HT smoothing choke faulty in Ekco A23 radio

[indigo.girl]

Hi Everyone,

I'm working on an Ekco A23. Have carefully started it up on a lamp limiter as usual and then progressed to full mains. Rectifer valve (V5 / AZ31) seemed to be working a bit high with 310 VAC (rather than 275V) and I was getting 360V delivered to the reservoir cap and smoothing choke. There was 342V on other side of choke which seemed about right (Z=550ohms, so with 18V across the choke this means its drawing 32mA of current which about the same all the valve anodes mA added up).

I tested all the valve voltages and was getting zero voltages at the grid for the output valve (V4 EL33) and zero volts at the anode for V3 (EBC33). I traced this problem back to R27 - part of the HT potential divider that feed both valves - this was a 10K resistor that had gone open circuit.




So I replaced it with a 10K resistor - but hadn't considered the power rating and so when I switched on it burnt out. . I didn't have a 10K high power resistor so I decided that as a test I would simply connect a wire in place of R27. The R27:R28 potential divider was 10K:68K so R27 was only going to reduce the voltage from 340V to 300V. So with R27 replaced with a wire I would expect to see 340V at R29 rather than 300V - this would be approx enough to test the radio.

However, upon switching on I now only get 40V HT! The voltage across the HT choke is now 140V (180V before and 40V after).

So- has the resistor burn out now damaged other parts of the circuit? Has the choke now gone faulty (note: it has been getting very hot all throughout my testing and the glue(?) in between the sheets has been bubbling a little). C35 also looks dodgy (I spotted a drip of something on its case). My plan is to order and replace the electrolytics and the key resistors feeding V3 and V4 (R27, R28, R29, R21). However, before I get spending money I want to check this choke issue isn't fatal. If the choke is kaput is there anything else I can use in its place? Can I just get away with the reservoir and smoothing caps?

Any thoughts much appreciated
Nicola

[ms660]

Ten out of ten for honesty....

I would replace the 10k resistor again (R27) It doesn't have to be high wattage, 1/2 watt should do, replace C35 with the nearest value you have to hand and of sufficient working voltage, make sure the voltage tapping on the mains transformer is on the high tap, not sure what that is but I'm guessing 250 volts, power up and check the voltages again, the choke might have survived.

Lawrence.

[turretslug]

In a situation like this, with a feed resistor doing more than one task- here feeding both the output valve's screen grid and the EBC33 1st AF amp anode resistor- it's wise to run through all the implications with a bit of finger-to-the-wind maths. An output valve screen grid typically takes a few mA in normal operation, equating to a few tens of volts drop across 10k. In other words, somewhat more than the 10-20V that would result from just the typical 1-2mA anode current of a radio 1st AF amp stage. The anode current of a tetrode/pentode is strongly influenced by the screen grid voltage and it's pretty common to find that the output valve in most domestic radios is run rather close to its dissipation limit anyway- that's only commercial sense, really- and increasing the screen grid voltage by probably 40-50V is likely to quickly get it rather hot, bumping up the anode current somewhat is also stressful to choke, mains transformer, output transformer.... Add that to the fact that the HT secondary voltage seemed a bit high anyway, and EL33s and decent smoothing chokes aren't getting cheaper. In many domestic set output stages, the screen grid was connected straight to smoothed HT, saving an R and an electrolytic C- the fact that this set didn't do that suggests that the EL33 needed that dropped screen voltage rather badly.

I know the temptation to crack on, and I've occasionally done a quick get-it-going lash-up- but I've also occasionally regretted the tedious consequences of taking a short-cut! Very much hope that doesn't sound like any sort of ticking off, we probably all need a practical case of something that came back and bit us to make us wary in future Good luck with the set, it otherwise sounded promising.

[Chris55000]

Hi!

Unfortunately the contretemps in the h.t. supply might have partially stripped the emissive-coating off the rectifier filament - those AZ31s with directly-heated cathode filaments don't take kindly to being over-run!

Before you condemn it, disconnect the h.t. lead from the choke to the rest of the receiver circuits, get hold of a standard 15W "pigmy" filament lamp (NOT energy-saving nonsense!) and try this between the choke and chassis - if it's dim with the bulb on it's own it's another AZ31 time I'm afraid - however you MIGHT get away with a 5Y3/5Z4 or similar as a test - the pinout's the same!

If the AZ31 has survived the onslaught you can use the pygmy lamp in series with the choke as a limiter and go on with conventional fault-finding in the AF & O/P stages!

Chris Williams

[Chris55000]

PS!

