Alle Beiträge von excogitation

Auspuffkrümmer (Grauguß) schweißen

Autos sind Prozesse. Leider.
Heute habe ich angefangen mich dem Riss im Abgaskrümmer zu widmen.

Weil ich wissen will ob ich es schaffe Guß zu schweißen und weil ich es nicht einsehe 300€ für den neuen Krümmer zu blechen
und mich dann mit der Einstellung des Wastegates (1/10mm) herumzuschlagen, bzw noch weniger Lust habe  650€ für die Turbo-Krümmer-Kombi hinzublättern.

Wie ich schon letztens unter dem Auto festgestellt habe gab es noch eine zweite Stelle aus der der Kleine gerußt hat – am Flansch zur Abgasrückführung hat eine Schraube gefehlt.

Es ist von unten aber ums verrecken nicht möglich diese Schraube einzudrehen.

Das kommt nicht annähernd so schlimm rüber _wie_ verrußt der Smart wirklich ist.

Hier der Riss … liegt natürlich super zum schweissen … not.

Um dem Verzug beim schweissen entgegenzuwirken habe ich eine 10mm
Eisenplatte entsprechend gesägt und gebohrt (15min):

Beide Rissenden habe ich angebohrt, um die weitere Rissausbreitung zu hemmen.

Dann den Riss aufgedrehmelt um dort später die Schweißnaht einzubringen.

Die NiFe 55 Elektroden habe ich vor ein paar Wochen in Ebay besorgt.

Der weitere Plan sieht jetzt so aus:

Mit Gasbrenner den Krümmer auf Temperatur bringen (wider dem Schweißverzug, Rissbildung),
dann mit dem TIG Schweißgerät (mit dem Elektrodenschweißen bin ich leider noch nicht
warm geworden) und den NiFe55-Kern als Schweißzusatz die Nut füllen.


Na gut, das könnte bestimmt auch schöner aussehen … ob’s hält wird die nächste Zeit zeigen.
Einstellungen: ca 200 Amp Gleichstrom, >20l/min Argon



Den Mutternsitz habe ich ein wenig freigeschliffen (Akkuschrauber + Schleifstein).


Nachdem eine der Entlastungsbohrungen nicht vollständig durchgeschweisst war (2mm fehlten) habe ich kurzerhand von innen nochmal geschweisst.


Die Schweißzusatzzuführung gestaltet sich nicht ganz einfach:

Aber das Ergebnis kann sich sehen lassen.

Wie man an den Bildern oben sieht habe ich versucht den Flansch abzuschrauben (eine neue Dichtung dafür war auch im Dichtsatz dabei), allerdings konnte ich die Inbusschrauben (die im verrosteten Zustand immer streiken) trotz einsägen eines Schlitzes auch nur ansatzweise zum Drehen bewegen. Also besser als 3 Schrauben ausbohren, Gewinde nachschneiden – es ist dicht, also darf es so bleiben.

Was mich viel mehr irritiert ist dass der Flansch für den Endtopf sich kurz nach dem herausdrehen der Befestigungsschraube verschoben hat (vor dem Schweißen).

Das zeigt sich auch an der Schraube:

Lässt sich aber beheben:

^^ soll man nicht machen – ist aber „nur“ GG und geht wohl wenn man entsprechend gefühlvoll vorgeht.


Nach weiterem Materialabtrag passt auch die Krümmermutter wieder an ihren Platz.


Foto nach dem Einbau kommt evtl. noch.

engraving an anodized aluminum laptop lid

Over at shackspace I used the 50W CO2 Laser to personalize my latpop lid.

I think it turned out really well. The aluminium oxidizes and thus turns white.
Settings used: scanning with 500mm/s @100%.

As template I created a two color pixel image with 1200dpi resolution.

“Scratch the surface of most cynics and you find a frustrated idealist — someone who made the mistake of converting his ideals into expectations.” – Peter M. Senge

USB charge station

Needing a multitude of chargers for ebookreaders, tablets, smartphones, USB-power-banks, flashlights just didn’t feel right and so I first used a cheap 10 port USB hub from aliexpress – but kept having problems with a good electric connection. As it turns out the hub doesn’t have any nibs on the contacts.

Buying a brand name (D-Link) active (so you can inject the necessary power) USB hub seemed the way to go – just to learn that it needs a connection to a computer to power up the 7 ports.

Some circuit bending took care of the issue:

3d printer – the plexiprinter

A few weeks back I decided it was time to get another 3d printer up and running over at shackspace … again.

I used the Fab@Home Model 1 that I brought like 2 years ago as a base, hooked up the 4pi board that Kliment gave me during the somewhat frustrating first 3d printer build experience.

The extruder ist one of the first commercially available direct drive extruders from qu-bd’s kickstarter campaign.

^^ see that? No heated bed!

I thought I could get this working in a few hours but after ttb had hooked up the endstops and I had built an adapter for the extruder it turned out the 4pi board was broken. *bummer*

So I threw it in a corner and started looking at the electronics the next day or so.

