Monday, 28 October 2013
Chimney sweep
Swept the main part of the chimney today. Needed 4 rods an allen key and a bit of silicon
Saturday, 19 October 2013
web cam
Installed the web cam today. Viewing inside the router firewall is easy. It's taken several attempts to get the viewing on the net and was only possible with a friends help.
Got a nice view of the garden, drive, mountains and a drain pipe.
This is not live. Email me for the camera address.
Got a nice view of the garden, drive, mountains and a drain pipe.
This is not live. Email me for the camera address.
Monday, 7 October 2013
DIY Wheel alignment
There are many ways to align your wheel and the normal method involves 30 minutes and a few quid. Not in France. €40 for a check and another €30 to adjust. And that's just a basic 2 wheel check. A 4 wheel check is well over the ton.
There are a few DIY tools in the market; Gunsons have done a simple gauge for years, and there is a new one using mirrors and a laser. Both are around £60/70.
For simplicity, if your cars wheelbase is the same front to back, you can set a line the same distance out from the wheel centres front/back and measuring the distance string to tyre and adjusting it until you get the desired difference.
I've used another method as the Xsara wheelbase is different front/back. You measure the distance from the same point on the front tyres left to right, from the rear and front of the front tyres. This method assumes the rims aren't bent and the tyres are pretty identical and that your wheels are equally at fault. I was pretty sure the tracking was out on the car as the outer edges were wearing and there was a bit of steering wheel wobble. This is toe-in or positive toe (http://www.etyres.co.uk/glossary-tyre-terms?term=toe-inout) I measured the 2 distances and found 151.5mm towards the rear and 149.5 towards the front, confirming the toe-in. I'd squirted wd40 on the adjustment nuts a few weeks back and they came straight off. 20mm AF on the nut, 18mm (odd size) on the rack arms. I finally found out the toe figures which are 0mm. I started with 1/4 turn each side and this swapped the measurements over! 1/8 turn back got the numbers to 151mm.
I've been studying the original / classic Dunlop system and will probably try to make a modern version with a web cam and/or a laser.
There are a few DIY tools in the market; Gunsons have done a simple gauge for years, and there is a new one using mirrors and a laser. Both are around £60/70.
For simplicity, if your cars wheelbase is the same front to back, you can set a line the same distance out from the wheel centres front/back and measuring the distance string to tyre and adjusting it until you get the desired difference.
I've used another method as the Xsara wheelbase is different front/back. You measure the distance from the same point on the front tyres left to right, from the rear and front of the front tyres. This method assumes the rims aren't bent and the tyres are pretty identical and that your wheels are equally at fault. I was pretty sure the tracking was out on the car as the outer edges were wearing and there was a bit of steering wheel wobble. This is toe-in or positive toe (http://www.etyres.co.uk/glossary-tyre-terms?term=toe-inout) I measured the 2 distances and found 151.5mm towards the rear and 149.5 towards the front, confirming the toe-in. I'd squirted wd40 on the adjustment nuts a few weeks back and they came straight off. 20mm AF on the nut, 18mm (odd size) on the rack arms. I finally found out the toe figures which are 0mm. I started with 1/4 turn each side and this swapped the measurements over! 1/8 turn back got the numbers to 151mm.
I've been studying the original / classic Dunlop system and will probably try to make a modern version with a web cam and/or a laser.
Wednesday, 2 October 2013
tech note - SCS sentinel pro 500 gate controller
This note refers to the PCB CTP50-2 (2007 approx)
The manual for this product if fairly useless and does not describe what the (spare) terminals are for, or what all the dip switches are for. Even though an emergency stop circuit is provided, this is not documented. A cynic would suggest this is all deliberate, as in large letters on every page is the premium rate phone number to call for assistance.
