YARPS, finally some progress!

So I have almost all of the parts (yeah I know) for my YARPS power supply. As part of the order I got myself a prototyping breadboard.

I've knocked up one half of the yarps power supply, ie the positive side, and hooked it up to my old model railroad transformer which has a 15V output.

Parts still to be ordered are the dual secondary 15V transformer and some 50K trim pots. However I had enough to make one side of the dual rail supply.

It didn't smoke, and the CRCRC part of the PS works very nicely Pics attached at the end.

I need to do a further test with BC560c's with different HFE's to see whether this affects the output voltage (my suspicion is that it will). I put a 36K resistor (which I had at hand) in place of the 50K pot and only got 8.6V even though spice said I should get 9.13V with that value resistor so something isn't working as per the sim, the use of a pot certainly seems to be required to get the desired voltage as the standard formula simply doesn't work at all.

Edit 24/4/2011 I tried three different hFE transistors today and although I measured differences it could partly be due to settling (I suspect as the LT317AT heats up it varies a bit). However there does seem to be a trend.

I first tested with hFE of 395 and voltage started to settle around 8.63V, then I tested wit hFE 438 and the voltage started to settle at around 8.60V last I tested with Hfe 346 and the voltage started to settle at around 8.64V. This was rather unscientific though as I did not let the device cool completely between tests, and did not measure after an exact amount of time either... there is a very small difference (around 40mV), but hardly anything to worry about, I will however be using matched transistors for the positive and negative sides of the reg, just because I can

Pics attached are as follows:

1st Ripple right after the rectifier across the 10,000uF cap (20mV/div)
2nd Ripple across the second (4700uF) cap after the first 3r3 resistor (5mV/div)
3rd Ripple across the 3rd (4700uF) cap (2mV/div)

I thought the trace on the 3rd pic was showing a lot of noise so stepped up the timebase until I could see what is in the fourth pic

4th shows interference across third cap (1uS/Div) This appears to be approx 1.1Mhz

after seeing this I decided to see what the scope was picking up with the probes floating (just sitting on the carpet).

5th picture is the scope probe not connected to anything. 2mV / div 1uS/div

I had a suspicion this could be noise coming in on the TV arial and contaminating the ground (as I get bad ground loops when the TV arial is connected) so I took the tv arial out of the wall socket and it was much better.

6th pic is scope probes floating with TV arial disconnected. (as you can see quite a difference, so noise is coming in over the mains and into the scope probably via the ground. Not sure if this means I need to service the PS of the scope...

The 7th pic shows the measurement across the 3rd cap with the TV arial disconnected (quite a difference to the third pic!!!) I'm not sure if the scope is picking up noise or if it is coming from the transformer, as both were plugged into the same mains circuit as the TV set. the measurement was again 2mV/div 10 ms/Div

The reduction in ripple from directly after the bridge to the input to the regulator is basically from ~110mV / div to less than 2mV/div. Load for these tests was approximately 125mA (which is less that what the real load will likely be, but the best I could do with the low wattage resistors I had available.

8th pic (the business end) is of the reg output (2mV/div 10mS/div) No sign of the ripple but certainly still signs of the noise. I was getting tired and didn't think to take another measurement at 1.1Mhz of the output of the reg. Will have to do that later.

One thing that was a bit odd was that the waveform on the scope across the third cap was jumping quite a lot, it was moving 1 or 2 divisions up and down, I also noticed that pretty much all of the measurements were not symmetrical around the centre line (even though with the scope probe shorted it was right on the centre line). Will need to get to the bottom of that.

Anyway it was a useful exercise for me to do, and it shows that my Overkill on the caps certainly does what it is intended to do (ie get the ripple to something pretty small before the reg). Hopefully soon I will have the transformer and pots, (and some more breadboard jumpers, as the ones that came with the breadboard were only just enough to do one side). I would like to do the negative side as well before starting to solder to my verro board.

The 9th pic is with the earth of the scope on the earth of the board, and the probe connected to a nearby earth point on the board. as you can see it centres on the scope, but there is still a hint of noise...

Oh and the final pic is one of the breadboard, as you can see I didn't cut any component leads, so it is far from ideal from a layout perspective I used a bulldog clip to clip a heavy steel pc blanking plate to the reg for the tests as I don't yet have the heat sinks for the regs either.

Attached Thumbnails