Soldering a LGA-16 MEMS part

7 views
Skip to first unread message

Skye Sweeney

unread,
May 19, 2020, 1:09:53 PM5/19/20
to nashuarob...@googlegroups.com
Some time ago I soldered an LGA-16 MEMS IMU to a protoboard. This part has no visible leads and is just 3mm by 3mm with 16 pads underneath.  I have recently found out that some of the internal sesnors are FUBAR. talking with the vendor they think I have either shocked it too much or fried it during soldering. I have new parts on order and will want to solder on replacements this weekend. But how?

My original method was solder paste, stencil, and a hotplate to reflow. It seemed to happen quickly so I do not think I cooked it. But I might have thermally shocked it by bringing it up to temp too fast. My hot plate does not have a thermal controller on it to get the exact profile the vendor wants.

So I could just try it again the same way, try hor air reflow, or "Yee Old Toaster Oven". Anyone have success soldering something like a tiny 3/6/9 axis MEMS motion sensor?

--
-Skye Sweeney

Jay Francis (Reactive Technologies)

unread,
May 19, 2020, 1:36:24 PM5/19/20
to nashuarob...@googlegroups.com
Yes - I've done similar parts.

I use my toaster oven (stencil + paste).

I've also done hot air, but it can be hit or miss.  Hot air works great for removal though.

Another thing to watch for, although I've only had issues with the NeoPixel LEDs, is the need to bake out moisture before soldering.   The part specs should say if it is recommended/required, and what you need to do.

--Jay
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nashuarobotbuilders/CACR%2B-fxdZ6oF-i9F4YB1807HL3WEwyope8fFwqMzQh6UZuLXjw%40mail.gmail.com.

Skye Sweeney

unread,
May 19, 2020, 2:25:58 PM5/19/20
to nashuarob...@googlegroups.com
My parts were delivered hermetically sealed. Popcorning should not be a problem.

I may buy a toaster and give that a try for the second go around.


To view this discussion on the web visit https://groups.google.com/d/msgid/nashuarobotbuilders/64401113-3759-e467-2fcf-42fc77ea4267%40ReactiveTechnologies.com.


--
-Skye Sweeney

Skye Sweeney

unread,
May 19, 2020, 4:50:28 PM5/19/20
to nashuarob...@googlegroups.com
Jay,
From the demo you gave a few years ago, I remember something of a recipe you used.
Turn on to medium for a few minutes. 
Stick in boards. 
Wait a minute or two.
Turn up to 'Broil"
Wait for solder to melt.
Kill power.
Wait 10 seconds
Open door
Wait 3 minutes.
Pull from oven.

Is that about right?
--
-Skye Sweeney

Jay Francis (Reactive Technologies)

unread,
May 19, 2020, 5:30:28 PM5/19/20
to nashuarob...@googlegroups.com
My current process (with my “profiled” toaster oven) is:

Put the boards in at room temperature (one, or a dummy board, will have a type k thermocouple kapton taped to the PCB, read temperature with a meter).

Set oven to Convection Bake, max temperature.

Start the oven

When I hit 185C I start a timer
I turn off the oven and open the door when I hit 220C

Carefully/slowly remove the boards (whole rack or tray).  They are still flowing above 175C ish, but you need to cool them down

Set them on top of the oven to cool the rest of the way

Note: It should be about a minute between 185C and 220C. Too fast is bad... too slow is bad... you can get pickier based on the solder/chip profiles.  You should “profile” your oven’s ramp rate from start to 185C as well and may need to make adjustments.  My oven happens to ramp close to recommended for the paste I use (just lucky...).


On May 19, 2020, at 4:50 PM, Skye Sweeney <sk...@fll-freak.com> wrote:



Skye Sweeney

unread,
May 19, 2020, 7:47:41 PM5/19/20
to nashuarob...@googlegroups.com
Thanks. I guess I need to order a toaster oven and a thermocouple!

-Skye 

Dana Kelley

unread,
May 20, 2020, 8:07:16 AM5/20/20
to nashuarob...@googlegroups.com
Skye,

I wonder if a build plate with a insulating cover on a 3D Printer will work.

Using G-code you could control the temperature precisely, add a digital thermometer on the pcb to track from the outside. Then modify G-Code accordingly.

