volt/ammeter

Some time ago I bought a bench top power supply, which turned out to be not one of my better investments as the volt/amp dispaly on the front died. Yes I should have sent it back, but I decided to fit another display. The voltage display works fine, but the ammeter reads 0.55A when it should be reading 1.75. It's wired up as shown except that the red wire (Vcc) is fed from a separate 16v supply as instructed to make it work on <3V (the transformer had 2 secondaries)

Anyone got any ideas, or have I bought a duff display as well?

I've just noticed that the bridge recifier on the power board is running very hot as well, this really wasn't one of my better buys! I'm not giving up that easy. Next step a new power board.

Hi Duncan,

I'll jump in and risk losing some body parts.

Ammeter should be in series with the load.

Also I can't see the unit being happy with the red & yellow wires joined.

Regards, Bear..

Hi Duncan, what happens if you connect the thin red wire to the supply DC 4 - 30volt, instead of a separate power supply. (OK, I read the reason why you would want a separate supply, but just to test it out).

Funnily enough, my 2 units arrived by post today, a little late, but have arrived. just need time to try them out.
BobH

The red and yellow leads are so you can measure a different voltage id desired, no hazard using it to measure the supply voltage.

Shouldn't GND be connected as well?

I suspect the answer is that you should use a ceramic screwdriver to turn the little potentiometer marked Iadj until it reads correctly...

Neil

Grizzly, the ammeter is in line with the load, the I+ and I- wires carry the current and there is a shunt on the board behind the connector. The yellow wire is the voltage you are measuring and the red wire is the voltage supply to the display electronics. You have to pwer this from a separate supply if you want to measure <3v.

Neil, the ground wire is connected. I've tried twiddling Iadj and can only get it to chnge over the range 0.3 to 0.4 A when it is actually passing 1A. I'm going to take the display out from this unit and test it on another known to be OK psu, as I just noticed that the big series transistor which is controlling the voltage is also getting hot at 1A and the psu is rated at 2A

Edited By duncan webster on 06/10/2018 15:27:23

So I've now connected the ammeter to another psu and load, but powered the display electronics from the suspect psu and it seems a lot nearer, within Iadj territory. Next step is to junk the power control board and make a new one from a 317 and a power transistor. There won't be much left of the original at this rate.

At least your comments above helped me to know I hadn't committed some obvious goof

Cut the Red wire..........Anybody fancy a pint. I'll get me coat.

Hi Duncan,

Thanks for the update, I suspected I was pushing my luck.

I trust the case/enclosure is OK.

Good luck, Bear..

A few thoughts: -

1. Has the shunt in the display replaced(displaced) the old shunt in the power supply or is it in series with it?

2. Are the grounds of both supplies connected together?

3. Have you put a separate meter in series with the display to check the current, if it's different then it's probably duff.

Don't know why but this thought immediately popped up in my ancient grey matter...

Black to red
Green to brown
Blue to .…!!

I think it's from my very early apprentice days...

George.

Things have moved on a bit. It appears that dumping what I thought was the display board seriously disrupted the 'power' board causing it to get very hot. I also found several more issues:

1. when I first got it it didn't have a fused plug and so was illegal. I should have sent it back, but it was easier to fit a proper plug
2. it wasn't CE marked
3. the smoothing capacitor after the transformer/rectifier is only 3300 uF, this seems inadequate for a 2A power supply, will result in high ripple.
4. The transformer was supposed to deliver 20V, but off load it delivered 26V. I know that off load you get more than on load, but 30% seemed like a lot to me.
5. when I examined the power lead there was copper showing through the insulation. Ok not through the outer sheath, but still not impressed.

I've finished up dumping everything apart from the case, a fuseholder and one of the potentiometers and fitted a new (old) transformer I had in stock, a new rectifier with a huge capacitor I had in stock, a 5V linear regulator for supply to Arduinos, and a buck converter to give a separate 1.5V to 18V supply. It's all working fine except that the new display shows amps about 20% higher than my Fluke. I've also compared it with a moving coil ammeter. Yes I've tried twiddling the Iadj pot, not enough change. Next step is to experiment with a second shunt in parallel with the built in one to reduce the volts drop and presumably the displayed current, but I might just live with it for a bit, bored with this job!

The new transformer was 14/9/0/9/14, so I've left it as centre tapped and used 2 separate rectifier/smoothing cap set ups for the linear and buck circuits. Why? Because I could! At least it is usable now.

When I've made some supports for the cap I'll post some photos, not fit for human consumption at present with the cap supported on cable ties.

