otl
Thanks again Hans - any other tips? - Will report on results once I have completed mods. Even with the 'hum' the difference is quite noticeable over SE or PP with Xfrmr output! Have been very happy with a SE El34 in tetrode mode with cathode feedback winding for some time, but now......
Regards,
David.
Thanks again Hans - any other tips? - Will report on results once I have completed mods. Even with the 'hum' the difference is quite noticeable over SE or PP with Xfrmr output! Have been very happy with a SE El34 in tetrode mode with cathode feedback winding for some time, but now......
Regards,
David.
Hi,
Have just completed addition of 4 x seperate Bias supplies and was (at first!) suprised that after balancing the output pairs to 200ma, and looking to zero any volts across the speaker terminals, 0.2v was the minimum possible no matter how the balancing was performed. I then realised that this voltage was due to the driver stage being grounded via the L.S. Removing the driver valves removed this 'offset' and zero volts was obtained. Obviously all OTL's using this technique will have the same result. Anyone who has experience of OTL's and this 'offset' know if it might cause any problems please?
Regards,
David.
Have just completed addition of 4 x seperate Bias supplies and was (at first!) suprised that after balancing the output pairs to 200ma, and looking to zero any volts across the speaker terminals, 0.2v was the minimum possible no matter how the balancing was performed. I then realised that this voltage was due to the driver stage being grounded via the L.S. Removing the driver valves removed this 'offset' and zero volts was obtained. Obviously all OTL's using this technique will have the same result. Anyone who has experience of OTL's and this 'offset' know if it might cause any problems please?
Regards,
David.
I then realised that this voltage was due to the driver stage being grounded via the L.S.
Yes, you need to use bleeder resistors of different value at the positive and negative main supply to balance out the current trough the driver stage. I have used 10k at the positive side and 20k at the positive side, this achieves almost perfect balance and i can easily get DC at the output to 10mV or less. Without these resistors the balance will also be different when output power is large, (output DC voltage shift).
Note! I don't use the Ciuffoli driver circuit, (I think that it is not an optimal design), so maybe you need different values of resistors, but the principal is the same, (using unequal bleeders to balance out the phase splitter current)
Regards Hans
Hello Hans,
Have just found an earlier OTL thread in which you mention the combining of the + & - sets of HT rails from both channels. The reduction in hum is certainly worthwhile - effectively doubling the capacitor values - any idea why this wasn't done originally? - after all the drivers use the same HTrail for both channels. You also mentioned in your last post to this thread that you didn't consider the A.C driver stages to be 'ultimate' - can I ask why please? Using a high gm driver, parralled with relatively high current might normally(?) be considered a good option for triodes. I would appreciate your thoughts.
Many thanks,
David.
Have just found an earlier OTL thread in which you mention the combining of the + & - sets of HT rails from both channels. The reduction in hum is certainly worthwhile - effectively doubling the capacitor values - any idea why this wasn't done originally? - after all the drivers use the same HTrail for both channels. You also mentioned in your last post to this thread that you didn't consider the A.C driver stages to be 'ultimate' - can I ask why please? Using a high gm driver, parralled with relatively high current might normally(?) be considered a good option for triodes. I would appreciate your thoughts.
Many thanks,
David.
any idea why this wasn't done originally? - after all the drivers use the same HTrail for both channels.
I think it was suggested by Ciuffoli from the beginning but dropped later, it will work but it will be more difficult to get good DC balance and there will be some crosstalk between channels. I have tried both ways but now I am using separate caps and Diode bridges for each channel as the DC offset on the output is more stable that way.
Regarding the driver there are several misconception regarding the 6C33 flying around on the internet, among those are that it should have very high input capacitance and have higher than normal gridcurrent. Due to this gridcurrent and high input cap some advocate to use a beefy driver.
However the 6C33 doesn't have significantly higher gridcurrent than any other power triode, this is evident for instance by the fact that 100kohm is recommended and is low enough as grid leak and with this value it is not possible to measure any significant shift in gridvoltage due to gridcurrent. The input cap due to Miller is not very high either as the gain from Grid to anode is so low in this circuit, the total Miller cap is actually much less than for a 300B in a cathode grounded amp.
Due to the reasons above it is not necessary to use a beefy driver, especially not when using a split load inverter as this has so low output impedance and therefore the poles created by the Miller cap and the split load inverter will be very high, in the MHz range. The driver only need to be able to give current to drive the gridleak and the Miller cap.