This model, along with the cheaper gate controllers, try to detect the gate fully open and gate fully closed positions by sensing the motor current; a stalled motor will draw a lot of current. Unfortunately, unless your gates are made extremely solidly your gates will bend and flex. In my case, steel gates made out of 30mm, 1.6mm box section and welded together, were flexed along the 1.5m mounting height by around 25mm, and the system still didn't recognise a closed gate. These gates are mounted onto the pillars with plastic mounts which were promptly dislodged; the gates had to be removed and the bits re-fitted.
To try to kill the motors a 0.68R resistor in parallel with the motors stops the gates; in normal running it works a treat. I later found the ADC lines from the current sensing resistors (R5/6) and ran +5V through a 4k7 to (R15/12) After a lot of messing around, I realised the system uses a timer to run the motors, so a variance in supply volts means they over/under run. On battery it's even worse, even though it detects loss of mains and does apply some adjustment. Basically, even if you fudge the learn sequence and simulate over-current, it will quite quickly lose track of where it is and bend the gates again. If power is lost for any reason, the system assumed the rams are fully out and if the rams were fully in, it will try for many seconds to push the rams in.
There are 8 spare terminals on the pcb and these didn't seem to do anything. I then read the sliding door model 400 manual and noticed this uses a limit switch. Trying again I found the connections for fully open and fully closed on both gates. NB the learn feature does not look at the limit switches. If the rams are fully out (switches properly made) and power is lost, on power up the system will only operate mot1 in the opposite direction.
A 12V SLA battery can be put on terminals FS1+ FS2-
So 2 new internal switches are now fitted to simulate over current (on adc) An emergency stop is also fitted (not shown in manual) A NC switch on terminals 14/15.
To set up the gates, install limit switches and run the learn feature (P2)
set rams 80mm from fully in
press p2 an be ready to hit mot1 over current. The ram will move another 20mm then the other will go. Press the other over current
the rams will now be sent out. Press the correct o/c when they get into position. Note if one ram only moves half way and then stops, you didn't hit the o/c button quickly enough on ram2.
done.
The proper aim is to find the correct limits, but I set up the rams for full travel and use the limit switches to stop them. Rams move a tad after limit switches are reached.
If the gates aren't mounted, or you don't care, just let the system learn on its own which will doubtless mean it send the rams fully in/out.
The transformer is a multi-tap 100VA toroid of dubious quality. It consumes 9W doing nothing with a pretty high 'copper loss'. Red (FS3) is 0V, black (FS4)12V, yellow 15V, green 20V. Note a SLA charger can be made using the 15V tap and a L200 regulator.
AC quiescent is 9W, +3W for charging. DC (battery) is 30mA (leds on, 4mA less off) 2A typically both rams running, 11A when jammed.
Removed the 7805 reg. System took 5mA @ 12V. If 5V (23mA no leds) is externally maintained on the pcb during mains fail the system will not notice. You have at least 4s to return power once you request gate open/close, so if you get a small efficient mains 5V psu, you could wire that into the pcb and then switch the mains onto the toroid when required. A small psu will take <1W.
If running at 5V, a 14V/6A SMPS on ac terminal will run the motors, or a 14V/4A with a 15000uf cap to ease the starting current.
The battery charger circuit is a joke and sends 13.75V through a 10R resistor to taper charge the SLA battery - around 83mA for half charged battery. After mains fail, charging current is only restored once the battery voltage drops below 9V or around 2.5V below the recommended minimum. If the battery is momentarily disconnected on mains return, charging recommences. The smallest relay seems to control this and was not investigated particularly.
The remotes are keeloq and use HCS301.
25/10/13 put a small 5v psu to directly power the electronics, with a relay to turn on the toroid when the system needs it. Overnight consumption shows this at 4-5W, which is odd for a 0.03A x 5V = 150mW consumption.
The manual for this product if fairly useless and does not describe what the (spare) terminals are for, or what all the dip switches are for. Even though an emergency stop circuit is provided, this is not documented. A cynic would suggest this is all deliberate, as in large letters on every page is the premium rate phone number to call for assistance.
