In p-p audio power amps ....Do 6C33 power tubes behave better using cathode (auto) bias or fixed bias ? Despite my vendor matching these tubes (understandably not keen on the idea); over time I' finding ther'e gradually drifting out...of characteristics. I've heard this is a common problem (is it ??) even though the tubes is still relatively new.......I've tried using cathode bias to
equalise which helps thd at lower pout........but when compared to 88's and 34's in fixed bias these are way more reliable over time.
This 6C33 tube was designed bomb proof....so whats on the move ?
replies welcome.
richj
equalise which helps thd at lower pout........but when compared to 88's and 34's in fixed bias these are way more reliable over time.
This 6C33 tube was designed bomb proof....so whats on the move ?
replies welcome.
richj
I am designing a 6c33 amp but I haven´t played with them yet.
From all the info I have picked up on them
1.High gm means diffcult to match tubes and they tend to drift.
2.Experienced users suggest a long break in time first with heaters and then powered up.Then they stabilise.
3.Cathode bias is safer of course.However depending on bias level you must get alot of dissipation on the cathode resistor.eg 200ma x 70v= 14W.So you really need a big aluclad heatsinked resistor there ideally.
From all the info I have picked up on them
1.High gm means diffcult to match tubes and they tend to drift.
2.Experienced users suggest a long break in time first with heaters and then powered up.Then they stabilise.
3.Cathode bias is safer of course.However depending on bias level you must get alot of dissipation on the cathode resistor.eg 200ma x 70v= 14W.So you really need a big aluclad heatsinked resistor there ideally.
protos said:
Why not make a current source such as LM317 with a large resistor to dump this current into; the LM317 is a floating device so with 300 ohms load this should work OK; the voltage setting resistors can easily be calculated from the formula's.
With 2000 muF on the input and maybe output this works fine.
I have had it running (on EL34's at 50 mA without a sink for years; here of course yes the resistor and regulator must be heatsinked.
I have found it makes a really rich sound - basically because absolute current matching makes the output trannie behave better.
success albert
In my experience it is only a problem if the anode dissipation is close to max and if the tubes has not been burned in. I have used more than 30 6C33C and never had a problem with drift, some tubes are used for more than 5 years.
Anode dissipation shoul be less than 40W and the burn in procedure is: heater only for at least 1 hour, follow that by burn in with at least the same current as is used in the circuit and the same or higher anode dissipation for at least one hour. Heater must always be switched on 30 seconds before anode voltage.
Regards Hans
protos said:
In fact I use it as a variable cathode resistor. This will enerate the bias; it can be combined if you like with a (large) bit of fixed bias (e.g. 50V), if you like, then the CCS will generate less heat and get less Vtotal (in this case 70V-50V=20V left over to maintain in a tight grip.
I hope this diagram (sorry, only thing I have is Powerpoint) is clear.
albert
As for me, I have some thought about using of 6C33C:
1) NEVER use it with fixed bias.
2) NO SENCE to use it with low plate dissipation or one heater - there a a lot of another tubes to get less power.
3) If you want output power as much as possible (15-18W) - use it with 200mA current and 270-300v anode voltage.
But if you want the better sound with a little bit less power (11-13W) - use it with 280-300mA and 200-220v at the plate.
So, we MUST implement autobias in any case. This can be done in a few ways:
1) traditional way - cathode resistor with capacitor.
You need a powerful resistor at the heatsink (P=~20-22W) and el.cap.
2) MOSFET at the cathode - the same power but no El.Cap.
3) Small cathode resistor (current sensor) and servo loop to the grid with OP, mossfet pt another tube(s).
But there is another way, that I'm working now (exaclty - output stage is ready, and I'm working with the driver/input stage).
We call this method "autofix", becouse it has all advantages of the autobias and also a fixed bias, and some additional features.
In a few words - we put autobias resistor after the rectifier bridge, before El.Cap at the plate power supply. Then - through diode we get the negative voltage from the point where this resistor is connected to the bridge.
After - we put RC-filter and get the negative voltage that depend of the lamp current? even highet, so I put the trimmer and obtained the posibility to control the lamp's current.
Cathode of the 6C33c is connected to the "-" of the power supply El.Cap - this is the ground.
Advantages:
1) no resistor, no capacitor athe cathode. Cathode is sitting on the ground.
2) Power supply El.Cap charged from the rectifier through the resistor - less peak current.
3) the value of the bias resistor is 1/2 of the traditional autobias resistor - > half of the heat dissipation.
4) Possibility to change the bias with trimmer.