Fortunately the 4pi board is open hardware (schematic on github) and so I set out to find what’s broken.
As it turns out there was an inductor mechanically broken off the board and I _only_ had to find one with the matching value. To get it done I simply unsoldered a lot of inductors from scrap pcb’s and measuring all with a LCR meter.

^^ _quick_ fix.

The next issue I ran into was the extruder that didn’t extrude properly. The problem was that no PLA would come out of the nozzle after a few minutes of printing. Tightening the screw that presses the material against the gear, increasing the stepper current as well as some cussing didn’t solve the issue.

Trying to manually push the PLA through the heated extruder let to the conclusion that the PLA was getting soft too far up in the supply screw and then sticking to the wall. No force whatsoever could then push it through the nozzle.

It then turned out that Phil also encountered the problem with his qu-bd but couldn’t overcome it and so his delta went to the corner again.

We thought of a fix and a few days later I manufactured a supply screw into which you can insert a teflon tube to preven the filament from getting soft too early.

^^ the original was in stainless steel (so I also did it in stainless – is that necessary?) – but the 2.5mm hole isn’t that much fun 😉

Next there was the missing steps issue occuring only in y axis.

^^dickbutt test sample

I fiddled with motor current, speed and acceleration settings, even checked whether the y-axis stepper motor had degraded magnets … but could only get rid of the missing steps by reducing the speed to a whopping 9mm/s (~180rpm).

^^ the comparison measurement show that the the tested motors were all equally good/bad.

Still not perfect but we’re getting there.

K2 Kickboard ev mod – analog to pdm controller

After I avoided programming my own microcontroller to get the thing rolling in as little time as possible (sidetracking kills a lot of drive) – I finally decided it was time to take it on.

Actually the realisation that the cheap servo tester assumes the poti position when powering up as middle position convinced me that it was necessary to write my own speed controller – since I can’t (or at least not realiably/comfortably) use the braking functionality of the alien power sytem esc.

Also this way I can – or at least I can try to – write my own soft-start routine so people don’t fall off the back.

@phrewfuf encouraged me to program the arduino nano I tucked in a box some months back for when the time was right 😉
Since he didn’t send me the code I just set down myself today and took the first steps.

Also I had to wildlarize the freaking busted poti I pulled out of a part drawer. If you’re  not familiar with the term:

„widlarize“ (invented by Bob Widlar): You take the bad component to the anvil part of the vice and beat it with a hammer। It will make you feel so good, for no damaged parts will appear in your latest application।

It sucks so hard wasting time and money with  broken components that I consider not reusing old parts – or at least not the parts lying around shackspace 😉

Here is already another poti hooked up to an analogue pin linked to the pin the servo is connected to:

#include <Servo.h>

//onboard led
int ledPin = 13;

//servo pin
int escPin = 3;  // servo signal connected to digital pin 3
Servo escServo;

int potiPin = 7; // potentiometer wiper (middle terminal) connected to analog pin 3
int potiVal = 0;     // variable to store the value read from the poti

void setup()
  pinMode(ledPin, HIGH);
  //initial Servo Position

void loop()

  potiVal = analogRead(potiPin);    // read the input pin
  potiVal = map (potiVal, 0, 1023, 0, 180); //convert input range to servo range



Next up I need want to get interrupts working.

I claim first driving 20 screens with 10 gpu’s in a single computer

*gnarf* … it seems I wrote this 2 months ago and never published it to the blog …

Used hardware:
5x Nvidia NVS 440 dual gpu, dual head -> 2xDMS-59 -> 2xDVI
2 A$$rock X79 Extreme 7 mainboards never buy Asrock. I mean it.
1 P9X79_WS mainboard
1 i7 3930k CPU (because I wanted it … not because it was necessary)
1 SilenX EFZ-120HA5
1 Enermax Liberty ELT500AWT 80Plus
4x 4GB (Quad-kit) Mushkin Redline DDR3 1866MHz
2x Noctua NF-P12 PWM silent fans

20 resuscitated Samsung 940B/BF/T screens
20 DVI-cables from 1.5m to 5m, 20 VESA mounts, 4×5-port socket ports

For now:

And a video … from a few weeks back

Build log/pictures and details will follow.

P.S.: And then there’s this *jealousy dripping out of my eyes* … for ~20k$.
(and it doesn’t even give you any trouble with MS OS gpu limits since there are only 4 driving 24 screens)

[Idea] Get it done platform

So lets just throw it out in the hope someone else does it.

Else I … and makefu have to 😉

I would like to have a plattform that works somewhat similar to kickstarter – but different.
The focus would mainly be on open source software – but maybe also tangible products.

Let’s say you use a piece of open source software – e.g. Openwrt and you stumble upon a broken DYNDNS package. Now you may still be able to set up dynamic DNS through a script – but it would also be nice to have the luci-app-ddns package working as

So on the get it done platform I would state the wish for the luci-app-ddns package to be fixed and throw in 20€ for someone to do it.

If others would also like to get it fixed they could also pledge some money.
Now the more money is on the table – the more likely it would get for someone – or even groups of people to start working on getting it done.