Teensy 4.0 Breakout Kit

@MM - instructions will be handy - there was a recent note about those close pins to battery holder needing trimmed/attended to in some fashion.

Interesting - you guys got one already - USPS says something to be here 12/12, tomorrow from MA.

...
But right now that's academic because disaster hit!
...

That is very unfortunate DerekR :) I lost One T4 - but that was having the IDE running at 960 MHz OC for a T4 with heat sink - and plugging in a fresh T4 without one … and sending a sketch not being aware of that SPEED setting. It was i2c ssd1306 display so it sat powered up and reprogrammed as I opened the driver and read code for settings and reprogrammed a time or two - all the while it was running hot. That one happened to be a Purple OSH T4.

As far as long term use and T4 stability - I have on my desk the three 1062 Beta units PJRC evolved - all still functional since Feb to June? The first in Jan 2019 was the 1052 - and I put that away working. So the three here I use and trust and have left running for days or weeks one way or another over 6+ months show no troubles. Then I have original TallDog Beta , the Tindie Ult Breakout ( used the least ), another T4 on a PJRC Purple display test board , one on a FRDM4236 breakout - so just 7 of them in my sample set - but all have been working fine since two of those came from first production run before they were released. They have been through many hours of USB Host testing, various Display testing.

I worked them hard in Beta and only issue was solved during beta where external power on pins was preventing proper shutdown - 7 Rx and Tx lines of 3.3V during testing, and that was a one wire change Paul made with no issues other than that.

Not sure what it could be there giving bad behavior … T4 Is allergic to 5V, and running overclocked can cause heat damage. Have you tried the 15 second Restore? Holding Program button 15 secs watching RED LED for it to flash then release and it will restore base level settings within the MCU - I have had to do that at times it seemed when displays or other things in development during Beta could set it odd.
 
Well, I remembered that I'd seen T4's at the local electronics store (You-Do-It Electronics in Needham, MA), so I jumped in the car and got the last one! So back in business tomorrow morning.

A few more comments on the build:
1) There was absolutely no issue with soldering around the battery holder. Just watch the direction of the iron as you solder.
2) The SD socket was much easier to solder than beta 1, as was the flex-cable socket (even though it can't be used at the moment.)
3) The flex-cable was MUCH easier to solder (with no damage as far as I can see).
4) The socket kit is great. My only comment is that I have some low-profile female headers that allow the male pins to be inserted more deeply than the ones in the kit. (Not sure where I got them, either Adafruit or You-Do-It).
5) My construction order was a) SMT devices on both sides of the board, b) switches, etc, c) flex-cable, USB & SDA sockets, d) battery socket. Then the headers....
6) Look closely at how you will solder the pins on the sockets to the pcb, and which one you will solder first. Use the supplied solder pads to solder the metal enclosures of the sockets to the pcb (for strain relief).

I'll post photos tomorrow after I've built up the new T4, and "blinked" all the pins.

Derek
 
is there someway to test the breakout with power before mating with the T4? seems that if the assumption is the power of the T4 is somehow inadequate [which i doubt] then using a generic power supply to test the breakout for shorts prior to mating might help. i'm not an electrical engineer so i can't offer a test circuit but just seems to me there should be a way to apply power to the breakout just to test for any shorts.
 
I just got the kit from Loglow/talldog. Thanks.

You're welcome!

I suspect I may need to wait until the detailed instructions are done.

They're in-progress. I'm experimenting with Dozuki's free offering (think iFixit) for step-by-step instructions with photos.

I was playing around with the PCB, and I dropped the coin cell battery holder. It took me some time to find it as it rolled to the furthermost corner under the desk at work and it blends in with the industrial carpet at work, and I will need some longer grabbers to get it.

This happens to me all the time. And every time I think "I'll be more careful, so this never happens again" and then it always happens again.

