Remember These?

[Cressy Snr]

PCBs have been fabricated and are making their way across from China; ETA middle of next week.

Blimey that was quick!

[Ray P]

Blimey that was quick!

Wouldn't like to delay your return to musical heaven!

[Cressy Snr]

Green, cloth-covered, single core, 18AWG, 600V rated pushback wire.
White PVC/cloth covered, single core 20AWG, 600V rated wire
Fender lamp bracket, amber jewel and bulb.
2 off 9-pin ceramic noval sockets for PCBs

Bits are arriving now. Steampunk plan is no go. Mid 20th century modern is the current vibe to match the speakers.

[Cressy Snr]

Bits are arriving now. Steampunk plan is no go. Mid 20th century modern is the current vibe to match the speakers.

Something like this:

But darker wood.

[Ray P]

I like the cut of your jib sir.

I'm not as adventurous and mine will be black and silver mainstream style...

[Cressy Snr]

Two input/driver PCBs arrived today, courtesy of Ray.

No excuse for not starting now.

Thanks for those.

[Ray P]

You're welcome Steve. I hope they help with your build and just shout if I can help.

[Cressy Snr]

The "Not OTL" breadboard has been removed finally from the room, in preparation for being plundered for transformers, chokes, bridge rectifiers and caps. I'm back to the solid state situation for now but as they say, "No pain, no gain"

I'll populate the driver boards tomorrow, and start making holes in metal at the weekend.

[Ray P]

...start making holes in metal at the weekend.

Steve, here are the dimensions relating to the PCB that you'll need, (all in mm, viewed with the row of pads for inputs/outputs at the top).

1 - Including the output caps section

Board size: 84 x 97
Monting holes: 4 x 4 from each corner (i.e. on a pitch of 76 x 89)
Centre of valve base: x=42, y=61

2 - without the output caps section

Board size: 84 x 70.75
Monting holes: 4 x 4 from each corner (i.e. on a pitch of 76 x 62.75)
Centre of valve base: x=42, y=34.75

Is that OK?

One further thought as forewarned is for-armed. On one occasion in the past, with a PCB using the same 9pin valve pad layout I found that the pins of some valve bases wouldn't fit into the pad holes. If you find this I recommend using something like a dremel to thin down the pins rather than taking the, seemingly, easier route of drilling out the holes; if you drill out the holes you'll destroy the through hole plating and soldering the bases to the PCB will get a lot harder and will risk lifting pads - guess how I know.

[Cressy Snr]

Thanks Ray. That'll save a lot of messing about.
I offered the sockets up to the boards yesterday and as luck would have it, the socket tags fit in the holes fine. Nevertheless, thanks for the heads up on that possible issue.

[Ray P]

No prob. I can send you a drawing for use as a drilling template if it helps?

[Cressy Snr]

No prob. I can send you a drawing for use as a drilling template if it helps?

Hi Ray, I'll probably manage thanks

Breadboard is now bereft of bits and in the outside loo.
I'm now poring over the layout. Can't afford to mess that one up.

[Cressy Snr]

All the parts have now arrived from RS.
I've decided that this is going to be an integrated amp with five inputs rather than just a power amp, which would require a passive pre to do all the signal routing. There may be a fly in the ointment at the 'mo though.

I had a conversation with Nick yesterday re the TL783 based constant current sink and I am a bit dubious about using it.
It may be something about nothing, and as Nick pointed out, RS spec summaries are not always right, but I hadn't noticed before that the minimum input voltage for the TL783 is 21.5V (according to RS).

The Texas Instruments application notes don't appear to mention an input voltage, but if instead, and with a bit of reading of Morgan Jones about solid state CCS' in general, we take the RS, 21.5V figure as meaning that the the TL783, being a high voltage regulator, requires a minimum of 21.5 V across it in order to operate correctly, then we are way short of that requirement when using it in the cathode of the 13E1.

Let's look back at my experiences with the breadboard and note that I had the 13E1 cathode voltages wandering about early in the project and put it down to grid current caused by too high a grid resistor value. Indeed, this wilful wandering was reduced by lowering the grid resistors to 39K, but it wasn't eliminated entirely. If we extrapolate this, it is possible that there were two fault conditions going on at the same time: too high grid resistors and a CCS not operating correctly. One fault got eliminated but not the other.

Sod's law dictates of course that this possible problem only made itself known after I'd dismantled the fecking breadboard,
otherwise I could have shoved an LM317T in the cathode which only requires 3V or so across it to start regulating, to see what happened.

This feeling I have that the TL783 might not be suitable for a 13E1 cathode CCS, when used in a single tube OTL config is further reinforced when we look at the bias voltage that the 6C33C develops when being run at 475mA, in Wolfgang's circuit:

In his cct, the 6C33C develops 24V across the TL783 when it is run at 475mA (within the 21.5V min voltage spec) but the 13E1 only develops 12V for the same current (way below minimum for said TL783).

I could of course be wrong and my logic is notorious for being completely off the mark and I hope I am wrong, but why the feck do these things only show up when you think you've got the job cracked?

I have a couple of scruffy looking LM317s and half-a-dozen LT1086-CM I could try but these would need zener protection.

[Nick]

Reading the device sheet I see that the electrical parameters are given at a in to out voltage of 25v unless otherwise noted, and the lowest value used is 10v. So it will certainly work with only 10v across it. The most important entry I would suggest is:

"Due to the dropout voltage and output current-limiting characteristics of this device, output current is limited to less than 700mA at input to output voltage differentials of less than 25V."

Looking at the relevant graph min differential at 500ma looks to be 10v, but remember in this case it needs to cover the range of output voltages on the cathode follower.

Personally if I was going to do something I would use either a single high voltage depletion mosfet or a pair of them or one depletion and one enhancement mosfets as a cascode. Or a C4S or any other number of constant current loads. I dont think it needs to be that good a constant current as at AC at least you are strapping a 8R resistor across it.

[Cressy Snr]

Cheers Nick.
I have a few DN2540, Paul gave me years ago, but these will only stand 500mA constant draw, so they'd be right on the edge like most things where OTLs are concerned.