My first time at trying to design a tube amp from scratch. I have modified a few other designs with good success. Have an EE degree but havent used it in decades. So I come here asking those who have done much more than me with tubes to give this design a "sanity check"
Thanks
Thanks
An externally hosted image should be here but it was not working when we last tested it.
I'll second Pieter on this, and also remark that I have a lot of experience with the GM70. This is not a first timer tube.
To sing they need 1kV - 1.2kV at in excess of 100mA - not trivial to safely and reliably get these voltages, for example I use stacked electrolytics in some of my designs when there is not room for film. They lived 5 years even with all of the reasonable precautions I took to assure that the voltage distributed evenly across the 3 x 220uF/450V caps with voltage equalization resistors I used. It didn't work. The replacements won't last any longer and will soon be replaced with 70uF/1300V film caps from Clarity.
For full power they need 200Vpp minimum and if you entertain thoughts of > 15W output some A2 capability is needed in the driver. They demand lots of grid current.
Pieter and a few others can make output transformers that will survive long term at these voltage, many transformers will fail over time due to the high voltage (both AC and DC) present on the primary winding.
I would use RC or LC coupling or my preference IT drive - in the event of a driver failure (it happens!) a catastrophic chain of events will not be initiated resulting in destruction of GM70, output transformer, and HV supply.
3A of clean low noise 20V for the filaments is not easily achieved, but at least Rod Coleman has provided a solution with his filament CCS.
I never published my amp design because it is both a work in progress and still non-optimum from my standpoint, and also extremely hazardous in the wrong hands.
To sing they need 1kV - 1.2kV at in excess of 100mA - not trivial to safely and reliably get these voltages, for example I use stacked electrolytics in some of my designs when there is not room for film. They lived 5 years even with all of the reasonable precautions I took to assure that the voltage distributed evenly across the 3 x 220uF/450V caps with voltage equalization resistors I used. It didn't work. The replacements won't last any longer and will soon be replaced with 70uF/1300V film caps from Clarity.
For full power they need 200Vpp minimum and if you entertain thoughts of > 15W output some A2 capability is needed in the driver. They demand lots of grid current.
Pieter and a few others can make output transformers that will survive long term at these voltage, many transformers will fail over time due to the high voltage (both AC and DC) present on the primary winding.
I would use RC or LC coupling or my preference IT drive - in the event of a driver failure (it happens!) a catastrophic chain of events will not be initiated resulting in destruction of GM70, output transformer, and HV supply.
3A of clean low noise 20V for the filaments is not easily achieved, but at least Rod Coleman has provided a solution with his filament CCS.
I never published my amp design because it is both a work in progress and still non-optimum from my standpoint, and also extremely hazardous in the wrong hands.
I think I get it, I ignore the trans impedence and bias the voltage on the grid as in there was 0 ohm load... ?? so that would put it in the -100 range to get 80ma idle current
Closer to -80V to -90V from my experience for that operating point. It varies significantly over tube life and also from tube to tube. It also depends to a significant extent to which filament polarity you reference the bias - with cathode bias I would consider a pair of 50 ohm resistors (big power ones!) to create a psuedo center tap for the cathode resistor.
Note that reducing plate current does not seem to appreciably improve tube life which is short but glorious. I'd run them pretty hard as a result. I average about 1200hrs between replacements which is 20% better than claimed. Emission drops pretty drastically late in life and they start to sound a bit nasty.
I put a fuse on the cathode side in case something happened to tube and bias goes up. This should somewhat protect the OPT.
Also used an Aurduino to monitor some voltages thru a voltage divider network which also does the duty of startup via relays/thermistors. If cathode voltage falls outside the range, it removes power in milliseconds.
Read Magz thread on "Midlife Crisis", lots of good information on building these HV supplies.
I did cap coupled, IT coupled and plate/cap coupled, the best setup by far was the IT but I spent a few bucks on a good IT. The second best (very close) was a dowdy plate choke and cap on the driver. I settled with 90ma 1200v and a Monolith Magnetics 8K transformer, the amp sounds wonderful and smokes many commercial amps I've used in my system over the years. It was also no slouch using a cheaper Hammond 6.5K OPT. If you want to do two stage its going to be difficult but there's a couple guys on the forums that used a HV pentode with some decent gain. This would save you from building two difference power supplies. If your handy with software I would download LTSpice and model a couple circuits. I can give you a link to my google drive and save you some trouble on downloading all the drivers which also has several schematics I've done over the years. Just shoot me a PM
Also used an Aurduino to monitor some voltages thru a voltage divider network which also does the duty of startup via relays/thermistors. If cathode voltage falls outside the range, it removes power in milliseconds.
Read Magz thread on "Midlife Crisis", lots of good information on building these HV supplies.
I did cap coupled, IT coupled and plate/cap coupled, the best setup by far was the IT but I spent a few bucks on a good IT. The second best (very close) was a dowdy plate choke and cap on the driver. I settled with 90ma 1200v and a Monolith Magnetics 8K transformer, the amp sounds wonderful and smokes many commercial amps I've used in my system over the years. It was also no slouch using a cheaper Hammond 6.5K OPT. If you want to do two stage its going to be difficult but there's a couple guys on the forums that used a HV pentode with some decent gain. This would save you from building two difference power supplies. If your handy with software I would download LTSpice and model a couple circuits. I can give you a link to my google drive and save you some trouble on downloading all the drivers which also has several schematics I've done over the years. Just shoot me a PM
Last edited:
At plate voltages like this, the GK-71 tubes (in triode mode) perform better.
Graphite plate GM-70 like around 1000V plate voltages.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.