Preliminary schematics: http://altor.sytes.net/audio/se33/se33.gif
(caution: this file is changed time to time, due to development process, and may have some not critical mistakes in values, etc., but the idea - should be clear).
1) NEVER use it with fixed bias.
2) NO SENCE to use it with low plate dissipation or one heater - there a a lot of another tubes to get less power.
3) If you want output power as much as possible (15-18W) - use it with 200mA current and 270-300v anode voltage.
But if you want the better sound with a little bit less power (11-13W) - use it with 280-300mA and 200-220v at the plate.
So, we MUST implement autobias in any case. This can be done in a few ways:
1) traditional way - cathode resistor with capacitor.
You need a powerful resistor at the heatsink (P=~20-22W) and el.cap.
2) MOSFET at the cathode - the same power but no El.Cap.
3) Small cathode resistor (current sensor) and servo loop to the grid with OP, mossfet pt another tube(s).
But there is another way, that I'm working now (exaclty - output stage is ready, and I'm working with the driver/input stage).
We call this method "autofix", becouse it has all advantages of the autobias and also a fixed bias, and some additional features.
In a few words - we put autobias resistor after the rectifier bridge, before El.Cap at the plate power supply. Then - through diode we get the negative voltage from the point where this resistor is connected to the bridge.
After - we put RC-filter and get the negative voltage that depend of the lamp current? even highet, so I put the trimmer and obtained the posibility to control the lamp's current.
Cathode of the 6C33c is connected to the "-" of the power supply El.Cap - this is the ground.
Advantages:
1) no resistor, no capacitor athe cathode. Cathode is sitting on the ground.
2) Power supply El.Cap charged from the rectifier through the resistor - less peak current.
3) the value of the bias resistor is 1/2 of the traditional autobias resistor - > half of the heat dissipation.
4) Possibility to change the bias with trimmer.
Preliminary schematics: http://altor.sytes.net/audio/se33/se33.gif
(caution: this file is changed time to time, due to development process, and may have some not critical mistakes in values, etc., but the idea - should be clear).
What do you base this on? It works perfectly with fixed bias!
Sounds very inefficient, have you measured output power at these settings?
In my SET I used 220V and 200mA and it gave a genuine 13W output power using 600 ohm anode load. It is an amplifier designed by Dakesue of San-ei which is the first who designed an amplifier with 6C33 outside the eastern block. He and many others believes that the 6C33C sounds better at around 200V and should not be used at as much as 300V. If you use a 6C33C at 60W anode dissipation it will drift, with ~40-45W max it will never drift, (assuming it is burned in correctly).
Regards Hans
tubetvr said:
On the experience.
If You use it with 40-45W on the plate - yes, but if so, why you use 6C33C ?!? Use 6C41C, 300B or other.
Yes, of couse.
13W - it possible only in A2, in A1 - I'm not belelive.
You can also get 30W in this schematic:
http://altor.sytes.net/ko/se.html
Yes, I've wrote that with 200v and 300mA the sound is better.
It will never drift with 60-65W anode dissipation, if you use autobiar or "autofix" and 6C33C-B.
With 6c33C (w/o "-B") - it possible.
Konnichiwa,
I find several interesting elements in your schematic.
Before I ask questions, just one note. If you place the filter elements (R4 & the mosfet gyrator with inverted polarity) into the negative line of the supply, rather than the positive one you can add their voltage drop to the bias voltage, increasing efficiency by reducing the input resistor value further.
I am rather interrested in your driver circuit.
I have some 6E5 here and I have noticed the kind of ultralinear circuit you use before, but I have so far lacked the time to investigate either ti any significant degree.
Would you mind sharing the attained performance of your driver circuit? (gain, linear output swing etc.) And if you have any, subjective comparisons would be welcome too.
Sayonara
altor said:
I find several interesting elements in your schematic.
Before I ask questions, just one note. If you place the filter elements (R4 & the mosfet gyrator with inverted polarity) into the negative line of the supply, rather than the positive one you can add their voltage drop to the bias voltage, increasing efficiency by reducing the input resistor value further.
I am rather interrested in your driver circuit.
I have some 6E5 here and I have noticed the kind of ultralinear circuit you use before, but I have so far lacked the time to investigate either ti any significant degree.
Would you mind sharing the attained performance of your driver circuit? (gain, linear output swing etc.) And if you have any, subjective comparisons would be welcome too.
Sayonara
In general I think it is a bad idea to use max allowed power dissipation with any tube be it 6c41, 300B, 6C33 or any other.