Can't offer much other than sympathy but:

• Off-load an unregulated DC supply will tend to rise to peak AC dropping back to RMS as soon as a load is applied. 26V from a 20V transformer may be reasonable.
• Increasing the 3000uF capacitor should be done with care - the bigger the capacitance the greater the switch-on surge. The existing rectifier diodes may not be man enough.

Dave

Peak value of rectified sine wave is about 1.4 x RMS. This is independent of transformer regulation, which is typically about 10% for a decent sized unit.

1.4 x 20V = 28V

Knock of two diode drops for a full wave rectifier 0.6 x 2 = 1.2V

That leaves you 26.8V off-load with no other losses.

20V with a 2A max current is a minimum load resistance of 20/2 = 10 ohms.

Time constant for 10 ohms and 3,300 uf is 0.0033 x 10 = 0.033 seconds.

The ripple frequency of rectified 50Hz AC is 100Hz, or of 0.01 seconds peak to peak.

0.033 is > twice the peak to peak time, which is the normal citeria for smoothing capacitors for non-critical applications.

Sorry Duncan, it seems to have been reasonably specced from here!

Neil

2 amps from a 3300uF cap gives a rate of change of voltage of 606 V/sec, which is 6v every cycle. Seems like a lot to me, 23%, but I bow to your wisdom. It did have a voltage control following so I don't suppose a bit of ripple matters.

Hi Duncan,

I've managed to get it into an online SPICE simulator.

I've assumed that as you measured 26V off-load (or thereabouts) that the off-load peak voltage is 28V to allow for the diode drops (some values in the above circuit have been changed). You can see the output voltage is fluctuating between 26V and 21V. bearing in mind that this represents an output current of ~23.5V/10 or 2/35A, then there should be sufficient headroom for a linear regulator at a 2A load.

I'm most impressed with your simulation, it gives almost the same ripple as my sum. However, the 26V was AC RMS, so with a cap I'd expect 35V DC which would have been challenging for my buck converter. When loaded with  2 off 21W 12V bulbs in series the voltage dropped to 19V, a 27% drop, well above the 10% suggested by SOD.
I found another transformer about the same size which had 14/9/0/9/14 secondaries, and measuring across the 2 off 14 tappings gave 28.1V off load, 25.6 V when drawing 1.7A .This is less than 10% drop. Connecting as centre tap with just 2 diodes into a way too big capacitor gives 18.9V, which is close to the 19.1 you'd expect, but then connecting the bulbs the volts drops to 15.1 at 1.35A, which is too low.

I should have given up then, but I wanted to understand why the DC volts drop was so much higher then the AC. I suspect it might be to do with the fact that the transformer is now effectively giving out short pulses into a very low impedance (capacitor) and so when it is conducting the current is high and so the voltage drop through the windings is high, but if anyone has a better explanation I'm all ears. I forgot to measure the AC volts whilst the DC was connected, but did attach my scope I really don't understand the vertical drop. Again if anyone has an explanation I'd be grateful. At least part fo the joy of making things is learning something new. I did try another set of diodes, didn't help, if anything it got worse. There is very little ripple on the DC, not surprising with my enormous cap. I've now given up, one transformer too high, one too low. I'll buy a complete 24V switch mode supply and connect the existing buck converter to that. I have a linear regulator for the 5V.

Edited By duncan webster on 12/10/2018 00:46:09

Edited By duncan webster on 12/10/2018 00:47:40

I forgot to mention that te ammeter behaves a lot better on this set up, still reading about 20% high, but I've had a play with shunting the shunt and it can be sorted.

This is all quite relevant to me at the moment, as I find myself needing a hefty 'power pack' for my astronomy, with the need for various supplies with different capabilities.

I could just package up a PC power supply in an outer case, then hang various things off it, but it would be better if I could get something giving about 18V - 20V at 10amps and then hang an assortment of buck and PWM controllers off it.

I've got some humungous toroidal transformers which I think have 2 off 35V secondaries. You're welcome to one if it's any use. At a pinch I could take one to MIdlands show. If these sound vaguely interesting I'll dig them out from their hidey hole and check the output. The advantage of toroidal is that if you're careful you can unwind some of the secondary turns to lower the output volts. I've even been known to add another winding when I wanted another very low power secondary.

Thanks Duncan, but as it will get a lot of lugging around I think I will look for a switch-mode PSU to suit - I may still have something lurking under the bench suitable despite the recent clearout...

Neil