The primary of an old TV frame-output transformer with the secondary ignored will make a useful trial smoothing-choke if you've got one - these were wound for a peak frame-output valve current of at least 70-80mA and had to have plenty of inductance to enable enough scan-power at the end of scan to be obtained.

Chris Williams

[indigo.girl]

Originally Posted by turretslug

In a situation like this, with a feed resistor doing more than one task- here feeding both the output valve's screen grid and the EBC33 1st AF amp anode resistor- it's wise to run through all the implications with a bit of finger-to-the-wind maths. An output valve screen grid typically takes a few mA in normal operation, equating to a few tens of volts drop across 10k. In other words, somewhat more than the 10-20V that would result from just the typical 1-2mA anode current of a radio 1st AF amp stage. The anode current of a tetrode/pentode is strongly influenced by the screen grid voltage and it's pretty common to find that the output valve in most domestic radios is run rather close to its dissipation limit anyway- that's only commercial sense, really- and increasing the screen grid voltage by probably 40-50V is likely to quickly get it rather hot, bumping up the anode current somewhat is also stressful to choke, mains transformer, output transformer.... Add that to the fact that the HT secondary voltage seemed a bit high anyway, and EL33s and decent smoothing chokes aren't getting cheaper. In many domestic set output stages, the screen grid was connected straight to smoothed HT, saving an R and an electrolytic C- the fact that this set didn't do that suggests that the EL33 needed that dropped screen voltage rather badly.

I know the temptation to crack on, and I've occasionally done a quick get-it-going lash-up- but I've also occasionally regretted the tedious consequences of taking a short-cut! Very much hope that doesn't sound like any sort of ticking off, we probably all need a practical case of something that came back and bit us to make us wary in future Good luck with the set, it otherwise sounded promising.

Oh damn - so it looks like my bypassing of R27 was mostly likely the problem rather than the initial burn out of the replacement R27. A self inflicted wound!! How annoying.

I've found some more spec data in the manufacturers datasheet which says that AZ31 should deliver 300V and 57mA to the choke. The choke is 550ohms so with 57mA running through it that means a voltage drop of 30V. So the smoothing HT voltage should be 270V (I was originally recording 340V). The R27:R28 potential divider would have dropped this 270V HT voltage to 230V for the feed resistors for V3 anode and V4 grid. So by bypassing R27 I was feeding 340V rather than the spec of 230V!! So no wonder its blown stuff. Hindersight is a wonderful thing...wish I'd spotted the more comprehensive manufacturer datasheet...lesson learned.

OK - so now I have a new set of problems:

1) AZ31 only giving 180V at filament (spec says 300V)

2) The choke has a voltage drop of 140V across it (rather than the 30V expected). This reduces the HT voltage to a meagre 40V. But more worryingly means that 250mA is being drawn through it (rather than the 57mA spec). Why is so much current being drawn? (btw the measured ohms is 650 ohms so the choke is still close to spec resistance). Is this strange effect on the choke a direct consequence of my faulty AZ31 or is it indicative of a faulty choke. (I'm trying to assess how many components are broken and work out whether its worth fixing still).

[TrevorG3VLF]

The choke may not change resistance significantly even if faulty.

Find out where that excess current is going. The set does not have series heaters, just pull the output valve out. Once you have solved the short, the voltages should come right.

[Herald1360]

The choke problem may be the winding shorting to the frame. Try disconnecting it and checking each end to the core as well as end to end resistance.

[indigo.girl]

Originally Posted by Chris55000

Before you condemn it, disconnect the h.t. lead from the choke to the rest of the receiver circuits, get hold of a standard 15W "pigmy" filament lamp (NOT energy-saving nonsense!) and try this between the choke and chassis - if it's dim with the bulb on it's own it's another AZ31 time I'm afraid

All may not be lost. I have tried your idea Chris about using a lamp (I only had a 42W energy saving one unfortunately) in series with the choke and going to chassis.

The bulb lights up well and I can measure 275V AC at the AZ31 filament (spec levels!). There is 233V DC before the choke and 145V DC after the choke. So I'm hoping that this means the AZ31 is actually OK.

[indigo.girl]

Originally Posted by ms660

Ten out of ten for honesty....

I would replace the 10k resistor again (R27) It doesn't have to be high wattage, 1/2 watt should do, replace C35 with the nearest value you have to hand and of sufficient working voltage, make sure the voltage tapping on the mains transformer is on the high tap, not sure what that is but I'm guessing 250 volts, power up and check the voltages again, the choke might have survived.