But anyway, before dropping it, I noticed that was a rather tight fit between the battery holder and pins 21-22 and pins 3-4 that might be problematical if you solder the pins in first and then solder the battery holder on. I suspect the best way to attach is to put the female socket pins through the PCB, but don't solder them (or at least pins 3, 4, 21, and 22). Then attach the coin cell holder and do surface mount soldering. Finally solder in pins 3, 4, 21, and 22, taking care not to melt the battery holder.

I think it's best to do all the pins first and do the battery holder last. After soldering, I use diagonal cutters to trim down pins 2-5 and 20-23 so they're as flush as possible. Then I reflow those 8 pins with the iron. Then, I install the battery holder. Seems to work well so far.

My kit had a small coin cell battery in it. As you transition towards selling it on tindie.com, I wonder whether you will need to have different versions one one has the battery in it (that can be mailed easily with ground transportation to the lower 48 states), and one without a battery for outside the USA or to Alaska/Hawaii (and have the people source a battery locally). Now that battery is so small, I doubt it would be an issue, but regulations are regulations. Or possibly just not include the battery, and make evrybody source the CR1220 themselves.

This is good to know. I'll consider what to do about this. I wonder if there's a lower-limit to the amount of lithium? Or if the restrictions are for lithium-ion but not plain old lithium? I'll have to do more research about this, but thanks for the heads-up.

I'm wondering whether the U1 and U2 SOIC mounts might be better to list the actual chip rather the U1/U2, to allow somebody to make sure they are soldering the right part.

That's a good thought, but there's not much room. I'll see what I can do.

There are 2x2 male and female headers, along with one 14-pin male/female header and one 12-pin male/female header. I assume the 2x2 headers are for VIN, Ground, Not used, and VUSB. If so, we probably need to pull out the pin left of the VIN pin. Is this correct?

Yup, you got it. The silkscreen lines on the top side of the breakout give a hint about the suggested header arrangement.

At present, I have no plans to use CAN, but I might want to use the CAN pins. Should I just not solder in the MCP2558FD?

Yeah, you can just leave off U1, R1, C2, and C3.

The socket kit doesn't have any parts for bringing out the micro-SD pins. I assume this is by design? How would you use the socket kit?

Correct. You would still solder the flex cable onto the underside of the Teensy. Everything else (including the other side of the flex) uses sockets.

I would imagine there are a lot of people like me that aren't as comfortable doing surface mount soldering that having a board with the SMT parts soldered would be useful.

I've already quoted this out, and it's not prohibitive. I'm even considering offering all kits like this. It'd be a little annoying to have options for CAN (or not) and perhaps the extra pullups (or not), but it might still work out. I'm interested in your thoughts about this.
 
I've assembled the new board - thanks very much for including me. It went very well, much smoother than the first Beta - took about 3.5 hours with quite a bit of time spent working out a sensible construction order. BUT - I hit a snag at the last step, ie as I was mating the T4 to the breakout. I found it impossible to insert the flex cable into its socket! I used the header kit, and the 2x2 female socket on the breakout prevents the flex cable from entering the socket properly, and although I struggled with it for quite a while I was not able to close the latch. I think the answer is to take out the 2x2 female socket and remove the unused corner (with a Dremel?)

Thanks for the report, and I'm glad to hear that it mostly went well. I'm very curious to see photos of the socket issue, because I just assembled one with the socket kit and it went pretty smoothly. Now I'm going to go back and re-examine it closely. I know the exact spot you're talking about though.

No need for a Dremel for those pin headers. A sharp utility blade is all you need.

But right now that's academic because disaster hit! I had assembled a brand new T4 to populate the breakout. It was working fine just on its own, dangling on a USB cable. I checked the breakout thoroughly for shorts - all ok - plugged in the T4 and,,, nothing. The Teensy is completely gonzo :( !!! There is no longer any 3.3v (yet 3.3v to ground is open on the breakout.). This is the fourth or fifth T4 that I have lost due to random failure of the 3.3v supply!!! (I leave my projects running 24/7 and the failures have all happened when I am nowhere near them, usually after a few weeks of running!!!), and all are running behind surge protectors. Makes me wonder if the T4 is truly ready for prime-time. I don't have any more spare T4's, so I'm now out of business for several days at least. Grrrr...

