Hi Folks.
I am starting an OTL project with the Atmasphere input/driver schematic and 4 6C33 in the circlotron.
I am planning to use one driver tube for each pair of 6C33 as they are difficult to match.
Now there could be a problem with setting the bias then, but I believe I have found a solution.
The attached schematic shows driver - output.
It is a little crowded where the switches and meters are, so I will try to explain.
There are two switches one that puts the amp in SET bias mode and the other that switches SET_BIAS mode between the tube pairs.
In SET_BIAS mode one of the driver tubes are disconnected from b+ so only one pair of 6C33 is drawing current and the two meters are monitoring the current through each of the tubes. The two output terminals are switched to ground. You then set the bias for this pair to 240 mA each.
Do the same for the other pair.
When set to normal operation mode, the meters are monitoring the current draw in 1 of the tubes in each pair and the other 2 tubes in the pairs are automatic regulated by the opamps (cut off freq about 1 Hz).
Pictures are of the top plate, the capacitor bank and schematic
I am starting an OTL project with the Atmasphere input/driver schematic and 4 6C33 in the circlotron.
I am planning to use one driver tube for each pair of 6C33 as they are difficult to match.
Now there could be a problem with setting the bias then, but I believe I have found a solution.
The attached schematic shows driver - output.
It is a little crowded where the switches and meters are, so I will try to explain.
There are two switches one that puts the amp in SET bias mode and the other that switches SET_BIAS mode between the tube pairs.
In SET_BIAS mode one of the driver tubes are disconnected from b+ so only one pair of 6C33 is drawing current and the two meters are monitoring the current through each of the tubes. The two output terminals are switched to ground. You then set the bias for this pair to 240 mA each.
Do the same for the other pair.
When set to normal operation mode, the meters are monitoring the current draw in 1 of the tubes in each pair and the other 2 tubes in the pairs are automatic regulated by the opamps (cut off freq about 1 Hz).
Pictures are of the top plate, the capacitor bank and schematic
Attachments
We faced a similar problem with our Novacron amplifier, which also employs 4 power tubes.
We also used 4 sections from a pair of 6SN7s, one section for each 6C33. To set the bias, a current meter is placed across the speaker terminals (this is done with a switch).
The bias controls are all set fully counterclockwise. In this setting, the power tubes don't conduct. The controls are arranged in pairs, one of a pair matched with its opposite. So when you bring up the bias on the first tube (on scale reading), you can then bring the meter back to zero by bringing up the bias on the opposing tube. You then do this again for the second pair of tubes.
Its a good idea to set the bias low initially as the 6C33s tend to see a bit of drift as they warm up. So we repeat the procedure again after an hour of warmup.
BTW I strongly recommend that you add a ring of cooling holes around the output tube sockets!
We also used 4 sections from a pair of 6SN7s, one section for each 6C33. To set the bias, a current meter is placed across the speaker terminals (this is done with a switch).
The bias controls are all set fully counterclockwise. In this setting, the power tubes don't conduct. The controls are arranged in pairs, one of a pair matched with its opposite. So when you bring up the bias on the first tube (on scale reading), you can then bring the meter back to zero by bringing up the bias on the opposing tube. You then do this again for the second pair of tubes.
Its a good idea to set the bias low initially as the 6C33s tend to see a bit of drift as they warm up. So we repeat the procedure again after an hour of warmup.
BTW I strongly recommend that you add a ring of cooling holes around the output tube sockets!
I was around the bias setting as you are describing , but I think I have better control over non matched sets the way I showed, and I don´t have access to enough tubes to find four pairs..
I am going to use forced air from the bottom of chassis, but probably I think I will follow your advise with holes round the 6C33. Thanks.
Attachments:
Now transformators are in place. WOW That's a lot of iron!!
Transformator covers mounted.
I am going to use forced air from the bottom of chassis, but probably I think I will follow your advise with holes round the 6C33. Thanks.
Attachments:
Now transformators are in place. WOW
That's a lot of iron!!Transformator covers mounted.
Attachments
I understand- we didn't do a lot of matching FWIW and it worked fine.
BTW, a very good idea to mount the sockets as you describe, but its not enough- add the cooling holes! The tube runs really hot and so do the sockets, which is a good part of why they fail, and the wiring on them fails due to heat too!
Hmm.
A lot easier the way you did the biasing. I will try that first and if it causes to much trouble, I can always change it. And I will make the holes too.
Thanks a lot for your input I really appreciate it.
You are quite right and as I said I appreciate it very much. The thing is I have already build one OTL with a single set of 6C33 and servo control of one of the 6C33 so you only have to adjust and monitor one tube and it works perfectly, so I would like to implement this to this amp with two sets of 6C33. But it is more complicated so maybe I should skip it after all..
I am not quite sure what you mean?
The problem is when there is an unbalance (DC) at the output, which of the two tubes is to be corrected?
I was thinking about having a servo correcting one of the tubes when unbalance occurs and then another servo balancing the other tube with the first. But this is two servos and they will probably interact.
