Some questions about the GU-17

[ECV02]

Hi all,

I was recently reading up on Darryl's "Brisk 4" amplifier and decided I want to build one of my own. The main difference is that I want to use the full  output capabilities of the GU-17 (300v plates, probably 220ish on the screen). Looking at the translated datasheet, I have a couple of (probably dumb) questions about this particular tube.

Q1) The datasheet quotes a 12w maximum plate dissipation. Since this is a double tetrode, is this 12w rating for both plates, or is it 12w per plate? Darryl cited 7w per anode, but I can't find anything that corroborates that.

Q2) The datasheet also quotes a -80 volt bias on the grids. I'm no expert, but -80v seems like an absurdly cold bias. In my limited knowledge this would certainly bias both tetrodes into cutoff over the full 360° of the driving signal.

Q3) Can ANYONE find plate curves for this tube? I can only find screen and mutual characteristics. IIRC there is some way to plot plate curves from the mutuals, but my memory is fuzzy.

Any help would be greatly appreciated.

Thanks,
E

[PRR]

1) sheet says Dissipation anodes, which maybe means both.

2) the thing is the size, and heater power, of an EL84. You can't have 24W in that size bottle. I think it is 12W total, which you can take as 6W each. In cool-bias audio service instead of steady transmitting work, and at $4/pop, you can round this to 7W per plate.

The -80V condition is surely Class C RF, which is not useful for audio.

Taking the Class C 300V 11W out 12Wpd condition, the input power is almost 24W and that is 80mA.

The Ia/Vg1 graph on the sheet I found is clearly farked (awful lot of "0.1A" marks). I would be thinking -30V to -35V, start cold.

There ARE tubes with better data, and specifically aimed at audio amps. While $4 is a jaw-drop price, I can see more than $20 of bench-work to determine a happy operating point. And I think it wants a higher load Z and lower B+ voltage than most "real" audio amps, so finding iron will be tough.

[ECV02]

Thanks PRR. Helpful as always.

On the topic of a high primary impedance, could this possibly be solved by turning this into a parallel push-pull amp? I figure matching the tubes wouldn't be much of an issue; I could put one tetrode of one tube and one of the other tube on either side of the OPT. I may resort to some SMPS designing for B+. Cathode bias of course, that way I could get away with 6 watts idle on each plate.


On the topic of the output power in audio service, how much power should I expect to get out of this? As you said, 11 watts is only for Class C, so the tube could be run far harder under those conditions. Looking at the Brisk 4, I'm thinking maybe 6 watts per tube? With 2 tubes this would yield ~12 watts, perfect for a low power studio/local gig amp.

[gcwills]

You may be interested in my work with  QQE03-12 valves here: https://www.guitargear.net.au/discussion/index.php?topic=25648.0

[ECV02]

Thank you gcwills. That looks like a nice build. A question, what did the primary impedance of the OPT work out to? If I can get that number I have a feeling the rest wouldn't be too difficult. I've got a preamp chalked up already. Going for a real AC/DC sounding amp here.

[gcwills]

The output transformer was a 100v line transformer reflecting a 8K plate to plate load impedance.

[DummyLoad]

use a pair of them each bonded together as a single tetrode then you have a 14W tube. retube output stage for 8 bux... nice! with a 5K load at 250V B+ datasheet states 14W output.

bump B+ to 300V and get 17.5W output with a single tube. run a pair at 250V B+ with 25W 5K OT with 1/2 speaker (2.5K reflected) load, should make around 25-28W. see datasheet attached. run a single tube with 300V B+ with deluxe reverb 6.6K OT, but Vg2 becomes a concern - you need to bleed off 100V for the screens.

250V plan seems less messy. 14W with 250V to anodes 200V to screens - 5K load is common - or use 10K (6AQ5) OT with 1/2 speaker load.


--pete

[DummyLoad]

i overlooked one critical parameter in the datasheet in the previous post, namely the class AB conditions in those datasheets: it seems that page is for AB2 operation - e.g., they want g1 current to make that stated output. estimate about 25-30% or thereabouts less power output in AB1. my apologies for that oversight.

datasheet for AB1 conditions attached.

for 250V B+ i'd use the parts listed below.

OT - http://www.hammondmfg.com/pdf/EDB1750E.pdf

PT - http://www.hammondmfg.com/pdf/EDB269EX.pdf

--pete

[kagliostro]

I was just considering the same thing because I was reading this datasheet

https://frank.pocnet.net/sheets/183/q/QQE03-12.pdf

decidedly more easy to read the Philips english version   

BTW

for guitar amplification is bad to use AB2 instead of AB1 ??

Franco

[DummyLoad]

I was just considering the same thing because I was reading this datasheet

https://frank.pocnet.net/sheets/183/q/QQE03-12.pdf

decidedly more easy to read the Philips english version   

BTW

for guitar amplification is bad to use AB2 instead of AB1 ??

Franco

for guitar AB1 - for HiFi for maximum power and plate efficiency, AB2. AB2 needs careful design so as not to exceed current rating of g1 of the power tube.

IMO, don't bother with AB2 for guitar or HiFi.


--pete

[kagliostro]

Many Thanks Pete

Franco

[PRR]

With C-R coupling into a grid, you can't get into AB2 (grid current) mode. The one-way grid-current charges-up the coupling capacitor which de-biases the amplifier away from grid current. AB2 needs direct-coupled or choke-coupled grid drive. This costs more. We don't do it.

And probably the way an OVER-driven cap-coupled amplifier retreats from grid current is part of Guitar Sound.

[kagliostro]

Many Thanks also to you PRR


Franco