Oh no! I'd love to know what happened. I just burned out a T3.6 myself, and I'm dying to know exactly what fried it.

I'll send a photo of the flex-cable interference if that will help.

Yes, please do.
 
Well, I remembered that I'd seen T4's at the local electronics store (You-Do-It Electronics in Needham, MA), so I jumped in the car and got the last one! So back in business tomorrow morning.

Oh, you're lucky. We had one RadioShack here in western MA for a while. Then it closed and re-opened as BRW Electronics (I think that's what it was called?) which was only open for a year or two before closing. I don't even know where to buy regular solder locally if I run out.

1) There was absolutely no issue with soldering around the battery holder. Just watch the direction of the iron as you solder.

Did you do holder-before-pins, or pins-before-holder?

2) The SD socket was much easier to solder than beta 1, as was the flex-cable socket (even though it can't be used at the moment.)

Great! I had the same experience. The new SD sockets are nice. They're brand-name, so a bit expensive, but that's ok. The generic flex sockets are actually quite a bit easier to solder than the similar brand-name ones were, so generic it is.

3) The flex-cable was MUCH easier to solder (with no damage as far as I can see).

Good. I think lining it up with tape helped me a lot. The new flex cables are the correct thickness (0.33 mm) for the flex sockets too, so they shouldn't pull out or slide around like the first ones did.

4) The socket kit is great. My only comment is that I have some low-profile female headers that allow the male pins to be inserted more deeply than the ones in the kit. (Not sure where I got them, either Adafruit or You-Do-It).

The slightly-non-standard-ness of headers is pretty annoying. I don't use generic ones in the socket kit for this very reason, they're all high-priced Samtec parts. Did the different headers cause any trouble?

5) My construction order was a) SMT devices on both sides of the board, b) switches, etc, c) flex-cable, USB & SDA sockets, d) battery socket. Then the headers....

That sounds about the same as me. For the socket headers, I assemble the male ones (and the flex cable) on the Teensy, then mate all the female sockets onto the bottom of the Teensy. Then I solder the flex connector onto the breakout. Then I insert the other end of flex cable (dragging the Teensy behind it) into the flex connector, lock it, and finally I position the Teensy with all the headers attached through the breakout holes and solder them all up.

6) Look closely at how you will solder the pins on the sockets to the pcb, and which one you will solder first. Use the supplied solder pads to solder the metal enclosures of the sockets to the pcb (for strain relief).

Yes, the SD socket has a bunch of enclosure solder points, and the flex socket has two slightly hidden ones.
 
is there someway to test the breakout with power before mating with the T4? seems that if the assumption is the power of the T4 is somehow inadequate [which i doubt] then using a generic power supply to test the breakout for shorts prior to mating might help. i'm not an electrical engineer so i can't offer a test circuit but just seems to me there should be a way to apply power to the breakout just to test for any shorts.

For what it's worth, the PCBs do (supposedly) go through electrical testing via automated flying probe machines at the board house, in order to weed out any defective ones before they're shipped.
 
For what it's worth, the PCBs do (supposedly) go through electrical testing via automated flying probe machines at the board house, in order to weed out any defective ones before they're shipped.

ok. but how about after the assembly? i was thinking more along the lines of trying to verify if the assembly process introduced any shorts or other problems.
 
MichaelMeissner said:
I was playing around with the PCB, and I dropped the coin cell battery holder. It took me some time to find it as it rolled to the furthermost corner under the desk at work and it blends in with the industrial carpet at work, and I will need some longer grabbers to get it.

This happens to me all the time. And every time I think "I'll be more careful, so this never happens again" and then it always happens again.

