks “as I have alluded, I am a little confused as to where you are going with your economic and sales model." While I did a basic explanation, I think it’s time to spend some time presenting my long term goals with this – including major commitment points (including financial commitments), and potential exits (and reasons why I would exit). I’m going to spend the next few days pulling this together, so keep tuned. I’m aiming at Thursday or Friday post.

Beta Tests

I understand that I’ve confused several followers on several points. Some people who emailed me about beta involvement have offered money, or hardware. I can’t stress enough that I am not accepting money or hardware for the beta. I will be loaning the hardware to beta testers and will eventually reclaim it.

To clarify first off. An alpha test is where it’s been debugged by the developers and any other people who have been directly involved with the product development. A Beta test involves people who have generally considered an end user, they may be tech savvy or early adopters, who test and are willing to accept, work around, and report bugs and issues. They work with the developer to be fix problems and ensure the product is what consumers expect.

I still have to alpha test the V4 once it’s built. This should get it ready for beta, however the beta units will have to be built up so this will take a few weeks. I’ve compiled a list of people, their bike hardware, etc and I will be working to get things ready for them in the coming months.

The circuit board and strain gauges will be enclosed and able to deal with some light weather, but no major testing (such as IPX67) will have been conducted. The people who are local can have their units installed by myself, or a local bike shop if they prefer. Any non-local beta testers will have the crank shipped to them in original SRAM (or FSA potentially for BB30 / 386) packaging. Duration or testing is expected to be about about a month. Once the test is over I’ll pay for return shipping or for local testers I’ll replace the users original cranks. I’m mainly looking for verbal feedback, but files recorded via head units could be emailed to me which will aid in my development. I’ll be following up throughout the entire test. Probably weekly at pre determined times or via email.

Simple, right?

Boards Shipped and Cost Analysis coming soon

Shipped! They were very quick at turning around the depanelizing of the other 92 orders and getting these in the post.

Obviously nothing huge in terms of news until I get these and try to reflow but I’m still here.

In other news DCrainmaker has an update on the stages powermeter. It seems they’ve dialled it in a bit. Honestly, while I think it’s as good in terms of accuracy it seems it’s holding up very well to scrutiny by several seasoned riders of the meter. For two stain gauge pairs and what I’ve proven to be 30 dollars of circuit boards on what I estimate to be a 30 dollar crank arm (I’ve sourced SRAM Rival’s for approximately $100 CAD + shipping) so these are 600 dollars for R&D / lights on / profit. In my world profit is really just R&D money.

This leads me to make comment on what I’ve seen in the industry in terms of pricing. While powermeters are getting more popular, all one has to do is spend 10 minutes searching craigslist / kijiji / etc to find that locally they are rare. So rare that I’ve seen someone in Alberta want 1500 for the SPIDER from a Quarq S975. Locally in Ontario it’s only recently that I’ve seen up to 4 Quarq units available and 2 - 3 SRM.

While looking for a test unit I spent some time and eventually got one reasonably priced which I plan on selling come Sept. These aren’t high volume products currently. This leads me to potential pricing structures, motivated by a person who emailed me asking me straight out “why do I [you] think power meters are so expensive”. I paused, and it felt like an epiphany. It’s not the parts, it’s the people!

I plan on posting soon about what it really costs, based on what I’ve learned today, to build and Sell a powermeter, and not take a loss. It’s harder than one thinks and I’m sure not everyone who were successful on being funded on Kickstarter or Indigogo have thought on this. I’ve read about a few who had no idea how to scale production from tens to hundreds, and were bunt (sometimes repeatedly) on the way. This has been the reason why you aren’t allowed to just show a 3D rendering of a product anymore for Kickstarter. Too many people who haven’t deliver or were months to over a year away.

Stay tuned and look forward to my post this week going to be called, “Good, Cheap, and Fast – Pick two”

Boards Ordered

Firstly I want to give a huge thanks to Paul Archer. In my day job my employer has signs everywhere that say “Verify your design inputs”, and to that end we have a process of independent verification for everything. Paul has been that sounding board and verifier in the last couple of weeks, and without him I could have spent a lot more time dealing with issues of this board. I won’t say it’s perfect but it should work well.

I’m using a Texas instruments reference design for an PIFA style antenna that’s been extended. I plan to tune it with a 2.4GHz spectrum analyser by measuring power output. I’ve seen this technique several times.

