You must use an external 1k5 pullup or the internal bootloader will not work on USB when you select to start it with BOOT0/1. There is no possible workaround in this startup mode, which is the only startup that works no matter what the content of the flash rom looks like. Believing that your bootloader will never be overwritten is just calling for trouble that can not be solved with an USB cable.
Community rework of standard maple and maple mini.
(181 posts) (14 voices)-
gbulmer
Moderatorala42 - That's very helpful. Thank you for the information.
So pulling the 'USB_DISC' pad to ground would work on the more recent boards (Orone-mini-S8D and Orone-mini-S8H).
Did you make any circuit changes to make it easy to do?
I assume, it would be better to pull the 'DISC' pad down via a resistor, say 1K, and not direct to ground to protect the STM32F's USB_DISC pin. That could be done by external circuitry.
Posted 2 years ago # -
On the aeroquad32 board I took a 220Ohm resistor for the pulldown, as i did not wanted to pull it down directly. I made also made tries with 330 and 470 Ohm IIRC, but wanted to be on the safe side, which means lower resistor value. A 1k pulldown to a CPU pin with a 1k5 pullup to 3.3V is a bad idea to pull a line down.
Using a CPU pin to pulldown the DP+ line without using an inverting transistor inverts the logic used in the libmaple boot loader and USB code, so they have to be modified, which does not hurt so much when using a F3 CPU.Posted 2 years ago # -
@ala42, thanks for the info. I agree, a hardware-only solution to get into the built-in DfuSe bootloader is fail-safe and thus preferable.
@gbulmer, sure, the easy access to the DISC trace like on the S8C/S8H does the trick. An adjacent ground hole would be nice indeed, there seems to be enough space for that. So one could use a 2-pin male header + female jumper, or 2-pin female header + ala42's 220 Ohm resistor as male jumper without the need of any external circuitry.
@gbulmer, whether/how much you have to alter the mask for laser-cutting depends on the mylar thickness and the cutter settings. I scale the pad areas to 70-80% for a 0.125mm mylar sheet thickness to get the right amount of solder paste onto the pads. Additionally I add an offset (important for the smaller pads mainly) which accounts for the lasers line-cut thickness. The latter depends on the lasers settings, e.g. combination of laser intensity and speed (I've not yet found ideal settings for the 0.5mm pads, needs some more experimentation...)
Posted 2 years ago # -
gbulmer
Moderatorala42 - I'll try to be clear.
The Orone-mini circuit for the USB 1.5K pull-up resistor (R-up) uses an inverting 'digital transistor' (it has on-board 10K resistors, saving space) to maintain compatibility with the Maple-mini bootloader. When the USB_DISC pin is low, R-up is 'on', pulling-up D+.
So the STM32F's 'USB_DISC' signal is driving the base of the transistor, and is not directly pulling up R-up. Hence the USB_DISC pin is only pulling down a 10K resistor, and the transistor is doing the work to pull up R-up. So a 1K pull-down resistor pulling the USB_DISC signal to ground might be sufficient, though using a smaller resistor would help.
The 'DISC' thru-hole pad (which gives access to the USB_DISC signal), which is available on Orone-mini S8D and S8H, could be shorted to ground to switch R-up on, and hence make it easier to use the manufactured-in STM32F303 bootloader.
I am asking for feedback on small modifications to Orone-mini:
1. Put a ground pad, on 0.1" header pin grid, next to the 'DISC' pad. This should make it even easier to switch on R-up by shorting the 'DISC' pad to ground, for example with a jumper on two header pins, a switch, or a bit of wire.
2. Also add a 220R-1K resistor in series with USB_DISC signal and the 'DISC' header pad. The purpose is to protect (current limit) the STM32F pin if the DISC header is shorted to ground by the user (maybe using a switch or jumper), and software accidentally drives the pin high. This seems like a "nice to have", so I will probably not do it as space is so tight.Posted 2 years ago # -
gbulmer
Moderatorsoycamo - a solder stencil would be lovely. Thank you for the offer. However, I live in the UK, so the postage may be quite expensive.
I can get use of a laser cutter, so tips on how to make a solder paste stencils using a laser cutter would be even more valuable!The experiments Chris did were getting close for larger pitch, but I don't think it's 'cracked' for anything as fine as the 0.5mm pitch.
I'd like to 'finish' the PCB enough to make it easy to use the STM32F303 too. Then I will be very interested in a solder stencil. The only outstanding idea that I am considering is modifying the PCB to make it easier to use the manufactured-in USB bootloader. I think that should be done by Monday, at the latest.
Posted 2 years ago # -
gbulmer
Moderatorventosus - okay, I'll try to add the ground pad on the 0.1" grid.
"I scale the pad areas to 70-80% for a 0.125mm mylar sheet thickness to get the right amount of solder paste onto the pads"
So the hole in the template is smaller than the copper pad? I don't think we've experimented with this enough. I'll talk to Chris on Monday. Maybe this is an area to explore."Additionally I add an offset (important for the smaller pads mainly) which accounts for the lasers line-cut thickness."