Yeah, unfortunately it is far enough under the steel case wall that I can't get to it, and I have problems getting back up if I crawl under the desk. So I ordered a replacement. And they come in a pack of 15, but fortunately it wasn't that much. I looked at the reviews on amazon.com, and there was a review on one of these, by a feller named Michael Meissner. Yeah, I bought a bunch in March, but I didn't remember buying them, and I also didn't put them in the drawer labeled 'coin cell battery and holders', so I need to buy them again (and likely find the originals just after picking up the package). FWIW, when I bought them originally, I had plans of gluing the holder to the sides of the stacking header, but I discovered that the CR1220 holder was still bigger than I wanted, so I put them back into the someday bin for future projects.

But as I look at it, main those surface mount components are small. I've just done my first surface mount solder (flash memory chip in a SOIC package), but these resistors and capacitors are so much smaller. I'm wondering if I will be able to see them using magnification and to use my iron which only has a medium sized tip.

And I've already dropped two of the smaller components as well (hopefully it was only 2 and not 3). Fortunately these were on a hardwood floor.

I also wonder about trying to identify which is the resistor and capacitor (and if you have different variants, which is which), since the package is so small. I suspect when you work with them all of the time, you come to know them by sight, but not everybody is at that level.

In terms of pre-soldered boards, I tend to think you don't want to have too many variations in what you stock, so you want to solder both the CAN and USB host chips in, and have a solder jumper that is either cut or joined to enable them. IIRC, one of them has to be attached before you put the other components in, so you have to make the decision ahead of time.

For the resistor pads, I suspect the easiest is just leave the pads as is, and if a person can't do the surface mount soldering, they can always attach the through hole resistors like they've done before. Alternatively have the resistors on the board and have a solder jumper to connect them. This likely means you have to have fewer resistors.

It would be helpful if you could post a few pictures showing the fully built board, and another with the Teensy in the socket connections. Yeah, having step by step video would be great, but at least having a picture would allow me to figure out the orientation and such.
 
Received, much thanks.

Ordering extra T4 boards Friday. Will likely wait until detailed instructions are available so as to test those as well.

I do have a usb microscope.
 
just got my kits as well. and I like what I'm seeing so far. very nicely done! maybe tomorrow I'll get time to work on one and get to testing.. I work a lot now. so not as much free time :)
 
@loglow -

Dan - I hope I'm wrong but I suspect there's a mismatch in the order of the traces on the flex-cable and the pcb trace to pin 36 at the flex cable socket :(

This morning I got the replacement T4 constructed and working on the breakout. I also shaved down the 2x2 socket with an Xacto knife and the cable slipped into its socket right away. The Teensy worked immediately without any problem. That's all good!

So my first test was to run "blink" on all pins. They all worked EXCEPT pin 36. So of course I thought it was something I had done and set about trying to track it down. Everything seemed to be ok: I resoldered all connector pins, checked continuity on the flex cable, looked at all connections with a high power magnifier - all good. (However, in the process I accidentally tore the pads off the flex cable, luckily I had the spare...)

Then I noticed that there is apparently a hardware problem: a mismatch (at the flex-cable socket) between the pin 36 signal from the underside of the T4 and the pcb trace to the breakout's pin 36. The two are off by one pin at the socket and as a result there is no connection from the T4 to pin 36 on the breakout and that is why "blink doesn't work. I've checked with an ohmeter and by following the pcb traces.

Sorry to be the bearer of bad news, please tell me I'm completely wrong, and it's all ok!!!
Derek
 
Last edited:
Kits Look Awesome - Based on Derek's note looking before soldering ….

Using DVM to ring out the Flex Cable:: CLK comes out on CMD? - then staring at it ...

The PCB on connect end shows CLK between D+ and D- but on the cable the D- and D+ are routed straight side by side - there is no via and crossing to put CLK to the middle to match the PCB silkscreen - that leaves CLK on the other side - where CMD is shown?

That puts CMD up on 3rd pin down labelled D-?

@loglow -

Dan - I hope I'm wrong but I suspect there's a mismatch in the order of the traces on the flex-cable and the pcb trace to pin 36 at the flex cable socket :(

...

Sorry to be the bearer of bad news, please tell me I'm completely wrong, and it's all ok!!!
Derek
 
BA0F03FB-F5AF-4464-8CCE-97B56E4B1A4B.png

Yeah, I'm seeing exactly what Defragster is describing. Looks like the error may only be on the flex cable?
 
