would this work? SE EL84 circuit

[Alex Kitic]

Originally Posted by bikerhifinut

Mind you I am still puzzling as to why this SE amp in my collection used a weedy 85mA transformer, it just doesnt make sense to me. the output stage uses a 150R with a 47uF bypass on the cathode and when I crunch the numbers I get a current pushing 50mA which multiplied for 2 EL84 is over 20% more than that supplied by the transformer.
I had a crazy idea that maybe the labelling on the transformer was 85ma per 280V winding, in which case its well specced at 170 mA?
A.

I assume that your Armstrong amplifier is probably a model 44 or model 55. They mostly differ in the tuner section, but whichever the model, this meant the tuner section was as well powered by the amplifier, i.e. by this same transformer.

This is the link where I found the schematics:
http://www.vintage-radio.net/forum/s...24&postcount=8

The high tension secondary winding seems to be CT, which means that if the rating was 85mA, both windings were that same rating, i.e. 560V CT at 85mA. But it does make very little sense, since there are lots of other tubes to be powered besides the two output tubes which probably consume most of the high voltage current. To me this looks rather like a case of miss-reading... and the transformer should be powerful enough for the RH84 without doubt.

Originally Posted by bikerhifinut

...hope that I can find room for the 6.3V 1A filament transformer underneath...

What would be the purpose of the filament transformer? The original transformer should already have one filament winding - and if you want you could use the EZ81 rectifier (as in the schematics linked) on the same filament winding (although I do not particularly like that arrangement, it works and should not cause problems).

If you would rather go for the best sound, than the choice would be an additional 5V 2A filament transformer to be used exclusively for the 5R4 or 5Y3 you might want to install in your RH84.

[Alex Kitic]

Originally Posted by bikerhifinut

I have just done a bit of googling the RH84 and it does have a good worldwide following.
And I've just unearthed a few LM317's, I think it's an omen.

Well, any news? Did you try anything yet? I am surely not the only one interested to know the followup...

[bikerhifinut]

Nothing yet Alex,

A poor excuse I know but a very demanding full time job (and then some) and other demands on my time have precluded any serious construction. Scratch Projects like this take me literally weeks to complete. Most of which is the chassis/cabinet work as I have to use very basic tools and facilities.
However, I will be using your circuit rather than the 3-3 as I do like it's simplicity.
It looks fairly easy to hardwire without using too much in the way of tagstrips.
I'm away for the next 10 days and I hope to get it started then. Unless work gets too busy.

Andy.

[Alex Kitic]

But of course, Andy, you don't need to apologize for having a job
I would like to have one, too...

Just let us know about your progress and ask if in trouble with any choice you might need to make.

[bikerhifinut]

Hi all and Alex,
Made a start.
So far the hardware is on the top plate which I have made from a sheet of reinforced double sided Fibreglass PCB. It hasnt distorted or bent too far so fingers crossed. I'll probably put a few extra supports in around the transformers to take the stresses out.

I've gone for a basic hardwire, using a busbar earth. I realised just now that the ground lift resistor may need moving a inch further back to keep the reference as close as possible to the psu.
I got the heater wiring in around the edges without too much trouble, tightly twisted solid core wire was used, and I referenced the CT on the heater winding to the signal ground at the HT CT. I took the ECC81 heater feed from one of the EL84's hoping that would work ok. The EL84's have their own independent heater wiring from the LT winding down each side of the chassis.
After that it got a bit more tricky as I puzzled out the wiring. I am being careful to ensure the earth returns are made in the correct order. It is starting to look a bit untidy so I hope I dont introduce any problems.
I have used the LM317 CCS but havent bothered with the 21v zeners in favour of the 2.2k screen grid feed.
does anyone have an idea of the power dissipation in the screen grid resistor? I haven't got any 2.2k in the toybox bigger than 1/4 watt, a quick calculation suggests about 150mW is lost in the screen grid resistor which would suggest I would get away with 1/4 W resistors but its a bit close for comfort in the long term.
My other query for ALEX is this:
If I find the mains transformer is getting stressed, how does it affect the sound of the amp if I tweak the CCS down a few mA, say to around 42mA. I appreciate I will have less power but I fear I will be moving the operating point a bit too far.
I'll check in with progress reports, I can only do this build in the odd free hour at the moment.
If it shows promise, I fancy a go at the more powerful SE version with a KT88.

Andy.

[Alex Kitic]

Hi Andy,

So far so good, your build looks promising. Everything seems to have been done by the book.

1) g2 issue: the screen grid will most probably rraw an average of 6mA, thus even a 1/4W can do for a short while. But you should really bother with a 21/22V zener, because that will bring major improvement to your amp, particularly in combination with the CCS under the cathode.

2) If your power transformer is not up to the task, you will notice it as overheating (everything above 70 degrees C is overheating in my book, although many commercial amps actually run hotter than that. The second sign will be too low B+, and/or decreasing B+ with operation time (after a couple of hours you notice that the power transformer is too hot, and B+ is lower than initially.

If that happens, there is nothing to do but change the power transformer: shaving off a few mA will not cure the problem, but it will kill the sound. With this amp, too much voltage can help a little bit but is not beneficial, while too little current can destroy output power and performance.

Keep us informed on your progress!

[bikerhifinut]

thanks Alex.
A friend has offered a meatier mains transformer if the japanese one isnt good enough, but I have the one filched from an old Armstrong stereo radiogram anyway, and will have to then live with the rectifier heater off the same secondary as the others. I have a 1 amp CRT heater transformer too and that would fit under with a bit of a fiddle so thats an option. But for now I shall try the present job.