Using printer head cooling fan you could also decrease cooling time and perhaps control it fairly well.

Dana


Skye Sweeney

unread,
May 20, 2020, 8:14:40 AM5/20/20
to nashuarob...@googlegroups.com
Possibly with a different heaed bed. I have on of those PCB heaters and it would not get anywhere hot enough to melt solder.
Might have to replace the normal bed with a small size silicone heater that could melt solder.

There are kits available to turn your toaster into a reflow machine with temperature profiles and he like. 
Some of those kits are just about as expensive as a Chineese dedicaed reflow oven.

To view this discussion on the web visit https://groups.google.com/d/msgid/nashuarobotbuilders/1289188048.1390996.1589976433564%40mail.yahoo.com.


--
-Skye Sweeney

Jay Francis (Reactive Technologies)

unread,
May 20, 2020, 8:47:08 AM5/20/20
to nashuarob...@googlegroups.com
My oven is an Oster (Sunbeam Products) TSSTTVDG01-A

-Probably- exact oven is not available now, but you can still look up specs.
 
I also use it for vacuum forming (large enough for a 12" x 12" frame).

--Jay

Dana Kelley

unread,
May 20, 2020, 9:14:50 AM5/20/20
to nashuarob...@googlegroups.com
Hmmm.... Low end reflow machines are under $250. To expensive for occasional use. And do they work well? I persoanlly find working out this type of issue tedious and not where I want to spend too much effort. However, trial and error can be very frustrating.

A cannibalized 3D printer provides extruder temeperature sensor, output controls for heating element and fan. The outputs could control relays that drive the power. GCode is easy to write.

Of course the real goal is to solder the parts on the pcb and not develop a reflow oven!

I'll be getting to this level or PCB development in the future. Right now I'm using through-hole mostly with tiny modules. Moving all components to the PCB is the eventual goal.

Jay Francis (Reactive Technologies)

unread,
May 20, 2020, 11:25:17 AM5/20/20
to nashuarob...@googlegroups.com
The most important part of the process is the ramp rate. 

If your oven's natural ramp rate is close to the recommended rate for the paste you're using, you can do the whole process manually with an unmodified oven and a good temperature probe.  Note: if it is close to the recommended ramp rate, adding a PID controller won't help matters unless you add insulation or additional heating elements (see ramp rate too slow below).

If the natural oven ramp rate is too fast, poke holes in the case.  Or, add a PID controller.

If the natural oven ramp rate is too slow, add insulation, add heating elements, or get a more powerful oven.

On the cooling side... Most ovens will retain heat better than the cool down ramp rates required.  Opening the door works (manually or automatically).  Or you can get fancier...

--Jay

May 20, 2020 at 9:14 AM
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
Possibly with a different heaed bed. I have on of those PCB heaters and it would not get anywhere hot enough to melt solder.
Might have to replace the normal bed with a small size silicone heater that could melt solder.

There are kits available to turn your toaster into a reflow machine with temperature profiles and he like. 
Some of those kits are just about as expensive as a Chineese dedicaed reflow oven.



--
-Skye Sweeney
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nashuarobotbuilders/CACR%2B-fxc4czVTuh50-hSySBwMPaAt2j%3D-uFUQCdAO3JqBBJK%2Bw%40mail.gmail.com.
May 20, 2020 at 8:07 AM
Skye,

I wonder if a build plate with a insulating cover on a 3D Printer will work.

Using G-code you could control the temperature precisely, add a digital thermometer on the pcb to track from the outside. Then modify G-Code accordingly.

Using printer head cooling fan you could also decrease cooling time and perhaps control it fairly well.

Dana


On Tuesday, May 19, 2020, 07:47:43 PM EDT, Skye Sweeney <sk...@fll-freak.com> wrote:


Thanks. I guess I need to order a toaster oven and a thermocouple!

-Skye 

On May 19, 2020, at 5:30 PM, Jay Francis (Reactive Technologies) <J...@reactivetechnologies.com> wrote:

My current process (with my “profiled” toaster oven) is:

Put the boards in at room temperature (one, or a dummy board, will have a type k thermocouple kapton taped to the PCB, read temperature with a meter).