Lawrence.

Hi Lawrence - if the circuit was working properly with a 270V HT line and drawing sensible currents would you have expected at 1/4W 10K resistor to have burn out? I'm wondering whether if I find the source of the problem and voltages and current return to spec then I can get away with a standard resistor at R27. Would you have expected a 1/4W resistor to blow in that position?

[turretslug]

"42W energy saving" probably equates to around 10-13 real electricity watts. I'm not au fait with the details of the inverter circuit in these, but if they're made for 100-240-ish V worldwide usage with a bridge rectifier input, then it sounds as if it would be drawing something a bit less than 100mA from the 145VDC it's getting out of the choke. Don't run it that way for too long.... for the sake of rectifier, choke and mains transformer... though things don't sound disastrous so far.

Have you checked C35?- if it was badly electrically leaky, that could explain much of the previous woe, and could have instigated the failure of the original R27.

I would use a 1/2 or (preferably) 1W resistor for R27.

[indigo.girl]

Well - I've been lucky tonight! I decided to put in another 10K R27 (still 1/4W) and run it again but this time with the bulb between the choke and the rest of the circuit as suggested. Before doing it I replaced C35 (a 4uF @350VWkg) with a 10uF @400VWkg). Once removed I could see the C35 was badly damaged and leaking out the ends.

I turned it on and measured valve voltages and I was getting close to spec values (lower as expected due to the bulb in there). Buoyed with success I took out the bulb and ran the circuit proper. Success! All the spec voltages achieved. So far I get just crackled from the radio and no stations but I injected an audio signal into the volume pot and go audio out the speaker. So the basics are there - just need to check out the receiver tomorrow.

Phew - so looking back the initial fault was R27 open circuit so no current was getting to the output grid or the V3 anode. Hence the higher than expected values at V2 and V1 too. When I bypassed R27 with a wire that put the full HT of 340V across C35 (working voltage only 350V) and so it blew it making it go short circuit. This effectively meant that the full 340V HT was running straight to ground through C35 drawing masses of current from the choke, rectifier and transformers and causing the scary voltage readings! Luckily not fatal to the rectifier valve or the choke - but I think I've been lucky this time...



PS- R27 gets very hot to touch to I will be replacing with a 1W power rating

[julie_m]

42W "Energy Saving" sounds like a Halogen lamp -- compact fluorescents usually go up to 20W and LEDs to 10W. So it will at least have a positive temperature coefficient.

[turretslug]

I expect that you're spot on there- I was thinking of the duplicitous and optimistic equivalent "watts" often used in marketing CFLs.

[ms660]

Originally Posted by indigo.girl

Hi Lawrence - if the circuit was working properly with a 270V HT line and drawing sensible currents would you have expected at 1/4W 10K resistor to have burn out? I'm wondering whether if I find the source of the problem and voltages and current return to spec then I can get away with a standard resistor at R27. Would you have expected a 1/4W resistor to blow in that position?

First a correction to my earlier post, should be 1 watt, sorry about that, got the practical value mixed up with P diss.

Here's how to work out the wattage, power dissipated in a resistor = Isquared*R, so for R27 (10k) first determine the current flowing through it ignoring V3 anode current and V4 screen current for the moment, R27 (10k) is in series with R28 (68k) that's a total resistance of 78k connected across the HT line, ohms law applies, so with an HT line of 270 volts the current flowing through R27/R28 (78k) is approx. 3.5mA, the same current flows through both resistors so the current through R27 = 3.5mA.

R27 also supplies V3 anode and V4 screen grid, an approximation of those currents based on the Trader sheet total up to 3mA, that added to the 3.5mA is 6.5mA, so 6.5mA is flowing through the 10k resistor (R27) and P = Isquared*R, so 0.0065A squared*10,000 = 0.422 watts dissipated.

A 1/4 watt resistor job would not be suitable, a 1/2 watt job would be ok but it would be just about at it's rated limit, a 1 watt job would be a good choice.

Lawrence.

[Chris55000]

Hi!

You're getting to the home straights now - if the AZ31 can light up a 42W E.S. lamp it will have adequate emission for the set - only other tip I'd like to suggest (now RS & Farnell will supply individual consumers on a pre-paid basis) is to use the power metal-oxide PR01/PR02 resistors for replacement purposes - they don't look too far out of place in an old set and I've found them to be well within the 5% tolerance marked on them - typically they're within 1% usually! - a 10K PR01 for the top resistor and a 68k PR02 for the bottom will be perfect for your set and not too expensive even in 5's!

Chris Williams