Confused about a search for KEL84 upgrades

[Paul Barker]

Unable to log into WAD forum and anyway its going off line isn’t it.

Why Im posting I was considering the circuit for my EL38 PP aspirations.

But apart from most of the open source info from people upgrading their kel84 I found load of amateurs comments, which Im not inclined to take with as much as a pinch of psalt.

So I found no comments about triode mode, and most comments about converting it to pentode mode. But no mention of adjusting the global feedback for pentode mode. Which in a few different amateur posts was seen as an upgrade.

I also found the recommendation to change the opt’s for stereo 20

It all read rather odd to me because the kel84 I heard, I think at the first or perhaps only wad fest I attended. In stock condition it was the best sounding wad amp I had heard up to that point. Later when I heard Greg’s push pull 300b I heard something phenomenal when he played organ music at Owston.

So when I embark on my EL38 probably Williamson esk. Im thinking Triode connected which the earliest Williamson was, I thought. Correct me if Im wrong about that.

[Paul Barker]

I think I know why non of the WAD guys have tried triode connected, or if they had they would probably get 3.4 watts. A kit builder who bought a 10 to 15 watt amp wouldnt have efficient enough speakers. Also the reason they migrated from ul to pentode.

But I found this Russian data sheet for triode connected AB, I also ran it in vatdiy and got even more power. But probably still not powerful enough to what they are used to having had the ul form already.

I think to benefit from triode connected you have to use a bigger beam tube/ pentode. Like the KT88 or my EL38.

So triode connected for a kel 84 is a non starter.

[Paul Barker]

OK so Mo asked me to fix his KEL84 many-years ago ;its been on mothballs all this time.

Oh, it needed a new ECL86 bythe way, MO had Yellow Writing Mullards by-the way. I bought two Mullards one yellow one white. Both sound great. In standard form the amp sounds just right exactly as I remember them from WadFests. Seems a shame to tinker with it.

The volume pot needs addressing. Maybe servisol will reform. Might not. Another thought is shunt it with a 20k resistor(poor mans stepped attenuator). KEL84 is excessively sensitive for modern times.

Ive looked at how it looks on the charts in UL as it comes. I dont appreciate the ratio of UL I know its close to Mullards idea. It doesnt look good to me, I would have spec’d 40% or more. There is too much 3rd for my SE prejudiced mind. I hate 3rd is a T shirt I should wear.

But it sounds a good amp from stock so I can only learn from this that the small amount of global feedback is cleverly chosen.

My job is to improve it for Mo.

First step try triode output operation and remove global feedback. I think 5.5 watts is as far as I want to push it, where the 3rd is only 1%. Thought for later are change input valve to ECC83. Get some dominant 2nd invited in, the concertina will then be the other half.

After that I may be done.

Im not interested in capacitor rolling or resistor rolling. Im after changing it. I have nothing bad to say about the opt Id be happy if someone had wound me that. Looking at past posts about upgrades a few people tried and liked Leak stereo 20 opts. Well those secondhand to day £400 each. Sowter will wind you some for £300 plus. I wouldnt bother. The transformers aren't in need of improvement.

If you want to make the amp a stereo 20 dig deep they are worth £900 today in working order. Horses for courses.

[Paul Barker]

Efficient enough put the factory for my speakers. My thoughts on operating in triode mode.

[Paul Barker]

Sounds loads better than the Cambridge it’s sat on!

[Paul Barker]

Fortunately there are triode conneted curves for the pentode section of the ecf80.

So to get a triode VA with 2nd h and not much odd harmonic I can keep the existing concertina phase splitter, triode connect the pentode the load resistor 32k is 3.7 x the triode operated ri of the pentode section, which has a mu of 44, which I think works.

There would have had to occured a lot of pcb surgery to mount a ecc81 or ecc83 in the place of the ecf80. Fortunately some radio hams had traced the curves for triode mode.

This free’s me up to just lift one leg of the components I dont want and the global feedback. Operate the output valves as triodes for 5.5 watts of lush triodes. 1% 3rd from the output, but also amplifying the 2nd of the VA whatever that is. This input valves as is the only possible way to get close to the se triode sound.

I think it is worth the lower power. I must order some servisol to clean up the volume pot.

[Paul Barker]

Fortunately there are triode conneted curves for the pentode section of the ecf80.

So to get a triode VA with 2nd h and not much odd harmonic I can keep the existing concertina phase splitter, triode connect the pentode the load resistor 32k is 3.7 x the triode operated ri of the pentode section, which has a mu of 44, which I think works.

There would have had to occured a lot of pcb surgery to mount a ecc81 or ecc83 in the place of the ecf80. Fortunately some radio hams had traced the curves for triode mode.

This free’s me up to just lift one leg of the components I dont want and the global feedback. Operate the output valves as triodes for 5.5 watts of lush triodes. 1% 3rd from the output, but also amplifying the 2nd of the VA whatever that is. This input valves as is the only possible way to get close to the se triode sound.

I think it is worth the lower power. I must order some servisol to clean up the volume pot.

Then if the coupling caps are polypropylene I might get some pio’s for it.

[Paul Barker]

Hoping a kel84 owner or two might notice this thread and comment .

[Paul Barker]

Couldnt find triode curves for the pentode section of the ecf82 but the pentode curves dont look a mile apart so I may order a pair to try for experimenting. Same pinout, and triode section should concertine too.

[Paul Barker]

Modern equivalent of servisol. When two Allen Bradley 33k resistors arrive I’ll commence triode conversion and clean the pot.

[Mike H]

Just to mention, of the ECF80 and its ilk, the triode section is pretty useless for anything except as an RF oscillator, which of course was its intended use. Something about the distance between its grid and its cathode, makes it a poor amplifier. I was using ECF82 for the Maplin Newton pre-amp and just about made the triode section work as a cathode follower, but there was some loss of Voltage amplitude output. In hindsight the pentode section is probably best configured as a triode, for audio use, if as pentode it wants to oscillate in the Megahertz range, if allowed. Not surprisingly as its intended use is typically first FM IF stage out of a FM tunerhead. See Leak Troughline etc. for application examples.


If I was doing Newton again it would be with ECC83 and 82, more suited to audio range.

[Greg]

Paul, I’ve responded to your email request. Your email address as linked to this forum did not work so I have sent to the address I have for you in my contacts. Hope it works.