Set oven to Convection Bake, max temperature.

Start the oven

When I hit 185C I start a timer
I turn off the oven and open the door when I hit 220C

Carefully/slowly remove the boards (whole rack or tray).  They are still flowing above 175C ish, but you need to cool them down

Set them on top of the oven to cool the rest of the way

Note: It should be about a minute between 185C and 220C. Too fast is bad... too slow is bad... you can get pickier based on the solder/chip profiles.  You should “profile” your oven’s ramp rate from start to 185C as well and may need to make adjustments.  My oven happens to ramp close to recommended for the paste I use (just lucky...).


--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nashuarobotbuilders/C9FAF04A-6D79-49B4-919D-34182E1D23E3%40ReactiveTechnologies.com.
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nashuarobotbuilders/3DF2F7C9-97F3-4CC3-85B8-041C0DB60EE9%40fll-freak.com.
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nashuarobotbuilders/1289188048.1390996.1589976433564%40mail.yahoo.com.
May 19, 2020 at 7:47 PM
Thanks. I guess I need to order a toaster oven and a thermocouple!

-Skye 

On May 19, 2020, at 5:30 PM, Jay Francis (Reactive Technologies) <J...@reactivetechnologies.com> wrote:

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nashuarobotbuilders/3DF2F7C9-97F3-4CC3-85B8-041C0DB60EE9%40fll-freak.com.
May 19, 2020 at 5:30 PM

Skye Sweeney

unread,
May 20, 2020, 12:03:36 PM5/20/20
to nashuarob...@googlegroups.com
The after markey kit I was looking at (Controleo?) had a servo to open the door to let heat escape.
But that kit with a toaster was more than a cheap Chineese model.

As for ramp rates, if the profile given by the chip maker differes than that of the paste, what so you do?
I guess make sure your paste needs don't violate the chip and if so get different paste?
If it does not violate, use the chip profile?

What ever happened to DIPs, sockets, and wire wrapping? :)


To view this discussion on the web visit https://groups.google.com/d/msgid/nashuarobotbuilders/5EC54BDA.4050909%40ReactiveTechnologies.com.


--
-Skye Sweeney

Dana Kelley

unread,
May 20, 2020, 1:42:04 PM5/20/20
to nashuarob...@googlegroups.com
Sky,

Jay's suggestion seems doable and straight forward.

Is there an easy way to measure and record the temperature curve automatically. https://www.tutorialspoint.com/arduino/arduino_temperature_sensor.htm This sensor works up to 250C GP103G4F. Mouser has it for $5.

Dana



Jay Francis (Reactive Technologies)

unread,
May 20, 2020, 2:12:38 PM5/20/20
to nashuarob...@googlegroups.com
I hope it’s doable and straightforward - it’s how I’ve been doing it for years ;-)

On May 20, 2020, at 1:42 PM, 'Dana Kelley' via Nashua Robot Builders <nashuarob...@googlegroups.com> wrote:



Jay Francis (Reactive Technologies)

unread,
May 20, 2020, 2:16:39 PM5/20/20
to nashuarob...@googlegroups.com
Stopwatch, notepad, temperature meter is all you really need.  Measure at 15, 30, or even 60s intervals is enough resolution.

On May 20, 2020, at 2:12 PM, Jay Francis (Reactive Technologies) <J...@reactivetechnologies.com> wrote:



Jay Francis (Reactive Technologies)

unread,
May 20, 2020, 2:20:10 PM5/20/20
to nashuarob...@googlegroups.com
I usually stick with the paste profile - most chips should tolerate that.  If they don’t, I guess you need to change paste.




On May 20, 2020, at 2:16 PM, Jay Francis (Reactive Technologies) <J...@reactivetechnologies.com> wrote:



Skye Sweeney

unread,
May 20, 2020, 2:21:18 PM5/20/20
to nashuarob...@googlegroups.com
I think if I were to McGuiver something, I would likely use a K-type thermocouple that is rated for much higher temps as it would have other uses possibly. Sparkfun makes an I2C/SPI thermocouple amp for aboit $15.But the real problem is that a toaster oven or my old hot plate have very high thermal inertia and low ability to change temperaturs quickly. So as jay was saying, a PID loop is lost if you don't have the authority to change temperature quickly. Commercial single board ovens use an IR bulb for the heat souce. The Industial ones seem to use multiple heat zones and a variable speed conveyor to create the temp profiles. But then again I don't have a spare 5K to spend on soldering a $3 part!