I’ve realized today while checking footprints that very small components will have a “drop factor”. Essentially, 0402 parts and the BAL-nRF1D03 components will be hard to hold and align. So much so that I suspect that I will need a 3:1 ratio. Drop 2 – 3 and get one installed.

I do like the Purple boards. I’ll likely do a trace antenna design this week and also have it made in case the PIFA doesn’t work out. I am having trouble though making sure that I have the correct impedance match for a trace. Seems 1.6mm FR4 would require 2.81mm wide trace in 1oz copper according to the calculators, but most designs I’ve seen are much narrower. I’m not 100% sure who to trust. Antenna’s still feel like black magic thus far.

There is a 10 pin expansion header with 7 GPIO outputs that can be configured to any type of communications bus in addition to the 10 pin SWD programmer header. It has built in RTD sensor inputs for never having to zero your power meter. An RTD is a temperature sensor, and this will be used to measure each strain gauge setup. This is similar to Rotor Power. This won’t be fully implemented initially as it’ll require calibration for every power meter and I will have to sort that process but it will be.

Circuit Board Designed–95%

Circuit Board design with the exception of the antenna and all respective parts (except the nRF51422 which I already have in my possession) have been ordered. They should arrive Tuesday or Wednesday which will allow me to check to make sure all the parts fit the pads I’ve outlined on the board and make any needed edits. During this time I’ll design the antenna.

Currently the board is 26mm x 32mm. It might have to get wider to accommodate the antenna slightly as I do no wish to use a meander trace. I want either a trace antenna or an IFA. I might make separate board with a ceramic antenna.

I’ve been given the suggestion to include an SMP connection such that a network analyser could be used to tune the antenna. I am going to try the technique where you make the antenna too long and cut it back. I’ve read about people tuning with a basic RF spectrum analyser with this method which is cheaper for me. (RF spectrum = 200 dollars, network analyser = thousands).

I need to spend some time on the silk screen labelling. It’d be an amateur mistake not to spend more time on this as it really helps clarify what the parts are for future debugging.

Total cost of all associated parts is somewhere in the vicinity of 30 – 40 dollars. However, I’ve been quoted at prototype assembly at 70 – 90 dollars a board. More motivation to finish the reflow toaster.

Kicad–Schematic 90%–PCB 15%

Some slow progress, but it’s coming along. The schematic is about 90% done, but the PCB has barely been touched. I’ve just been doing some trials making sure it all translates well in Kicad from schematic to PCB layout. Below you can see the footprint layout compared to a CR2032. This is quite small. The trace PCB antenna will come off from the integrated balun from ST micro to the right. You can see the peculiar five dot layout next to the nRF51422 QFN footprint. This is ST micros integrated Balun. It’s 2mm^2.

To give you an idea on where this is going to reside I have added the picture below. The area just behind / below the chainrings seems sufficient for the circuit board. However two major hurdles.

1. Battery Placement
2. Antenna Placement

Coin Cell for scale. Note: I’m showing a CR2032, but I plan to use a CR2450 620mah cell. Slightly bigger physically.

Once I get the schematic and board close to finalization I will order the parts from Digikey. Most of the parts are very inexpensive except the ADS1248. This is considered the heart to the design in my mind.

When I order strain gauges I’ll also be ordering platinum RTDs. Each RTD will sense one of 3 strain gauge locations. (Top and Bottom for Right leg sensor, one for left leg sensing).

This is going to be close to a “NEVER ZERO” power meter. Rotor is pretty much there from what I’ve heard, but mine will be priced much more competitively once out of the beta. FTSH header include for reprogramming / hacking via the CMSIS-DAP.

Hoping to have the rough PCB by the end of the weekend, and a 3D concept of the enclosure which will be 3D printed. The enclosure will likely just resemble a circular battery holder by the crank arm and a simple covering over the board and strain gauges. Nothing fancy.

Progress Update and Commentary – With Finn and Jake

Anyone who knows me personally know that I’m a big Adventure Time fan. I love the nonsensical nature of the show as it provides a nice relaxing escape. Lately I’ve been gathering all the pieces for my Trek Madone 5.2. It’s not really a 5.2 as it is not going to be kitted out with Shimano (Ultegra) gear. It will receive my SRAM Rival group set that was purchased a few years ago and put on my GT GTR Series 4 which had originally came with Sora / ST-2200 shifters. I’ve tried to keep my bike with my ability level – however a Madone 5 Series is well beyond my need. I promise this is going somewhere. Just stick with me after the break.