This model, along with the cheaper gate controllers, try to detect the gate fully open and gate fully closed positions by sensing the motor current; a stalled motor will draw a lot of current. Unfortunately, unless your gates are made extremely solidly your gates will bend and flex. In my case, steel gates made out of 30mm, 1.6mm box section and welded together, were flexed along the 1.5m mounting height by around 25mm, and the system still didn't recognise a closed gate. These gates are mounted onto the pillars with plastic mounts which were promptly dislodged; the gates had to be removed and the bits re-fitted.
To try to kill the motors a 0.68R resistor in parallel with the motors stops the gates; in normal running it works a treat. I later found the ADC lines from the current sensing resistors (R5/6) and ran +5V through a 4k7 to (R15/12) After a lot of messing around, I realised the system uses a timer to run the motors, so a variance in supply volts means they over/under run. On battery it's even worse, even though it detects loss of mains and does apply some adjustment. Basically, even if you fudge the learn sequence and simulate over-current, it will quite quickly lose track of where it is and bend the gates again. If power is lost for any reason, the system assumed the rams are fully out and if the rams were fully in, it will try for many seconds to push the rams in.
There are 8 spare terminals on the pcb and these didn't seem to do anything. I then read the sliding door model 400 manual and noticed this uses a limit switch. Trying again I found the connections for fully open and fully closed on both gates. NB the learn feature does not look at the limit switches. If the rams are fully out (switches properly made) and power is lost, on power up the system will only operate mot1 in the opposite direction.
A 12V SLA battery can be put on terminals FS1+ FS2-
So 2 new internal switches are now fitted to simulate over current (on adc) An emergency stop is also fitted (not shown in manual) A NC switch on terminals 14/15.
To set up the gates, install limit switches and run the learn feature (P2)
set rams 80mm from fully in
press p2 an be ready to hit mot1 over current. The ram will move another 20mm then the other will go. Press the other over current
the rams will now be sent out. Press the correct o/c when they get into position. Note if one ram only moves half way and then stops, you didn't hit the o/c button quickly enough on ram2.
done.
The proper aim is to find the correct limits, but I set up the rams for full travel and use the limit switches to stop them. Rams move a tad after limit switches are reached.
If the gates aren't mounted, or you don't care, just let the system learn on its own which will doubtless mean it send the rams fully in/out.
The transformer is a multi-tap 100VA toroid of dubious quality. It consumes 9W doing nothing with a pretty high 'copper loss'. Red (FS3) is 0V, black (FS4)12V, yellow 15V, green 20V. Note a SLA charger can be made using the 15V tap and a L200 regulator.
AC quiescent is 9W, +3W for charging. DC (battery) is 30mA (leds on, 4mA less off) 2A typically both rams running, 11A when jammed.
Removed the 7805 reg. System took 5mA @ 12V. If 5V (23mA no leds) is externally maintained on the pcb during mains fail the system will not notice. You have at least 4s to return power once you request gate open/close, so if you get a small efficient mains 5V psu, you could wire that into the pcb and then switch the mains onto the toroid when required. A small psu will take <1W.
If running at 5V, a 14V/6A SMPS on ac terminal will run the motors, or a 14V/4A with a 15000uf cap to ease the starting current.
The battery charger circuit is a joke and sends 13.75V through a 10R resistor to taper charge the SLA battery - around 83mA for half charged battery. After mains fail, charging current is only restored once the battery voltage drops below 9V or around 2.5V below the recommended minimum. If the battery is momentarily disconnected on mains return, charging recommences. The smallest relay seems to control this and was not investigated particularly.
The remotes are keeloq and use HCS301.
25/10/13 put a small 5v psu to directly power the electronics, with a relay to turn on the toroid when the system needs it. Overnight consumption shows this at 4-5W, which is odd for a 0.03A x 5V = 150mW consumption.
Subscribe to:
Posts (Atom)