I think I will try my hot plate again but characterize the profile with a "Thermapen" thermometer. Then I can adjust the control knob by hand to try to match the needed profile. If that fails, I guess I need to suit up in PPE and go to Wallmart to buy a cheap toaster oven! 

To view this discussion on the web visit https://groups.google.com/d/msgid/nashuarobotbuilders/1248968448.1537690.1589996521136%40mail.yahoo.com.


--
-Skye Sweeney

Skye Sweeney

unread,
May 22, 2020, 12:24:17 PM5/22/20
to nashuarob...@googlegroups.com
So last night I ran a quick test wih my hot late. 
This is not a food service hot plate but rather an industrial/scientific one.
The hot surface is 9*9 inches and a good 1/4" thick aluminum.
It goes up to 315C.
It was not hard to get it up to the recommended 217C in the time specified.
It was hard to get it to cool down. In ten minutes it only dropped a few degrees.
There is a pretty large thermal mass in that table.
I think I will need to either build a wind tunnel, or raise he board off the plate to get the needed cooling rate after reflow.

New chips arraived in the mail this morning. 
My wife calls me to tell me:
"A package from DigiKey has arrived and it is marked perishable. Should I put it in the refridgerator?"

I guess these parts (subject to popcorning) may not be hermetically sealed like the last batch.


--
-Skye Sweeney

Dana Kelley

unread,
May 22, 2020, 3:14:35 PM5/22/20
to nashuarob...@googlegroups.com
I would try raising it off the platform. How quickly must it cool down?

Skye Sweeney

unread,
May 22, 2020, 3:32:59 PM5/22/20
to nashuarob...@googlegroups.com
You need to approach "liquid" (217C) at less than 3C/second.
From "Liquid" you need to hold for 10 to 30 seconds. 
Then drop at LESS than 4C/second to 150C. After that they don't have constraints.
So a sudden yank from the heat after "liquid" is not a good idea.

So if I put a shallow pan of sand on my hot plate, then take that off after it goes liquid, that might have the right temperature rate.
I seem to remember that my neighbor has a kid with a sand box and it gets dark at night....

To view this discussion on the web visit https://groups.google.com/d/msgid/nashuarobotbuilders/928966938.2384351.1590174869949%40mail.yahoo.com.


--
-Skye Sweeney

Dana Kelley

unread,
May 22, 2020, 4:24:11 PM5/22/20
to nashuarob...@googlegroups.com

The Sandbox Caper... And wearing a mask is now expected... you're all set for an evening of intrigue.

So it needs to drop to 150C, or drop 67C in 16-17 seconds. That's fairly quick.

Dana Kelley

unread,
May 26, 2020, 8:46:09 AM5/26/20
to Nashua Robot Builders
So.... just curious on how this panned out...
On Wednesday, May 20, 2020, 08:14:42 AM EDT, Skye Sweeney <s...@fll-freak.com> wrote:


Possibly with a different heaed bed. I have on of those PCB heaters and it would not get anywhere hot enough to melt solder.
Might have to replace the normal bed with a small size silicone heater that could melt solder.

There are kits available to turn your toaster into a reflow machine with temperature profiles and he like. 
Some of those kits are just about as expensive as a Chineese dedicaed reflow oven.

On Wed, May 20, 2020 at 8:07 AM 'Dana Kelley' via Nashua Robot Builders <nashuarob...@googlegroups.com> wrote:


--
-Skye Sweeney
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.
May 20, 2020 at 8:14 AM
Possibly with a different heaed bed. I have on of those PCB heaters and it would not get anywhere hot enough to melt solder.
Might have to replace the normal bed with a small size silicone heater that could melt solder.

There are kits available to turn your toaster into a reflow machine with temperature profiles and he like. 
Some of those kits are just about as expensive as a Chineese dedicaed reflow oven.