The problem is when there is an unbalance (DC) at the output, which of the two tubes is to be corrected?
I was thinking about having a servo correcting one of the tubes when unbalance occurs and then another servo balancing the other tube with the first. But this is two servos and they will probably interact.
I'm no electronics guru so I have no.idea if this would work but it may get your thought process going, if you have a servo for each half of the circlotron, what if you measure the voltage differential between the grid and cathode of each tube on that side of the circlotron, set up a method of setting a parameter that a normal voltage at idle is 23-27V difference, outside this and indicate this tube is going out of spec. You could possibly also monitor the input signal so it only operates at no music signal or when you push a button to check and correct, I personally like Ralph's option for simplicity, reliability and its out of circuit when listening, everything in your amp affects its sound, so the less sand the better for me. There is likely a way to servo it, but is it really worth the effort to achieve when it's just a couple minutes at listening start and check as you flip the vinyl.....all part of the experience!
Lets face it, life is about experience!
Lets face it, life is about experience!
In a single set circlotron auto bias system, one of the tubes have a bias set potmeter and you measure over a cathode resistor the current draw. Then a servo controls the other tubes bias so the output is zero. The servo has a very low cutoff freq. (1 Hz)so you can leave it in with music playing.
The problem is when you have TWO sets of tubes. If the tubes are not matched, you have to autobias each of them separately, but how will you know which tube to correct, when the output drifts away from zero?
That is the problem
The problem is when you have TWO sets of tubes. If the tubes are not matched, you have to autobias each of them separately, but how will you know which tube to correct, when the output drifts away from zero?
That is the problem
No each of the two tubes in one side of the circlotron have trimmers. You would have to match the tubes in order to have one trimmer to control both.
Remember when you set the bias for these two tubes, you must short circuit the output or the tubes will not conduct.
And the bias setting is not at the grids of the 6C33, but the driver tube.
I want to make a system that auto sets the bias on the two tubes that dosent have bias trimmers in their drivers and at the same time corrects DC fault at the output.
Remember when you set the bias for these two tubes, you must short circuit the output or the tubes will not conduct.
And the bias setting is not at the grids of the 6C33, but the driver tube.
I want to make a system that auto sets the bias on the two tubes that dosent have bias trimmers in their drivers and at the same time corrects DC fault at the output.
So the "Beast" is up and running now.
The schematic of the amplifier looks deceivingly simple, but it is rather complicated all the same, when all the peripheral circuits has to be counted in:
Bias, Meter circuit and switching, inrush delay and a long delay of the HT for the 6C33 tubes (3-5 min) to avoid cathode stripping.
Not yet ready to comment definitely on the sound quality, but it sounds fine.
Nu hum or noise whatsoever.
The schematic of the amplifier looks deceivingly simple, but it is rather complicated all the same, when all the peripheral circuits has to be counted in:
Bias, Meter circuit and switching, inrush delay and a long delay of the HT for the 6C33 tubes (3-5 min) to avoid cathode stripping.
Not yet ready to comment definitely on the sound quality, but it sounds fine.
Nu hum or noise whatsoever.
Attachments
Hi
I can now tell you this:
This amplifier is sounding extraordinarily good
It needed a very long time to break in and still needs warming up time of about 20 min before it really sings.
And I haven´t even started to change tubes or resistors to better types .
It is made of good components and material, but nothing exotic.
I have some NOS 7n7 tubes underway and some sockets to make them work in the amp.
The amp is very transparent (the first couple of weeks too bright actually but now very neutral) extremely fast and dynamic and has an ability to resolve even the most complex music pieces.
And depth and soundstage out of the ordinary.
I have in house as references the following amps:
Goulmund Telos 350
My_ref Fremen edition
Hypex 180 - 400 and 700
Pass F5
The amazing FET circlotron
A 6C33 OTL from Hans Bejner
Rhoender FC-100
among others...
The beast is now my reference and the amp that sits in my main stereo system now!
Thanks Ralph for sharing the original Atmasphere M60 schematic that is the basis for this amp and for your support to all us , who actually tries to build tube OTL circlotrons.
I can now tell you this:
This amplifier is sounding extraordinarily good
It needed a very long time to break in and still needs warming up time of about 20 min before it really sings.
And I haven´t even started to change tubes or resistors to better types .
It is made of good components and material, but nothing exotic.
I have some NOS 7n7 tubes underway and some sockets to make them work in the amp.
The amp is very transparent (the first couple of weeks too bright actually but now very neutral) extremely fast and dynamic and has an ability to resolve even the most complex music pieces.
And depth and soundstage out of the ordinary.
I have in house as references the following amps:
Goulmund Telos 350
My_ref Fremen edition
Hypex 180 - 400 and 700
Pass F5
The amazing FET circlotron
A 6C33 OTL from Hans Bejner
Rhoender FC-100
among others...
The beast is now my reference and the amp that sits in my main stereo system now!
Thanks Ralph for sharing the original Atmasphere M60 schematic that is the basis for this amp and for your support to all us , who actually tries to build tube OTL circlotrons.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.