Okay, I am pretty sure we've diligently done that.Lots of hacker spaces, colleges and schools have laser cutters. So if we can identify the right techniques to make laser-cut solder stencil, then I'd hope that a lot of folks could successfully assemble PCBs using SMD parts and 'toaster oven' technology. These are the sort of folks the 'large part' Orone-mini is intended for.
Posted 2 years ago # -
@gbulmer, my foils are on the thicker side, thus the scaling, to not get too much paste on the pads.
This method looks really promising: http://www.spikenzielabs.com/blog/?p=669
They use the raster-edging mode to do the fine-pitch masks instead of the vector-cutting mode to reach a higher resolution, I'll give this a try, too.Posted 2 years ago # -
siy
Member@gbulmer, there is a way to create stencils from aluminium cans (a lot of beverages are sold in such cans). It requires established dry film photoresist process and there is a little trick to make foil cut from can flat, but otherwise process is very simple and straighforward. And quite fast, because even in ferric chloride etching aliminium takes much less time than copper. Thickness of the resulting stencil is about 0.15mm which is good enough for most cases.
Posted 2 years ago # -
@gbulmer @sly the laser cutter I've used works as a printer, so I've seen good results by setting it to "print" at 1000dpi. You can use old x-ray sheets from dentists' offices. My friend has used parchment paper (!) with okay results. In Eagle you just use the "paste" layer and shrink the pads. I have done this with Photoshop but you can also do it using the design rule check thing.
Posted 2 years ago # -
gbulmer
Moderatorsiy - Making the solder paste stencil using etching is a good suggestion. However, I am concerned on a few areas:
1. I would prefer to use a technique that is easy for hardware beginners to use without my involvement. I actually think more people have access to laser cutters than to etching equipment (in the UK). I think folks often use pre-coated photoresist which doesn't need developer. AFAIK, dry film photoresist uses a developer. Is that correct?
2. Might using cans as raw material a bit too challenging for schools? I would be worried that folks would cut themselves. Is there a good way to make handling the can easy and safe?Using thin aluminium sheet with a nice rounded edge might be okay.
Posted 2 years ago # -
gbulmer
Moderatorsoycamo - "the laser cutter I've used works as a printer, so I've seen good results by setting it to "print" at 1000dpi. You can use old x-ray sheets from dentists' offices. "
I am not clear what "print" might mean.
Are you using the laser cutter to cut through the material (I'd call that "cutting"), or just exposing "x-ray sheets" without cutting (I'd call that "printing")?
If you're only exposing it without cutting, how do you make a stencil with holes?I can use Eagle to generate the stencil, as Gerber's, png, pdf, and dxf.
The DRC->Masks dialogue is designed to set the masks parameters (solder resist and via tenting, and solder paste stencil) for direct generation of the appropriate layers, and I do not know a reason to use photoshop.It is a while ago since Chris did the experiments, so I may be wrong.
I think we left the DRC->Masks->Cream at their default 0/0/0 values.
Making the stencil apertures 70% smaller would be a good thing because it helps retain the 'web' of stencil material between those laser cut holes.Posted 2 years ago # -
@gbulmer, most laser cutters have two modes: (1) vector cutting and (2) raster engraving.
In cutting mode, the laser cutter acts as a plotter. It takes a vector graphic as input and the laser head moves along those lines, circles, etc, like a plotter.
In engraving mode however, the laser cutter acts as a printer. It takes a raster graphic as input and engraves your image line-by-line with a given resolution onto your medium. If the medium is thin enough, you'll just engrave through it...
I've been in my FabLab yesterday and tried out the method mentioned in my previous post, and the results were superb. For the bigger apertures, it's fine to use the vector cutting mode (as its much faster). For smaller apertures (e.g. 0.5mm) you want to use the raster engraving mode at the highest resolution.
These are the settings I've found to be ideal for an Epilog Legend (60W) for 0.12mm mylar stencils:
Raster engraving for the 0.5mm pitch apertures:
- Resolution 1200 DPI
- Speed 30%
- Power 18%Posted 2 years ago # -
gbulmer
Moderatorventosus - "... In engraving mode however, the laser cutter acts as a printer. ..."
Okay, if that is what "printing" meant, I understand.
I don't agree with the nomenclature, though. I happily call the two effects cutting/plotting or engraving, or even raster drawing.I did wonder if the old dental x-ray film was being exposed by the laser, and then used in some sneaky way.
We tried using the laser cutter with filled rasters, but found using it as a vector cutter worked better, but still not 'perfect'. So your results are very helpful.
Thank you for the values. I'll try to give them a try next week. It would be great to find a way to make low-cost solder paste stencils.
I created a DXF for the 'cream' layer, and there are holes in it for the 'lugs' for the Molex USB-mini socket. That is a defect. So I need to fix my library part.
Posted 2 years ago #
Reply »
You must log in to post.