Instead you need a driver that can give high output voltage and do that at low distortion, the solution is to use lower current and to optimise the operating point for symmetrical and low distortion.
If you want more details send me a mail.
Regards Hans
Just a note pertaining to tubetvr's post 12 above:
'It would be OK for a symmetrical output stage e.g a SS amp where ripple from the power supply is balanced out but in this stage the balance is not so good as the PSRR is different in the 2 sides, looking from the power supply it is a cathode follower and a ordinary cathode grounded stage and they have quite different PSRR. This is one of the few disadvantages of serial connected OTL amps compared to a circlotron.'
It isn't very hard to fix the unbalanced-PSRR problem of the stacked-
type OTL output stage. Look at Broskie's (TubeCAD) articles Output
Stage PSRR enhancement or study any of the numerous OTL designs
to be found on his site:
http://www.tubecad.com/
These "Aikido" techniques work well whether you use a split-triode
or differential phase inverter - by injecting rail noise into the in-
verter stage so as to achieve cancellation. And yes, it works in
practice. And is much cheaper and easier than using chokes.
(As Hans might know, we designed B Jansson's Transcendent-
based "T12" OTL using these tricks!) Recommended.
Morgan
'It would be OK for a symmetrical output stage e.g a SS amp where ripple from the power supply is balanced out but in this stage the balance is not so good as the PSRR is different in the 2 sides, looking from the power supply it is a cathode follower and a ordinary cathode grounded stage and they have quite different PSRR. This is one of the few disadvantages of serial connected OTL amps compared to a circlotron.'
It isn't very hard to fix the unbalanced-PSRR problem of the stacked-
type OTL output stage. Look at Broskie's (TubeCAD) articles Output
Stage PSRR enhancement or study any of the numerous OTL designs
to be found on his site:
http://www.tubecad.com/
These "Aikido" techniques work well whether you use a split-triode
or differential phase inverter - by injecting rail noise into the in-
verter stage so as to achieve cancellation. And yes, it works in
practice. And is much cheaper and easier than using chokes.
(As Hans might know, we designed B Jansson's Transcendent-
based "T12" OTL using these tricks!) Recommended.
Morgan
Hello Morgan,
Thank you for the info. - have just downloaded the article on PSRR - looks like a 'good read' but may take a while to digest I think! As you can tell I am new to OTL's but am learning fast due to help from the group and especially from Hans who has been most enlightening. Thanks to all.
Regards,
David.
Thank you for the info. - have just downloaded the article on PSRR - looks like a 'good read' but may take a while to digest I think! As you can tell I am new to OTL's but am learning fast due to help from the group and especially from Hans who has been most enlightening. Thanks to all.
Regards,
David.
Hi,
Have just digested the John Broskie 'Noise reduction' article as per. Morgans previous mail. I decided to try the fix on the phase splitter stage of my OTL as there was a low hum level - you had to listen close up to hear it, but.... in search of perfection! I added the mod. to one channel only as I only had a single capacitor to hand at the time. The effect was quite unexpected - no hum! on the modified channel and even more startling - minimum hum on the un-modded channel! I even removed the components as I really didn't believe it! (of course the hum returned!) Thanks greatly to you Morgan for pointing me in the right direction! I really didn't believe you could teach old dogs new tricks! A must for any noisy OTL! Very highly recommended.
Thanks again,
David.
Have just digested the John Broskie 'Noise reduction' article as per. Morgans previous mail. I decided to try the fix on the phase splitter stage of my OTL as there was a low hum level - you had to listen close up to hear it, but.... in search of perfection! I added the mod. to one channel only as I only had a single capacitor to hand at the time. The effect was quite unexpected - no hum! on the modified channel and even more startling - minimum hum on the un-modded channel! I even removed the components as I really didn't believe it! (of course the hum returned!) Thanks greatly to you Morgan for pointing me in the right direction! I really didn't believe you could teach old dogs new tricks! A must for any noisy OTL! Very highly recommended.
Thanks again,
David.
Thank you for trying it. You're the third person to try it successfully,
that I know of (including myself). Well, forth counting the originator,
or rather perhaps populariser, of the idea: Mr B. himself.
This trick should also generally improve PSRR symmetry, with the effect
that output stage asymmetry or "DC on the output" on high-power
peaks should be lessened (if present). The limitation is in the size of
the caps, which would have to be rather large to work down to near
DC.
Morgan
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.