I read that in Class A Push-Pull OPT the quiescent currents cancel - hence no core saturation issues. I don't understand how that is done and looking for enlightenment.
In Class A both output tubes will be biased to approx full power and that DC current must flow though the OPT. There is typically no dual rail power (i.e. B+, 0, B-) so the OPT core is continually energised between B+ and 0v via both tubes. How do the currents 'cancel' in this case - both tubes are on and passing the same B+ derived current. Is the magic all in the windings?
I'll happily take a YTFM if someone can suggest an easily digestible tome to consult on OPT design.
In Class A both output tubes will be biased to approx full power and that DC current must flow though the OPT. There is typically no dual rail power (i.e. B+, 0, B-) so the OPT core is continually energised between B+ and 0v via both tubes. How do the currents 'cancel' in this case - both tubes are on and passing the same B+ derived current. Is the magic all in the windings?
I'll happily take a YTFM if someone can suggest an easily digestible tome to consult on OPT design.
This is true for any tube push-pull output stage, not just class A. The plate currents
go through the windings in opposite directions, so the core flux cancels if they are equal.
Remember that both currents go into the center of the primary, and then out the ends.
https://www.ampbooks.com/mobile/amp-technology/amp-patent-7173488/push-pull-net-current-zero.gif
Last edited:
Thats great -- so I'm I'm understanding correctly.... it's not the fact that current continually (or not) flows in the core that causes saturation issues, it's the level of magnetic flux generated in the core at any time. Opposite windings generate opposing flux and its the flux that cancels not the currents.
Are you sure that is applies to all push pull amps? When operating in say the Class B section of an AB amp the current is only going though one winding - the other is 'off'. Hence there will be a flux 'imbalance', leading to an increase in net flux, possibly reaching saturation.
Are you sure that is applies to all push pull amps? When operating in say the Class B section of an AB amp the current is only going though one winding - the other is 'off'. Hence there will be a flux 'imbalance', leading to an increase in net flux, possibly reaching saturation.
Yes, in a class AB amp, when only one of the tubes conducts briefly on a peak,
this averages out when the other does the same. Heavy LF material can cause problems.
Another problem, but can be easily addressed:
The push pull transformer laminations do not have an appreciable air gap.
So the no-signal (quiescent currents) of the output tubes need to match.
Just a couple of milli-Amps of imbalance can cause the transformer to saturate early.
It can distort the bass, and also inter modulate the bass with the treble frequencies.
Using individual self bias and well matched output tubes; or adjustable fixed bias, with careful adjustment, can give a good balance of current.
Some 'matched tubes' may not have been checked for match at the same voltages and currents that your amp uses.
The push pull transformer laminations do not have an appreciable air gap.
So the no-signal (quiescent currents) of the output tubes need to match.
Just a couple of milli-Amps of imbalance can cause the transformer to saturate early.
It can distort the bass, and also inter modulate the bass with the treble frequencies.
Using individual self bias and well matched output tubes; or adjustable fixed bias, with careful adjustment, can give a good balance of current.
Some 'matched tubes' may not have been checked for match at the same voltages and currents that your amp uses.
That indicates the core is very sensitive to DC with respect to saturation - esp as typical operational currents in Class A push pull output stages are an order or magnitude higher than that and a 10% imbalance would not be difficult to achieve by tube variation alone, let alone winding inaccuracy. I've been trawling the net to find a 'study' on the impact of DC on OPTs - something that numerically shows how saturation varies with DC with the real world examples, rather than general theory. Interestingly there is very little info on it given its importance.
I agree, OPTs are not new and are well understood but... relatively few understand them well! There are lots of discussions on many OPT aspects but there seems comparatively little documentation (ex graphs) re real world studies out there on DC effects. Happy to self study and will dig out the tomes you mention - many thanks. Any other direction gratefully received.
I've built a KT88 PP classA amp biased at 80mA with torroidal OTP. Initialy the psu was not heavily filtered and I was relying on the differential function to cancel the ripple. That was happening when the tubes' current was slightly off balance (about 2-3mA). There was not any problem for the OTP. Later I modified the psu resulting on very low ripple so now there is no noise even if the idle currents are imbalanced. This allowed to bias the tubes exactly at the same idle current again without any problem for the OTP. No overheating, no bad sound.
when you get to read Crowhurst, you will see a lot of graphs and charts...
today there is the nomex paper that can be used, this is relatively new...
and lots of bobbins too that you can use to make OPT's....
Meno van der Veen is the man for toroids...
Patrick Turner has done a superb job of putting up a webpage.....education+diy
the website will answer all your questions, you only need to invest time reading and learning....
