audio-talk

Boards arrived this morning.

Thanks Ray.

You're very welcome Steve, I hope they help.

Looking forward to this adventure.

Ray

In the absence just yet, of the components needed for the Aikido boards, I've done a "back of a fag packet" design for the 13E1 circuit, using components and valves I have in.

Not used my ECC99s yet, even though I bought them years ago (used 5687s instead) I thought I'd work up a driver stage based on them.

13E1 is choke loaded by a parallel pair of Triad Magnetics C-40X, 600mA, 320mH, 10 Ohm DCR chokes.
Grids are biased with a negative supply. Adjusting the bias for 2.5V across the chokes will give the required, half an amp of bias current. Whether it'd hold steady or not is another matter entirely.

Whichever method I use in the end, this is the layout:

I'll be up cycling one of the discarded cases from the push-pull monoblocks.
1KVA HT transformer on the right. Heaters and bias transformer on left.

The toroids will all go inside the case, so the top will be uncluttered.
Of course, there will be a new top plate and the case will be resprayed.

Not using a current source and using fixed bias will remove all the DC stability the original output stage had.

I'll use the current source then. No point in making problems; not when the solution to instability is so inexpensive and practical.

Right,
So I've decided it would be an act of folly, not to use the TL783 regulator, it is also, however the case that I have zero experience of the mathematics behind thermal analysis. So you can guess what's coming.

How do I size the heat sink that each TL783 will mount to?
I want to put an order in for bits tomorrow.

I've got no personal experience with tl783 but I'm almost sure(snigger) you will treat it the same as any plastic critter....i.e

voltage differential x current through will give you a watts radiated..so look for a sink with the correct spec.

sinks come with a c/w rating, i.e degrees C dissipated for each watt attached.

an example:

if you've got 125v across it with 500ma through it then you must lose 60 odd watts. a sink with a c/w rate of 0.5 will go to 30 degrees above ambient.

Cheers Ed.
At 150V on the plate, the 13E1 will bias at about 15V in the cathode when it is passing 500mA. So if we have 15V dropped across the TL783 at 500mA, then we are dissipating 7.5W, so the heatsinking won't be too onerous. Or am I missing something important?

New schematic below:

Steve, Wolfgang reminded me to size the heatsinks generously so that they stay at a stable temperature to avoid thermal issues making the amp unstable; he's a lot more experienced in this stuff and will be more specific.

For my 6C33C I purchased a Modushop Dissipante chassis that has two heatsinks, both rated at 0.45C/W so I should be fine!

Watching.


I can feel the heat from here.

Cressy Snr wrote: ↑Sun Nov 11, 2018 6:05 pm At 150V on the plate, the 13E1 will bias at about 15V in the cathode when it is passing 500mA. So if we have 15V dropped across the TL783 at 500mA, then we are dissipating 7.5W, so the heatsinking won't be too onerous. Or am I missing something important?

Dissipated heat by the regulator is Vcathode-1,3V times current. With a heat sink for 10-15W you would have some headroom if you decide to set the current a little higher out of some reason.

On the grid stopper: In the case of the 6C33 10k was the absolute minimum in order to avoid noise. I started at 1k because I always try to get the grid stopper as low as possible, especially with OTLs, or avoid them completely for buffers with gain.

Wolfgang wrote: ↑Sun Nov 11, 2018 7:57 pm
Dissipated heat by the regulator is Vcathode-1,3V times current. With a heat sink for 10-15W you would have some headroom if you decide to set the current a little higher out of some reason.

On the grid stopper: In the case of the 6C33 10k was the absolute minimum in order to avoid noise. I started at 1k because I always try to get the grid stopper as low as possible, especially with OTLs, or avoid them completely for buffers with gain.

Thanks Wolfgang,
At least I'm in the right ballpark. I'll derate the heatsinks about x2 as you suggest.
From looking at his published circuits, Bruce also tends to use large grid stoppers on his output stages, so it is, as you say, something to think about if you've found the same advantage.

So having acquired some more 13E1s yesterday, I'm set up with enough glassware to make this OTL amp a realistic proposition. I had two 13E1 already, but I remember one of them being distinctly dodgy in my old "Alexandra Palace" amplifier, which was a conventional transformer amp.

After a couple of months, the bias on the dodgy valve wouldn't hold, even with cathode resistors. I think it was probably gassy, as, on first power up, it had eaten most of its getter within a couple of hours. Also its grid resistors had been too high in value, which didn't help matters, but I was pretty wet behind the ears in those days and had no idea they were wrong.

Lack of stability was the main reason why I ended up ditching the amplifier. Pity as it (during the short time it had worked properly) had sounded wonderful, and at Owston driving a pair of Tannoy Devon speakers, belonging to Mr I, it had been a match made in heaven.

An order will go in for the rest of the components on Monday, so I'll be able to get started next weekend on it.

It's house alterations during the week and amp building at weekends. That's plenty to keep me occupied.

Got the order in for the bits, which are coming tomorrow; all except for the fooking heatsinks for the TL783s
Bloody thing has suddenly gone onto back order, despite the system saying they were in stock.
Delivery of that line number is not now scheduled until week commencing 17 December. I really wish they would warn before taking your money.
PITA but not the end of the world as there will be plenty to be getting on with, including a case to be built by our Ant.