...

Yeah, I'm seeing exactly what Defragster is describing. Looks like the error may only be on the flex cable?

Happy :) / Sad :( :: Glad I read the lines right - so small. If @Loglow can confirm and correct easily it seems all in that cable.

I started on the PCB and #36 does connect the two DAT pads and likewise the pin #37 does go to the two CMD pads. So hopefully easier to get the FlexCable confirmed and redone and shipped as needed.
 
Mine arrived to our PMB in town today. So far it looks like I should wait to do anything with it, until this is resolved.
 
I also am seeing what defragster and etmoody3 see on the flex cable..
Code:
teensy side       pcb side 
dat1         ->   dat1
dat0         ->   dat0 
clk          ->   d-       this 
d-           ->   clk       and this signal need to be swapped..
d+           ->   d+
cmd          ->   cmd
dat3         ->   dat3
dat2         ->   dat2

will look at the pcb later and ring it out with a voltmeter to make sure the connections to the sd card seem correct. looks like a simple flex pcb correction would fix it, or a adjustment to the pcb itself I'm only seeing 2 signals swapped..
 
I also am seeing what defragster and etmoody3 see on the flex cable..
...
will look at the pcb later and ring it out with a voltmeter to make sure the connections to the sd card seem correct. looks like a simple flex pcb correction would fix it, or a adjustment to the pcb itself I'm only seeing 2 signals swapped..

Good info and plan - I stopped short of further testing - hoping the fix will be evident to @loglow - but now would be the time to ring it all out. Never did the same for the initial beta board - but there USB Host D+/D- were good and have been tested to work where SDIO didn't after soldering. Indeed usable USB Host would not be the case with this cable with those lines out of place.

I just got 3 fresh T4's at $18 OSH pricing from an Adafruit 'black' sale - but don't want to render them less than ideal with 7 others already on my desk. And two others I haven't yet made time to solder to FRDM boards already set aside.
 
okay so I just checked the pcb, and I cant see anything wrong with it, no shorts to any other sd card pins. so fixing the flex would fix the sd card I do believe. I might see if I can patch the pcb to fix the issue on one of these, waiting on 2 fresh t4's from osh park first though.
it looks fixable on the bottom of the pcb. dont think it could be fixed on the flex pcb though. if I dont solder the d- or d+ I can run a jumper to clk on the teensy.. hmmm...
 
Morning all
Mine arrived yesterday late in the day. As @KurtE said - think going to wait on the assembly until the issue is resolved.
 
Yeah... so... this is why I need to stop ordering PCBs at 4 AM...

Now that I see it, I know exactly why I made this mistake too!

I think I'll keep the flex cable as-is and fix this on the breakout PCB, like this:

Teensy_4_0_Breakout_T.png

Thanks guys. I'll send each of you a few new boards next week.

Also, if you need any additional components, just send me a note.
 
Glad you can see a way out - and have a note to avoid in future - good/painful lesson.

I saw a 5V PIC based KS with 32x32 RGB panel that I gave input to make Teensy capable - the maker added a spot for it and routed to spec and made sure the 4 pots on it went to 3.3V and not 5V when running Teensy - after I discovered on the demo unit he sent from Australia . He did the work - and one last PCB edit and the board rerouted and fixed 5V went back to the pots, so to use the board I had to cut traces and run wires to feed the pots - not fun or pretty for using a T_3.6 - or even a 5V aware T_3.x where the pots peaks at 3.3V :) I hacked one of the two boards I got and then left them alone :(
Glad we had Derek's heads up trouble report to get a deeper look before any more parts got soldered away.

Odd - I thought from reading here the Flex cable would have been the easier/cheaper part to remake (and ship out)- though that would have crossed the CMD over the USB line - or maybe had other issues.

At a glance it looks like the new battery holder has similar tab spacing to the initial side mount battery holder to make it interchangeable?
 
Back
Top