Skip to main content ↑ ↓

Messages

This section allows you to view all Messages made by this member. Note that you can only see Messages made in areas you currently have access to.

Messages - hlipka

4
Tools of the trade / Re: L-Mark LK-320P heat shrink/tube/label printer
Thanks for that tip - it worked like a charm:
[attachment=0]
I did place a small string (about 5 to 10 mm) of glue right where header and wire meet, then just pushed the label over it. No need to rotate or fill out the label, it just needs to stick. When the glue cooled down, I used the hot air station (set to 180°C) to shrink the labels. While doing so the glue starts to melt again slightly, and due to the shrinking label it nicely fills out the inside.
5
General discussion / Free cable build (no PCB this time...)
OK, due to the mishap with sending out the stuff from the free coupons I waited a little bit longer than usual. In addition to the Buspirate probe labels I got a "Make Your Own Cable Kit" sample set. Cables is something one can never have enough on an electronics workbench :)

So this is what I got in my mail:
[attachment=2]
Looks nice. Time to take it out of the bag:
[attachment=1]
Oh, one of the wires is trying to escape! What I was always wondering (since its never stated): is the cable big enough to make use of all crimp housings or do I need to choose? Turns out it has just the right amount of wires.

I like the feel of the cable, the wires feel thicker than the normal ones in a ribbon cable.

Time to mount it together:
[attachment=0]
Biggest difficulty: decide which colors should go together. Unfortunately one cannot choose the order of the colors in the ribbon cable - it would have been cool to have a black or red wire at the end of each cable assembly to see whats pin 1. I managed to get one assembly with red and one with black at least (blue and green are also OK I think).

Thanks again to DP!
6
General discussion / Re: high speed low latency wireless connection between two f
With "shared medium" I mean that you are not alone in the frequency spectrum. It doesn't matter whether someone is used your actual wifi connection - the moment someone else is sending in your frequency range you are jammed. If you don't handle that in accordance to what wifi is doing expect that to happen often. The 802.11* standard define how wifi clients need to share the medium, but it still can mean your throughput suffers. 8And the 5GHz band only buys you some time, it will be crowded too in the future - 802.11ac already uses 160MHz bands)
You can speed up quite a bit by using UDP instead of TCP, but then the connection might be lossy. (But OTOH the TCP guarantee of not losing data means that data is re-send, adding a delay).
Btw: did you already check whether the FPGA supplier of your choice has already a ready-made solution? E.g. Xilinx has a large number of IP blocks available.
7
Tools of the trade / Re: L-Mark LK-320P heat shrink/tube/label printer
So thanks to the Free PCB coupon program I got a set of white labels. The sure look nice too even without the cable:
[attachment=1]
(Ok, one of the "3.3V" prints looks wrong, but apart from that its really nice)
My plan was to use them on my other two Buspirate probe cables. I have one with IC hooks attached and a Sparkfun cable (with female headers). It turns out that it doesn't work:
[attachment=0]
I don't think the labels will shrink that much, so I'm stuck. No "free label build" this time :(

Any ideas? I might use real headers instead. Or trying to stuck something inbetween the cable and the label. That might also work with the IC hook cable then.
8
General discussion / Re: high speed low latency wireless connection between two f
Don't do that. Every standard wireless system out there that can handle such data rates (802.11n and 802.11ac come to mind) will have latencies larger than what you are looking for. In fact, even Gigabit Ethernet comes only with 0.01ms for the pure transmission.
And don't forget that any wireless transmission uses a shared medium, so you need to be prepared to a sudden connection loss just because someone near you is downloading a movie.
You can probably roll your own schema, but transmitting 100Mbps on a (lets say) 20MHz frequency band is not exactly trivial (and even with a 100MHz band its difficult).
Use a cable (heck, without GbE even then its difficult :).
9
Project development, ideas, and suggestions / Re: Need help from analog gurus. 78L08 output at -0,9V
Two things come to my mind: first, can you test the power supply without anything else connected to it? That could rule out other parts of the circuit as the reason. Second, please post a schematic of what you did (and verify that your PCB actually matches the schematic). And maybe third: the 8xx regulators don't need capacitors, can you test without them?
11
Bus Pirate Support / Re: Buspirate doesnt work kb50 chip super hot help!
The KB50 is the 5V voltage regulator (a MIC5205-50). When its getting hot, something is either drawing to much current (as noted already) or something else is wrong. Aside from shorts, check the capacitors - maybe one of them is shorted or gone dead in other ways (e.g. if the ESR is too high the LDO might oscillate).
12
General discussion / Re: Precise analog voltage from PWM
Actually you should use a RC-Filter. Calculate its cut-off frequency to be high enough to follow the voltages changes you need, but low enough to well under your PWM frequency (by a factor of at least about 10). Then the capacitor can be much smaller than the 330µF you have now.
I suspect what you are seeing here are the effects of different internal reistances of your PIC GPIO pins - it works in push-pull-configuration, and there charging and discharging will probably work at different currents.
13
Tools of the trade / Re: L-Mark LK-320P heat shrink/tube/label printer
Got my cable from the giveaway today. They surely look nice - thanks to DP!

