Electronics advice

[malc-c]

OK, not sure if I've just been unlucky and sods law has struck or one of these LED dimmers that everyone has cottoned on to has caused the problem... but my 13.8v regulated power supply has just gone pop, one what is looking to be a promising evening, with clear skies all around. The strange thing is the fuse didn't blow, but it's now kicking out 20.3v rather than 13.8v and none of the components are showing signs of damage associated with the large pop that came out of it. OK it's some 10 years old, but up until now has been fine, even providing the amps when charging 12v RC Battery packs.

Up until now my power distribution was via a small box containing 5 phone sockets all wired in parallel - 12v in and thus 12v out to the HEQ5, DC focuser and when needed, two dew heaters. I've now modified this to include two of the cheap LED dimmers discussed in one of the other threads, the only modification being the removal of the screw terminal block and 4 wires soldered directly to the board. These have been wired in parallel to the input, with two LEDs connected across the outputs to indicate the output to the heaters. These LEDs dim up and down nicely, so all seems OK.

Thinking that maybe my soldering wasn't up to spec I checked everything out and couldn't find any direct shorts, or anything obvious, and as far as I'm aware the fuse should blow if a direct short has been applied across the PSU. So I've tried measuring the resistance to see if this would indicate a short. With everything connected (scope, heaters, camera but obviously no power) the meter read 10.41K ohms across the input. From my school days, ohms law stated that current in amps is volts divided by ohms. So under normal circumstances this would be 13.8 / 10410 ? - if so this equates to a very low amperage of 0.00133 Amps so that shouldn't of caused the PSU to go pop

I've done some basic testing, connecting the DVM between the positive feed and setting it to amps, with just the dimmers connected it drew around 0.018A (still kikking out 21v) - however connecting the two dew straps to either the dimmer outputs or direct to one of the paralleled outputs didn't increase this current draw. If I turn the dimmers down the current drops to 0.002A so they seem to be having some effect. I don't want to risk connecting the HEQ5 to the 21v supply as I don't want to risk damaging it (although it may of been damaged when the PSU went pop- who knows), so wondered if any of you guys can offer some suggestions as to what to try, other than replace the psu which will have to wait until the end of the month when I get paid.

[pete_l]

and none of the components are showing signs of damage associated with the large pop that came out of it.

Check if any of the smoothing electrolytics have bulges in their cases. The bulges could be at the bottom, where the devices are attached to the PCB and therefore be difficult to spot. If so, they'll need replacing.

[malc-c]

Pete, thanks for the reply. I'll check them out, however I would these play a part in the regulation process ? - When I turn the psu off the voltage decreases gradually rather than down to zero in a blink of an eye

[faulksy]

the led dimmers that your on about doesnt control amps/resistance they control volts which is were there name comes fom pwm, sounds like your power supply, cant you connect it to a 12v dc supply and then check your out going voltage

[Fish-Man]

they control volts which is were there name comes fom pwm

That's not quite true.

PWM Stands for Pulse Width Modulation.

This means that the unit sends out a square wave at the voltage designated on the unit. So it would be something like "fully on" for 80% of the cycle and "fully off" for 20% of the cycle.

When you turn the control knob down, the unit sends out narrower and narrower square waves, so it would now be something like "fully on" for 20% of the cycle and "fully off" for 80% of the cycle.

In effect the item connected, is being switched on and off very quickly (normally about 25kHz) but always at the same voltage.

HTH

Regards,

Paul

[malc-c]

the led dimmers that your on about doesnt control amps/resistance they control volts which is were there name comes fom pwm, sounds like your power supply, cant you connect it to a 12v dc supply and then check your out going voltage

There is a relationship between resistance (load), voltage supplied, and the current that will drawn. With a dew heater connected, and with a PWM (actually they are more probably PPM (pulse proportional modulated) output, the frequency would be to fast for a normal DVM to read, so regardless of the pulse width the meter would see it as if the voltage was applied 100% of the time. So I would of expected a change in current drawn when the heaters were connected. The two indicator LEDS across the outputs vary in brightness as the knobs are turned which would indicate the dimmers are functional.

