PDA

View Full Version : Took apart the AP1500 PSU


Risky
25th August 2004, 10:31
It appears that the votage selecting switch is simply switching fixed resistors, so it would be possible to mod this to have the control on the front panel or even use a Rheostat for completely variable voltage 12-24V.

I'm going to have to get out my multimeter and investigate.

Pug
25th August 2004, 13:42
Nice one Risky. ^_^ That's one less warranty to worry about :p ;) :D (j/k)

Got pics? (haven't modded my screwdriver to do mine yet).
use a Rheostat for completely variable voltage 12-24V.
I'm not an EE genius admittedly but would imagine it would have to be a hefty pot if you did. :blink: :unsure:

Risky
25th August 2004, 13:59
Nice one Risky. ^_^ That's one less warranty to worry about :p ;) :D (j/k)

Got pics? (haven't modded my screwdriver to do mine yet).
I'm not an EE genius admittedly but would imagine it would have to be a hefty pot if you did. :blink: :unsure:

I took some pics but they'll have to wait till tonight to upload.

I'm not sure it would be a big pot as its running a regulator and this should just be the reference voltageto control the output. Alas its going to be a sod to take apart the board as the huge heatsink is in the way.

Oh and for bits I bought a Draper set of security bits in Dyas fairly cheap. Every weirdo bit in the world it seems.

Pug
25th August 2004, 14:32
I took some pics but they'll have to wait till tonight to upload. Nice one mate. Look forward to it. ^_^

I'm not sure it would be a big pot as its running a regulator and this should just be the reference voltageto control the output. Alas its going to be a sod to take apart the board as the huge heatsink is in the way. :blink: /me watches Risky's reply sail over my head... :unsure:
Told you I wasn't an EE. :p :D

Oh and for bits I bought a Draper set of security bits in Dyas fairly cheap. Every weirdo bit in the world it seems. Ideal. Don't think we have a Dyas round here though...
/me reverts to screwdriver modding. :D

[OT] btw, did you have 12mm tubing still? - I got a mail earlier from someone asking about some and couldn't remember if yours was 10 or 12mm.

Risky
25th August 2004, 15:47
:blink: /me watches Risky's reply sail over my head... :unsure:
Told you I wasn't an EE. :p :D
Neither am I, but I read teh BiT electonics forum a lot!

Ideal. Don't think we have a Dyas round here though...
/me reverts to screwdriver modding. :D
http://www.google.co.uk/search?hl=en&ie=UTF-8&c2coff=1&q=security+bit+set&btnG=Search&meta=cr%3DcountryUK%7CcountryGB and leave your poor screwdriver alone!

[OT] btw, did you have 12mm tubing still? - I got a mail earlier from someone asking about some and couldn't remember if yours was 10 or 12mm.

12 from res to pump to rad, 10 thereafter. I have some 12mm tube spare. But its nylon and you know what that's like to work with. From teknobots if you're curious. My 10mm stock is Legris Semi-Rigid Nylon and Flexible Polyurethane Polyether Tubing, and lots of it :eek:

Starbuck3733T
25th August 2004, 18:40
What's the current (Amps) draw on the AP1500 at full voltage (24VDC I assume?)

With motors its bound to be better than an 1 AMP, if not 2. that's 48W of juice at max and NOT something you want to be trying to do with a linear regulator or rheostat. Unless it has some sorta thing where you supply it a reference voltage (for reference only, little to no current draw) you'll be wanting to look at PWM for driving the motor. PWM'd 24V would be a pretty slick thing. Easy enough to do with a 555's output controlling a few MOSFETs in parallel (so you can handle the amps that beast needs). You'd still have to heatsink them, but i'd still require less stuff than the equivalen linear regulator (BAD IDEA).

I'm an EET, not an EE supergenius.

Risky
25th August 2004, 19:29
Afaik its 18W max so <1A.

Now this will be a little easier with a picture, but evn that isn't so easy.

The board is almost completely covered with the biggest heasink you could fit in the case. That is cooling one device with three pins. The line from the middle one runs throught the sector switch which varies the amount of resistance on the line. The resistors are rather standard looking fellows - no big ugly things fwiw.

The curse is that the heasink seems to have been bolted on befor ethe device was soldered or the ony way to detach it (and thus look at the board properly, is to unscrew it with a pointed pliers - possibly a one way process :eek:

I can would out the connections by shining a light through the board to see the tracks and a lot of squinting, however in any case the component count is pretty low. so its not a complex beast. However there are some very think tracks on the back in plcaes which might indicate the high current sections, possibly.

For the record the spec for the adaptor is:
Input voltage : 12 - 13.8Vdc
Output voltage : 12 - 24Vdc, regulated
Output current : 2500mA max

Starbuck3733T
25th August 2004, 22:49
Afaik its 18W max so <1A.
[... snip ...]
I can would out the connections by shining a light through the board to see the tracks and a lot of squinting, however in any case the component count is pretty low. so its not a complex beast. However there are some very think tracks on the back in plcaes which might indicate the high current sections, possibly.

