Good to hear that! So the negative feedback was forcing a resonant peak between the ot and the parafeed cap?
Thats the way it appears to be.
I have just been emailing Gary Pimms, the designer of the CCS I am using. After checking the CCS function more carefully, it turns out it is only partially working and not achieveing anything like its real potential. I anticipated this might happen in the design stage, so got extra components to build a less sophisticated version using just resistors. I will be building this in the next few days. Should clean up the bass no end.
Will keep you posted.
Shoog
I have just been emailing Gary Pimms, the designer of the CCS I am using. After checking the CCS function more carefully, it turns out it is only partially working and not achieveing anything like its real potential. I anticipated this might happen in the design stage, so got extra components to build a less sophisticated version using just resistors. I will be building this in the next few days. Should clean up the bass no end.
Will keep you posted.
Shoog
I worked out something today, and its not good.
Not only do Valves have AC resistance, but they also have DC resistance. Since the TT21 is designed to handle much higher currents than the 807, it has a much lower DC resistance. Infact it is about 1/3 of the 807.
This has the practical effect when making a cascode that they will act as a DC voltage divider. Hence the TT21 will consume about 1/3 (in my case 190V), and the 807 will consume about 2/3 (in my case 520V).
As far as I can assertain this is fairly fixed and so it is very difficult to force the TT21 to consume its required 400V.
Conclusion they do not make good partner. When designing a cascode circuit DC resistance must be taken into account at the very start.
I am having doubts that it will be possable to make the TT21 work satisfactorily in this circuit.
Shoog
Not only do Valves have AC resistance, but they also have DC resistance. Since the TT21 is designed to handle much higher currents than the 807, it has a much lower DC resistance. Infact it is about 1/3 of the 807.
This has the practical effect when making a cascode that they will act as a DC voltage divider. Hence the TT21 will consume about 1/3 (in my case 190V), and the 807 will consume about 2/3 (in my case 520V).
As far as I can assertain this is fairly fixed and so it is very difficult to force the TT21 to consume its required 400V.
Conclusion they do not make good partner. When designing a cascode circuit DC resistance must be taken into account at the very start.
I am having doubts that it will be possable to make the TT21 work satisfactorily in this circuit.
Shoog
If I was you and have another TT21 on hands, I would eliminate the 807... or if you've got another 807... I thought that this kind of configuration would need the same tube for the CCS and the output one - but seeing your work I believed it would be great, I'm not an expert in that kind of topology. (just a thought: putting zeners to make the appropriate voltage drop?) I guess you have to dump the idea of the amp... and use similar tubes. Maybe swapping them around?
If I can make the TT21 behave in a more pentode fashion, then the situation should improve as the rp should go up, and hence the voltage division should become more equal.
Shoog
Shoog
It seems ok... but pentode man... for a geekie output topology like this I would stick to triodes... Even directly heated ones! Have you thought about that?
Pentodes will always make better current sources than triodes, by a factor of their difference in gain. Also Pentodes as Cathode Followers are apparently very linear.
Good news, my guess that making the CCS TT21 behave more like a pentode would improve the situation. In a quick lashed up test I wired up the screen differently(dropping the transistor capacitance multiplier which was acting as a bottle neck), and hay presto the CCS is consuming the magic 400V. Might even get Gary Pimms CCS working.
It shore has been a pain in the *** getting this far, but hell I now know one hell of a lot more about Pentodes, transformers, Current sources. Learning on the hoof is the only way to really learn.
Shoog
Good news, my guess that making the CCS TT21 behave more like a pentode would improve the situation. In a quick lashed up test I wired up the screen differently(dropping the transistor capacitance multiplier which was acting as a bottle neck), and hay presto the CCS is consuming the magic 400V. Might even get Gary Pimms CCS working.
It shore has been a pain in the *** getting this far, but hell I now know one hell of a lot more about Pentodes, transformers, Current sources. Learning on the hoof is the only way to really learn.
Shoog
I was a little premature. Altering the screen supply has helped a little, but not as much as I thought. The CCS tube (TT21) is now dropping 250V when it was dropping 190V, which shows its slightly more pentode like. The extra voltage drop was in a resistor that went high.