--
-Skye Sweeney
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.
May 20, 2020 at 8:07 AM
Skye,

I wonder if a build plate with a insulating cover on a 3D Printer will work.

Using G-code you could control the temperature precisely, add a digital thermometer on the pcb to track from the outside. Then modify G-Code accordingly.

Using printer head cooling fan you could also decrease cooling time and perhaps control it fairly well.

Dana


On Tuesday, May 19, 2020, 07:47:43 PM EDT, Skye Sweeney <s...@fll-freak.com> wrote:


Thanks. I guess I need to order a toaster oven and a thermocouple!

-Skye 

On May 19, 2020, at 5:30 PM, Jay Francis (Reactive Technologies) <J...@reactivetechnologies.com> wrote:

My current process (with my “profiled” toaster oven) is:

Put the boards in at room temperature (one, or a dummy board, will have a type k thermocouple kapton taped to the PCB, read temperature with a meter).

Set oven to Convection Bake, max temperature.

Start the oven

When I hit 185C I start a timer
I turn off the oven and open the door when I hit 220C

Carefully/slowly remove the boards (whole rack or tray).  They are still flowing above 175C ish, but you need to cool them down

Set them on top of the oven to cool the rest of the way

Note: It should be about a minute between 185C and 220C. Too fast is bad... too slow is bad... you can get pickier based on the solder/chip profiles.  You should “profile” your oven’s ramp rate from start to 185C as well and may need to make adjustments.  My oven happens to ramp close to recommended for the paste I use (just lucky...).


--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.
May 19, 2020 at 7:47 PM
Thanks. I guess I need to order a toaster oven and a thermocouple!

-Skye 

On May 19, 2020, at 5:30 PM, Jay Francis (Reactive Technologies) <J...@reactivetechnologies.com> wrote:

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.
May 19, 2020 at 5:30 PM
My current process (with my “profiled” toaster oven) is:

Put the boards in at room temperature (one, or a dummy board, will have a type k thermocouple kapton taped to the PCB, read temperature with a meter).

Set oven to Convection Bake, max temperature.

Start the oven

When I hit 185C I start a timer
I turn off the oven and open the door when I hit 220C

Carefully/slowly remove the boards (whole rack or tray).  They are still flowing above 175C ish, but you need to cool them down

Set them on top of the oven to cool the rest of the way

Note: It should be about a minute between 185C and 220C. Too fast is bad... too slow is bad... you can get pickier based on the solder/chip profiles.  You should “profile” your oven’s ramp rate from start to 185C as well and may need to make adjustments.  My oven happens to ramp close to recommended for the paste I use (just lucky...).


--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.


--
-Skye Sweeney

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.


--
-Skye Sweeney


--
-Skye Sweeney

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.


--
-Skye Sweeney

Skye Sweeney

unread,
May 26, 2020, 9:06:13 AM5/26/20
to nashuarob...@googlegroups.com
Thanks for asking, but unfortunately, little to no progress,

Rather than spending quality time in my lab, I elected to pull my back and so spent quality time lying flat in bed.
I did take the opportunity to buy a thermocouple and PC interface to characterize my hotplate before I do the soldering next weekend.
I also spent more time trying to get the IMU chip (ICM20602) I do have running. 
I get the feeling Invensense does not care a lot about their documentation.
They claim there is an interrupt for reaching the watermark on the FIFO, but they fail to document the interrupt enable register. 
It is only an 8 bit register, but none of those bits seems to generate the interrupt on watermark. 
They use terms like GDRIVE that they never define. 
They talk about "burst mode" to access the FIFO but never explain what that is.
In open source code for the this chip I see people writing to undocumented registers with comments like "Had to do this to get the chip to work".
In that same code I also see no one trying to access the chip by interrupt and lots of comments saying that the FIFO gets corrupted after the first 24 of 1000 samples.

I think the added precision on paper of this part is not worth the missing information on paper.
I may solder the part, just to say I did and then throw the whole thing away for a more run of the mill part.
Live and learn.

To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
May 20, 2020 at 8:14 AM
Possibly with a different heaed bed. I have on of those PCB heaters and it would not get anywhere hot enough to melt solder.
Might have to replace the normal bed with a small size silicone heater that could melt solder.

