The Ukrainian post is working quite well. I recently ordered a box of GM70 from there and it was here in no time
Indeed, and everything is tracked at every step. Two weeks delivery that includes customs checks etc is very impressive, even without taking into account the current situation.
I want to build a single ended amp that can deliver 25W+ with reasonable distortion, but without having to use >800V B+.
Great! As you probably know this tube has been modeled on SPICE if you want to run simulations before building the amp:
https://www.diyaudio.com/community/threads/vacuum-tube-spice-models.243950/post-7079225
With Tubelab's UNSET configuration you can probably get around 40 Wrms at 5% THD in SE mode, if you can find a transformer capable of that power on the full audio range.
https://www.diyaudio.com/community/threads/vacuum-tube-spice-models.243950/post-7079225
With Tubelab's UNSET configuration you can probably get around 40 Wrms at 5% THD in SE mode, if you can find a transformer capable of that power on the full audio range.
Last edited:
Yup, it was me who asked for the model 🙂
I'm looking to start with something simple, EF184+Mosfet+GU-72 in triode mode, with a Hammond 1627 or 1630, 520V B+.
I will look at the UNSET configuration, thanks.
You can have a look here: https://www.diyaudio.com/forums/tubelab/340856-unset-coming.html
The basic concept is to drive the output tube from its cathode through a p-mosfet source follower, and give local feedback to the tube from anode to g1 thorugh a voltage divider. g2 is connected to a low impedance fixed bias optimized for the chosen loadline.
The advantage is that you get triode-like curves, but g1=0 is shifted on the left of the plot like a pentode curve (so you swing more, so you get more power), and you share the cathode current with the source follower that is usually set at at some tens of volts, so you can increase the current flowing through the tube, then getting more power from it.
Lab tests have shown efficiencies very close to 50% at 5% THD considering the power dissipated by the tube and the source follower, and of course higher than 50% if we consider the tube only. GU72 seems a great candidate for that circuit.
The basic concept is to drive the output tube from its cathode through a p-mosfet source follower, and give local feedback to the tube from anode to g1 thorugh a voltage divider. g2 is connected to a low impedance fixed bias optimized for the chosen loadline.
The advantage is that you get triode-like curves, but g1=0 is shifted on the left of the plot like a pentode curve (so you swing more, so you get more power), and you share the cathode current with the source follower that is usually set at at some tens of volts, so you can increase the current flowing through the tube, then getting more power from it.
Lab tests have shown efficiencies very close to 50% at 5% THD considering the power dissipated by the tube and the source follower, and of course higher than 50% if we consider the tube only. GU72 seems a great candidate for that circuit.
I've been looking at the circuits, and definitely looks interesting and simple, I will try it. Having said that, I'm not overly impressed by distortion levels shown in that thread. But yes, it seems to be able to extract more power from the same tube than a classical design design.
This is a steep tube (15-23 gm) and running it in tetrode mode with CFB output transformer could do wonders. Never used a tetrode though, not sure what would be the best way to deal with the kinks and if it would be wise to rely on NFB to linearize them.
GU72 is certainly a cool tube (correction: a really hot tube, more on that later...)
I have a pair and some sockets on the shelf. No topcaps, so I had to make my own connectors from copper strips when ran one of them on a breadboard years ago. I think I stopped at 560V / 110mA or so which produced enought heat to melt the solder joint between the anode wire and my DIY top cap connector, resulting in an evaporated screen grid resistor.
I got around 15W in triode connection with a 3,2k plate load and quite a bit of CFB. The CFB saved the day there, as 3,2k results in a rather steep loadline.
I´m thinking about returing to these tubes in the future, together with a pair of LL1688/85mA OPTs I bought almost two decades ago when I was young enough to be comfortable with the idea of having transmitter tubes with kilovolt power supplies in my hobby room😎
The transformers cab be wired for either 9,2 or 5,5k/8R, both should work with a trioded GU72 giving different numbers for output power, distortion and damping factor. The current rating could probably be pushed a bit too as I would'nt get near the full 30W rating of the transformers, being intended for 845 and similar big bottles.
A question for anyone who has actually built something with GU72: The cathodes are internally connected to the heaters, will AC heating cause hum problems?
I have a pair and some sockets on the shelf. No topcaps, so I had to make my own connectors from copper strips when ran one of them on a breadboard years ago. I think I stopped at 560V / 110mA or so which produced enought heat to melt the solder joint between the anode wire and my DIY top cap connector, resulting in an evaporated screen grid resistor.
I got around 15W in triode connection with a 3,2k plate load and quite a bit of CFB. The CFB saved the day there, as 3,2k results in a rather steep loadline.
I´m thinking about returing to these tubes in the future, together with a pair of LL1688/85mA OPTs I bought almost two decades ago when I was young enough to be comfortable with the idea of having transmitter tubes with kilovolt power supplies in my hobby room😎
The transformers cab be wired for either 9,2 or 5,5k/8R, both should work with a trioded GU72 giving different numbers for output power, distortion and damping factor. The current rating could probably be pushed a bit too as I would'nt get near the full 30W rating of the transformers, being intended for 845 and similar big bottles.
A question for anyone who has actually built something with GU72: The cathodes are internally connected to the heaters, will AC heating cause hum problems?
Excellent question, I was unpleasantly surprised a week ago while looking at the pinout, and now I'm wondering whether the UNSET setup will be able to drive the parasitic capacitance of a heater transformer or an SMPS.
@jcalvarez the issue has been addressed time ago. It has been suggested to reference the heater winding to the cathode through a 10k resistor.
@Fuling in case of doubt, do you have a 12V SMPS to be further filtered to supply the heaters in DC?
@Fuling in case of doubt, do you have a 12V SMPS to be further filtered to supply the heaters in DC?
I may have misunderstood what you are suggesting, the heater winding is connected to the cathode, because the heater is connected to the cathode internally.
@jcalvarez It's me that misunderstood the question, as I was sure your project was with GU70 and not GU72.
A dedicated transformer for heaters should be better, but a transformer winder will be needed to determine the details.
A dedicated transformer for heaters should be better, but a transformer winder will be needed to determine the details.
I'm hoping that the tube and the p-channel mosfet follower can drive a relatively high capacitance anyway. I already bought the tubes, no options but to try.