When an amp is biased so cold that it runs in B for any significant part of the half wave, the Q-current is already lowered enough to compensate for the difference in current between windings. Remember you have to lower the Q-current to get to run in B at all. In many cases you don't get there till you're over 30%-50% power and above that there isn't much room left. And as far as hearing any supply ripple during B cutoff from one tube.... at high volume, for part of a cycle?.... If you can't hear any hum when the amp is idle, you won't notice any in the music.
Last edited:
correct.
you need to consider AC and DC currents independently and remember that the inverted nature of the primary windings that causes the DC to offset also causes the AC to add due to the fact that the AC signals are also inverted by the phase splitter. In effect for AC the two halves of the primary are in parallel so it doesn't matter if there is signal though one (class B) or both (class A) halves the flux in the core remains the same.
The is a loose statement of lenz's law which put another way says that current delivered to the load does not add appreciable flux to the core and indeed all transformers (power, PP and SE) rely on this. However this does not relate to the discussion at hand about the DC offsetting and AC summing in a PP transformer.
dave
Since flux in a ferromagnetic material is a highly non-linear function of MMF (magneto-motive force) it is always better to think of MMF adding up or cancelling, rather than flux. (MMF in an iron circuit adds up correctly but flux does not.)
Here is one practical example of what can happen with unbalanced currents in the push pull output transformer:
There is a graph in the manual of the Heathkit W-5M amplifier of distortion, versus frequency, versus mis-adjusted bias (which results in unbalanced output tube plate currents). The W-5M has KT66 output tubes. Even though there is negative feedback to reduce distortion, the bias still needs to be adjusted so that the quiescent plate currents are matched.
The KT66 has a transconductance of about 7mA/Volt. A miss-adjusted bias voltage of 0.5 Volt (versus bias voltage that is adjusted properly to exactly match the tube currents), will have an unbalanced current of 3.5 mA.
The distortion at 2 Watts, due to unbalanced plate currents of 3.5mA, has the following distortion results:
20Hz 1.75%
50Hz 0.75%
100Hz 0.25%
Seemed like more than I might expect for an amp that uses negative feedback.
For an amplifier that does not have negative feedback, it could have even higher distortion due to unbalanced plate currents in the output stage.
There is a graph in the manual of the Heathkit W-5M amplifier of distortion, versus frequency, versus mis-adjusted bias (which results in unbalanced output tube plate currents). The W-5M has KT66 output tubes. Even though there is negative feedback to reduce distortion, the bias still needs to be adjusted so that the quiescent plate currents are matched.
The KT66 has a transconductance of about 7mA/Volt. A miss-adjusted bias voltage of 0.5 Volt (versus bias voltage that is adjusted properly to exactly match the tube currents), will have an unbalanced current of 3.5 mA.
The distortion at 2 Watts, due to unbalanced plate currents of 3.5mA, has the following distortion results:
20Hz 1.75%
50Hz 0.75%
100Hz 0.25%
Seemed like more than I might expect for an amp that uses negative feedback.
For an amplifier that does not have negative feedback, it could have even higher distortion due to unbalanced plate currents in the output stage.
It works like this (attached): the center tapped primary is connected series aiding, as seen from the placement of the phase dots. The total current enters the center tap and splits. This means that the current is flowing into one dotted side, and out the other. If the current flows this way, the DC mag flux cancels if there is the same number of turns on both sides of the center tap (always the case for balanced output). The primary of the OPT therefore responds to differential current only. No current differential, no signal and no DC magnetization.
Attachments
I'm trying to understand what is the right thing to do. Given that the OTP primary windings most probably won't be absolutely matched then if a pair of unmatched tubes is used how they should be biased; for hum cancelation they have to be at different idle currents. Or it would be better to bias at equal currents meaning using only matched tubes - perhaps matched to the transformer's windings mismatch(?) - or simply low ripple psu?
They never are since that would imply a coupling coefficient of 100%, and that's an impossibility. However, it's not worth worrying about with audio frequency OPTs that use high perm cores. The inherent imbalance is negligible.
Most practical OPTs, especially the E-I cores, can stand some DC imbalance. If you're not using cathode bias, then the best is to adjust both halves of the PP pair to draw the same plate currents under no signal conditions. These days, hum cancellation shouldn't be an issue. Back in "the day" high capacitance capacitors that could stand up to high voltages weren't so available. Not so much these days. Even if you use hollow state power diodes, you can still add LC ripple filters. That adds expense, but we're DiY-ers and don't have to answer to the accounting dept.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.