Is it so that your own amplifiers are limited in power because the driver doesn't give enough voltage swing?, FYI my amp goes purely in A1 without a hint of gridcurrent. It is well known, (at least in Japan where 6C33C was first used outside Soviet union) that this operating point, (220V, 200mA and 600ohm load) will give 13W output power using a good output transformer.
Do you even run the tubes above max allowed dissipation? it is max 60W.
Based on my own experience using more than 30 6C33C mostly in OTL amplifiers but also in a SET my advice is to burn in the tubes correctly and run them at no more than 40-45W anopde dissipation, this will give very long life and no drift problems.
I have never needed to change any 6C33C with up to almost 8 years use in my SET and more than 6 years in my own OTL and the OTL amps I have built for others, in addition I have never even needed to adjust the bias in the OTL amp. It seems that Viktor at BAT have the same experience and his advice is the same, (careful burn in and no more than 40-45W dissipation).
Regards Hans
As I said before - it depend of the tubes.
I have some tubes with "-B", that work at 80W withiut any drift during 5-6 hours, in the "autofix" topology.
I have also one tube without "-B', than drifd if the plate power is more then 58W in any topology - as in "autofix", as in autobias.
In any case, the tube like 6C33c, tha were intended to use in the power supply of the Soviet Jet-Fighter are very reliable.
Many yesrs ago, I've worked in the military electronic development in the USSR
So it has no sence to use in not at the full power.
I have heard a lot of the 33'based OTL amplifiers, but never heard a good sound from OTL.
Maybe, OTL Amps with the good sound are exists in the univers, but I was not happy to meet them
You are right, I just only told that to my mind - no sence to use 6C33C not in the full power.
The advantage af the 6C33C is the low plate impedanse, so the output transformer could be build very easy.
My latest trafo has 7Hz-70kHz Freq.range (at -3dB) with very good phase characteristic:
http://altor.sytes.net/AUDIO/SE33/SE_33.htm
You are right, but I have no high voltage "p" mosfets, but have a huge quantity of the 200-800v N-mosfets
Driver circuit - is the thing that I'm working now (not every day, because I'm very busy and tied at my job)
I've tryed some variations in the resistor values, tube voltages and current.
The previous results that I have (with autobias - i/e/ RC in the cathode) were:
anode power supply: 370v
cathode current: 30mA
plate voltage: 185v
Gain: 65
Frequency range (-3dB): 4.2Hz-527.3 kHz
Max output swing: 309v p-p (109V RMS)
Signal to noise releted to the input: 74dB.
Output impedance: 2.7k
All measuremens has been done at the 50k load after 2.2uF film cap.
Now I change some values, got Gain=60, the sound better.
Next step - I'll try to eliminate cathode resistor+capacitor, using the same "autofix" topology as at the output stage, but 2 channels commonly and the bias with the 3v Li-cell.
I've herd some drivers with cell, and was not satisfied, but it were not with 6E5P, may be this tube will sound.
It is not my intention to argue but my experience is very different from yours. I used 6C33C in a SET until I designed my OTL and now I never want to go back to a SET again.
Don't get me wrong, 6C33C SET sound very good and make even the worst CD or record sound better but compared to an OTL you loose a lot of details. I liked the sound of my SET but never actually realised what I missed until I built my OTL. I must also say that I have experience of many SET amps with different tubes and many different manufacturers but I still more or less have the same impression, (the loss of details and the ability to make bad recordings sound better).
For me my OTL gives a much clearer sound than any SET I have heard but still no hardness or distortion in the sound, some people say that a good OTL sound like a mix between a traditional tube amp and a good quality transistor amp and I think that is a good description.
Regards Hans
BTW what load impedance do you use in your SET?
analog_sa said:
No, they use fixed bias at the output stage and alseo fixed bias at the driver, with the "trik" to eliminate input capacitor.
My "Hybrid Zen" (Nelson's Zen3 with the tube input stage) also sounds very good, but the sound of the tube SE - is different.
With OTL you need very special speakers, due to the low power and high output impedance. And in any case - you will lose the "bass".
In the current design - 500 Ohm.
No parallell tubes, only one pair working in AB1 with ~0.2A bias current, you can press them a bit more but then you would probably sacrifice reliability, for test purpose I have run the amplifier at full output power for 24 hours without any ill effects. In normal operation the power dissipation is much lower of course and doesn´t exceed ~40W which is why it works so well.
You can't and any reasonable OTL works in AB1, two 6C33C tubes in inverted Futterman configuration give ~10.5ohm output impedance and then you use feedback to reduce output impedance to desired value.
Regards Hans
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.