Commentary

While I’ve trained with what I’ll call “Virtual Power”, which is using the regression curves for my Kurt Kinetic for 3 winters now, I haven’t experienced outdoor power until now. It is a very different experience having those numbers in front of you on a road. It’s teaching me it’s hard to zone in on a specific number and stay there. For my weight (100kg, I’m a big, dense boy) I’ve found that if I don’t push 400-500 watts then I don’t really go up even the smallest of hills very quickly. I recently read in the Wattage group that you really need a year of playing before you can seriously use this data. I’d say I’m not a novice, but I am not intermediate with data.

Having used this “Virtual Power” for a few years I thought I would be decent at evaluating power meters and what makes them good or bad. I’m a much more technical person, and as such I consistently forget that the coaches and athletes who use these tools while not knowing really how they work really. They don’t understand what would make one better than the other, which is why most people blindly say one is better than the other without understanding the technical merits of one over the other.

I’m not making any evaluations on products here, but I’ve been reading quietly in a lot of forums. In a previous post I showed how strain gauges, even in thermally compensated designs, can drift – very repeatedly drift as a function of temperature. A “good” auto-zero algorithm would adjust the zero based on this temperature – the holy grail, the platinum standard, etc, of power meters which wouldn’t require the user to do so much interacting. I don’t know the algorithms used by Powertap, SRM, Quarq, Rotor Power, etc but I do know that I am incorporating an auto zero adjustment because it will make it more accurate.

Rounding full circle to how I began, I will soon have a bike of very high level equipped with a tool that I think should be bought long before other bike upgrades but most people won’t get a power meter until they have a 4000 dollar bike. I am in no way capable of fully utilizing the performance of such a high quality frame. I’m nervous about this frame but not for the reasons you would expect. I’m nervous because showing off my prototype on it may illicit feelings that this is going to be another 2000 dollar plus product.

My goal in trying to create a lower cost power meter is to allow those aluminum framed, entry level, Shimano 105 / SRAM Apex/Rival / Campy Athena people who want to train more effectively to do so. It’s the 1500 – 2500 dollar bike people. The people who’s bike is equivalent to the price of one of these training tools currently. I think I will keep my 2008 GT GTR frame such that when I show off the new prototypes I can show them working on both my old used aluminum frame and my new carbon bike.

Progress Update

Seriously, I want to give a huge shout out to the people who made Arduino possible. What they did was translate the cryptic language of microcontrollers to something more easily digestible and along the way made some simple tools. I’ve been stuck for two plus weeks on “Basic Code”. I’ve been stressed out because it’s been improbably annoying to replicate the millis() function of the Arduino which is necessary for calculating cadence from any pulse type sensor. It had to do with existing “Timer.c” library that came in the reference code. It was quite annoying.

This has throw off the schedule somewhat, however I believe I have it sorted finally. However, in sorting this out I found out that the reference code doesn’t work like I thought it did and is therefore less power efficient than I thought. This comes as a surprise, though I am seriously behind on a couple of tasks. The Gyro Evaluation still shows up and I still want to do it, but it’s hard to make it happen. The priority now is the current measurements. I have to get this done this weekend now that the cadence calculation is almost worked out. At least I’m 3 weeks ahead on ordering cranks.

I really wanted to be doing PCB design this week which is why I took three days off work – such that I could do more work. Who knows, I might get to start on it yet.

Kicad–Another Learning Curve

For my initial circuit boards I used Cadsoft Eagle. In my undergrad I tried it and I very much disliked it due to it feeling “dated”, but I was busy doing mechanical engineering. As a result I decided to push my EE-like ambitions out, and moved on with my life.

When I was hit with the need to design PCBs I had found that the hobbyist community had already strongly developed around Eagle. To be honest, I’ve never been sure why. I mean, in the 6 years I haven’t been using Eagle it hasn’t changed – well, 6 years ago it crashed a bit, and now it doesn’t. But the interface looks identical and as a result I’ve had to change my Solidworks / Ansys mind-set towards a very different workflow.

However, today, I hit on a legal dilemma. CADsoft Eagle’s free version is for non-commercial works only. If I create the circuit board for my beta group with Eagle, then I have to reproduce it later in another package if I wanted to sell it. It’s better to bite the bullet and do this up front. I don’t have a specific timeline built up, but I plan to start ordering parts within a week (resistors, capacitors, inductors, battery holders, etc) and designing up the enclosure (which will be produced after I get a new PCB up and running).