There is a problem with the supply, in that it's no longer regulated to 13.8v, and appears to be kicking out the full 21v DC that the rectifier is supplying. I'm hoping that the control board in the HEQ5 hasn't hone pop either, but as I don't have any other 12v supplies (all the other switch mode supplied kick out around 16-19v ) I don't want to risk further damage. My gut feeling is that it was co-incidence, in that the unit being old was on its way out. Just gutted it happened on such a perfect night for imaging.

[RogerTheDodger]

It's probably the voltage reference in the PSU that went pop, hence the same output voltage as the unregulated voltage. As the transformer and rectifier sound like they are ok, you could replace the rest of the electronics quite cheaply if you are handy with a soldering iron.

[Gina]

PC CPUs give a regulated 12v as well as other voltages (5v, 3.3v, -5v and maybe -12v). Only trouble is, 12v is at the bottom end of the SW mount power requirement and may show a power-low indication (flashing red LED). This is what I've read and haven't tried 12v on my mount (I bought a 13.8v analogue regulated supply from Maplin for the job).

[malc-c]

It's probably the voltage reference in the PSU that went pop, hence the same output voltage as the unregulated voltage. As the transformer and rectifier sound like they are ok, you could replace the rest of the electronics quite cheaply if you are handy with a soldering iron.

Well I've designed and built this so fairly confident with a soldering iron

If you can suggest where to start looking... apart from a couple of transistors the only other device on the PCB is a T0220 package, the markings are not clear, but it looks like UTC 29088D, the output tranny has D717 the rest are rubbed out, but it looks like Y ** 2A

Google hasn't been much help..

The good news is I found a 12v (actually kicking out 14v) psu so I tested the HEQ5 and all is well ! - Phew !!

[Gina]

That's good news Malcolm I repaired a PC PSU many years ago - a bad solder joint as I recall.

[malc-c]

Hi Gina,

Oh those were the days.. the only thing to watch with converting the modern ATX psu's are that you have to tie the 5v line to load in order to get the 12v line to work... I'm not 100% sure, but in the days of flying indoor model RC helicopters, many a charger wen pop when this 5v line failed for some reason... not sure why though, but I do have a nice 400W computer PSU spare at the moment, however I do prefer the linear regulated supplies as a bench supply for an observatory

[malc-c]

Well I've replaced the capacitors on the board, with values as close to the originals (the two large cap were 3300uf, the replacements are 4700uf) but still outputting 21.4v. The circuit seems fairly basic, my guess is that the T0220 device is a variable voltage regulator, with the descrete components (and a couple of trim pots) used to tune the volatge to 13.8v, with the D717 transistor handling the current load. My guess is it's the To220 device that's gone pop.

[RogerTheDodger]

Sounds like it. I'd scrap the electronics and re-use the transformer and rectifier to build something like this http://ludens.cl/Electron/Ps20/Ps20.html but with less 2N3055's on the output. I've build loads like this for CB/HAM friends.

[Gina]

You can get far more efficient transistors than the ubiquitous 2N3055, in TO220 cases. Handling 2 or 3A for the mount is easy. I would suggest an LM317T for the control amplifier (voltage regulator) backed up with a power transistor to increase the current. LM317Ts are rated at 1.5A. You can provide over-current protection for the PSU too. I would also recommend an over-voltage crowbar with fuse to protect the output from over-voltage since you want to power expensive kit. I could design the circuit if you like and find links to source the components. The LM317 series only uses two resistors to set the output voltage. It has built-in current limiting which can be used to limit the output in conjunction with a piggy-back power transistor I believe.