There are kits available to turn your toaster into a reflow machine with temperature profiles and he like. 
Some of those kits are just about as expensive as a Chineese dedicaed reflow oven.



--
-Skye Sweeney
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
May 20, 2020 at 8:07 AM
Skye,

I wonder if a build plate with a insulating cover on a 3D Printer will work.

Using G-code you could control the temperature precisely, add a digital thermometer on the pcb to track from the outside. Then modify G-Code accordingly.

Using printer head cooling fan you could also decrease cooling time and perhaps control it fairly well.

Dana


On Tuesday, May 19, 2020, 07:47:43 PM EDT, Skye Sweeney <s...@fll-freak.com> wrote:


Thanks. I guess I need to order a toaster oven and a thermocouple!

-Skye 

On May 19, 2020, at 5:30 PM, Jay Francis (Reactive Technologies) <J...@reactivetechnologies.com> wrote:

My current process (with my “profiled” toaster oven) is:

Put the boards in at room temperature (one, or a dummy board, will have a type k thermocouple kapton taped to the PCB, read temperature with a meter).

Set oven to Convection Bake, max temperature.

Start the oven

When I hit 185C I start a timer
I turn off the oven and open the door when I hit 220C

Carefully/slowly remove the boards (whole rack or tray).  They are still flowing above 175C ish, but you need to cool them down

Set them on top of the oven to cool the rest of the way

Note: It should be about a minute between 185C and 220C. Too fast is bad... too slow is bad... you can get pickier based on the solder/chip profiles.  You should “profile” your oven’s ramp rate from start to 185C as well and may need to make adjustments.  My oven happens to ramp close to recommended for the paste I use (just lucky...).


--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
May 19, 2020 at 7:47 PM
Thanks. I guess I need to order a toaster oven and a thermocouple!

-Skye 

On May 19, 2020, at 5:30 PM, Jay Francis (Reactive Technologies) <J...@reactivetechnologies.com> wrote:

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
May 19, 2020 at 5:30 PM
My current process (with my “profiled” toaster oven) is:

Put the boards in at room temperature (one, or a dummy board, will have a type k thermocouple kapton taped to the PCB, read temperature with a meter).

Set oven to Convection Bake, max temperature.

Start the oven

When I hit 185C I start a timer
I turn off the oven and open the door when I hit 220C

Carefully/slowly remove the boards (whole rack or tray).  They are still flowing above 175C ish, but you need to cool them down

Set them on top of the oven to cool the rest of the way

Note: It should be about a minute between 185C and 220C. Too fast is bad... too slow is bad... you can get pickier based on the solder/chip profiles.  You should “profile” your oven’s ramp rate from start to 185C as well and may need to make adjustments.  My oven happens to ramp close to recommended for the paste I use (just lucky...).


--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.


--
-Skye Sweeney

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.


--
-Skye Sweeney


--
-Skye Sweeney

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.


--
-Skye Sweeney

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nashuarobotbuilders/670ade15-6643-4c25-adea-10a45f766065%40googlegroups.com.


--
-Skye Sweeney

Dana Kelley

unread,
May 26, 2020, 9:29:27 AM5/26/20
to Nashua Robot Builders
Hope your back to normal soon!

When using the MPU6050 it hung up a lot. The IC2 would just not respond in mid transfer.
Searching the internet I found lots of complaints concerning Ivensense secrecy and poor documentation.

Always trying to guess how to make it work is a true hacking approach that leads to too much frustration.

Maybe someone can get it to work, but how reliably?
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.
May 20, 2020 at 8:14 AM
Possibly with a different heaed bed. I have on of those PCB heaters and it would not get anywhere hot enough to melt solder.
Might have to replace the normal bed with a small size silicone heater that could melt solder.

There are kits available to turn your toaster into a reflow machine with temperature profiles and he like. 
Some of those kits are just about as expensive as a Chineese dedicaed reflow oven.



--
-Skye Sweeney
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.
May 20, 2020 at 8:07 AM
Skye,

I wonder if a build plate with a insulating cover on a 3D Printer will work.

Using G-code you could control the temperature precisely, add a digital thermometer on the pcb to track from the outside. Then modify G-Code accordingly.

Using printer head cooling fan you could also decrease cooling time and perhaps control it fairly well.