[attachment=1]
The labels feel rather stiff, this makes the pins much easier to hanel in my opinion (better grip on them). Unfortunately I didn't remember that I have the Sparkfun version of the BP - so the header is reversed and none of the label match :( SO I will try to carefully open the connector and reverse the cable.
The pins itself look cheap, OTOH. They look like the got just cutted out of a large sheet of metal:

[attachment=0]
(Compare with normal header pins as reference) I'm not sure how much of a problem that is, but it surely looks strange.
14
CPLD programmable logic / Re: Review - new NumatoLab FPGA boards
I can suggest looking at the books from FreeRangeFactory. There are several eBooks taking on topics such as digital design in general, using VHDL for PFGA programming and one about using the Xula boards (its for the Spartan3A boards, but its the most hands-on guide).
15
Breakout boards / Free PCB build - PICkit3 clone on QFP prototyping board
I finished this a while ago, and promised to write about it...

I got the 0.8+0.5mm xQFP protoboard from the Free PCB drawer. Since I wanted to get a PICkit3 for a while, this looked like a perfect fit. So I looked up the schematics, ordered what I needed, and started work.

[attachment=0]

It took longer than expected, and in the end I had just one silly mistake costing me about two months looking for it (just two swapped wires:( I did write a complete description of the while PICkit3 clone project, including more images and the complete schematic I used in the end.

So, what about the protoboard? After all, this was not just a prototype but a real, complete project.

I like that it has a large number of different footprints and connection options available.
But if they are not used, they are just in the way. I did not use the SSOP footprint,
and it takes much room. The footprint for the crystal was on the wrong side of
the QFP footprint for me (one needs to be lucky to have it at the right place).

Having the footprints for the push buttons in the middle of the board means
they are most likely covered by wires. So maybe they should go further to the edges.
This would also move the prototyping area more closely to the QFP footprint.
Maybe it should go to the center, with all the other stuff around it. There
might also be space to save by not using through hole connections for all parts. Maybe
just using SMD pads for connecting wires to the crystal can reduce the overall
footprint in half, without sacrificing usability.

One thing constantly ticking me off was that the numbering on the QFP footprint
differs by side. Since the MCU comes in 0.5 mm pitch (as do most QFP parts I found)
it goes to the bottom side of the PCB, whereas anything else goes to top. But there
the layout is reversed, so Iavoided looking at the printed QFP pin numbers and
always needed to count in reverse. Thats really annoying. Next time I would rather
use the single-footprint version of the board just for that reason (even if it
comes without USB footprint).

I liked the availability of a 3.3 V power rail in all the different places around the board.
But since the PICkit uses different voltages (3.3V in some places, and a switched
voltage rail called V<sub>DDint</sub> in other places) I could have used a second voltage rail.
I suppose the same might be true for other projects, needing 5 V and 3.3 V.

( ! ) Fatal error: Uncaught exception 'Elk_Exception' with message 'Please try again. If you come back to this error screen, report the error to an administrator.' in /var/www/dangerousprototypes/forum/sources/database/Db-mysql.class.php on line 696
( ! ) Elk_Exception: Please try again. If you come back to this error screen, report the error to an administrator. in /var/www/dangerousprototypes/forum/sources/database/Db-mysql.class.php on line 696
Call Stack
#TimeMemoryFunctionLocation
10.01572446552session_write_close ( )...(null):0
20.01602578168ElkArte\sources\subs\SessionHandler\DatabaseHandler->write( )...(null):0
30.01602578944Database_MySQL->query( ).../DatabaseHandler.php:119
40.05962717704Database_MySQL->error( ).../Db-mysql.class.php:273