A

[bikerhifinut]

It's wired up.

Having a real issue with oscillation/motorboating.
If I remove the ECC81 first stage valve it stops.

So the main problem lies there.

I had to do a lot of tweaking with the PSU as the transformer shoves out 330V AC off load instead of it's rated 280V. Even under load the rectified DC off the EZ81 is 370V! all the voltages are raised, not surprising as the mains tx is specced for 220V mains. so 10% over on the heaters ain't a surprise.

I Junked the CCS on the EL84 cathodes as the measurements I got made no sense at all. 1.5V at the cathode suggests theres an issue. So I went with a 270 ohm cathode resistor for now which shows between 9.5 and 11V depending on the EL84, and which channel. Thats a current of around 38-40 mA, but the series resistor I used on each HT rail to the output valves seems to hold the HT at around 305V. I can tweak it a bit to get the 315 as per the circuit.

This Oscillation is a real sod. Although it can be heard on both channels, If i short out the input on the RH Channel it stops dead.
I did put i grid stoppers of 220 ohm carbon composition on the ECC81's and wonder if they might be to blame in some way?
The frequency can vary from classic 1hz to a higher frequency sounding like the motorboat is at full throttle so to speak.
I am wondering if the ECC81 which is a used valve from a trusted source, its a CV No thing with 12AT7 also stencilled on it, could perhaps be unhappy with its operating conditions.
Any ideas?

Andy.

[Alex Kitic]

As usual, for some reason I did not get a notification that there is a new post in the thread...

1) Oscillation / motorboating:
This is not motor-boating, but rather plain and simple oscillation. You said yourself that you do not get this when the input is shorted, which shows it is normal behavior. I always short the inputs of my amps when testing.

Of course, once you connect the source (preamplifier, passive preamp with CD connected, etc.) you will not have any oscillation and everything will function properly. If you continue to have some oscillation while listening to music (or cannot hear the music because of the oscillation) than there must be something wrong with your wiring.

To that sense, any ECC81/12AT7 garden variety will be good enough for the task (electrically), particularly if reasonably good testing or NOS. For the best sonic results I recommend a good driver tube, something like CV4024 or 6201.

2) Transformer issues
Obviously the transformer is powerful enough, which is a good thing. But the fact that you are having too high voltage on the secondaries is mostly related to the difference in primary voltage. I guess you are in the UK so your AC mains should be around 240V, while the transformer seems to be made for 220V. 240/220=1.0909 thus 305.45 on the secondary, which probably becomes at least 10% more without load... I would say "everything seems just fine" - but for the mains voltage, which incidentally on the long run might or might not damage your transformer for good.

Now there are several ways to solve the issue:
a) voltage bucking transformer - get some transformer with a 18-24V secondary and use it in voltage bucking connection, like shown on this picture:



Basically, you wire the secondary in series with the primary: in phase it will boost the voltage (I use one such configuration at home to get 230V since my mains is stable 223V) while in opposite phase (i.e. the other way round, as shown in the picture) it will buck the voltage, or reduce it. You need to measure your mains voltage to be able to define exactly the reduction needed, but I expect that you a secondary between 18 and 24V will do just fine.

Since the RH84 consumes around 80W of mains power (actually less), at 220V that is approximately 360mA. Thus the secondary of the transformer needed for the task must be able to supply at least 360mA - we are talking here about a 10VA transformer (i.e. very small).

b) You can burn the excess voltage on the primary with a resistor. If the excess voltage is 20V and current draw is 360mA, a 7.2 ohms resistor will do (an example: in practice there is no ready made 7.2 ohms resistor and you will obviously need to calculate your case here). The resistor will probably dissipate 7.2W, thus I would recommend using at least an 11W unit... but most probably the best solution in long term would be to use a nice heatsinked 25W resistor and forget about it once it is installed.

While I like the voltage bucking transformer solution, maybe the resistor is much simpler and cheaper to do?

I would not recommend leaving the primary at too high voltage and reducing the voltage on the HT secondary... if there is need to reduce it after getting the primary on the right voltage, there are beneficial ways to do it - like placing a resistor between the CT (center tap) and ground, or placing one resistor in series between each end of the HT winding and the respective anode: not only will either of these solutions lower the voltage to a required B+, but will not damage the sound (actually, it will improve the sound).

3) CCS issues
You have obviously wired the LM317 incorrectly. It is a common mistake, regardless of the correct designation in the schematics. This happens because the LM317 does not have the same distribution of pins as the 780x types, and the schematics symbol...



The LM317 in the schematics is a major improvement compared to the cathode resistor, both in sonic terms (which some might not be able to understand or will have difficulty assessing) and in reliability terms. So much so that I am not using cathode resistors to set the bias of power tubes any more. Thus you should by all means wire your LM317s correctly and use them.

4) You are not mentioning any zeners... as I said, that is also a major improvement over the plain resistor (so much so that I am not using resistors for g2 any more: the difference is literally appalling).

OK, let us know where this gets you!

[Alex Kitic]

BTW, once you solve the primary voltage either by transformer or resistor, all the voltages should be fixed, including the heaters.

If after that you need to lower the B+ because it is too high (EZ81 is rather efficient, using a 5R4 would give you much lower B+ but it would be a different pair of shoes altogether), you should apply the trick with the resistor between CT and ground.

This resistor needs to be calculated and tried for best results. It will need to be rather high power (W) since it will work for both anodes (double voltage drop) and will also operate in the DC domain (in total: 4x power dissipation in respect to what you might imagine that you need).