Dana


On Tuesday, May 19, 2020, 07:47:43 PM EDT, Skye Sweeney <s...@fll-freak.com> wrote:


Thanks. I guess I need to order a toaster oven and a thermocouple!

-Skye 

On May 19, 2020, at 5:30 PM, Jay Francis (Reactive Technologies) <J...@reactivetechnologies.com> wrote:

My current process (with my “profiled” toaster oven) is:

Put the boards in at room temperature (one, or a dummy board, will have a type k thermocouple kapton taped to the PCB, read temperature with a meter).

Set oven to Convection Bake, max temperature.

Start the oven

When I hit 185C I start a timer
I turn off the oven and open the door when I hit 220C

Carefully/slowly remove the boards (whole rack or tray).  They are still flowing above 175C ish, but you need to cool them down

Set them on top of the oven to cool the rest of the way

Note: It should be about a minute between 185C and 220C. Too fast is bad... too slow is bad... you can get pickier based on the solder/chip profiles.  You should “profile” your oven’s ramp rate from start to 185C as well and may need to make adjustments.  My oven happens to ramp close to recommended for the paste I use (just lucky...).


--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.
May 19, 2020 at 7:47 PM
Thanks. I guess I need to order a toaster oven and a thermocouple!

-Skye 

On May 19, 2020, at 5:30 PM, Jay Francis (Reactive Technologies) <J...@reactivetechnologies.com> wrote:

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.
May 19, 2020 at 5:30 PM
My current process (with my “profiled” toaster oven) is:

Put the boards in at room temperature (one, or a dummy board, will have a type k thermocouple kapton taped to the PCB, read temperature with a meter).

Set oven to Convection Bake, max temperature.

Start the oven

When I hit 185C I start a timer
I turn off the oven and open the door when I hit 220C

Carefully/slowly remove the boards (whole rack or tray).  They are still flowing above 175C ish, but you need to cool them down

Set them on top of the oven to cool the rest of the way

Note: It should be about a minute between 185C and 220C. Too fast is bad... too slow is bad... you can get pickier based on the solder/chip profiles.  You should “profile” your oven’s ramp rate from start to 185C as well and may need to make adjustments.  My oven happens to ramp close to recommended for the paste I use (just lucky...).


--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.


--
-Skye Sweeney

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.


--
-Skye Sweeney


--
-Skye Sweeney

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.


--
-Skye Sweeney

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuilders+unsub...@googlegroups.com.


--
-Skye Sweeney

Skye Sweeney

unread,
May 26, 2020, 9:47:29 AM5/26/20
to nashuarob...@googlegroups.com
I have sent several emails to Invensense and since I am not buying hundred of thousands or millions of these parts, the answers I get are pretty terse.

I have had much better luck with ST and although the Random Walk Error is not as good, I found the documentation to be superb. 
I may just have to live with the RWE and wait for them to release the next generation parts. 
I hope these new parts will include a HW integrator to offload much of the math off my CPU.
Fingers crossed.

To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
May 20, 2020 at 8:14 AM
Possibly with a different heaed bed. I have on of those PCB heaters and it would not get anywhere hot enough to melt solder.
Might have to replace the normal bed with a small size silicone heater that could melt solder.

There are kits available to turn your toaster into a reflow machine with temperature profiles and he like. 
Some of those kits are just about as expensive as a Chineese dedicaed reflow oven.



--
-Skye Sweeney
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
May 20, 2020 at 8:07 AM
Skye,

I wonder if a build plate with a insulating cover on a 3D Printer will work.

Using G-code you could control the temperature precisely, add a digital thermometer on the pcb to track from the outside. Then modify G-Code accordingly.

Using printer head cooling fan you could also decrease cooling time and perhaps control it fairly well.

Dana


On Tuesday, May 19, 2020, 07:47:43 PM EDT, Skye Sweeney <s...@fll-freak.com> wrote:


Thanks. I guess I need to order a toaster oven and a thermocouple!

-Skye 

On May 19, 2020, at 5:30 PM, Jay Francis (Reactive Technologies) <J...@reactivetechnologies.com> wrote:

My current process (with my “profiled” toaster oven) is:

Put the boards in at room temperature (one, or a dummy board, will have a type k thermocouple kapton taped to the PCB, read temperature with a meter).