[malc-c]

Sounds like it. I'd scrap the electronics and re-use the transformer and rectifier to build something like this http://ludens.cl/Ele.../Ps20/Ps20.html but with less 2N3055's on the output. I've build loads like this for CB/HAM friends.

seems logical..... it wouldn't be too bad if the manufactures had used standard recognised components (2n3055, LM314 etc) so replacements could be sought and replaced, but with parts that are hard to trace and have some of the identifying markings erased that makes life difficult. Thanks for the link. I have most of those components in the hobby box, other than the high wattage resistors. The maplin unit is rated at 3A continuous, and there are no VA ratings on the transformer, so have no idea if the transformer used in all the Maplin units of this type are the same (ie 7A rating) and it's just the number of power transistors used that restrict the current rating, or that they use different transformers in each of their 3A, 5A or 7A rated supplies that come in this case design.

[malc-c]

You can get far more efficient transistors than the ubiquitous 2N3055, in TO220 cases. Handling 2 or 3A for the mount is easy. I would suggest an LM317T for the control amplifier (voltage regulator) backed up with a power transistor to increase the current. LM317Ts are rated at 1.5A. You can provide over-current protection for the PSU too. I would also recommend an over-voltage crowbar with fuse to protect the output from over-voltage since you want to power expensive kit. I could design the circuit if you like and find links to source the components. The LM317 series only uses two resistors to set the output voltage. It has built-in current limiting which can be used to limit the output in conjunction with a piggy-back power transistor I believe.

Gina,

That would be cool, I have an account with RS, so any items I don't have can easily source...

[Gina]

Gina,

That would be cool, I have an account with RS, so any items I don't have can easily source...

OK I'll see what I can do

[pete_l]

LM317Ts are rated at 1.5A.

If you want something with a larger current capacity, have a look at the LM338T that is rated up to 5 Amps. The only issue being that you do have to get rid of all the heat the linear regulator will create. If your input voltage is 21V and you're dropping that to 13.8, then at 5Amps you need to remove 36 Watts of heat.

BTW, a good, cheap source for the LM338 is www.bitsbox.co.uk - very reasonable P&P at £1.50

[malc-c]

Pete, thanks for the suggestions. Some nice examples in the data sheet. What puzzles me is that given 13.8v seems a "standard" for bench power supplies, that there isn't already a similar chip where the input could be up to 25vdc and it would provide 13.8v out using a minimum of a couple of caps and a blumming big heatsink !

[pete_l]

Well the 78xx range is fairly common - and available in a wide range of output currents. Remember that with any of the 3 pin linear regulators, you can set whatever output voltage you want (within reason) by placing the right value zener diode between its COMMON pin and *real* 0V. So just choose the biggest honkin' 7805 you can lay your hands on and get an 8.2V zener and off you go to 13.2V!

Fer example:

http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=0CN4CEBYwAA&url=http%3A%2F%2Felectroschematics.com%2F6027%2F7805-regulator-ic-circuits%2F&ei=R_MwUIDlIOah0QWm84DgDA&usg=AFQjCNG9B78v8lGI0Nl3OkNDqDDvS_H6eQ&sig2=hAawFRdEWDZziZrWno_LKQ

[malc-c]

Pete,

That's sounds good, especially if an over voltage crow bar is added... I've just measured the rectified DC from the bridge and its 23.9v, so I assume I would be better off using a 7812 and a 1.2v Zener so that there is less heat to dissipate ?

[Gina]

As I see it there's still the same heat to dissipate. (Vin - Vout) / amps

[pete_l]

As I see it there's still the same heat to dissipate. (Vin - Vout) / amps

Yes, that's right. One thing yo can do is have a two stage regulator. First, drop the 21V supply down to (say)17V and then drop the 17V down to your working voltage. It's still necessary to dissipate the same number of Watts, but you can do it over two components and have each on their own heatsink.

[malc-c]

The only thing that worries me with all this is the equipment I'm connecting it to. If this was juts to power a home project then I would be more comfortable, but the cost of replacing the main control / driver board on my HEQ5 would be beyond my means.

I'm opting for one of these

£34 from the suppliers.That is unless Gina or Pete come up with an easy circuit that has both over voltage and short circuit protection

[pete_l]

Looks pretty tasty.