GM70 lifetime
- Thread starter ziocalepino
- Start date
Well, is difficult to be 100% sure. Soviet datasheets were quite conservative, and tubes were built like tanks.
But it also depends on what do you want to do. Class A? Class AB? Push-pull? Single ended? What output power are you targeting? The easier you go on the GM-70, the longer it will last. 1000h is really an absolute minimum.
But it also depends on what do you want to do. Class A? Class AB? Push-pull? Single ended? What output power are you targeting? The easier you go on the GM-70, the longer it will last. 1000h is really an absolute minimum.
GM-70 is a thoriated tungsten filamentary triode. If the tube is new, or in good condition, it may be worth running the filament at a a slightly reduced voltage, to slow down the exhaustion of the active layer. This can yield a very useful increase in lifetime.
The attached diagram is from the EIMAC app-notes on the subject. You can adjust the filament voltage, and check the gm, power out, etc and use the lowest heating power that maintains performance.
The attached diagram is from the EIMAC app-notes on the subject. You can adjust the filament voltage, and check the gm, power out, etc and use the lowest heating power that maintains performance.
Attachments
Very interesting corves, thank you! What do they mean by "properly" derated?
> What do they mean by "properly" derated?
It means that the filament voltage can be reduced by a small amount, at the beginning of the lifetime. The nominal reduction is 5%, but it can be more or less; but the test for proper derating is: does the application circuit still function properly? In an audio amplifier, the test might be that the gm is still sufficient for the amp, and that the required output power is still available.
This only works for thoriated tungsten filaments. With coated filaments, reducing the filament voltage too much can shorten the lifetime.
It means that the filament voltage can be reduced by a small amount, at the beginning of the lifetime. The nominal reduction is 5%, but it can be more or less; but the test for proper derating is: does the application circuit still function properly? In an audio amplifier, the test might be that the gm is still sufficient for the amp, and that the required output power is still available.
This only works for thoriated tungsten filaments. With coated filaments, reducing the filament voltage too much can shorten the lifetime.
Some members that use GM70 had informed around 2K hours without Softstart or delay, with a proper Softstart circuit it will last much more.
However this tube is much more robust than 805, 845 at a $20each
Hi FullRangeMan, where have you read this?Some members that use GM70 had informed around 2K hours without Softstart or delay, with a proper Softstart circuit it will last much more.
However this tube is much more robust than 805, 845 at a $20each
Many thanks,
Erik
Many thanks for this post Rod, I am putting together a GM70 amp and will try this trick.
Hi Erik! 2000 Hrs should be within reach, if the tube was stored carefully.
The data sheet gives good starting information.
The first line « долговeчности / Dolgovechnost » means Lifetime, and the data is MINIMUM 1000 hrs, according to Technical [test regime] Specification CA3.310.003 TU1.
Second Line «критиeрии долговeчности / Kritieriy Dolgovechnosti» means end-of-life is judged to be when the cathode emission is down to 0.6A
Third Line «гарантии Cpок хранeния ... условиякх »
- says that the storage time guarantee is three years.
The fact the the GM-70 is a thoriated tungsten filament offers a good opportunity to extend the lifetime for audio usage:
The data sheet allows ±5% of nominal voltage for the filament, but if instead you can stabilise the voltage to -5%, you should be able increase the lifetime according to the guidance of the EIMAC graphic in post #7.
And further, given that audio should not need anything near to 0.6A of peak emission, it may be possible to reduce the voltage even further on the GM-70. If it sounds as good as at 19V, it should work fine.
As the tube wears out, or does not sound right, you can readjust the voltage upward, to draw the last emission capability out of the filament.
If you try this, please report back with any results!
Thanks for the explanations Rod,
I am starting with a rather minimalistic design, but it includes an adjustable simple CCS to feed the cathode, so dialing in 19V is feasible. This also makes it a slow start.
Now another question. As I am quite respectful of the HV on this amp, I decided to make a minimalistic HV PS. It consists of a 330V AC secondary feeding a doubler consisting of 2x 40uF/900V DC link caps and 2x 6kV diodes https://www.tme.eu/en/details/by6-dio/tht-universal-diodes/diotec-semiconductor/by6/. The positive output is directly connected to a 66uF/1300VDC DC-link cap, a 15H/100mA choke is used between the negative output and the 66uF cap. 3x 220k/1W (500V DC rated) resistors in series as bleeder.
Now the question: HV will come up instantly, while the CCS fed cathode will heat the valve slowly - will this be a problem?
An easy addition for a slow start-up would be a damper diode (like 6CL3) in the negative leg...
I am starting with a rather minimalistic design, but it includes an adjustable simple CCS to feed the cathode, so dialing in 19V is feasible. This also makes it a slow start.
Now another question. As I am quite respectful of the HV on this amp, I decided to make a minimalistic HV PS. It consists of a 330V AC secondary feeding a doubler consisting of 2x 40uF/900V DC link caps and 2x 6kV diodes https://www.tme.eu/en/details/by6-dio/tht-universal-diodes/diotec-semiconductor/by6/. The positive output is directly connected to a 66uF/1300VDC DC-link cap, a 15H/100mA choke is used between the negative output and the 66uF cap. 3x 220k/1W (500V DC rated) resistors in series as bleeder.
Now the question: HV will come up instantly, while the CCS fed cathode will heat the valve slowly - will this be a problem?
An easy addition for a slow start-up would be a damper diode (like 6CL3) in the negative leg...
Thoriated tungsten filaments are not coated (unlike oxide filaments), and the startup sequence may not be overly important at anode voltages of 900V or so.
The best way to control anode current at startup is by means of the grid bias. Make sure the grid bias starts first, and you could force it to start up at a more negative voltage than for normal running, if you want to ramp the anode-current gently.
The best way to control anode current at startup is by means of the grid bias. Make sure the grid bias starts first, and you could force it to start up at a more negative voltage than for normal running, if you want to ramp the anode-current gently.
Thanks for the reply. Just these days I was reading Kevin's posts about the GM70, and he wrote that he is replacing the GM70's every 1000 hours, to avoid nasty surprises.
Do you have a more specific link?
Thanks! Erik