Set oven to Convection Bake, max temperature.

Start the oven

When I hit 185C I start a timer
I turn off the oven and open the door when I hit 220C

Carefully/slowly remove the boards (whole rack or tray).  They are still flowing above 175C ish, but you need to cool them down

Set them on top of the oven to cool the rest of the way

Note: It should be about a minute between 185C and 220C. Too fast is bad... too slow is bad... you can get pickier based on the solder/chip profiles.  You should “profile” your oven’s ramp rate from start to 185C as well and may need to make adjustments.  My oven happens to ramp close to recommended for the paste I use (just lucky...).


--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
May 19, 2020 at 7:47 PM
Thanks. I guess I need to order a toaster oven and a thermocouple!

-Skye 

On May 19, 2020, at 5:30 PM, Jay Francis (Reactive Technologies) <J...@reactivetechnologies.com> wrote:

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
May 19, 2020 at 5:30 PM
My current process (with my “profiled” toaster oven) is:

Put the boards in at room temperature (one, or a dummy board, will have a type k thermocouple kapton taped to the PCB, read temperature with a meter).

Set oven to Convection Bake, max temperature.

Start the oven

When I hit 185C I start a timer
I turn off the oven and open the door when I hit 220C

Carefully/slowly remove the boards (whole rack or tray).  They are still flowing above 175C ish, but you need to cool them down

Set them on top of the oven to cool the rest of the way

Note: It should be about a minute between 185C and 220C. Too fast is bad... too slow is bad... you can get pickier based on the solder/chip profiles.  You should “profile” your oven’s ramp rate from start to 185C as well and may need to make adjustments.  My oven happens to ramp close to recommended for the paste I use (just lucky...).


--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.


--
-Skye Sweeney

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.


--
-Skye Sweeney


--
-Skye Sweeney

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.


--
-Skye Sweeney

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.


--
-Skye Sweeney

--
You received this message because you are subscribed to the Google Groups "Nashua Robot Builders" group.
To unsubscribe from this group and stop receiving emails from it, send an email to nashuarobotbuil...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/nashuarobotbuilders/2adb0768-0da2-4431-b16b-7c8f75aec9ab%40googlegroups.com.


--
-Skye Sweeney

Skye Sweeney

unread,
Jun 7, 2020, 8:46:27 PM6/7/20
to nashuarob...@googlegroups.com
For closure, I can give an update.

With Jay lending me a thermocouple (while mine arrives from AdaFruit) I was able to characterize my industrial hot plate well enough to create a process to approximate the solder thermal cycle. I bring my hotplate up to what is claimed in 475F and wait for the heat to stabilize. Then I place the circuit board on a thin sheet of aluminum (to even the heat) and drop that on the hot plate on standoffs. The standoffs allow the PCB to come up to an even 150C as requested. Then I remove the standoffs and the temperature ramps up to 200C (needs to be between 183 and 230C) as it reflows. Then 30 seconds later I pull the PCB and the aluminum plate off the hot plate and put in on a metal can. It then cools down at around 1C per second to prevent thermal shocks.

Using this procedure, I soldered the second IMU chip to a breakout board. Unfortunately this chip does not work any better than the first.

So I have spent a few hours checking for mistakes on my part. I found a few but nothing that explains the weirdness.
So then I started double checking the real simple stuff. Like after power on do the register values match the documented defaults?
In fact they do not. Not even close. The differences are in biases to be applied to each axis.
The documentation firmly states these will be zero on reset when in fact they are significant values.
My coding style had me set these registers to zero in my initialization routine (belt and suspenders).
But that blows away some kind of factory offsets for each axis.
If I do not reset those values, the data values start to make sense.

The documentation is just riddled with these types of errors. I already found errors in how they describe the FIFO system including then stating the wrong size for the FIFO. Makes you wonder if they ever proofread the document. Also makes you wonder if any one else is using this part. When I complained that an entire register definition was missing, they assured me it was not. I got the feeling they felt that way because no one else has complained. Grrrr.


[trimmed]

Reply all
Reply to author
Forward
0 new messages