Kicad is a free and open source alternative with no commercial restrictions as far as I can see. I’ve been doing a tutorial to get my bearings and it’s not too different from Eagle – except it looks and feels 5 – 8 years newer. The workflow and buttons are similar but like anyone who switches from one 3D software to another: the combinations of CTRL, ALT, Shift + Mouse L/R/Middle + movement are all different. After using one piece of software a lot, say ANSYS Classical, then moving to Solidworks, then to Ansys Workbench, means you are constantly confusing yourself on how to navigate a simple space. It really makes you feel a bit like an idiot – luckily nobody really notices.

To be honest, I wish I had access to Altium Designer as Nordic already has reference schematics and layouts in these. However, Altium Designer is not a cost feasible solution ($7245 US) and I’m sure the learning curve would be even worse. I’m sure an experienced designer would be able to make use of the design software better than I could. However I’m at best a Mechatronics kind of guy.

Code Mostly Ported and Other Things

I’ve been working in the background on things that most people will find dull – more coding. No major breakthroughs, but no breakthroughs are needed at this point. I’ve ported about 65% – 70% of the code and what is left to do is below. More after the break.

SPI, Buttons, Toasters, and More

Just a small update today on what I’ve been working on.

• nRF51422 – Talking to the ADS1247 on the SPI bus, figuring out the reference code, and reading and writing pins.
• Looking into Gyro’s. They have very very high current consumption. 4.2 – 6.4ma. This is much to high. I can’t find any stats on start up time. Sleep mode doesn’t reduce current below 2ma. Still too high.
• Strain Gauges – Not much new, but I’m hoping I can find 1000 ohm gauges instead of 350 ohm. This will save power
• Toaster Oven – I’ve decided to make a reflow oven. This should speed testing new circuit boards. I have an SSR on order, and a thermocouple and sensor chip on order from Adafruit. Let’s see if I get dinged with Tarrifs.
• I’m interested in going to the ANT+ Symposium in Sept, but looks like I’ll have to upgrade to Alliance Membership. $1500 dollars a year is a little hard to swallow right now. If I’m in the market for beta testers by June I’ll sign up then and just suck it up.
• Beta Testers – If you’re a cyclist in Ontario or know someone who might be interested in testing my next prototypes have them drop me a line at either kwakeham@gmail.com or kwakeham@accuity.com. Oh no, I gave away my new company name! Website coming in a month.
• I’m thinking about making the board, or a variation of it, available to the hacker community. The plan would be to have space that other sensors (Accelerometer, Gyro, Magnetometer, General Purpose Analog Input with Gain [ADS1247/8]). Idea for now. ANT+ for all I say (but become an Adopter to get the ANT+ Key of course at www.thisisant.com )
• Below is some testing. I’m sure all are becoming familiar with my laptop layout.
• While I might pick up a BTLE sensor Tag from TI, I have no ambition to pursue BTLE at this time. Cost mainly. ANT+ will require FCC + $1500 membership + $750 device registration / testing. BTLE requires $22500. I am not joking. I became a BTLE adopter and I can barely understand the documentation and only have the vaguest ideas how it works compared with ANT+. Right now that’s a learning curve I am not willing to accept. If Accuity gets off the ground, has people test some prototypes, then I’ll reconsider.

ADS1248 / ADS 1247 Received and Mounted

I got my breakout boards today for the 24 bit Sigma-Delta ADC’s. Soldered and conductivity tested. Now to wire them to an Arduino and a strain gauge. Approximately 3ma for the ADS1247, 2.048 reference means 5.85ma to each 350 ohm strain gauge setup. Still a little high, but we’ll see. Proto-Advantage was super quick in sending with regular snail mail. I live close to them (Hamilton, ON I think) so it’s not surprising.

V3–Power Meter–The Complete How To

For the current visitors to the blog, it's finally here! For new visitors welcome to the complete how to. What I am going to explain start to finish how to make and advanced left and right measuring power meter. The quickest and easiest way to figure out how to put your own together or find out how other power meters work such as SRM, Quarq, Cycleops or Power2max is watch the videos in sequence below.
Left to Right V1, V2, V3 of my power meter. Questions? kwakeham@gmail.com

V3–Video and Tutorial Coming Soon

I’m working on a video that details the whole process of instrumenting and building the power meter as well as explaining the concepts. I’m starting from scratch in the video. How power is calculated, how rotational velocity measured, and most importantly, and in-depth, how torque is measured via strain gauges as well as how a strain gauge works. Below are some screenshots of the upcoming video. Currently I’m planning on shooting the table top demo of V3 and the on bike video to conclude the raw footage capture Sunday (Jan 13th, 2013). I might have it edited together in the evening for posting Monday or Tuesday.