Still I am now changing the current source such that it has 2K of resistance in its tail, which the current source monitors. This should stiffen it by a factor of five. It also serves to consume a 100V before it hits the 807 plate. The 807 should now be running at 400V, which gives it a dissapation of 20W. All this is outside the range of the TT21 datasheet, so its going to be lucky if it hits the needed 50mA of constant current. Running it at any less current will mean that the 807 will belly out on deep bass (which is exactly what its doing now with only 42mA of current).
I think I have a really good understanding of how all the bits of this circuit is working now, so it should be plain sailing from here on in. The sensable thing to do would be to ditch the TT21's, sell them at a good profit. Replace them with 807 CCS's which would work in Gary Pimms circuit for better performance. Still its a challenge now to get this damn thing working. If I do then the TT21's should last damn near forever, been so undertaxed. The only issue is will they be introducing distortion?
Shoog
Still I am now changing the current source such that it has 2K of resistance in its tail, which the current source monitors. This should stiffen it by a factor of five. It also serves to consume a 100V before it hits the 807 plate. The 807 should now be running at 400V, which gives it a dissapation of 20W. All this is outside the range of the TT21 datasheet, so its going to be lucky if it hits the needed 50mA of constant current. Running it at any less current will mean that the 807 will belly out on deep bass (which is exactly what its doing now with only 42mA of current).
I think I have a really good understanding of how all the bits of this circuit is working now, so it should be plain sailing from here on in. The sensable thing to do would be to ditch the TT21's, sell them at a good profit. Replace them with 807 CCS's which would work in Gary Pimms circuit for better performance. Still its a challenge now to get this damn thing working. If I do then the TT21's should last damn near forever, been so undertaxed. The only issue is will they be introducing distortion?
Shoog
I tried everything with the TT21 current source. Every time I set things so that it should have been generating more current, the voltage across it shrank back and the current stayed at 42mA. I was all set to take my own advise and ditch the TT21, substituting another 807 or 6L6.
I decided to try the one thing I hadn’t yet tried yet. I had been running the 807 in Pentode mode with a separate stiff screen supply. I decided to try it in triode mode. So I strapped it in Triode mode and hay presto, the voltage across the TT21 expanded to 280V, it started delivering 52mA. This dropped a further 112V across the current sensing resistor, which left me with the magic 360V across the 807.
Sound, much smoother than in pentode. The best it has sounded. Infact I listened to 3 CD’s there and then. The scoped up output looks perfect and distortionless. The downside was that I was peaking out at about 3Watts (I think), with was not enough to drive my inefficient test speakers. Also I couldn’t tell whether when pushed hard, I was hearing speaker distortion, or amp distortion. Not quite the 10Watts I had set out to achieve, but I think it will be adequate on my Sensitive Tannoys. Seems terribly wasteful to build a 50Kg amp that only delivers 6watts out for about 200Watts in.
Shoog
I decided to try the one thing I hadn’t yet tried yet. I had been running the 807 in Pentode mode with a separate stiff screen supply. I decided to try it in triode mode. So I strapped it in Triode mode and hay presto, the voltage across the TT21 expanded to 280V, it started delivering 52mA. This dropped a further 112V across the current sensing resistor, which left me with the magic 360V across the 807.
Sound, much smoother than in pentode. The best it has sounded. Infact I listened to 3 CD’s there and then. The scoped up output looks perfect and distortionless. The downside was that I was peaking out at about 3Watts (I think), with was not enough to drive my inefficient test speakers. Also I couldn’t tell whether when pushed hard, I was hearing speaker distortion, or amp distortion. Not quite the 10Watts I had set out to achieve, but I think it will be adequate on my Sensitive Tannoys. Seems terribly wasteful to build a 50Kg amp that only delivers 6watts out for about 200Watts in.
Shoog
Shoog said:
That tube is the Il Duce of all Kt88's.......it's worth a fortune....put in in a standard p-p or SE design.....don't ditch it!
tube lover richj
- Status
- Not open for further replies.