For the record the spec for the adaptor is:
Input voltage : 12 - 13.8Vdc
Output voltage : 12 - 24Vdc, regulated
Output current : 2500mA max

Well then, that's not too horrible. I'd start poking it with a multimeter (when its off) to find out what goes where. your assumption of thick traces = power is most likely correct.

The 3 pin devices may very well be a FET (or MOSFET) doing switching regulation. Too bad disassembly may equate to death :(

Risky
26th August 2004, 00:17
Well I 've just drawn out the schematic on paper, but without knowing the exact parts or values yet.

The IC is a Richtek DC-to-DC Converter Control Circuit (http://www.richtek-ic.com.tw/Product/Docs/DS34063A-08.pdf)

The control switch is varying the potential at one pin of the IC "Feedback Comparator Inverting Input". This is held on a divider from the output and the effect of the truning the control up is drop the potential on that pin. however it is never grounded and the is a fixed resistor in line with the variable value from the switch.

.....aaargh main box has just died and i need it now... runs.....

Risky
26th August 2004, 02:15
OK here's the pics:

http://www.veryrisky.co.uk/webimages/stepuppsu03.jpg
Lid off and them diodes
http://www.veryrisky.co.uk/webimages/stepuppsu08.jpg
The pcb tracks
http://www.veryrisky.co.uk/webimages/stepuppsu09.jpg
That FET or MOSFET?
http://www.veryrisky.co.uk/webimages/stepuppsu11.jpg
The IC and some bits
http://www.veryrisky.co.uk/webimages/schematic.jpg
And my effort at a schematic from that.
EDIT: There is an error there. The connection above the section circled in red is not to the power line there, but to the output line after the diodes (top right).


I think you can see that it should be possible to bypass the existing switch and replace.

fivecheebs
26th August 2004, 09:35
I've been thinking about this also but unfortunately thats as far as i have got. Glad to see someone with a better idea of what they are doing do it first :) thanks risky, i hope you can discover an easy way to do it :)

Risky
26th August 2004, 10:07
Its looking good.

Whiel sitting around waiting for a power cut to be fixed this morning I got the multimeter out.

Now first, there's an error in my drawing above - the line from the control pin that connects to the VR is shown headign for the top raid to early - it actually connects to the output line at the far top right.

Now some values:

Power out to Input pin 10k ????? maybe its 1k??
Pin to VR : 600R
Variable 0-570R

So as far as I can see you just need a 570R pot which will only draw mA probaby desolder to 600k to cut the line to the variable element and connect that in series with the pot. :cool:

Risky
26th August 2004, 13:37
Oh make that a log pot, I think as the resistors got larger in each step. Unsurprisingly there are no 570R pots around, but a smaller one should do as you can always bypass the whole unit if you want 12V.

Now what I need is:

500R ish pot
4pDT switch if I want to bypass the whole thing and connect directly.

and from a longstanding project I might as well throw in a current transducer to measure the draw of the pump and wire up a spare led on the pump terminals to show power on the to the pump.

fivecheebs
26th August 2004, 14:57
well much of that makes little sence to a bozo like me ;) hehe but when i see some more diagrams and pics i hope ill start to get it.

Good work Risky :)

Risky
26th August 2004, 17:50
well much of that makes little sence to a bozo like me ;) hehe but when i see some more diagrams and pics i hope ill start to get it.

Well I need to test it but the solution is roughly
- Cut one trace
- solder on two wires
- attach (up to) 570R (log) pot between them.

fivecheebs
26th August 2004, 18:04
Thanks Risky, now i get ya :)

One question though, why "up to 570R"? I would have expected to use one 570 or larger.

Risky
26th August 2004, 18:18
Thanks Risky, now i get ya :)

One question though, why "up to 570R"? I would have expected to use one 570 or larger.

The switch part increases the resistance from zero to 570 for 12V-24V output. Increasing the resistance will increase the voltage at that "Feedback Comparator Inverting Input" which somehow varies the output and I though it best to keep it within the range of the current setup.

Of course a smaller pot will limit how much you can drop the voltage - if you wanted to ensure you can get 12V out then you coudl add a series resistor to bring the maximum resistance up to 570R and instead you woldn't be able to get to 24V. I think I'll go with the former approach but build in a switch to bypass the whole unit and send 12V to the pump if I want.

fivecheebs
26th August 2004, 18:34
OK mate, thanks. I can see how it works now :)

Risky
27th August 2004, 11:14
Since I'm at it rather than fiddling around with the existing LED, I will run one off the pump line in as that way it will only show if the pump has power.

Starbuck3733T
27th August 2004, 15:54
Well i haven't been paying attention for a bit - looks like you're well on your way. I shall cease my babble!

Risky
3rd September 2004, 11:42
I must get on with this but in the meantime I've worked out that since you can't get 530Ohm pots, the control would be a 1.1k resitor in parallel with a 1kLog pot. I might make a run to maplin later and give this a try.

fivecheebs
3rd September 2004, 14:01
Good luck mate :)