Overview. More pictures after the break.

TSSOP 20 Breakout? Help?

I’ve been reading up on the SAR (successive approximation, eg the ADS8321) versus Delta-Sigma (eg the ADS 1247). From what I can find out the Delta-Sigma’s are slower BUT they have much better noise rejection. I’m using an averaging technique with the SAR where the Delta-Sigma may not need an averaging technique or very little.

So if anyone knows of a TSSOP 20 breakout to DIP or DIL that’s available in Canada without a fortune in shipping let me know.

This could also reduce cost. The ADS1247 is approximately $13.50 CAD which eliminates two ADS8321 and AD623 and high precision resistors. I believe that is about $45 dollars. Major electronics could be < 40 dollars. Obviously, a power meter won’t be that – labour + other parts + casing + testing + overhead and equipment.

Still looking for an appropriate gyro on breakout board. I think I’ll eventually have to pick up a hot air rework station since I’m not ready to commit to a reflow oven – Gyros are only available in LGA style. If anyone in the Kitchener-Waterloo-Cambridge or Guelph area in Ontario that has one and willing to help out, let me know!

Hacking your own PCB

I mentioned a problem with the LM4140 reference voltage as a poor design decision. Adding a capacitor didn’t solve it nor did adding a resistor. Realizing I was unhappy with this I decided to hack the board and put on a voltage divider for the reference voltage.

I explained before but the short is you need a voltage reference in between the +ve voltage and the ground in order to have a single sided amplifier. This reference raises the differential voltage from the strain gauge.

I’ve attached a picture of the old reference resistor and the hack of two resistors and a bridge wire. It’s giving much better stability.

So why is this ultra precise high grade IC failing… because it’s source is a poor linear regulator on the Arduino Pro Micro. So this means that as the supply voltage drifts up and down that means that relative the reference it’s drifting. This is now ratiometric and even if the supply drifts the reference is still at the same “relative” voltage keeping it all nice and stable.

I seriously hate magnetic reed sensors. I dislike how I have to program the microcontroller. V4 will have a MEMs gyro.

I spent yesterday figuring out how would I program the nRF51422 if I switched to that for development. Kicking myself for missing out on the free board Nordic offered, but even more so for not getting the free day of training. That could be invaluable. However, my self learning has revealed that the nRF51422 is not as scary too deal with. Segger J-Link to 9 pin “JTAG” (not true JTAG as I understand it for the nRF51422 which actually uses a two pin arrangement to program, but it does contain the normal JTAG lines) is all there is to it.

A big issue with this is that if I don’t incorporate an OEM J-LINK means others can’t reprogram it. Mixed feelings on this. Let me know your thoughts? I haven’t decided if V4 will be a minor increment or more major. Feel free to let me know your thoughts via email.

V3 Progress

No pictures today. Just a bit of a text update.

I've received all the components to complete the V3 circuit board. Surface mount soldering by hand is generally difficult, but the AP2 modules are proving to be the absolute most stubborn pieces to solder that I've had the displeasure of attempting. This is a situation where you need to accept the old adage: "The right tool for the job". I am not using the right tool.

If you attempt to surface mount the AP2 module by hand here is something to keep in mind. There are small through holes for the contacts on the base, just big enough for a fine leg of a resistor to fit down. They also make great test points. If you test and cannot get the solder on your PCB to connect to the AP2 then put a wire down the hole tinned lightly with solder. Heat the wire and keep pushing it down, then alternate heating the pad on your board (the small amount exposed) and the wire while applying pressure downward. Eventually you'll make a connection. Check it with a multimeter. Once you've made a good connection heat the pad and then the wire and put the wire out -- it will pull solder up through the hole and seems to make a very reliable connection.

This is time consuming. I spent about 3 hours trying to surface mount two of these to my PCB and several connections would not connect. This technique takes several minutes per contact, so about an hour to get all the connections working. All of this could have been avoided if I had a reflow oven. Necessity is the mother of invention.