Yes, replacing oil filled caps and old electros as a first step is pretty well recognised as standard procedure.

C18 and C19 between active and chassis and neutral and chassis need to be Y rated. ( I believe X rated caps are for across active and neutral.)

I've got Y's on order but can only get 0.01 to replace the 0.001's, is the value critical?

Neutral is normally at earth potential, so the principal requirement is that the cap is able to withstand the voltage and what is riding on the mains.

I have miles 24KV Aerials that bring in a lot of RF. I normally have a 0.1mfd Neutral to Active, and 0.0047's from both to earth.

I see no point in going higher than 0.01mfd as the lower values are better for RF. I also have MOV's on the lines as we get a lot of lightning.

If you go too big you also start drawing higher unbalanced current, which is not helpful if you have an RCD , also the cap my get quite hot.

Marc

Neutral is normally at earth potential, so the principal requirement is that the cap is able to withstand the voltage and what is riding on the mains.

Good point.

I think both X and Y rated caps are designed to fail open circuit for safety purposes.

The caps have finally turned up and I've replaced all the old oil filled and electrolytic caps. For some reason I forgot to include the 0.03μF in my order to replace C3, the one known faulty capacitor. The only values I have that are close are 0.025μF or three 0.01μF, I decided to go with a 0.025. This resolved the missing 400Hz AF but it's now about 365Hz due to the different value cap. The 1000Hz is as close to spec as I can measure with the CRO.

Apart from the missing 400Hz AF now being back, replacing all the other caps made no difference to the RF output and if anything slightly increased the 50Hz ripple. I replaced C13, a 0.001 ceramic for no other reason than it was very easy to get to and is on the B+ to ground. Surprisingly this improved the shape of the output on range D but it's still clipped on ranges A, B and C.

I'm not too sure of my next move but I think I'd better check voltages again and compare to previous measurements. I'll also replace C14, it's easy to get at and it's on the output. If it was leaking could drag the output down. I'm starting to wonder if there could be a problem with the 12BH7.

Warren

Scraps, how did this end up?

I am getting around to recapping my LSG-11.

It still clips the output, I come back to it every now and then. New valves made no difference and apart from the selenium rectifier nearly everything has been replaced. I use it regularly for aligning radios and it seems to do the job okay. I think Marc might have been on the right track when he mentioned the voltages for these LSG11's are right on the limit for the valves. I did wonder about the voltage on the oscillator grid of the 12BH7 going too high and drawing grid current causing the clipping. I experimented with the value of the 15k resistor which made no difference.

I quickly checked out mine yesterday, and every cap is within spec and clean, and the twin 10μF filter cap is reading a low ESR. There's no sign of waveform clipping on the output -- I haven't looked either side of C12.

In this case I'm not inclined to fix what isn't broken, but I will probably replace the selenium rectifier even though it's showing 104 volts DC at the filter cap.

It would be interesting to put the CRO on the B+ now that everything has been replaced. New valves would have eliminated the cathode heater theory. Do make sure the chassis& metal case are mains grounded.

Clipping tends to result from overload, however I would check all resistors feeding the 12BH7 and Its wiring. The CRO should be able to see the ripple on B+. Just working on the knee of the operating curve which is not really shown at that voltage, is enough to cause an issue where one part of the cycle is amplified more than the other.

Modulation efficiency normally falls as the frequency increases.

If AC is getting into the modulation, it will put hum into the system so do make sure that an earth point shared with a heater has good bonding.

Low B+ if it is not an overload, can be a wrong (low) filter cap bad filter caps & new ones do fail, or the rectifier. You may have to add resistance due to the increased efficiency of the diode.

Marc

You may have to add resistance due to the increased efficiency of the diode.

... or string diodes in series to match the selenium voltage drop and perhaps add a resistor to limit in-rush on power-up.

My Tech-20 Clone of that says 120VAC out of the transformer, as a point of interest what voltage rating are the electrolytic's?

I sold the LSG-11 I had, but before it left another turned up. All of the oil filled caps were U/S & a divider resistor was cooked. Its OP was weak, so it has new valves.

This means that between it and the DSE one there is a chance of getting element & rail voltages. Some 450V Electrolytics are crap & I have had failures in a short time & I suspect that a daily driver that is coming back soon has them in it. One set I have had fail in two years, this lot will be 2008.

The one I have had Fox components the others, no: Different builders marked on them.

Marc

as a point of interest what voltage rating are the electrolytics?

The dual section electro is 150V.

Mine has been earthed by the previous owner and there's a 500V RFI cap between active and earth.

500V cap is not my idea of sensible; The "X" type mains approved cap has 275V~ on it, or 275 VAC and lots of standards association stamps on it like UL.

Get rid of it, or replace it. 0.1mfd across the mains I find useful & 0.01mfd from A to E and N to E. I have a lot of RF on the mains.

The dual section, if it is original, has likely passed its use by date and is liable to be pulling down the B+. I tossed the ones in the one here. As said, CRO will see the ripple.

There is no surge rating on most electrolytic caps here. As that has a metal rectifier there will be a surge that can get to twice EMF before the heaters do their thing. So the caps need to handle that. Several voltages have been phased out, so nothing under 300V & the closest may be 450V.

Marc

The B+ at the electro is as per the schematic that Scraps has (~105V). The electro capacitance reads okay and has a low ESR as I mentioned, but it's very old and due for retirement.

I will replace the RFI filter.

(As per the Interference thread, I too have RFI but I'm working on getting that problem eliminated.)

ESR is not what I am concerned with, as much as their leakage. As it is chemical there must be some or it will loose the oxide layer it needs to maintain polarity. I built a reformer on the end of the bench PSU mainly for newer caps that choose to go feral & leak when new. General formula 0.01CV for leakage in micro amps. Something like 10mA is not really acceptable.

My principal source of mains RFI is the fact that there are several Radio & communications transmitters around here (aside from Wallwarts & other electrical stuff) this gets picked up by the 24KV Utility aerials and then rides on the mains.

I have a range of capacitive filters on circuits to keep it out of the Radios, Computers & TV's, some with MOV's as well, to get rid of surges.

Marc